The Oxford Handbook of the Brazilian Economy (Oxford Handbooks) 9780190499983, 9780190600006, 0190499982

Citation preview

T h e Ox f o r d H a n d b o o k o f

T H E B R A Z I L IA N E C ON OM Y

The Oxford Handbook of

THE BRAZILIAN ECONOMY Edited by

EDMUND AMANN CARLOS AZZONI and

WERNER BAER

1

3 Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and certain other countries. Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America. © Oxford University Press 2018 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by license, or under terms agreed with the appropriate reproduction rights organization. Inquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above. You must not circulate this work in any other form and you must impose this same condition on any acquirer. Library of Congress Cataloging-in-Publication Data Names: Amann, Edmund, editor. | Azzoni, Carlos, editor. | Baer, Werner, 1931–2016, editor. Title: The Oxford handbook of the Brazilian economy / edited by Edmund Amann, Carlos Azzoni, and Werner Baer. Description: New York, NY : Oxford University Press, [2018] | Includes bibliographical references and index. Identifiers: LCCN 2017057398 | ISBN 9780190499983 (hardcover : alk. paper) | ISBN 9780190600006 (epub) Subjects: LCSH: Economic development—Brazil. | Brazil—Economic conditions. | Brazil—Economic policy. Classification: LCC HC187 .O975 2018 | DDC 330.981—dc23 LC record available at https://lccn.loc.gov/2017057398 1 3 5 7 9 8 6 4 2 Printed by Sheridan Books, Inc., United States of America

Contents

Acknowledgments  Contributors  1. Introduction  Edmund Amann and Carlos Azzoni

ix xi 1

PA RT I   H I STOR IC A L P E R SP E C T I V E S  2. The Colonial Economy  Flávio Rabelo Versiani

17

3. The Nineteenth and Early Twentieth Centuries  André Villela

40

4. Brazilian Structuralism  Joseph L. Love

63

5. Brazil’s Import-​Substitution Industrialization  Werner Baer

89

6. Experiences of Inflation and Stabilization, 1960–​1990  Fernando de Holanda Barbosa

105

7. Leviathan Captured: Neoliberalism as Solution and Problem in Brazil  124 Philippe Faucher 8. Growth Volatility and Economic Growth in Brazil  Jorge Arbache and Sarquis J. B. Sarquis

147

PA RT I I   M AC ROE C ON OM IC P OL IC Y A N D I N ST I T U T ION S  9. The Brazilian Development Bank  Luiz Ricardo Cavalcante

177

vi   Contents

10. The Evolution of Brazil’s Banking System  Gustavo S. Cortes and Renato L. Marcondes 11. Brazil’s Macroeconomic Policy Institutions, Quasi-​Stagnation, and the Interest Rate–​Exchange Rate Trap  Luiz Carlos Bresser-​Pereira

198

221

PA RT I I I   T H E P ROD U C T I V E SE C TOR S  12. Evolution and Sectoral Competitiveness of the Brazilian Manufacturing Industry  Paulo César Morceiro

243

13. The Agricultural Sector  Carlos José Caetano Bacha

266

14. Traditional Agriculture and Land Distribution in Brazil  Charles C. Mueller

288

15. Brazil’s Agricultural Modernization and Embrapa  Geraldo B. Martha Jr. and Eliseu Alves

309

16. Manufacturing, Services, and the Productivity Gap  Jorge Arbache

338

17. Energy in Brazil: Past and Future  José Goldemberg

358

18. Infrastructure  Edmund Amann, Werner Baer, Thomas Trebat, and Juan Villa Lora

377

19. Trade Policy from the 1930s to the Present  Simão Davi Silber

394

PA RT I V   B R A Z I L’ S R E G ION S  20. Regional Disparities  Carlos R. Azzoni and Eduardo A. Haddad

423

21. Brazil’s Northeast  Alexandre Rands Barros

446

Contents   vii

PA RT V   S O C IOE C ON OM IC DI M E N SION S  22. Changes in Income Distribution in Brazil  Rodolfo Hoffmann 23. The Development of Brazilian Education: A Tale of Lost Opportunities?  Claudio de Moura Castro 24. Anti-​Poverty Transfers and Poverty Reduction  Armando Barrientos 25. South-​South Cooperation for Social Development: Brazil and Africa Examined  Anthony Hall

467

489 511

535

26. Labor Market Development in Brazil: Formalization at Last?  Celia Lessa Kerstenetzky and Danielle Carusi Machado

552

27. Environmental Issues  Ariaster B. Chimeli

577

28. The Economics of Health in Brazil  Antonio Carlos Coelho Campino, Maria Dolores Montoya Diaz, and Flavia Mori Sarti

593

PA RT V I   B R A Z I L A N D T H E WOR L D E C ON OM Y  29. Brazil, the BRICS, and the Changing Landscape of Global Economic Governance  Peri Silva

621

30. Brazilian Trade and International Economic Prospects in an Anti-​Globalization Era  Donald V. Coes

645

31. The Evolution of Foreign Direct Investment in Brazil  Breno Augusto da Silva e Silva

664

32. Multinational Corporations from Brazil  Edmund Amann

682

viii   Contents

PA RT V I I   T H E C HA N G I N G ROL E OF T H E STAT E  33. The Rise and Fall of State Enterprises  Armando Castelar Pinheiro

701

34. Antitrust and Competition Policy in Brazil  Eduardo Pontual Ribeiro, Camila Pires-​Alves, and Luis Carlos D. Prado

718

35. Corruption Scandals, the Evolution of Anti-​Corruption Institutions, and Their Impact on Brazil’s Economy  Mariana Mota Prado and Lindsey Carson

741

Index 

769

Acknowledgments

In March 2016, during the early stages of this volume’s preparation, one of its editors, Professor Werner Baer, sadly passed away. For generations of scholars working on the political economy of Latin American and Brazilian development, Werner was a truly inspiring figure. From his early work on industrialization through to his contemporary research on socially inclusive growth, Werner not only kept pace with, but also set the agenda for research into an exciting, but surprisingly little understood, region. His work appeared in scores of highly cited journal articles, as well as several notable volumes including The Brazilian Economy, which eventually ran to six editions. Besides his published output, Werner, during his career at Vanderbilt University and then at the University of Illinois, mentored and encouraged dozens of younger colleagues and students, instilling in them an enthusiasm and knowledge that was to prove the foundation for numerous successful careers. Werner’s PhD students include a former governor of the Brazilian Central Bank, a former president of Ecuador, and several leading academics in Brazil and throughout the region. Werner was tremendously enthusiastic about this Handbook, and it proudly features some of his last published work. It was certainly a privilege for us to see this volume through to conclusion. In developing the chapters for this volume, the editors in July 2016 organized a workshop in São Paulo. This drew on the participation of contributors from Brazil, the United States, and Europe. We are grateful to the UK Department for International Development (DFID), the University of Manchester, and the University of São Paulo for the funding and administrative support that made this valuable event possible. Finally, we are extremely grateful to Christopher Lyon, from the University of Manchester’s Global Development Institute, for his assistance in copyediting and formatting the chapters. Edmund Amann and Carlos Azzoni, July 2017

Contributors

Eliseu Alves,  former President and current Presidential Advisor, Embrapa, Brazilian Ministry of Agriculture Edmund Amann,  Professor of Brazilian Studies, Institute of History, Leiden University Jorge Arbache,  Secretary for International Affairs, Brazilian Ministry of Planning and Professor of Economics, Department of Economics, University of Brasília Carlos R. Azzoni,  Professor of Economics, Faculty of Economics and Administration, University of São Paulo Carlos José Caetano Bacha, Professor of Economics, University of São Paulo at Piracicaba-​Esalq Werner Baer,  Professor of Economics, Department of Economics, University of Illinois at Urbana-​Champaign Armando Barrientos,  Professor of Poverty and Social Justice, Global Development Institute, University of Manchester Alexandre Rands Barros,  President, Datamétrica, Recife Luiz Carlos Bresser-​Perreira,  former Brazilian Minister of Finance and Professor of Economics, Fundação Getúlio Vargas–​São Paulo Antonio Carlos Coehlo Campino,  Professor of Economics, Faculty of Economics and Administration, University of São Paulo Lindsey Carson,  Adjunct Lecturer in International Development, SAIS, Johns Hopkins University Claudio de Moura Castro,  former Professor of Economics, Catholic University of Rio de Janeiro Luiz Ricardo Cavalcante,  Economist, Brazilian Senate Ariaster B. Chimeli,  Professor of Economics, Faculty of Economics and Administration, University of São Paulo Donald V. Coes,  former Professor of Economics, Department of Economics, University of New Mexico

xii   Contributors Gustavo S. Cortes,  PhD Candidate, Department of Economics, University of Illinois at Urbana-​Champaign Maria Dolores Montoya Diaz,  Professor of Economics, Faculty of Economics and Administration, University of São Paulo Philippe Faucher,  Associate Professor, Department of Political Science, University of Montreal José Goldemberg,  Professor, Institute of Physics, University of São Paulo Eduardo A. Haddad,  Professor of Economics, Faculty of Economics and Administration, University of São Paulo Anthony Hall,  Professor of Social Policy, London School of Economics Rodolfo Hoffmann,  Professor of Economics, Faculty of Economics Administration, University of São Paulo, Esalq; and University of Campinas

and

Fernando de Holanda Barbosa,  Professor Brazilian School of Economics and Finance (EPGE) Fundação Getúlio Vargas, Rio de Janeiro Celia Lessa Kerstenetzky, Professor of Economics, Department of Economics, Fluminense Federal University Joseph L. Love,  Professor Emeritus of History, Department of History, University of Illinois Danielle Carusi Machado,  Adjunct Professor of Economics, Department of Economics, Fluminense Federal University Renato L. Marcondes,  Associate Professor of Economics, Faculty of Economics and Administration, University of São Paulo Geraldo B. Martha  Jr., Researcher, Secretariat of International Relations, Embrapa, Brazilian Ministry of Agriculture Paulo César Morceiro,  Researcher, NEREUS, University of São Paulo Charles C. Mueller,  Professor Emeritus, Department of Economics, University of Brasilia Armando Castelar Pinheiro,  Coordinator of Applied Economics, IBRE-​FGV, Rio de Janeiro and Professor of Economics, Federal University of Rio de Janeiro Camila Pires-​Alves,  Federal University of Rio de Janeiro Luis Carlos D. Prado,  Federal University of Rio de Janeiro Mariana Mota Prado,  Associate Dean Graduate Studies, Faculty of Law, University of Toronto

Contributors   xiii Eduardo Pontual Ribeiro,  Professor of Economics, Institute of Economics, Federal University of Rio de Janeiro Sarquis J.  B. Sarquis,  Minister Counsellor, Embassy of Brazil, Paris, and Associate Researcher, London School of Economics Flavia Mori Sarti, Assistant Professor, School of Arts, Sciences and Humanities, University of São Paulo Simão Davi Silber,  Professor of Economics, Faculty of Economics and Administration, University of São Paulo Peri Silva,  Associate Professor, Department of Economics, Kansas State University Breno Augusto da Silva e Silva,  Professor, Department of Economics, Fluminense Federal University Thomas Trebat,  Director, Columbia University Rio Global Center Flávio Rabelo Versiani,  Professor of Economics, Department of Economics, University of Brasília Juan Villa Lora,  Researcher, Inter-​American Development Bank, Washington, DC André Villela,  Researcher, Fundação Getulio Vargas–​EPGE, Rio de Janeiro

T h e Ox f o r d H a n d b o o k o f

T H E B R A Z I L IA N E C ON OM Y

Chapter 1

Introdu c t i on Edmund Amann and Carlos Azzoni

Brazil’s rise to global prominence over the past 30 years comprises several dimensions. These include an increasingly activist posture on the world diplomatic stage, an influential sporting and artistic culture, and a pivotal role in the international struggle to preserve biodiversity and fight climate change. However, at the core of Brazil’s emergence as a major global player lies its vast and diverse economy. According to World Bank estimates, in 2016 Brazil possessed the seventh-​largest economy on earth, the largest economy in Latin America, and the third-​largest economy among the BRICS (Brazil, Russia, India, China, and South Africa) group of major emerging-​market nations. Reflecting on the sheer scale of the Brazilian economy alone cannot do full justice to its global significance. This significance arises as much from the policy lessons that Brazil can offer to the rest of the world as to the country’s pivotal role in the international economy. In terms of Brazil’s direct role in the global economy, it is hard to overestimate the country’s importance. As the process of globalization has unfolded over the past three decades, Brazil has become one of the major destinations for inward foreign direct investment (IFDI). Brazil stands alongside Mexico as the most important recipient of such flows in Latin America. The surge in IFDI has partly resulted from an ambitious privatization program that, during the 1990s, was the world’s largest. The same forces of economic liberalization that have propelled IFDI have also encouraged Brazilian enterprises to internationalize. As a result, Brazil’s stock of foreign corporate assets has increased by more than a factor of 10 since the early 1990s. Important though investment flows are, it is probably Brazil’s role in international production chains that has attracted the most attention. Having long enjoyed a significant comparative advantage in the production of natural resource–​based (NRB) products, Brazil has taken advantage of the global growth in trade, in particular the rise of China, to propel its exports. As a result of this process, exports of such products as soy, iron ore, and meat have rapidly accelerated to the point where Brazil is the world’s leading exporter of these and related product categories. The Brazilian export

2    Edmund Amann and Carlos Azzoni success story is not confined to commodities; Brazil is now the world’s largest exporter of regional jet aircraft and is South America’s largest automotive products exporter. As Brazil’s economy expanded through the 1990s and the first decade of this millennium, so did the country’s role in international cooperation and governance. Having played a central role in the creation of a regional customs union, Mercosul, in the early 1990s, by the 2000s Brazil’s economic governance role had become truly global. During this period it cofounded the BRIC group, jointly launched the Shanghai-​based New Development Bank, and was admitted to the G20 group of the World’s largest economies. As indicated previously, the global influence of the Brazilian economy also stems from the policy lessons that other countries have drawn. Up until the early 2010s, at least, Brazil became renowned for its socially inclusive development model (Amann and Barrientos 2016). The latter ensured that the fruits of rapid export-​fueled growth filtered through to the poor, resulting in poverty alleviation and a more even distribution of income. Such progress was achieved through a combination of conditional cash transfer (CCT) programs (notably the Bolsa Familía), macroeconomic stabilization, and buoyant labor markets. Learning from the Brazilian experience, a number of countries, notably in Sub-​Saharan Africa, began implementing CCT programs. The sense that Brazil serves as a laboratory for the testing of new economic ideas has roots that stretch back a long way beyond the creation of CCT programs in the first decade of the 2000s. In the 1950s and 1960s, Brazil was a pioneer in implementing the insights of the structuralist development economists. This gave rise to the rapid industrialization of the economy and a turning away from Brazil’s traditional focus on NRB product exports. Three decades later and facing hyperinflation and stagnating output, Brazil once again proved willing to put new economic ideas to the test, from neostructuralism (as in the Bresser anti-​inflation plan) to elements of rational expectations theory (in the design of elements of the Real Plan). In this sense, and irrespective of its absolute size, Brazil has long served as a focus of interest for economists from around the world. Although Brazil now clearly ranks among the world’s leading economies, it is equally obvious that it remains beset with serious difficulties, many of which, but by no means all, are of its own making. After more than a decade of rapid expansion, Brazil entered a period of economic slowdown and recession in 2012, from which it only appeared to be gradually emerging five years later. As growth vanished and the economy contracted, unemployment rose, placing the progress made in alleviating poverty in severe jeopardy. At the same time, rising prices started eating away at living standards. Eventually, though, the authorities managed to contain price pressures; certainly there was no return to the hyperinflation typical of the 1980s. What factors resulted in this reversal of fortune? The chapters that follow this introduction will, of course, give a much more detailed account of the issues involved. However, in essence the difficulties stem from unaddressed structural constraints, in particular low productivity growth, a failure to sufficiently diversify away from NRB production, and chronic underinvestment in infrastructure and education. Once global

Introduction   3 demand for commodities began to slacken (which occurred shortly after the global financial crisis of 2008–​2009), the Brazilian economy was left badly exposed as a critical source of demand for its products fell away. Had Brazil enjoyed a more diversified economy and a more responsive supply side, it would have been able to cope with the external commodities demand shock more adroitly. As it was, the economy proved incapable of responding rapidly and positively to changing global circumstances. Adding to the difficulties and impeding recovery has been an unprecedented political crisis. This crisis, centred on the so-​called Lava Jato (Car Wash) scandal, has focused on the diversion of funds intended for public sector contracts to political parties. In an illustration of institutional robustness, investigations have been rigorously prosecuted by the federal judicial authorities, resulting in the imprisonment of leading political and business figures. A parallel scandal centring on the illegal manipulation of the public sector accounts saw the impeachment of President Rouseff in mid-​2016. As regards the subject matter of this volume, Brazil’s economy, the corruption scandals have two important ramifications. First and most obviously, the political and regulatory uncertainty that they have brought have created an adverse environment for public-​and private-​sector investment. They have also deflected policymakers from the pursuit of reforms. Taken together, these developments have done little to accelerate the exit of Brazil’s economy from recession. Second, and perhaps more profoundly, the wave of scandals has thrown a harsh light on the nature of state-​business relations. As the investigations have unfolded since 2013, the enormous scale of the clientalism and graft surrounding the interaction of the state and Brazil’s leading enterprises has become abundantly clear. This development, at the very least, calls into question the continued sustainability of a critical component of Brazil’s development strategy over the years: the close articulation between state and business. As will be argued in subsequent chapters, the latter was central to the industrialization process and has remained firmly embedded in Brazil’s political economy despite economic liberalization and opening from the late 1980s onward (Alston et al. 2016, Chapter 2). Summing up our initial discussion, Brazil, then, remains a globally vital but troubled economy. Against this background, the purpose of this volume is to offer real insight into the Brazilian economy’s development in the contemporary context, understanding its most salient characteristics and analyzing its structural features across various dimensions. At a more granular level, this volume will set out to accomplish the following tasks. First, it will seek to provide an understanding of the economy’s evolution over time and the connection of its current characteristics to this evolution. Second, it will attempt to portray and explain Brazil’s broader place in the global economy and consider the ways in which this role has changed and is likely to change over coming years. Third, and very much reflecting contemporary concerns over its performance, the volume aims to provide an understanding not only of how one of the world’s key economies has developed and transformed itself, but also of the ways in which this process has yet to be completed. This will involve understanding the current challenges facing the Brazilian economy and the kinds of issues that need to be tackled for these to be addressed successfully.

4    Edmund Amann and Carlos Azzoni To accomplish these objectives, the structure of the volume is composed of seven parts. Part I  examines the evolution of the Brazilian economy over time, reflecting on such relevant issues as the colonial era, industrialization, and the emergence of a Brazilian tradition of economic thought. Part II then examines Brazil’s macroeconomic policy and institutions, focusing in particular on the evolution of macroeconomic policy approaches and public and private banking institutions. Next, in Part III, the focus becomes more microeconomic, with the emphasis placed on the supply-​side characteristics of the Brazilian economy. Thus, this section contains chapters analyzing the agricultural sector, industry, energy, and crucially, the infrastructure and productivity gaps. Part IV recognizes that an important—​though often overlooked—​dimension of Brazil’s economy stems from the country’s vast territorial dimensions and regional diversity, whether in terms of sectoral focus, income, or export orientation. Against this background, Part IV contains chapters on the regional distribution of economic activities and the Northeast, a region often characterized as Brazil’s poorest and least developed. Next, Part V examines the critical socioeconomic dimension. As previously indicated, Brazil has recently experienced a significant period of rapid, pro-​poor growth. Part V examines the background to this achievement and its future sustainability. Chapters in this section focus on such critical issues as human capital formation, the role of conditional cash transfer programs, health, and environmental issues. Part VI then moves on to consider the role of Brazil in the global economy. It features chapters on trade, foreign direct investment, and Brazil’s involvement in the BRICS group of major emerging economies. Finally, Part VII considers the changing role of the state. The evolution of business government relations forms an important focal point of this section, with chapters on privatization, competition policy, and corruption. To provide a broader context for the detailed analysis embedded in these sections, the remainder of this introductory chapter considers the recent development of Brazil’s economy, focusing on what we argue to be its incomplete transition and associated challenges.

1.1.  The Brazilian Economy: Incomplete Transition and Outstanding Challenges Since World War II, Brazil’s economy has undergone enormous structural changes (Ioris 2014). Collectively, these changes have repositioned Brazil in the global division of labor, have reconfigured fundamentally the role of the state, and have driven significant improvements in standards of living, even for the poorest groups of the population. Yet despite these achievements, there would be common acceptance that Brazil’s economic performance has fallen well short of its obvious potential. Rich in

Introduction   5 natural resources, endowed with a talented and entrepreneurial population, and largely insulated from global conflict, Brazil has still failed to enter the ranks of the world’s most developed nations. It remains firmly anchored among upper middle-​income economies and continues to contend with a highly skewed distribution of income and pockets of stubborn poverty. This stands in stark contrast to the experience of the East and Southeast Asian newly industrializing economies such as Taiwan and South Korea and, of course, China. In these economies, which have employed modernization and industrialization strategies not entirely dissimilar to those of Brazil, sustained year-​on-​ year growth in output and exports have driven an unprecedentedly rapid increase in gross domestic product (GDP) and living standards. As a result, China has become the world’s second-​largest economy while, for example, South Korea, now an Organisation for Economic Co-​operation and Development (OECD) member, has entered the ranks of the advanced industrial nations. All of this suggests two key points that require addressing in any comprehensive analysis of the Brazilian economy. The first centers on the fact that, notwithstanding the current crisis, significant progress has been made over the past six decades. In this connection, it is important to acknowledge Brazil’s very real economic achievements and consider how they were accomplished. The second point is to recognize that Brazil’s economic transformation, in particular concerning the issue of social inclusion, is clearly a work still in progress. Despite very real accomplishments, the Brazilian economy remains beset by chronic structural constraints. As recent events have proved, these constraints render it highly vulnerable to external shocks. The remainder of this section divides into two components. In the first, we concisely consider the evolution of Brazil’s economy in long-​term perspective, drawing attention to the policies and circumstances which have underpinned the considerable progress that has been made. In the second section we reflect on the nature of the Brazilian economy’s incomplete transition, highlighting the various issues and structural obstacles which will need to be tackled if the country is to embark on a path of sustainable, inclusive long-​ term growth.

1.1.1. The Development of the Brazilian Economy over the Long Term In a very real sense, the evolution of the Brazilian economy over the postwar period can be thought of a process of structural change that has experienced alternate periods of acceleration and retardation. Central to this idea of structural change lies the pursuit of desenvolvimento or, in English, development. It is significant that the Portuguese term, as opposed to its most common English translation, connotes a process of breaking from inherited circumstances, literally “dis-​involvement.” Successive generations of policymakers have sought, with various degrees of commitment and success, to engineer a rupture from a legacy dating back to the colonial era in which Brazil functioned

6    Edmund Amann and Carlos Azzoni as an essentially peripheral economy, supplying raw materials to the more developed countries (see Chapter 2). The first real signs of such structural change, as André Villela indicates in Chapter 3 of this Handbook, stemmed from the ramifications of the coffee economy that grew up in the state of São Paulo in the second half of the nineteenth century. Rapid coffee-​driven export growth fueled capital accumulation and significant immigration from Europe and eventually Japan. As this process unfolded, the new supplies of capital, labor, and skills provided the basis for a nascent industrial sector, which initially concentrated itself in the southeast of the country, close to centers of coffee production. By the start of the twentieth century, and without any formal development or industrialization strategy, Brazil had taken the first real steps toward structural transformation and a repositioning of itself in the global division of labor. This process may well have continued in an organic and unplanned fashion but for the traumatic international economic events unleashed by the Wall Street Crash of 1929. In the aftermath of the Crash, tariff barriers were elevated in Brazil’s traditional North American and European markets, while access to foreign capital was brutally attenuated. Responding to these events, the administration of President Getúlio Vargas engineered a decisive break from Brazil’s previous liberal economic model. A combination of currency devaluation and increased tariff protection under Vargas during the 1930s provided stimulus for the growth of nontraditional sectors. In addition, in the so-​called Estado Novo (New State), Vargas created Brazil’s first developmentalist administration, committed to modernization, structural transformation, corporatism, and a much extended role for the public sector in economic affairs (Baer 2014). The most iconic symbol of the new economic order was the creation of the Companhia Siderurgica Nacional (National Steel Company, CSN), the first major state-​owned industrial enterprise in Brazil. Many more such state-​owned enterprises (SOEs) were to follow in three decades following World War II, not least in the energy sector (Chapter 17). The end of World War II saw the beginning of a more fundamental and structured effort to transform the Brazilian economy and recast its role in the international division of labour. As indicated by Joseph Love in Chapter 4, the background to this was a revolution in intellectual thought surrounding the process of economic development. Unlike previous currents of economic thought, the latest wave, structuralism, was largely Latin American in origin. Central to the structuralist analysis was a contention that peripheral Latin American economies, unless they moved away from a traditional reliance on commodities exports, faced a future of stagnating growth and expanding external deficits. The reason for this lay in what Prebisch (1950) and Singer (1950) characterized as a secular decline in the terms of trade for exports of primary products (as opposed to exports of services or manufactures). The obvious remedy, therefore, lay in the pursuit of an industrialization strategy that would enable peripheral economies to break out of the low-​growth trap engendered by colonial era–​style patterns of engagement with the global economy. The intellectual insights offered by structuralism found a ready audience in successive Brazilian administrations during the first two decades after World War II. As highlighted

Introduction   7 by Werner Baer in Chapter 5, these translated rapidly into a planned process of industrialization, commonly referred to as import substitution industrialization (ISI), which involved, in essence, the substitution of domestic for foreign-​produced industrial products. This, it was hoped, would alleviate the external constraint, lessen vulnerability to external shocks, and drive an increase in growth and living standards. The policies involved in many ways represented a continuation of those initiated by Vargas in the 1930s. They comprised extensive deployment of tariff and nontariff barriers, managed exchange rates, the creation of SOEs, and the increased use of directed credit from state-​ run financial institutions. In many ways, the initial phases of ISI, which extended across the late 1940s and throughout the 1950s were a significant success. Chapter  5 points to the fact that GDP growth accelerated against a background of rapid structural change as new industries were established and the traditional NRB sectors gave way to industry and manufacturing as engines of growth. The period was epitomized by President Juscelino Kubitschek’s famous Plano de Metas (Plan of Targets), which in the late 1950s saw the foundation of South America’s largest automotive sector and the creation of Brasilía, the nation’s new futuristic capital. Yet, despite the structural transformation and undoubted dynamism of its early period, it rapidly became apparent that ISI was failing to achieve its key macroeconomic objective: the alleviation of the external constraint. In essence, the problem here stemmed from the fact that the ISI process was quite heavily import intensive and required substantial inflows of foreign investment (which in turn gave rise to interest and profit remittances). In order to ensure current account equilibrium under these circumstances, exports would have needed to grow rapidly. This in fact did not happen to a sufficient extent due to a combination of exchange rate overvaluation, a traditional export sector damaged by an adverse shift in the internal terms of trade, and the fact that the new industries, still ascending their learning curves and heavily protected, were not competitive at world market prices. At the same time, the expansion of demand, especially consumer demand, engendered by ISI posed a significant challenge to price stability given insufficient supply-​side flexibility and the wedge driven between domestic and world prices by protectionist policies. As a result, inflation became a very serious issue. Ultimately, by the early 1960s, these factors resulted in a slowdown in growth amid rising prices. Following its seizure of power in 1964, the military government attempted to revitalize ISI through a series of reforms designed to boost exports and increase investment. The reforms centered on capital market liberalization, managed devaluation, and selected trade liberalization. For a while, this proved quite successful and gave rise to the so-​called miracle years of rapid growth between 1967 and 1973. However, the trebling of world oil prices in 1973 and the associated advent of a global economic slowdown forced a reassessment of Brazil’s development strategy. The military government’s response, the Second National Development Plan of 1974, comprised an attempt to de-​emphasize export growth in favor of intensified ISI. It was hoped that through the substitution of imports in new sectors such as capital goods, the external constraint could be successfully tackled (Amann 2000). In fact, this did not prove to be the case, as Brazil’s current

8    Edmund Amann and Carlos Azzoni account deficit expanded at worrying speed during the second half of the 1970s and into the early 1980s. The first half of the 1980s marked a significant turning point for Brazil both in economic and political terms. After two decades in power, the military finally ceded control to a civilian administration in 1985, ushering in a period of democratic rule which has endured up until the present day. The transition away from military rule came at time of economic crisis. As Chapter  6 by Fernando Holanda indicates, the inward-​ looking economic strategy pursued through the 1970s and into the 1980s proved incompatible with price stability. By the mid-​1980s, hyperinflation had become a worrying reality. Meanwhile, the cost of servicing Brazil’s enormous external debt had become unsustainable following the unprecedented monetary tightening among the Western economies in 1979–​1981. By 1987, following Mexico’s example from four years earlier, Brazil defaulted on its external obligations. Under these circumstances, Brazil had no choice but to seek the assistance of the World Bank and the International Monetary Fund. As was the case right across the region, as Philippe Faucher argues in Chapter 7, the end result was the implementation of a policy set commonly characterized as neoliberal and almost universally termed “the Washington Consensus.” At its heart, the new policy regime sought to dismantle the protectionist structures of ISI through a progressive strategy of trade and investment liberalization (see Simão Silber, Chapter 19 in this volume). The idea here was to remove the wedge between domestic and world prices, exposing the domestic productive sector to external challenge and thereby stimulating competitiveness. This, it was hoped, would trigger export growth and contribute to an alleviation of the external constraint. At the same time, the investment liberalization part of the package—​especially as it related to privatization (see Armando Castelar Pinheiro, Chapter 33)—​would drive inflows of FDI, which would provide a potentially more stable source of financing. In Brazil, all these key elements of the Washington Consensus came to be implemented, starting in the late 1980s with President Sarney’s administration and a series of tentative trade reforms. This was followed by a much bolder package of trade liberalization between 1990 and 1994, instigated under Brazil’s first popularly elected president in almost 30 years, Fernando Collor de Melo. Under his successors, Itamar Franco (1992–​1994) and Fernando Henrique Cardoso (1995–​2002), liberal reforms continued apace with the creation of the world’s largest privatization program. This saw the sale of the telecommunications and much of the electricity, water, and sanitation sectors to domestic and especially foreign investors. Limited policy reforms also occurred in health and labor markets (Chapters 26 and 28). At the same time, steps were taken to increase market access by removing barriers that had previously prevented the entry of new participants. Thus, in the case of oil exploration, an ISI-​era edict preventing foreign investment in upstream activities was abolished in 1995. As these policy initiatives unfolded, FDI targeted at Brazil surged (Breno Augusto da Silva e Silva, Chapter 31 in this volume). Although the subsequent PT (Workers’ Party) administrations of Luiz Inácio Lula da Silva (Lula) and Dilma Rousseff (2003–​2016) were to prove less ideologically committed

Introduction   9 to Washington Consensus–​style policies, there was no real attempt to roll back any of the major trade, investment, or market reforms. Still, by the same token, nor was there any effort to extend their scope. This means, as Chapter 19 (by Silber) and Chapter 30 (by Donald Coes) reveal, that Brazil, by global standards, remains a relatively closed economy. As will be argued in the next section, this poses a fundamental challenge for an economy seeking to escape crisis and reignite growth. More than the ambitious package of trade and market reforms implemented in that decade, the 1990s are recalled as the critical period when hyperinflation was successfully tackled. After several previous failed attempts (see Chapter  6), the administrations of Presidents Franco and Cardoso finally implemented an effective stabilization package known widely as the Real Plan (see Chapter 11 by Luiz Carlos Bresser-​Pereira). The Plan was launched in 1993 and the currency it gave rise to—​the real—​is still in use today, a remarkable achievement given Brazil’s long history of monetary instability and failed currencies. Central to the Plan’s success was an eclectic mixture of orthodox measures (a fiscal adjustment) and heterodox interventions (progressive de-​indexation) allied to the creation of a new currency, pegged to the US dollar within an initially narrow band. In the latter sense, Brazil’s assault on inflation had much in common with that of other countries in the region during the 1990s. Argentina, for example, successfully tackled inflation by launching a convertibility plan that saw the peso pegged at parity to the dollar. The central principle behind these initiatives was to provide a set of clear policy rules that would be consistently followed, allowing the monetary authorities to build up credibility among economic agents. Allied to the forging of a new collective psychology surrounding the price level, the adoption of a new, comparatively strong currency directly forced inflation down. This direct impact resulted from the influence of lower import prices on general price formation in an increasingly open economy. The Real Plan initially proved highly successful, reducing inflation from four to single digits between 1993 and 1996. Growth also picked up, and there were notably positive impacts on the incidence of poverty and indigence. However, as Chapter 11 makes clear, the Real Plan was not without flaws. The initial band within which the currency fluctuated against the dollar turned out to be incompatible with the maintenance of external balance, highlighting a recurring theme: the inability to reconcile sustained growth with external equilibrium. Eventually, in 1999, the authorities were forced to bow to the inevitable, and the real was subject to a maxi-​devaluation. The policy mechanisms designed to ensure continued price stability were altered, with the emphasis switching to inflation targeting. Nonetheless, despite these serious issues, the real remains in place and inflation has kept within single digits for the great majority of the time. Indeed, at the time of writing, consumer price inflation had fallen to close to 3% per annum, a record low. In many ways, the Real Plan laid the macroeconomic foundations for the years of rapid growth and social progress that were to characterize the first two administrations of President Lula (2003–​2010) and the first administration of President Rousseff (2011–​ 2014). While inflation targeting kept inflation at bay, the Brazilian economy embarked

10    Edmund Amann and Carlos Azzoni on a period of expansion, the likes of which had not been experienced since the so-​ called miracle years of 1967–​1973. Amann and Barrientos (2016) suggest a number of elements came together, forming a “Brazilian Development Model” which ensured not only that growth would accelerate, but also that its fruits would be more evenly distributed than had been the case during previous boom periods. The factors propelling growth included a buoyant global market for commodities combined with an NRB sector that had undergone significant modernization and productivity growth (see Chapter 13 by Carlos Bacha; Chapter 14 by Charles Mueller; and Chapter 15 by Geraldo Martha and Eliseu Alves). As we have seen, price stability was ensured thanks to a broadly successful inflation-​targeting framework. Supplementing this pro-​poor element was the development of targeted poverty alleviation programs featuring, among others, the Bolsa Familía Conditional Cash Transfer initiative (see Chapter  24 by Armando Barrientos). As the Brazilian economy approached the end of the first decade of the twenty-​first century, its global image had been transformed. Investors and observers took note of its dynamism and its capacity to deal effectively with poverty and inequality. With this new international prominence, Brazil became increasingly active in global economic governance and development initiatives. As Chapter 25 (by Anthony Hall) and Chapter 29 (by Peri Silva) make clear, Brazil became intensely engaged in Sub-​Saharan African development and the creation of a cooperation agreement with the other BRICS economies: Russia, India, China, and South Africa. However, as will be argued in more detail in the next section, Brazil’s economic success during this period masked some serious structural issues. Among the most pertinent were a failure to diversify sufficiently export revenues beyond NRB products, poor productivity growth outside agriculture, and a weakening commitment to fiscal responsibility. By 2012–​2013 a secondary effect of the international financial crisis was making itself felt as commodity prices and demand fell sharply. Brazil’s capacity to cope with this external shock, which would always have been restricted given its NRB-​export dependence, was further inhibited by the outbreak of a wide-​ranging corruption scandal in 2014. As we have seen, although the president was not directly implicated, the fallout from the scandal resulted in the impeachment of President Dilma Rousseff in 2016. The combination of external economic shock and an internal political crisis proved highly detrimental to Brazil’s hitherto dynamic economic performance. In 2014 Brazil fell into a recession that was to last for two years, the longest in the country’s history. Between 2014 and 2016 the economy contracted by no less than 8%. By 2017 Brazil had emerged from recession, but the speed of the recovery appeared modest. Contemplating the rapid deterioration of Brazil’s economic performance since the beginning of the 2010s, one is forced to reflect on the factors that may have predisposed the economy to such a striking reversal of fortune. In the final section of this introductory chapter, we consider some of the key structural constraints that, we argue, must be overcome if Brazil is to avoid similar crises in the future and regain lost momentum.

Introduction   11

1.2.  Outstanding Challenges In considering the challenges that continue to afflict the Brazilian economy, it is important not to lose sight of the fact that, in one key sense at least, real progress has been made. During much of the period between the end of World War II and the late 1980s, Brazil was prone to repeated bouts of hyperinflation. However, since the early 1990s and the introduction of the Real Plan, relative price stability has become an established fact of economic life. In fact, since 1996 consumer price inflation on an annualized basis has remained in single figures with the exception of just one year: 2015. At the time of writing, inflation was hovering around 3% per annum, a level not markedly dissimilar to the OECD average. Unlike its neighbor Argentina, the end of the exchange rate–​ targeting regime in Brazil at the end of the 1990s did not trigger an uncontainable inflationary surge, indicating to some extent, the resilience of the authorities’ price stability strategy (see Chapter 11). Another impressive facet of Brazil’s macroeconomic progress has centered on the stability of its financial system. While it is true that the sector is characterized by high intermediation costs and an arguably excessive role for the public sector (see Chapter 9 by Ricardo Cavalcante; and Chapter 10 by Gustavo Cortes and Renato Marcondes), it has remained highly capitalized. Thus, Brazil was comparatively unaffected by the first-​ round effects of the international financial crisis of 2007–​2008. Unlike the United States and the United Kingdom, for example, no Brazilian banks came close to failure, and there was no requirement for government bailouts or subsequent quantitative easing. In many senses, the Brazilian experience during this period mirrors that of Canada: both countries reaped the benefits of highly capitalized, heavily regulated, and comparatively closed financial systems. While price stability and its pro-​poor effects should be rightly celebrated, the problem remains that financial stability and the successful containment of inflation appear not to have been reconcilable with rapid and sustained growth. In this sense, despite the progress of recent years, little has changed. As Chapter 8 (by Jorge Arbache and Sarquis J. B. Sarquis) shows, Brazil’s average growth, when viewed over the past two decades, has been somewhat disappointing. Indeed, it was lower than that experienced between the end of World War II and the crisis-​hit 1980s. In examining the growth record since 1990, Chapter 8 makes clear not only that average growth has been modest, but also that this average conceals marked volatility. Thus, for example, a growth surge in the immediate wake of the introduction of the Real Plan in 1993 could not be sustained beyond two or three years. Indeed, in 1998-​–​1999, Brazil entered a crisis that required the intervention of the International Monetary Fund (IMF). Growth did pick up in the following two decades, but even at the height of the last boom, in only one year, 2010, did growth, at 7.5% per annum, approach Chinese levels. Brazil then stood, of course, on the verge of the most prolonged recession since the interwar period. A central question that

12    Edmund Amann and Carlos Azzoni therefore must be addressed concerns the roots of this disappointing—​and volatile—​ growth performance over the long term. One obvious root cause of the historically high volatility of economic growth in Brazil lies in the country’s continued high reliance on NRB product exports. Given the fact that global commodity markets remain characterized by rapid and accentuated fluctuations in price and demand, it should not be surprising that Brazil’s recent growth record has been marked by such volatility. Chapter 30 suggests that patterns of trade and market liberalization pursued in Brazil over the past three decades have effectively increased that country’s comparative reliance on NRB product exports, therefore increasing the economy’s predisposition to growth volatility. Against this background, one pressing question emerges: Why did Brazil fail to diversify its export base given the implicit dangers of commodity overspecialization? An important part of the answer to this question must lie in problems of competitiveness, which have continued to afflict industry and other potential sources of “nontraditional” exports (Bonelli & Pinheiro 2016). The issue here is encapsulated in the productivity question. As Chapter 12 (by Paulo César Morceiro) and Chapter 16 (by Jorge Arbache) indicate, productivity growth in industry and services has struggled to accelerate, while in agriculture (see Chapters 13 and 15) the expansion of productivity has been impressive. Indeed, over the past 30 years, agricultural productivity growth in Brazil has comfortably outstripped that of both the United States and China. Part of the reason for Brazil’s superior performance in agriculture centers on its undeniable natural comparative advantages, whether in terms of land area or climatic conditions. However, as Chapter 15 points out, this is only part of the story: there has also been a program of aggressive investment in agricultural technology over a long period, which has stimulated productivity growth and has cemented Brazil’s position as a leading world producer in such products as soya, beef, and chicken. In industry, by comparison, modernization, technological upgrading, and productivity-​ boosting efforts have been more fragmented (Amann & Figueiredo 2012). While productivity growth and export success have been experienced in some product categories (e.g., aerospace), such achievements have proved to be the exception, rather than the rule. Therefore, as Chapter 16 suggests, raising productivity and competitiveness outside the NRB sector remains a critical policy issue—​and one that will need to be addressed if Brazil is to enjoy a more sustainable growth trajectory in the future. The sense that the lead sectors in the Brazilian economy have become centered on NRB activities is further reinforced when one examines the phenomenon of the home-​ grown multinational corporation (MNC). As Edmund Amann argues in Chapter 32, in a broad sense Brazil’s natural comparative advantages are reflected in the sectoral composition of its own multinationals. This raises questions surrounding the extent to which it is likely that fresh homegrown MNCs can emerge in the future, offering Brazil new and growth-​accelerating linkages with the global economy. Viewed through the lens of desenvolvimento, the pursuit of economic growth is only a means to a series of ends of which the most important must surely be poverty alleviation and social inclusion. As noted in Chapter 22 (by Rodolfo Hoffmann) and Chapter 24

Introduction   13 (by Armando Barrientos), Brazil has made considerable progress over the past quarter-​ century in reducing the incidence of absolute poverty and evening out the country’s notoriously skewed distribution of income. Yet the progress made thus far remains vulnerable to reversal; indeed, between 2014 and 2017 poverty indicators deteriorated in the face of a steep and prolonged recession. These recent developments highlight the fact that, perhaps contrary to common perception, the Bolsa Familía has not inoculated Brazil from poverty for once and all. Chapters 22 and 24 analyze the complex dynamics that have underpinned the progress which Brazil has made regarding poverty and inequality over the longer term. However, this analysis lays bare continuing vulnerabilities that will require sustained growth and appropriate allocations of public resources to properly address. In the light of recent developments, the maintenance of these conditions is clearly not assured. Perhaps the most important single factor needed to break intergenerational cycles of poverty and inequality centers on sustained improvements in human capital endowments. Yet, disturbingly, as Claudio de Moura Castro points out in Chapter 23, Brazil’s efforts in this direction have proven inadequate, certainly when viewed in international comparative terms. Underinvestment and underachievement in education among the poorer income groups not only render these groups less able to escape from poverty, but also clearly stand as an obstacle in the path of efforts to ramp up productivity growth. Thus, looking to the future, improvements in educational investment and attainment will be absolutely critical in promoting sustained, inclusive growth. Important though these considerations are, they should not obscure yet another long-​ standing feature that will require attention if Brazil is to realize its full potential: regional inequality. Chapter 20 (by Carlos Azzoni and Eduardo Haddad) and Chapter 21 (by Alexandre Rands) indicate the extent to which regional disparities in terms of wealth, concentrations of economic activity, and educational achievement remain vital issues in Brazil. The North and Northeast of the country remain much poorer than the South and Southeast, despite decades of regional policy and, until recently, a commodity boom that favored staple activities in Brazil’s economic periphery. As in other areas of Brazil’s economic reality, it is education that stands out as a prime candidate for investment and improvement if the regional question is to be successfully addressed. Intimately connected to both the growth and regional questions lie the issues of the environment and infrastructure. Regarding the former, there is no doubt that Brazil’s environmental track record has improved in recent years with slowing rates of deforestation and a strong institutional commitment to the global climate change agenda. Yet Ariaster B. Chimeli in Chapter 27 alerts us to the fact that much is left to be accomplished if Brazil’s economic growth is ever to be truly characterized as environmentally sustainable. Turning to infrastructure, as Chapter 18 (by Edmund Amann, Werner Baer, Juan Villa, and Thomas Trebat) demonstrates, lack of investment here—​ especially in the field of transportation—​has remained a consistent feature of the Brazilian economy for years. Not only has this inhibited the expansion of new export sectors and the growth of output more generally, it has also served to inhibit reliable and

14    Edmund Amann and Carlos Azzoni cost-​effective interregional linkages, whether in terms of transportation, power transmission, or telecommunications. Tackling the infrastructure deficit will remain one of the key challenges facing future administrations. However, in confronting this and other issues, policymakers will have to contend with a radically changed institutional and political landscape. As indicated in the previous section, the very nature of the business-​government relationship that has underpinned Brazilian development, from the ISI era to the present, is now the subject of unprecedented scrutiny and critique. The emergence of the Lava Jato scandal and the response by federal prosecutors have resulted in a wave of indictments and convictions across the business and political elites. By mid-​2017, the president himself had been formally accused, raising the possibility of yet another impeachment trial. Chapter 35 (by Mariana Prado and Lindsey Carson) analyzes the evolution of corruption in Brazil over time, the attempts made to contain it, and its potentially harmful economic impacts. What becomes clear is just how systemically corrupt practices have become woven into the fabric of business-​government relations. Looking ahead, it is clear that any resurgence of sustainable growth will require a new set of rules and greater transparency. This will imply, in all likelihood, a more arms-​length, formalized relationship between the state and the private sector. In other words, the corporatism that has continued to characterize Brazilian capitalism will need to be discarded or radically reconfigured.

References Alston, L., M. Melo, B. Mueller, and C. Pereira. 2016. Brazil in Transition: Beliefs, Leadership and Institutional Change. Princeton NJ: Princeton University Press Amann, E. 2000. Economic Liberalisation and Industrial Performance in Brazil. Oxford: Oxford University Press. Amann, E., and A. Barrientos. 2016. “Is There a Brazilian Development Model?” Quarterly Review of Economics and Finance 62 (November): 7–​11. Amann, E., and P. Figueiredo. 2012. “Brazil.” In E. Amann and J. Cantwell (eds.), Innovative Enterprises from Emerging Market Economies, 249–​299. Oxford: Oxford University Press. Baer, W. 2014. The Brazilian Economy: Growth and Development. Boulder, CO: Lynne Rienner. Bonelli, R., and A. C. Pinheiro. 2016. “Auge e declínio da indústria no Brasil.” In R. Bonelli and F. Veloso (eds.), A crise de crescimento do Brasil, 193–​221. Rio de Janeiro: Elsevier. Ioris, R. 2014. Transforming Brazil:  A History of National Development in the Postwar Era. New York: Routledge. Prebisch, R. 1950. “The Economic Development of Latin America and Its Principal Problems.” Economic Bulletin for Latin America 7: 1–​12. Singer, H. 1950. “The Distribution of Gains between Investing and Borrowing Countries.” American Economic Review, Papers and Proceedings 40: 473–​485.

Pa rt  I

H I STOR IC A L P E R SP E C T I V E S

Chapter 2

T he C ol onial E c onomy Flávio Rabelo Versiani

2.1. Introduction The colonial period—​from the arrival of the Portuguese in 1500 to independence from Portugal in 1822—​encompasses roughly two-​thirds of Brazilian economic history.1 It could not fail to influence the further development of the country’s economy. A basic question, which has been analyzed from various perspectives, poses itself naturally: Why has the path of growth of the Brazilian economy—​and of Latin American countries, in general—​diverged so markedly from that of other countries colonized approximately in the same period, in particular the United States and Canada? The beginning of that divergence is usually placed in the second half of the eighteenth century, and is associated with the onset of rapid productivity growth brought about by the industrial revolution. According to Angus Maddison’s estimates, the Brazilian per capita gross domestic product (GDP) would have been 106% of that of the present-​day United States in 1700; this percentage decreased to 51% in 1820. This suggests that the eighteenth century should be a focal point of attention in examining the evolution of the Brazilian economy.2 The emergence of industrialization in Western Europe in that period originated the pattern of international trade typical of the nineteenth century, characterized, schematically, by exports of industrial goods by some countries, especially in Western Europe and North America, and exports of raw materials and food products by less-​developed economies, especially in Latin America and Asia. A branch of literature, pioneered by Raúl Prebisch (1949) and influential in Latin America in the second half of the twentieth century, emphasized the consequences of such patterns of trade for the future development of countries exporting primary goods. The specialization of resource-​rich Latin American economies in primary production, even though advantageous in the short run—​as demonstrated by the theory of comparative advantage—​would have been detrimental to the long-​run development of those economies, barring them from full access

18   Flávio Rabelo Versiani to the productivity gains associated with industrial activities (on this literature and its extensions, see Baer 1962; Love 1994; Palma 2008). Following that approach, authors as Furtado (1968 [1959]) sought to investigate the reasons that Brazil had not followed the path of industrialization in the eighteenth century, taking advantage of the windfall of gold and diamond production in that period. Furtado’s answer stressed the lack of technical knowledge on the part of the Portuguese, resulting from the nonexistence of an industrial tradition in Portugal. A series of treaties with England, from the seventeenth century, had opened the Portuguese market to English manufacturers. The so-​called Methuen Treaty of 1704 consolidated the wine-​ versus-​textiles pattern of trade between the two countries, preventing Portugal from access to industrialization. Prado (1971 [1942]) adopted a more general line of interpretation, which has gained increased favor more recently. Prado pointed out that the overall orientation of Portuguese colonization in Brazil induced policies and institutions adverse to the development, in the colony, of productive activities other than those profitable to the colonizing country—​that is, the production of goods exportable to Europe. The influence of these policies and institutions was strongly imprinted on all aspects of Brazilian life, not only during the colonial period, but also to some extent after the country’s independence from Portugal. The colonial heritage had significant effects on independent Brazil’s economy and social structures, as well as on its institutions; it was a pervasive conditioning factor. Prado’s arguments have a modern touch, as they have points in common with the approach more recently adopted by economists and economic historians when examining the question of the income gap between nations. A common factor in those analyses is the central importance attributed to institutions, in line with Douglass North’s pioneering work (North 1981, 1990). From this point of view, institutional patterns, favorable or unfavorable to economic development, would be the main factor explaining the divergent paths followed by the economies of the Americas (Coatsworth 1993; Engerman and Sokoloff 1997, 2012; García 1993).3 Differences in institutions have been associated, in this literature, with the different conditions under which colonization took place. Some authors see institutional patterns as “inherited” from the mother country. Accordingly, British norms and practices, considered more conducive to economic growth than those prevailing in the Iberian countries, would have been a decisive factor of superiority of the British colonies in the present-​day United States, as compared to those in Latin America. As put by North, “[t]‌he evolution of North America and of Latin America differed radically right from the beginning, reflecting the imposition of the institutional patterns from the mother country upon the colonies and the radically divergent ideological constructs that shape the perception of the actors” (North 1990, 102). A different approach sees institutions as evolving from the type of productive activity predominant in the initial period of colonization, conditioned, to a large extent, by the natural endowment of productive factors. Where production for exportation—​such as sugarcane plantations—​prevailed, institutions less favorable toward economic growth

The Colonial Economy    19 would develop, especially with regard to property rights and the rule of law; the opposite would be true in regions where colonization was based on small, family-​size farms (Engerman and Sokoloff 1997, 2012; see also econometric studies such as Rodrik et al. 2004). Engerman and Sokoloff (2012), in particular, have argued that the divergent paths of development followed by the economies of North, Central, and South America were essentially defined by their initial factor endowments and climatic conditions. The natural environment determined the main type of productive activity in the process of colonization, and, in turn, influenced the institutional pattern that evolved in each case. In regions where natural endowments favored the development of large-​scale agricultural production for exportation, as in Brazil, land distribution was very unequal, and slave labor prevailed. Those conditions, in turn, would have caused large inequalities in wealth and in political power, leading to an institutional framework unfavorable to economic development. In what follows, we examine some institutional aspects of the Brazilian colonial period that seem particularly relevant in explaining later developments. It would appear that both inherited Portuguese institutions and those associated with geography and climate had a bearing on those developments. After an overview of the colonial period (section 2.2), the following sections deal with the relationship between government and the private sector (section 2.3); slavery (section 2.4); and land distribution (section 2.5). The final section presents some concluding remarks.

2.2.  The Colonial Period: An Overview In the first decades after their arrival in 1500, the Portuguese, more interested in the profitable Eastern trade, were mainly concerned with retaining possession of their newfound lands.4 Attempted occupation by other European powers—​especially the French, in this initial period—​was their main worry. From the mid-​sixteenth century, the favorable prospects of sugar production led to a gradual increase in European settlers (almost exclusively Portuguese). It is supposed that their numbers grew from less than 5,000 in 1550, to around 30,000 in 1600 (out of a total population of around 100,000), and 100,000 in 1700 (out of a total of around 300,000). These totals include subjugated Indians and African slaves, the latter in increasing proportion. Portuguese settlers and slaves were mainly concentrated in the northeastern sugar-​producing region (Pernambuco and Bahia captaincies) in the sixteenth and seventeenth centuries. Converted Jews (cristãos-​novos) were a significant proportion of the settlers in that period.5 In the eighteenth century the population increased markedly, as the gold rush provoked large-​scale migration from Portugal, as well as increased slave trade from Africa. This caused an upsurge in the population in the central-​southern part of the

20   Flávio Rabelo Versiani country, the main locus of mining activity. By the end of the colonial period, the population reached around 3.6 million, 30% of whom were African slaves, and was more or less evenly balanced between the Northeast and the Central-​South. Settlement in the Amazonian basin was still very limited. To administer the colony, Portugal attempted first to divide the territory into 14 capitanias (captaincies), granted to individuals who had distinguished themselves in the Eastern colonies, or had other kinds of court connection. The grantees were supposed to explore the captaincies with their own resources, or to raise capital under their own initiative. Only two captaincies had any economic success, mostly due to sugarcane cultivation: Pernambuco, in the Northeast, and São Vicente, in the South. All were eventually returned to the Crown. In 1550, a central government was established in Bahia, transferred, in the eighteenth century, to Rio de Janeiro. However, the captaincy governors remained the most influential representatives of the Portuguese authority throughout the colonial period. In relation to economic activities, the colonial period can be roughly divided into four parts, reflecting the main productive activities at different times. From 1500 to the 1550s, the exploration of brazilwood predominated. The decades from the 1560s to the 1690s were the era of sugar; during most of this period, Brazil was the world’s largest producer and exporter. Production expanded rapidly up to the initial decades of the seventeenth century, and at a slower pace after that. Available estimates show the number of sugar mills increasing from 60 in 1570, to 346 in 1629, to 582 in 1710. The majority of those mills were in Pernambuco, although Bahia and, later, Rio de Janeiro were also centers of sugarcane cultivation and processing. In the second half of the seventeenth century, increased production of sugar in the French and British colonies in the Caribbean, favored by colonial preference and proximity to Europe, caused decreased demand for the Brazilian product (Furtado 1968 [1959], Chapters 4–​6; Mauro 1997, 257; Schwartz 1984, 423). A third period, from the last years of the seventeenth century to the 1770s, was the gold cycle, the first significant strikes having occurred in the 1690s. Gold extraction reached its peak from the mid-​1730s to the mid-​1760s (close to 14 tons per year, in those years), declining afterward; by the end of the century, it was less than one-​third of that peak. The total amount of gold extracted in the eighteenth century was close to 900 metric tons, according to Noya Pinto’s estimate (Pinto 1979, 114).6 Furtado (1968 [1959]) argues that consumption and investment expenses in the sugar economy were mostly directed to imports owing to the highly unequal income distribution, which caused a high propensity to consume imported articles; investment expenses, mostly sugar mill equipment and slaves, also led to expenditures abroad. Consequently, there was no significant market for internally produced goods—​not even for agricultural foodstuffs, which were mainly cultivated within the confines of the sugar farms. On the contrary, the gold cycle brought about a substantial internal market. Gold extraction did not require large outlays; it was, in a way, a democratic activity, open to many individuals—​although, of course, not all of them were successful miners. Accordingly, income distribution was much less unequal. Consumption from

The Colonial Economy    21 the mining region fostered cattle raising and the cultivation of food staples in various areas to the north, west, and south of that region. The gold boom contributed significantly to economic integration of the country. The question remains as to why, in spite of such stimulus of demand, an industrial sector did not develop in Brazil at the time. This crucial point will be dealt with in the next section. After the decline of gold extraction, the fourth and final phase of the colonial period benefited from interrupted supplies of agricultural commodities from North and Central America, caused by the American wars of independence, the Napoleonic wars, and the Haitian revolution of 1791–​1804. On the demand side, the growing need of the English textile industry for cotton was also a factor in what has been called an agricultural “revival” in this period. The years from the 1770s to independence had previously been considered a period of economic stagnation (notably by Furtado 1968 [1959], Chapter 16). But more recent data for exports of agricultural goods in this period (sugar, rice, cotton, tobacco) have revealed a different picture (Alden 1984). The large number of slaves brought from Africa in these years also indicates increasing economic activity (see section 2.4).

2.3.  Colonial Government and the Private Sector Portuguese colonial administration had elements in common with the way that Spain administered its American colonies:  “a centralized form of government with power concentrated in the Crown [. . .] the institutions in the overseas colonies [being] denied any effective political voice” (Bernecker and Tobler 1993, 2). However, such similarity is more evident in the period after 1700. Before that, Portuguese control over private agents and institutions in Brazil was much less strict. In fact, in the first two centuries of colonization, private initiative had a central role. Quoting Leroy-​Beaulieu’s study on colonization, Gilberto Freyre writes that, in this initial period, there was a “complete absence of a regular and complicated system of administration” in Brazil; the colony was “not much governed” (peu gouverné, in the French economist’s words) (Freyre 1986 [1933], 26). Capital for building engenhos (sugar mills), which required considerable outlay, came mostly from European sources. The Portuguese Crown built a few mills in the beginning; in general, however, it mostly offered land grants and tax incentives in order to attract investors. Some captaincy grantees built engenhos, as in Pernambuco:  “Portuguese and foreigners with access to European credit were also among the first mill owners” (Schwartz 1985, 23). Transactions among private mill owners were frequent; in Bahia, “mills were actively bought and sold,” and the same was true in Pernambuco (Cabral de Mello 1997, Chapter 10).

22   Flávio Rabelo Versiani In Pernambuco, funds from converted Jews (cristãos-​novos), seeking to escape the Inquisition and attracted to a lucrative business, seem to have been important in financing the sugar business in the sixteenth and early seventeenth centuries. “The export sugar trade of Pernambuco was mainly in the hands of cristãos-​novos, in connection with Jews in Amsterdam and Hamburg. Many of the cristãos-​novos also owned sugar mills [. . .].” Cristãos-​novos were also present in the Bahia sugar business (Gonsalves de Melo 1996, 26, my translation; for Bahia, see Pinho 1982). In the southern captaincy of São Vicente, individual initiative was also prevalent. The numerous expeditions from São Vicente toward the interior of the colony, especially in the seventeenth century, with the purpose of capturing Indians as slaves, were carried out largely on the initiative of private individuals, without the prior approval of colonial authorities. In contrast to sugar production, which was restricted to those who could afford the initial investment in a mill and slaves, such expeditions—​bandeiras—​were democratically “accessible to all men, without distinction of class, function or profession.” These expeditions were instrumental in expanding the territory of the colony far beyond the boundary dividing Portuguese and Spanish possessions in the New World, established in the Treaty of Tordesillas of 1494 (Ellis 2003, 308, my translation). The colonial authorities, especially at the local level, were strongly under the influence of private interests. “[T]‌he colonial elites sought and found ways to make royal and municipal government responsive to their interests and goals” (Schwartz 1984, 498). The municipal councils (the Senados das Câmaras) were usually in the hands of locally dominant groups. As Freyre emphasizes, those councils “limited the power of the kings” (Freyre 1986 [1933], 27). In 1549, the Portuguese Crown decided to establish in Bahia a general government of the colony, aimed at centralizing administration and taking back some of the powers previously granted to the captaincies. However, this move toward centralization had a limited effect in the sixteenth and seventeenth centuries. An example is the frustrated attempt of the first governor general, Tomé de Sousa, to impose his authority on the Pernambuco captaincy, where sugar production was starting to gain importance. He met with tenacious resistance from the headstrong donatário, Duarte Coelho, who complained bitterly to King João III: his powers and privileges should not be curtailed, but rather increased. The king complied, and Pernambuco was excluded from the governor general’s jurisdiction, to the latter’s great resentment. The following decades saw a tug of war between the successors of Sousa and Coelho; by the beginning of the seventeenth century, there could be no doubt that the Coelhos had been victorious (Holanda 2003, 140–​141). During that century, “the Pernambucan governors [. . .] continued to free the captaincy from the power of centralized government in Bahia” (Dutra 1973, 60). Meanwhile in the other successful captaincy, the São Vicente region, the governor general’s authority would not prevail. The Jesuit missionaries wished to establish a village in the inland plateau, to facilitate their contact with the Indians, whose religious instruction they were intent on pursuing. Sousa refused them permission; he wanted to restrict colonization to the coastal area, in order to avoid diversion of productive resources away from sugar production. Nevertheless, the royal decision apparently

The Colonial Economy    23 favored the Jesuits, who were allowed to found, in 1554, what would become the present city of São Paulo (Holanda 2003, 144–​146). In fact, the governors general (who at times received the pompous title of viceroy) had their authority limited to the captaincy where they lived (Bahia, up to 1763, and then Rio de Janeiro); their privileges over the captaincy governors were little more than decorative (Holanda 2003, 154; Silva 1984, 482).7 In short, in the first two centuries of colonization, as long as the Crown received its taxes, “the planter class were [sic] given free rein.” This opened the way for an enlarged influence of powerful families and caudilhos (local chieftains), and eventually to conflicts among them, especially in the interior of the country. The seventeenth century, in particular, witnessed various struggles involving local caudilhos (Schwartz 1984, 499).8

2.3.1. Transformations in the Eighteenth Century This situation changed drastically in the beginning of the eighteenth century:  the Portuguese state became much more active and interventionist in the colony. Some writers have described this transformation in dramatic terms:  “The colonial power [.  .  .] suddenly changes its attitude [towards rural rulers] in the obvious purpose of dominating and triturating them”; “The steamroller of royal administrative centralization was sharply felt, in the eighteenth century” (Oliveira Viana 1973 [1920], 191; Avellar 1983, 52, my translations). The cause was the discovery of gold, which brought about the need to control its extraction, fight smuggling and, above all, to collect taxes. Significant gold mines were found in the last years of the seventeenth century; the discovery of diamonds, in 1730, further impelled the government to adopt strict controls. Those controls could occasionally take the form of despotic and arbitrary measures. Examples include the expulsion of all goldsmiths from mining regions, or a similar measure directed to all free blacks and mulattos; the imprisonment of friars present in those regions, with no definite function (they were the main suspects of promoting gold smuggling); a prohibition of sugar mills being built in mining areas; varied restrictions on commerce with other regions; and so forth. Many of these measures were ineffective, and some were revoked; but most were clearly disruptive of private economic activity (Holanda 1985, 277–​279, 290; Russel-​Wood 1984, 566–​567, 573–​574, 578;). A sign of augmented Portuguese authority was the creation of new royal captaincies and many vilas (municipalities) in the mining region in the first decades of the eighteenth century. New comarcas (judicial districts) were also established, and the presence of military forces was secured. Two companies of tropa de linha (professionally trained regular troops) were stationed in Minas Gerais in 1729 and 1739. Milícias (local auxiliary troops) multiplied throughout the mining areas (Russel-​Wood 1984, 560–​565). The weight of colonial authority increased further in the second half of the eighteenth century, during the reign of Dom José I (1750–​1777). In this period, the dominant figure

24   Flávio Rabelo Versiani in the Portuguese government was Secretary of State Sebastião de Carvalho e Melo, Marquis of Pombal. Sometimes described as an enlightened despot, Pombal tried to increase the gains that Portugal could extract from Brazil, by far its most important colony at this time. This meant additional strengthening and centralization of the Crown’s authority, with no concessions to local autonomy. The powers of town councils (Senados das Câmaras) were even further curtailed (Silva 1984, 479–​482). As mentioned earlier, gold extraction fell rapidly after the peak reached in the 1760s. The colonial authorities attributed such a decline to an increase in smuggling; accordingly, stricter procedures were adopted, in an attempt to guarantee the desired level of tax revenue. This was one of the causes of a frustrated insurrection in the captaincy of Minas Gerais, the main mining region, in the late 1780s. During this period, Portuguese authorities viewed with alarm the spread of small workshops producing various kinds of textiles, especially in Minas Gerais. This was considered a dangerous diversion of resources away from activities profitable to the Crown—​in particular, gold extraction. The development of a local source of income could also open the way to economic and political independence, as Antônio de Noronha, governor of Minas Gerais (1775–​1780) warned at the time with some apprehension. Martinho de Mello e Castro, the influential Portuguese secretary of state in charge of the colonies (1770–​1795), concurred with Noronha in his instructions to a new Minas Gerais governor in 1788: the installation of manufacturers in Brazil would bring forth “the gravest harm and most pernicious consequences” (Mello e Castro 1844 [1788], 18, my translation). Besides, the fall in gold production had caused a decrease in imports from England, as the persistent deficit in the Portuguese balance of trade with that country had previously been compensated by remittances of Brazilian gold. The fall in imports from England made possible an increase in textile production in Portugal; the Brazilian workshops were thus undesirable competitors. In view of all those threats to the colonizing country, an alvará (royal decree) of 1785 prohibited fabrication of all types of textiles in the colony, excepting only coarse cotton goods for slave clothing or for packaging (Silva 1984, 494 ff). As Prado noted, formal prohibitions, such as the 1785 decree, were only one aspect of a larger system of practices and institutions that guided the colonization of Brazil. That system fixed “the narrow horizons permitted to a colony established to supply a few agricultural commodities” (Prado 1971 [1942], 264). As noted earlier, those horizons had become much narrower in the eighteenth century, severely restricting the scope for private enterprise. Under such circumstances, the development of an industrial sector based on private investment would be a practical impossibility.9

2.3.2. Characteristics of Portuguese Colonial Administration The increasingly centralized and coercive nature of the colonial policy in the eighteenth century should be considered in light of the peculiarities of the Portuguese system of

The Colonial Economy    25 public administration. The legislation that would establish the guiding principles of colonial administration in Brazil was very confusing, and sometimes contradictory. It is worthwhile to quote Prado (1971 [1942], 349–​350) at some length in that respect: If we study the colony’s administrative legislation, we find a mass of enactments, subject to continual modifications often of a contradictory nature, which appeared to be entirely unconnected and pile up with no guiding plan whatsoever. [. . .] The administrative organs and functions mentioned in one place disappear in another or appear under different names and in different forms; persons to whom authority had been delegated often received their instructions in the form of letters which became law, and which frequently established new rules or a different allocation of functions and competence from those previously in force. When a new administrative organ or function was created, the law made no attempt to harmonize it with existent bodies or functions; minute and often contradictory instructions were issued for the matter at hand, only the immediate needs being provided for.

During the period when the marquis of Pombal served as secretary of state, attempts were made to rationalize the institutional framework of colonial policy toward Brazil, but with little success. As Silva notes, “the intricate web of authority and overlapping functions so characteristic of the old regime did not disappear” (1984, 481). Part of the explanation for this “chaotic jumble,” in Prado’s words, was that the legislation in force in the colony was in general merely copied from Portuguese laws and codes, with no attempt to adapt them to an entirely different context. Such chaos had at least two noticeable consequences. The first was a widespread tendency for legislation to be applied differently in different circumstances, to be enacted repeatedly, or simply to be disobeyed. For instance, a law of 1681 prohibiting the construction of a sugar mill too close to another (an attempt to restrict supply, in a period of falling prices) was sometimes ignored or circumvented; although formally in effect until the nineteenth century, the law was considered to be dead in 1800, only to be reinstated in 1802 by a royal provision. Legislation limiting the placing, by creditors, of liens on sugar mills was enacted at least six times, from 1636 to 1700 (Schwartz 1985, 195–​197). A second, and more important, consequence was that, as written norms were often confusing and difficult to interpret, local authorities had considerable leeway when taking decisions in particular cases. According to Prado (1971 [1942], 352), the duties of a captaincy governor, for instance, “[. . .] were never clearly defined, and his authority and jurisdiction always varied widely from captaincy to captaincy and from governor to governor. They varied, above all, in accordance with the personalities and capabilities of the men appointed to the office.” In the eighteenth century, regulations and restrictions applied almost exclusively to mining, and to persons related to that activity. But the great majority of the population was in other trades; in relation to those, local authorities were largely autonomous (Holanda 1985, 295).

26   Flávio Rabelo Versiani Such autonomy was enhanced by the frequent occurrence of local conflicts in the mining area. The increased number of authorities in the region—​whose functions and jurisdictions were, as mentioned, rather ill defined—​had to react quickly. Necessarily, they would often make decisions at their own discretion (Russel-​Wood 1984, 570–​571). The autonomy of local authorities, the fact that their decisions were often contingent on their personalities and inclinations—​all this fostered a situation clearly detrimental to the “rule of law.” Under such circumstances, it would undoubtedly be advisable for individuals affected by such decisions to be on good personal terms with the authorities. Thus, personal relations and contacts would be of primary importance; laws and rules would not be uniformly applied to all. This brings to mind an argument developed in Buarque de Holanda’s influential study on Brazilian historical evolution (Holanda 1989 [1936]). Holanda’s analysis puts great emphasis on the importance of personal ties in Brazilian society. He argued that an “ethic rooted in emotion” (109) prevails in Brazilian culture, leading to a “system of relationships built essentially on direct, person-​to-​person connections” (97). This would hamper a “rigid application [. . .] of any legal prescripts” (113), and would cause reluctance to “accept a supra-​individual principle of organization” (105, translations mine).10 This sort of behavior could lead, on the part of the authorities, to some difficulty in understanding “the fundamental distinction between the private and the public domains” (Holanda 1989 [1936], 105–​106). This notion relates to Max Weber’s concept of patrimonialism; a public official who confuses private and public spheres is typical of the “patrimonial state,” which is antithetical to the “bureaucratic state” characterized by impersonal regulations and juridical guaranties afforded to all citizens (Weber 1968, esp. Chapter 12). To some authors, notably Faoro (1958, 1975), patrimonialism—​essentially, confusion between private and public spheres, leading to the use of state institutions for private benefit—​has been a constant factor in Brazilian history, dominating the relations between the government and the private sector from colonial times to the present. This is not the place to discuss Holanda’s or Faoro’s arguments. But it may be argued that the need to develop personal ties with the authorities, in colonial times, especially in the eighteenth century—​a form of patrimonialism—​could have influenced, or reinforced, the type of behavior described by those authors.11 The relevant point to be stressed here, however, is that such a lack of general, impersonal rules and procedures certainly had a strong dampening effect on private initiatives in the eighteenth century. As Weber writes, “the patrimonial state lacks the political and procedural predictability, indispensable for capitalist development, which is provided by the rational rules of modern bureaucratic administration” (Weber 1968, 1095; emphasis in the original). Even in the absence of the effective opposition of the Portuguese authorities, therefore, industrial investment would have been a doubtful possibility at that time.

The Colonial Economy    27

2.3.3. Taxation The methods adopted to collect taxes in the colonial period were an additional source of uncertainty and a disincentive to producers. In the gold cycle, the main tax was the quinto, the fifth part of the gold extracted, which was property of the Crown. The form of collecting the quinto varied greatly, not only in different periods, but also from one region to another. At least a dozen methods were tried. Gold was extracted form alluvial deposits, using very simple techniques; getting a good strike was a very uncertain proposition. The frequent changes in the form of taxation added a further element of uncertainty. Capitation (commonly a head tax on slaves, but sometimes falling on all shops and businesses of the region) was often adopted, in the hope of minimizing losses through contraband. It was a particularly unfair form of taxation, since it had little relation to the income derived from gold extraction; disincentive effects would be considerable (Russell-​Wood 1984, 584ff.). The most general tax was the tithe, charged on all types of production. Collection was generally farmed to contractors, the dizimeiros. This involved additional distortions. Collection was often made at intervals of three years, and was evaluated at the prices of the third year, possibly higher than the three-​year average, thus increasing the value of the tax to be paid. The producer had to pay in money, which could be a serious problem if his production had not yet been sold, or if it sold at lower prices than those used in the evaluation. Should the producer be unable to pay, his property could be confiscated. Traveling through the province of Goiás, at the end of the colonial period, the French naturalist Saint-​Hilaire found that it was difficult to buy food provisions; fearing the dizimeiros, farmers cultivated only what their families would consume, or what they could in advance be certain of selling (Prado 1981 [1942]; Saint-​Hilaire 1975 [1847], 118). In short, the colonial system of taxation was a source of uncertainty as well as a disincentive. Gold and diamond extraction could eventually be very profitable to individual miners; however, the general environment of insecurity, uncertainty, and patrimonialism would effectively block any possibility of private investment leading to the diversification of productive activities.

2.4. Slavery Brazil was the main destination of the Atlantic slave trade. Data available in the last decade have enabled a better understanding of the time profile of the trade, especially in the case of Brazil.12 According to that data, of the estimated 10.7  million African slaves who disembarked in the Americas from the sixteenth to the nineteenth centuries, 4.9 million came to Brazil. Of those, 3.6 million arrived in the colonial period. The reasoning that relates institutional traits of the colonial period to factor endowments and climate conditions is certainly relevant in the case of slavery. When the

28   Flávio Rabelo Versiani good prospects of sugar production for export became apparent in the first years of colonization, the use of slave labor seemed a natural solution. Labor requirements of sugar production under the plantation system—​previously adopted by the Portuguese on the island of São Tomé—​were high. Attracting free laborers from Portugal, with its very small population, would be much too expensive, in comparison with the possibilities offered by enslaving Indians and by the African slave trade (in which the Portuguese were already experienced). The increasing profitability of the sugar business and the fact that Africans were more accustomed to working in agriculture made for a decrease in the utilization of enslaved Indians; by the end of the sixteenth century, African slaves prevailed. In the transatlantic slave trade, males outnumbered females, frequently in a proportion of two men for each woman. This was apparently influenced not only by preference on the part of buyers, but also by the relative price of male and female slaves in the African market. The gender imbalance caused the slave population to have a negative rate of growth; more slaves had to be imported merely to maintain the size of the labor force, which of course reinforced that imbalance. This process continued and intensified with the large demands for labor in the gold boom, and later in the agricultural revival of the final decades of the colonial period.13 Table 2.1 shows the number of slaves who disembarked in Brazil from Africa. It is noteworthy that, after the end of the gold boom, slave imports increased markedly. There are indications that the upsurge in agricultural production in those years, absorbing an intense flow of forced migration from Africa, involved not only an expansion of agricultural exports, as mentioned earlier, but also production for internal consumption. Agricultural production that had previously developed to satisfy demand from the mining regions found an alternative market in the city of Rio de Janeiro—​the main urban center of the colony—​after the end of the gold boom. Rio also became, toward the end of the colonial period, a hub of interregional commerce. Brown’s (1986) detailed study of internal trade centered in Rio de Janeiro in this period shows exchanges taking place between regions as far apart as Rio Grande do Sul, in the extreme South, and Pernambuco, in the Northeast. “Areas which were either inappropriate for export production or too far away to sustain high transport costs began to be drawn into an Table 2.1 Slaves Disembarked in Brazil in the Colonial Period No. of Slaves (Thousands)

Subperiod

Main Productive Activity

1561–1690

Sugar

1691–​1770

Gold

1,419

17,700

1771–​1821

Agricultural “revival”

1,552

30,400

Total

671

Annual Average 5,200

3,642

Source: Estimates in Voyages: The Trans-​Atlantic Slave Trade Database. http://​www. slavevoyages.org.

The Colonial Economy    29 internal network which supplied entrepôt cities and plantations” (Brown 1986, 670). Evidence in this direction is also present in works by Fragoso and Florentino (Fragoso 1992, 104–​106, 134 ff; Fragoso and Florentino 2001). It is common to find in the literature an association between Brazilian slavery and large-​scale agricultural production for export. Slave labor was certainly a central element in sugar or coffee plantations; however, evidence brought to light in recent periods has revealed that slavery played a larger role in the Brazilian economy and society. For one thing, the use of slave labor was widespread in practically all types of productive activity; “it became clear that slave labor was present in all areas of the economy, either producing for external or for internal markets” (Luna and Klein 2004, 198, my translation). People of all stations of life had slaves. An extreme example comes from a charity hospital in Rio de Janeiro, the Santa Casa de Misericórdia. The hospital received slaves for medical treatment, charging their owners for the cost. In some cases, however, the owner could not afford the payment, and left the slave in the hospital. The cost of maintaining those abandoned slaves strained the finances of the institution; it was decided, then, that slave owners classified as “indigents” would be exempt from payment for the treatment of their slaves. A provision to this effect was introduced in the hospital’s regulations (Soares 1958, 48). The idea that most slaves belonged to large holdings should also be subject to qualification. Sugar farms generally had many slaves (not unfrequently more than one hundred, especially in the earlier period), but average holdings were quite small, reflecting the large number of smallholdings. For instance, data from five localities in the gold-​mining region, in different points of the gold cycle in the eighteenth century, show that the average size of slave holdings is, at most, seven. Out of a total sample of 14,500 slaves, more than three-​quarters (76%) belonged to holdings smaller than 20, and more than one-​half (57%) to holdings of fewer than 10 slaves (Luna 1981). In the main sugar-​producing area of the province of Bahia, then the chief sugar exporter of the colony, a sample of more than 22,000 slaves in 1817 showed a similarly low average size of holdings: 10.4. Four in 10 slaves belonged to holdings smaller than 20, and one-​fourth to holdings of fewer than 10 slaves. Nearly one-​half of the sample (48%) corresponded to holdings of sugar-​mill owners, with a much larger average size (66 slaves per owner) (Schwartz 1982).14 Data still unpublished from research by the present author (in collaboration with José O. Vergolino), based on a sample of 258 inventories of descendants’ estates from all regions of the province of Pernambuco (the sugar-​production coastal zone, the cattle-​ raising and cotton-​producing backlands, and the capital city of Recife), in the period 1800–​1820, point in the same direction. Slave ownership was present in 90% of the inventories; the total number of slaves was 2,617, an average of 10 slaves per owner. The proportion of slaves in holdings of fewer than 20 (89% of the inventories) was 43%; that in holdings smaller than 10, 28%. The modal (most frequent) holding was that of just one slave. At the other extreme, 42% of the slaves belonged to only 11 individuals, each of them holding, on average, 100 slaves.

30   Flávio Rabelo Versiani Another example comes from a census of the São Paulo province in 1804. The number of slaves in the province was 43,949; the average holding was six. Fully 73% of the slaves were in holdings of 20 or less, and 50% in holdings of 10 or fewer slaves (Luna and Klein 2003, ­tables 5.2, 5.4). One important result of such wide diffusion of slave ownership from the colonial period is that most Brazilians considered slavery, until late in the nineteenth century, to be a normal, unquestioned aspect of daily life. According to Joaquim Nabuco—​himself one of the most vehement and combative opponents of slavery—​“[Brazilian] society, at all levels, had as much perception and conscience of the anomaly of slavery as it had of the motion of the Earth” (Nabuco 1975 [1897–​1899], 602). No doubt, this helps to explain why Brazil was the last Western country to abolish slavery. Abolitionism became an influential movement only in the 1880s; slavery was finally abolished in 1888. The social basis of support for the forced-​labor regime was not restricted to the Brazilian upper classes.

2.5.  Land Ownership Land ownership was highly unequal from the beginning of the colonial period; land was made available in large tracts (sesmarias), without payment, to those having the means to explore them. This system was in force until the end of the colonial period. Land concentration went hand in hand with wealth concentration and, certainly, concentration of political power in the hands of a few. In fact, it was a self-​reinforcing process. To obtain a sesmaria, the petitioner had to show that he was able to explore it; this essentially meant that he had to prove the possession of a sufficient number of slaves (or the capacity to buy them). In other words, he had to be a man of means (rarely, a woman). In addition, he had “to count on [. . .] personal relations, to make good his petition” (Holanda 1985, 296, my translation). Conversely, extensive land properties could open the way to increased wealth and increased power. In the words of Furtado, the structure of land property was no less than “a system of power” (1982, 107). Some authors have associated the pattern of land distribution in the colonial period, in countries like Brazil, with the predominant productive activity at the time. Engerman and Sokoloff (2012), in particular, have related the unequal distribution of landholdings to economies of scale in the production of certain agricultural products—​sugar, in the case of Brazil (Engerman and Sokoloff 2012, 39, 42 n.14, 331). However, a closer examination of historical facts shows that a causal relation linking sugar cultivation to large landholdings is a doubtful proposition in the Brazilian case. The typical size of sesmarias was much larger than could be justified by the technical requirements of sugar production. In fact, it would seem that economies of scale in sugarcane plantations were not large; there are many examples of cane being cultivated in relatively small plots.

The Colonial Economy    31 The supply of sugarcane to be processed in the typical sugar mill (engenho) came in large part from independent planters, the lavradores. Many engenho owners in the colonial period did not plant cane at all. In a report sent to Holland, in 1640, the Dutch councilor van der Dussen gave detailed information on the 166 engenhos existing at the time in four of the six captaincies under Dutch domination at the time (Pernambuco, Itamaracá, Paraíba, and Rio Grande do Norte). Sugarcane to be processed came mostly from lavradores (69.8% of the total). The average area of engenhos was rather small: 43.5 hectares; the area cultivated by each lavrador averaged 8.3 hectares.15 Those numbers should be viewed as approximations; Dussen himself suggests that production reported by engenho owners and lavradores should be increased by one-​third (Dussen 1981 [1640], 163–​164). However, even if the preceding areas are doubled (or tripled), they would still be quite out of proportion with the average area of sesmarias granted. Other findings point in the same direction. Schwartz finds that the average area cultivated by each lavrador in a large Bahia engenho in 1626 was 6.1 hectares. According to him, “there was no need of extensive holdings and lavrador agriculture was probably highly intensive” (Schwartz 1973, 163).16 The same author gives data for 35 Bahia engenhos in 1785, in which the average number of lavradores was 4.1. Of the total number or slaves, 40% belonged to lavradores, who possessed on average 10 slaves; the average holding of the engenho owners was 61 (Schwartz 1985, table 11-​1, 306). Lavradores were generally allowed to occupy a plot on the lands of the engenho owner, but some had their own land. The canes they harvested had to be brought swiftly to the engenho, otherwise the sugar content was reduced. The resulting sugar was divided between the engenho owner and the lavrador. The average number of lavradores per engenho, in various regions and periods, was three to five; they usually owned five to ten slaves (Cabral de Mello 1997, 430; Dussen 1981 [1640]; Schwartz 1985; Tollenare 1978 [1818]).17 Data for a later period also show sugar farms much larger than the cultivated portion. After visiting the important Engenho Salgado in Pernambuco in 1816, the French trader Tollenare verified that, out of a total area of 8,700 hectares, only about 370 hectares were effectively cultivated; woodland and pastures occupied the remaining area (Tollenare 1978 [1818], 56–​57). Considering the available evidence, it would be difficult to justify the size of land grants in the colonial period by reference to the land requirements of sugar cultivation. Noncultivated areas could be related, to some extent, to sugarcane production; pastureland was required for the numerous oxen needed on a typical sugar farm, firewood was necessary for the sugar-​mill furnaces, and so forth. However, the disproportion between the size of sugar farms and the cultivated area—​as in Engenho Salgado—​seems too large. It is necessary to look for another explanation. A careful study by Nozoe (2008) allows a close examination of sesmaria granting in the province of São Paulo from the sixteenth century to the nineteenth. This can be taken as a good sample of land distribution practices in the entire colony. It should be noted that sugarcane cultivation was important in the São Paulo area in the first period

32   Flávio Rabelo Versiani of colonization, and also during the “agricultural revival” of the late eighteenth century and the first decades of the of nineteenth. Nozoe’s sample of 1,367 documents of land concession (cartas de sesmaria) for the period 1568–​1822 shows that the average size of the sesmarias granted was 9,700 hectares (not far from the size of Engenho Salgado). Land distribution was highly unequal; the aggregate area of the 91 land grants larger than 50,000 hectares amounted to more than 4.5 million hectares, more than one-​third of the area given to all 1,367 grantees (13.3 million ha). The Gini index was 0.54, computed for sesmarias granted; but Nozoe argues that it would be much higher if computed for the total area received by individual grantees, since multiple grants to the same person (or to members of the same family) were common. As could be expected, individuals identified by titles such as dom, doutor, or desembargador (judge of the court of appeals) were granted larger than average sesmarias; the same was true for military officers and religious dignitaries. Some of the largest grants were given to men who had distinguished themselves as explorers of the interior and discoverers of gold mines. Nozoe also examines the structure of land distribution in São Paulo toward the end of the colonial period, based on a sample of 8,717 registries of rural properties from all areas of the province in 1818. He finds that the average size of landholdings (445 ha) was less than 5% of the average size of sesmarias granted, revealing a strong process of fragmentation, mainly through purchase. Having traveled through São Paulo in 1819, Saint-​Hilaire mentions the existence of about two hundred small engenhos in the region of Itu and Campinas; only in the largest of those did the number of slaves reach 20(Saint-​ Hilaire 1976 [1851], 110, 175). The “agricultural revival” apparently caused a wave of land acquisitions in São Paulo by sugar producers. The fact that engenhos were so small in this area is another indication that economies of scale were not important in the sugar business in the colonial period. The pattern of fragmentation of sesmarias in small properties (by purchase, and also by inheritance) at the same time that others were kept at the original size—​ causing an increase in the inequality of distribution—​seems to have been common. In Escada, one of the main sugar-​producing municipalities of Pernambuco, Eisenberg (1974, 130)  found in the 1850s (after independence) that half of the 84 sugar plantations had less than 995 hectares, while the 13 largest ones, with more than 3,000 hectares, accounted for fully 70% of the cultivated land. A similar picture is described by Mattoso, also in the mid-​nineteenth century, for the parish of Santiago do Iguape in the sugar area of Bahia; the average size of farms was only 313 hectares, but a single sugar estate, with 5,600 hectares, covered one-​fourth of the area of the parish. This confirms that small sugar farms were economically viable—​while a few powerful individuals maintained their prestige and power as large landholders (Kátia Mattoso, cited in Nozoe 2008, 213). On the other hand, the inequality of distribution increased markedly. The Gini index for land distribution in 1818 for the province as a whole was 0.84; in the cattle-​raising area, it was as high as 0.94. The 436 largest properties, 5% of the sample total, had no

The Colonial Economy    33 less than 71% of the total area. As Nozoe points out, a similar process of concentration occurred also in the Brazilian Northeast. No doubt, the consequences of the practice of granting large, sometimes immense sesmarias went far beyond the colonial period. But its initial purpose was unrelated to the needs of engenho owners, since, as argued earlier, the economics of sugar cultivation would not require such large tracts of land. The introduction and ample diffusion of that practice seem to have originated from a combination of three elements: the Crown’s desire to occupy its dominion; the sheer vastness of the available lands; and the inadequate transplantation to the colony of a formal system of land distribution conceived under very different circumstances. The desire to populate the colony was impelled, especially in the sixteenth century, by the perennial Portuguese fear of losing their colony to other powers, particularly the French (and in the seventeenth century, the Dutch). Considering the abundance of land, to provide a tract to anyone who wished to settle in it would seem a logical solution. This idea of granting land with the purpose of stimulating occupation of the colony was sometimes made explicit, as in a royal decree of 1590. Occupation was the main thing; whether or not the land would be cultivated was a secondary consideration (Costa Porto n.d., 66, 94). The granting of sesmarias was an old Portuguese institution, regulated since the fourteenth century. It was first established in a situation of scarcity of wheat, and, consequently, increasing prices. The authorities, verifying the existence of uncultivated land, would intervene, either forcing the owners to make their properties productive, or distributing land, as sesmarias, to those willing to cultivate it. The basic objective was to increase production; if the land was kept unutilized, it could be confiscated. The size of sesmarias was, in such a comparatively small country, not a particularly open question, and small properties were the rule. Following their time-​honored practice of adopting in Brazil the same policies and institutions of their own country, the Portuguese did not introduce any relevant change to the fourteenth-​century legislation on sesmarias. The requirement of land cultivation was maintained, but rarely was enforced; in any case, enforcement would have been difficult, given the size of the colony and the restricted human resources of the colonial administration. On the other hand, nothing was determined up to the end of the seventeenth century as to the size of sesmarias to be granted. The unsurprising result, as seen earlier, was the granting of huge stretches of land to influential persons. Extensive sesmarias, especially in good locations on the coast, were frequently fractioned to be sold in smaller tracts, or rented out. Even after a size limitation was fixed, in 1695–​1697, it was not always obeyed; larger sesmarias were still granted in the eighteenth century (Costa Porto n.d.; Nozoe 2008). We may conclude that the factor-​endowment argument, as a determinant element of land distribution, does not seem to apply in the Brazilian case. The highly unequal distribution of land in colonial Brazil was not a consequence of the decision by the Portuguese to implement sugarcane cultivation in their American possession, taking

34   Flávio Rabelo Versiani advantage of the very favorable natural conditions. The main cause was their desire to occupy—​the sooner the better—​a very extensive territory subject to the ambitions of other European powers, and the adoption to this end of an institution indiscriminately transplanted from the mother country. On the other hand, whatever its cause, the institutional framework that led to unequal land distribution certainly had adverse long-​run consequences in many ways. Inequality in wealth has a tendency to perpetuate itself, via the larger influence of the wealthier in government and in politics. Land distribution can also influence the structure of demand and, indirectly, the structure of production. Land occupation based on family farms—​as in the British North American colonies—​“will foster a more equitable distribution of incomes, which in turn fuels a demand for a broad range of goods and services [. . .], thus inducing investments in other sectors of the economy” (Garcia 1993, 73). Such stimuli would not have been present when land was handed out in vast sesmarias.

2.6.  Concluding Remarks The fact that the Brazilian economy could not benefit from the productivity gains made possible by the industrial revolution in the eighteenth century may largely be attributed, as argued earlier, to practices and institutions characteristic of the Portuguese colonial policy at the time. In this sense, it may be said that such policy was instrumental in establishing an initial divergence between the Brazilian growth path and that of the early-​industrializing countries of Western Europe and North America. In the postcolonial period, an elastic supply of land, a consequence of the colonial system of land distribution, tended to discourage initiatives to increase the productivity of land. The rapid expansion of coffee production in the second half of the nineteenth century and the first decades of the twentieth was brought about mostly by increases in cultivated area; planters had no incentive to invest in land productivity (Furtado 1968 [1959], Chapter 28). In general, the growth of agricultural production in Brazil, up to at least the mid-​twentieth century, was largely the result of area expansion, rather than increases in production per unit of area. The colonial distribution of land in large tracts also had a lasting effect on the supply of unskilled labor. As land was appropriated by a limited number of favored individuals from the early periods of colonization, the growth of the landless population was mainly absorbed in the large landholdings, especially cattle farms. Free laborers—​including ex-​ slaves—​settled in those farms under various institutional arrangements, commonly as cattle hands growing their subsistence crops in small lots. To the landholder, additional dwellers had near-​zero opportunity cost, given the vast size of the typical property; they were useful not only as a cheap source of labor but also, on occasion, as bodyguards in regions where conflicts were frequent and authorities scarce.

The Colonial Economy    35 Furtado (1982) argues that the existence of this growing mass of rural laborers, with very low productivity and income levels, exerted a long-​run downward pressure on urban wages, with important effects on the inequality of income distribution. In a broad perspective, the development of the postcolonial Brazilian economy up to the first decades of the twentieth century could be stylized in terms of the classical Lewis (1954) model. An elastic supply of labor would further disincentivize the search for productive methods that would increase general productivity. Slavery, a basic institution of the colonial period, was a dominant element in Brazilian life for longer than three and a half centuries, and left an enduring mark on various aspects of the country’s present-​day economic and social structure—​certainly, on the configuration of income distribution. Portuguese colonial policy, frequently confused and disorganized, increased the importance of personal ties in the relations between authorities and the public. This hampered an impersonal application of rules and laws, increased uncertainty, and was certainly unfavorable to private initiative. To what extent this established a pattern of behavior that persisted beyond the colonial period is an open question. To sum up, institutions of Portuguese colonization related to land, labor, and the public sector were decisive in shaping the development of the colonial economy, and influenced later developments in various ways and degrees. Some of those institutions—​ such as slavery and, in part, the system of land distribution—​were a consequence of the form of colonization adopted by the colonizing country. Nevertheless, institutions transplanted from Portugal, often with no attempt to adapt them to very different conditions, were also important in this context.

Notes 1. This chapter had the benefit of Stanley Engerman’s careful reading and detailed criticism. I  received also valuable comments and suggestions from Stuart Schwartz, Iraci Costa, Nelson Nozoe, José Vergolino, and participants in a seminar at the Universidade de São Paulo. I wish to thank all of them; of course, any defects and interpretations are my own. 2. Computed from numbers in Maddison (2001, Table  2–​22a). For explanations on how those numbers were determined, see Maddison (2001, 249–​250); Gallman (1972). 3. It may be noted that Adam Smith had already stressed the importance of institutions in the development of colonies: “political institutions of the English colonies have been more favorable to the improvement and cultivation of [. . .] land than those of [the Spaniards, Portuguese and French]” (Smith 1991 [1776], 509). 4. For more details on this topic, see Furtado (1968 [1959], Parts 1–​3). 5. Population estimates in this and the next paragraph draw on IBGE (1990, 30ff.); Marcílio (1984); Alden (1963); Simonsen (1977, 217); Quirino (1966). 6. To compare, the estimated amount of gold sent to Spain from its American colonies, in the period 1500–​1660, was about 180 tons; silver remittances were much larger, close to 17,000 tons (Hamilton 1970, 42). Considering the average bimetallic ratios at the time (Hamilton 1970, 71, 123), that amount of silver would be equivalent, in monetary terms, to something like 1,600 to 1,700 tons of gold. Those numbers indicate that Brazilian eighteenth-​century

36   Flávio Rabelo Versiani gold corresponded roughly, in value, to one-​half of the precious metals brought to Europe from the Spanish colonies in the sixteenth and seventeenth centuries. 7. The Count of Sabugosa, viceroy from 1720 to 1735, “repeatedly complained that governors of Minas Gerais failed to keep him informed of events in the mining areas and did not accord him due respect” (Russel-​Wood 1984, 569). 8. On conflicts among family groups, see Oliveira Viana (1973 [1920], Chapter 11). 9. A striking example of the restrictions imposed on the colony in the eighteenth century is the fact that a printing shop opened in Rio de Janeiro in 1747 was promptly closed by the Portuguese authorities—​lest it should produce subversive material. Only in 1808 was a printing establishment allowed to function in Brazil. In contrast, books had been published in the Spanish colonies since the sixteenth century (Holanda, 1969 [1936], 86). 10. For Holanda, the typical Brazilian would be a “ ‘cordial man’ in the etymological sense of the adjective; of the heart, heartfelt—​an individual to whom emotions are an essential attribute of behavior” (Holanda 1989 [1936] Chapter 5). 11. Analyzing the central role of the government in the Brazilian economy in recent periods and the practices of clientelism and political patronage in the relation between the government and the private sector, Lazzarini (2011, 40, 117–​120) refers to Faoro’s thesis, concurring that institutional traits inherited from the colonial period are hard to change. 12. The large database on the transatlantic slave trade (Eltis et al. 1999) has since been revised and enlarged, and is available in Voyages: The Trans-​Atlantic Slave Trade Database, http://​ www.slavevoyages.org (cited henceforth as Voyages). 13. On the economics and demography of Brazilian slavery, see Klein and Luna (2010). 14. The sample is for the Bahian parishes of Santo Amaro and São Francisco. 15. Sugarcane production was measured in tarefas; estimates of cultivated area are based on a ratio of 0.44 hectares per tarefa (Schwartz 1973, 163; see also Schwarz 1985, 113–​114). Dussen’s report has data on cane produced by 387 lavradores in 91 engenhos. The preceding proportion of cane produced by lavradores is based on twenty-​four engenhos for which production by engenho owners and by lavradores is given separately. The average area of engenhos was also computed from this smaller sample. Original data in Dussen (1983 [1640], 141–​176). 16. The author points out that average lavrador holdings were in fact larger, probably due to extensive tracts of non-​arable land, a system of fallowing, and so forth. 17. The large number of lavradores in Engenho do Conde (Schwartz 1973) was atypical.

References Alden, Dauril. 1963. “The Population of Brazil in the Late Eighteeenth Century: A Preliminary Study.” Hispanic American Historical Review 43 (2): 173–​206. Alden, Dauril. 1984. “Late Colonial Brazil, 1750–​1808.” In The Cambridge History of Latin America, Vol. 2, edited by Leslie Bethell. Cambridge: Cambridge University Press. Avellar, Hélio de A. 1983. Administração Pombalina. (História Administrativa do Brasil, Vol. 5). Brasília: Editora Universidade de Brasília. Baer, Werner. 1962. “The Economics of Prebisch and ECLA.” Economic Development and Cultural Change 10 (2): 169–​182. Brown, Larissa V. 1986. “Internal Commerce in a Colonial Economy: Rio de Janeiro and its Hinterland, 1790–​1822.” PhD dissertation, University of Virginia.

The Colonial Economy    37 Bernecker, Walter L., and Hans W. Tobler. 1993. “Introduction.” In Development and Underdevelopment in America, edited by Walter L. Bernecker and Hans W. Tobler, 9–​10. Berlin: W. de Gruyter. Cabral de Mello, Evaldo. 1997. Rubro veio: O Imaginário da restauração Pernambucana, 2nd edition. Rio de Janeiro: Topbooks. Cirne Lima, Ruy. 1990. Pequena história territorial do Brasil: Sesmarias e terras devolutas, 4th edition. São Paulo: Secretaria de Estado da Cultura. Coatsworth, John H. 1993. “Notes on the Comparative Economic History of Latin America and the United States.” In Development and Underdevelopment in America, edited by Walter L. Bernecker and Hans W. Tobler, 10–​30. Berlin: W.de Gruyter. Costa Porto. n.d. O sistema sesmarial no Brasil. Brasília: Editora Universidade de Brasília. Dussen, Adriaen van der. 1981 [1640]. “Relatório sobre o Estado das Capitanias Conquistadas no Brasil, apresentado [. . .] ao Conselho dos XIX na Câmara de Amsterdam, em 4 de abril de 1640.” In Fontes para a história do Brasil Holandês. I. A economia açucareira, edited by J. A. Gonsalves de Mello, 131–​252. Recife: Parque Histórico Nacional dos Guararapes. Dutra, Francis A. 1973. “Centralization vs. Donatarian Privilege:  Pernambuco 1602–​1630.” In Colonial Roots of Modern Brazil, edited by Dauril Alden, 19–​60. Berkeley: University of California Press. Eisenberg, Peter L. 1974. The Sugar Industry in Pernambuco: Modernization without Change, 1840–​1910. Berkeley: University of California Press. Ellis, Myriam. 2003. “As bandeiras na expansão geográfica do Brasil.” In História geral da civilização brasileira, Vol. 1, edited by Sérgio B. de Holanda, 300–​323. Rio de Janeiro: Bertrand Brasil. Eltis, David, Stephen D. Behrendt, David Richardson, and Herbert S. Klein (Eds.). 1999. The Trans-​Atlantic Slave Trade:  A Database on CD-​ROM. Cambridge:  Cambridge University Press. Engerman, Stanley L., and Kenneth L. Sokoloff. 1997. “Factor Endowments, Institutions, and Differential Paths of Growth among New World Economies: A View from Economic Historians of the United States.” In How Latin America Fell Behind; Essays on the Economic Histories of Brazil and Mexico, 1800–​1914, edited by Stephen Harber, 260–​304. Stanford, CA: Stanford Uuniversity Press. Engerman, Stanley L., and Kenneth L. Sokoloff. 2012. Economic Development in the Americas since 1500: Endowments and Institutions. Cambridge: Cambridge University Press. Faoro, Raymundo. 1975. Os donos do poder: Formação do patronatopolítico brasileiro. Rio de Janeiro: Globo. Fragoso, João. 1992. Homens de grossa aventura: Acumulação e hierarquia na praça mercantil do Rio de Janeiro (1790–​1830). Rio de Janeiro: Arquivo Nacional. Fragoso, João, and Manolo Florentino. 2001. O arcaísmo como projeto. Rio de Janeiro: Civilização Brasileira. Freyre, Gilberto. 1986 [1933]. The Masters and the Slaves:  A Study in the Development of Brazilian Civilization. Berkeley: University of California Press. Furtado, Celso. 1968 [1959]. The Economic Growth of Brazil. Berkeley: University of California Press. Furtado, Celso A. 1982. “Estrutura agrária no subdesenvolvimento brasileiro.” In Análise do modelo brasileiro by Celso Furtado, 98–​122. Rio de Janeiro: Civilização Brasileira. Gallman, Robert E. 1972. “The Pace and Pattern of American Economic Growth.” In American Economic Growth, edited by L. E. Davis et al., 15–​60. New York: Harper & Row.

38   Flávio Rabelo Versiani García, Daniel D. 1993. “Economic Growth and Stagnation in the Colonial Americas:  An Exploratory Essay.” In Development and Underdevelopment in America, edited by Walter L. Bernecker and Hans W. Tobler, 51–​87. Berlin: W. de Gruyter. Gonsalves de Melo, J. A. 1996. Gente da nação: Cristãos-​Novos e Judeus em Pernambuco, 1542–​ 1654. Recife: Massangana. Haber, Stephen (Ed.). 1997. How Latin America Fell Behind: Essays on the Economic History of Brazil and Mexico, 1800–​1914. Stanford, CA: Stanford University Press Hamilton, Earl J. 1973. American Treasure and the Price Revolution in Spain, 1501–​1650. New York: Octagon Books Holanda, Sérgio B. de. 1969 [1936]. Raízes do Brasil, 5th edition. Rio de Janeiro: José Olympio. English translation: Roots of Brazil. Translated by G. H. Summ. Notre Dame: University of Notre Dame Press, 2012. Holanda, Sérgio B.  de. 1985. “Metais e pedras preciosas.” In História geral da civilização brasileira, edited by Sérgio B. de Holanda. 259–​310. São Paulo: Difel. Holanda, Sérgio B. de. 2003. “A instituição do governo geral.” In História geral da civilização brasileira, edited by Sérgio B. de Holanda, 131–​143. Rio de Janeiro: Bertrand Brasil. IBGE—​Fundação Instituto Brasileiro de Geografia e Estatística. Estatísticas históricas do Brasil. Séries Estatísticas Retrospectivas, Vol. 3, 2nd edition. Rio de Janeiro: IBGE. Klein, Herbert S., and Francisco V. Luna. 2010. Slavery in Brazil. Cambridge:  Cambridge University Press. Lazzarini, Sérgio G. 2011. Capitalismo de Laços. Rio de Janeiro: Elsevier. Lewis, W. Arthur. 1954. “Economic Development with Unlimited Supplies of Labour.” The Manchester School 22 (2): 139–​191. Love, Joseph L. 1994. “Economic Ideas and Ideologies in Latin America since 1930.” In The Cambridge History of Latin America, Vol. 6, Part 1, edited by Leslie Bethell, 391–​460. Cambridge: Cambridge University Press. Luna, Francisco V. 1981. Minas Gerais: Escravos e senhores: Análise da estrutura populacional e econômica de alguns centros mineratórios (1718–​1804). São Paulo: Instituto de Pesquisas Econômicas. Luna, Francisco V., and Herbert S Klein. 2003. Slavery and the Economy of São Paulo, 1750–​ 1850. Stanford, CA: Stanford University Press Luna, Francisco V., and Herbert S Klein. 2004. “Economia e sociedade escravista: Minas Gerais e São Paulo em 1830.” Revista Brasileira de Estudos Populacionais, Campinas 21(2): 173–​193. Maddison, Angus. 2001. The World Economy: A Millennial Perspective. Paris: Development Centre of the Organisation for Economic Co-​operation and Development. Mauro, Frédéric. 1997. Portugal, o Brasil e o Atlântico, 1570–​1670, Vol. 1. Translated by M. Barbosa. Lisboa: Estampa. Mello e Castro, Martinho de. 1844 [1788]. “Instrução para o Visconde de Barbacena.” Revista Trimensal de Historia e Geographia 21 (April 1844). North, Douglass C. 1981. Structure and Change in Economic History. New York: W. W. Norton. North, Douglass C. 1990. Institutions, Institutional Change and Economic Performance. Cambridge: Cambridge University Press. Nozoe, Nelson. 2008. A apropriação de terras rurais na capitania de São Paulo. Livre-​Docencia dissertation, Universidade de São Paulo, Faculdade de Economia, Administração e Contabilidade, Departamento de Economia. Oliveira, Viana. 1973 [1920]. Populações meridionais do Brasil, Vol. 1. Rio de Janeiro: Paz e Terra.

The Colonial Economy    39 Palma, José G. 2008. “Dependency.” In The New Palgrave Dictionary of Economics, 2nd edition, 1335–​1340. London: Palgrave Macmillan. Pinho, Wanderley. 1982. História de um engenho do recôncavo, 2nd edition. São Paulo: Editora Nacional (Brasiliana, Vol. 374). Pinto, Virgílio Noya. 1979. O ouro brasileiro e o comércio Português, 2nd edition. São Paulo: Editora Nacional (Brasiliana, Vol. 371). Prado, Caio, Jr. 1971 [1942]. The Colonial Background of Modern Brazil. Berkeley: University of California Press. Prebisch, Raúl. 1949. The Economic Development of Latin America and Its Principal Problems. New York: United Nations Economic Commission for Latin America. Quirino, Tarcízio do R. 1966. Os habitantes do Brasil no fim do século XVI. Recife: Imprensa Universitária. Rodrik, Dani, A. Subrabanian, and F. Trebbi. 2004. “Institutions Rule:  The Primacy of Institutions over Geography and Integration in Economic Development.” Journal of Economic Growth 9: 131–​165. Russell-​Wood, A. J. R. 1984. “Colonial Brazil: The Gold Cycle, c. 1690–​1750.” In The Cambridge History of Latin America, Vol. 2, edited by Leslie Bethell, 547–​600. Cambridge: Cambridge University Press. Saint-​Hilaire, Auguste de. 1975 [1847]. Viagem à província de Goiás. Belo Horizonte: Itatiaia Saint-​Hilaire, Auguste de. 1976 [1851]. Viagem à província de São Paulo. Belo Horizonte: Itatiaia. Schwartz, Stuart B. 1973. “Free Labor in a Slave Economy:  The Lavradores de Cana of Colonial Bahia.” In Colonial Roots of Modern Brazil, edited by Dauril Alden, 147–​198. Berkeley: University of California Press. Schwartz, Stuart B. 1982. “Patterns of Slaveholding in the Americas: New Evidence from Brazil.” American Historical Review 87 (1): 55–​86. Schwartz, Stuart B. 1984. Colonial Brazil, c.  1580–​ c. 1750:  Plantations and Peripheries. In The Cambridge History of Latin America, Vol. 2, edited by Leslie Bethell, 421–​500. Cambridge: Cambridge University Press. Schwartz, Stuart B. 1985. Sugar Plantations in the Formation of Brazilian Society: Bahia, 1550–​ 1835. Cambridge: Cambridge University Press. Silva, Andrée M.-​D. 1984. “Portugal and Brazil: Imperial Re-​organization, 1750–​1808.” In The Cambridge History of Latin America, Vol. 1, edited by Leslie Bethell, 441–​468. Cambridge: Cambridge University Press. Simonsen, Roberto C. 1977. História econômica do Brasil (1500–​1820). São Paulo: Ed. Nacional. Smith, Adam. 1991 [1776]. The Wealth of Nations. London: David Campbell. Soares, Ubaldo. 1958. A escravatura na misericórdia. Editora Fundacao Romao de Matos Duarte: Rio de Janeiro. Tollenare, L. F. de. 1978 [1818]. Notas dominicais. Recife: Secretaria de Educação e Cultura. Voyages: The Trans-​Atlantic Slave Trade Database. Available at http://​www.slavevoyages.org. Weber, Max. 1968. Economy and Society. Edited by G. Ross and C. Wittich. Berkeley: University of California Press.

Chapter 3

The Ninet e e nt h and Early T we nt i et h Centuri e s André Villela

Eric Hobsbawm famously dubbed the 1789–​1914 period the “long” nineteenth century.1 With this he intended to highlight elements of continuity in the sociopolitical landscape of Europe, starting with the French Revolution and running all the way through World War I and the collapse of the “old order.” A case could be made for a similarly “long,” if slightly different, nineteenth century in Brazilian economic history. Such a period would start sometime in the early nineteenth century proper (perhaps in 1808, with the transfer of the Portuguese royal family to Brazil and the opening of the ports to friendly nations) and would extend into the early 1930s, at which point the collapse of the coffee industry brought down the First Republic and allowed for a so-​called shift in the dynamic center of the Brazilian economy (Furtado 1970, Chapter 32). Out went the leading role that the primary-​export sector had played since the early days of Portuguese settlement and in came the domestic sector—​the manufacturing industry, especially—​as the main source of economic growth.2 Within this long nineteenth century of Brazil’s economic history, one can identify two subperiods, which roughly coincide with the country’s political history, namely 1808–​ 1900 and 1900–​1930. The former covers the years in which Brazil served as the political and economic center of the Portuguese Empire (1808–​1822), the years of the Brazilian imperial period itself (1822–​1889), and the first republican decade. The latter represents most of the period corresponding to the First (or “Old”) Republic. What sets these two subperiods apart is the pace of economic growth. While for most of the 1800s the Brazilian economy certainly did grow in size, per capita gross domestic product (GDP) barely inched forward. Economic growth until the early 1900s, for the most part, was extensive and was enabled by rapid population growth in a land-​abundant and resource-​ rich setting. Productivity gains in this period were few and far between, faced with a host of obstacles. From the early twentieth century onward, however, some of these

The Nineteenth and Early Twentieth Centuries    41 obstacles were gradually removed, and economic growth picked up and became intensive, allowing for continued population expansion amid rising (if still low) per capita incomes. In a sense, then, at the beginning of the 1900s, Brazil gradually embarked on what economists call “modern” economic growth—​characterized by sustained increases in per capita GDP—​in a process that had been pioneered earlier by a few economies of the West and their so-​called offshoots.

3.1.  Brazil’s Long Nineteenth Century: An Overview The population of Brazil in 1798 has been estimated at 3.3 million.3 By the time of the first official census in 1872, the total population had reached approximately 10 million, implying an average annual growth rate of 1.5%. By the end of the imperial period in 1890, the Brazilian population had increased to 14.3 million, and in 1930 it was 34–​35 million, meaning that the average annual rate of population growth in these four decades had increased to 2.2% (Merrick and Graham 1979).4 While population figures for the pre-​census period in Brazil (that is, before 1872) are still subject to dispute, estimates of the size of the Brazilian economy in the nineteenth century carry an even greater degree of uncertainty. Several authors have attempted to provide a measure of the size of market-​based economic activity in Brazil before the twentieth century, including Contador and Haddad (1975), Goldsmith (1986), Leff (1982), and, more recently, Tombolo (2013).5 Although their estimates vary, they seem to agree that per capita GDP barely grew during the first half of the nineteenth century, and displayed a moderate rate of growth for the remainder of the imperial period.6 As shown in Table 3.1, from the early twentieth century onward—​as a result of a number of factors, to be analyzed later in this chapter—​the Brazilian economy began to display higher rates of per capita growth, on the back of the country’s engagement in the liberal international world order that would prevail until the onset of the Great Depression. To a large extent, Brazil’s role in the evolving international division of labor—​that of exporter of tropical commodities and importer of manufactured goods, capital, and labor—​determined the relative dynamism of its economy. During the period under review here (the early 1800s to 1930), coffee undoubtedly figures prominently in the story, alongside crops such as sugar, cotton, and (briefly, around the turn of the century) rubber. This export-​oriented sector comprised the more advanced part of the Brazilian economy, at a time when the manufacturing industry was still in its infancy. Throughout the imperial period and, in part, in the First Republic (1889–​1930), the behavior of this export sector (its share of GDP, growth over time, terms of trade, etc.) would have a great bearing on overall economic performance, although modern industry, from its inception at the turn of the century, took on an increasing role in providing the Brazilian economy with further sources of dynamism. However, the definitive shift in this process

42   André Villela Table 3.1 Brazil: Population and Per Capita GDP, 1820–​1930a

Year

Population (Million)

Per Capita GDP (1990 Geary-​Khamis Dollars)

Average Annual Rate of Growth (%)

1820

4.7 (1819)

683

1870

10.1 (1872)

713

1890

14.3

794

0.5

1900

17.3

672

–​1.7

1930

34–​35

1,048

1.5

0.1

a Population figures in IBGE (1990). Per capita GDP estimates from the Maddison.

Project Database, http://​www.ggdc.net/​maddison/​maddison-​project/​home. htm,2013 version (accessed December 26, 2016).

would have to wait until the Great Depression and the accompanying collapse of the coffee industry; at this point, the manufacturing industry and the domestic economy more generally took over from the export sector as the linchpin of the Brazilian economy. Throughout this long nineteenth century, agriculture was the mainstay of the Brazilian economy, although its share of GDP began to decline around 1900. As there are no statistics for agricultural output during the imperial period, one must rely on figures for exports in order to get a glimpse of the structure of the primary sector—​in particular, the one oriented toward the international market.7 As shown in more detail in Table 3.2, before 1830 sugar, cotton, and hides—​the main export items for most of the colonial period—​still prevailed, only to see coffee overtake sugar as Brazil’s top export item (a position it would hold for more than a century). Although it had been introduced into Brazil in the 1720s, it was only in the early nineteenth century, with a combination of political upheaval in Saint-​Domingue (theretofore the major world producer) and the narrowing of Cuba’s specialization toward sugar alone, that Brazil filled the void and established itself as the world’s main coffee producer and exporter.8 This capacity, in turn, stemmed from the country’s vast supply of land and labor (slaves), which drew into the chain of production and commercialization domestic and foreign capital, respectively. While coffee took over from sugar as the chief export item in Brazil, the latter, alongside cotton and tobacco, remained an important crop in the Northeastern region. To a great extent, the contrasting fortunes of these other crops—​whose output and exports grew at a much slower rate vis-​à-​vis coffee—​help explain the spatial shift in the center of the Brazilian economy that unfolded after mid-​century, from the Northeast toward the Southeast (Rio de Janeiro, São Paulo, and Minas Gerais). The emphasis that the literature places on export crops during the nineteenth century might suggest that they comprised the largest sector of the Brazilian economy in this period. But this would be misleading. Indeed—​and despite some dispute concerning the

The Nineteenth and Early Twentieth Centuries    43 Table 3.2 Commodity Distribution of Brazilian Exports, 1821–​1913 (%) Cacao

Coffee

Cotton

Hides

Rubber

Sugar

Other

Total

1821–1829

1.4

25.7

13.5

11.6

0.0

30.4

17.4

100.0

1830–1839

0.8

37.4

9.9

8.0

0.0

31.5

12.3

100.0

1840–1849

1.4

38.6

4.5

8.7

0.0

32.1

14.7

100.0

1850–1859

1.1

48.2

6.2

7.7

2.1

21.6

13.2

100.0

1860–1869

0.8

47.0

17.9

5.7

3.3

11.5

13.8

100.0

1870–1879

1.0

55.8

10.6

4.7

5.3

12.8

9.9

100.0

1880–1889

1.7

60.1

4.2

3.2

10.9

11.4

8.5

100.0

1890–1899

1.6

68.0

1.7

2.8

13.8

3.8

8.3

100.0

1900–1909

3.0

53.6

2.4

4.1

25.5

1.2

9.3

100.0

1912–​1913

2.4

61.0

2.2

5.9

19.8

0.1

8.6

100.0

Source: Online appendix in Absell and Tena-​Junguito (2016).

true share of exports in GDP (to be discussed in the next section)—​there seems to be little doubt that domestic agriculture employed the bulk of the labor force in nineteenth-​ century Brazil. The backward technology—​and, hence, low productivity—​of this type of activity, combined with its sheer size, helps explain the low rates of GDP growth mentioned earlier. This state of affairs would slowly begin to change at the dawn of the republican period, as railway expansion brought significant productivity gains to the economy, allowing for higher rates of per capita GDP growth from the turn of the century onward. Structural change would also contribute to higher rates of growth from the early twentieth century onward. In 1900 agriculture comprised approximately 45% of the Brazilian economy, with the services sector accounting for a similar share, whereas industry amounted to just over 10% of the whole. Three decades later, these shares had changed to 36%, 49%, and 15%, respectively (Bonelli 2003). The increasing share of the manufacturing sector in the economy, enhanced market integration, and the general expansion of capitalist relations in both goods and factor markets all combined to produce—​for the first time in Brazilian history—​high and sustained per capita GDP growth. Throughout the nineteenth century, slaves represented a significant—​ though declining—​share of the Brazilian population and workforce. Between 1800 and 1856 (when the final, illegal, landing took place), some two million African slaves entered Brazil, out of a total estimated 3.2 million brought to the Americas in the same period.9 Starting in mid-​century, immigration of free European labor increased, amounting to an estimated flow of some 220,000 individuals between 1851 and 1870. This was followed by a similar number of immigrants in the 1870s, half a million in 1881–​1890, and a peak of more than 1.1 million in the final decade of the nineteenth century. Between 1900 and

44   André Villela 1930, another 2.3 million immigrants would enter Brazil, mostly from Southern Europe, but also from Germany, Central Europe, Japan, and the Middle East.10 If the supply of labor to the Brazilian economy was kept perfectly “elastic” by means of both forced and free immigration, access to the two remaining factors of production would have to rely on a combination of nature and legislation in the case of land, and the willingness and/​or capacity of Brazilian and foreign entrepreneurs in the case of providing capital in order to fund investments in railways, ports, and public utilities, and to finance government debt. Prior to 1822, access to land in Brazil took the shape of either a royal grant of a tract of land of up to 50 square miles (sesmaria) or outright squatting (posse). The sesmaria system would be discontinued in 1822 with political independence from Portugal, and for the next three decades land policy would remain in legal limbo. This did not stop, however, the march of the coffee frontier northwest from the city of Rio de Janeiro along the Paraíba River valley, mostly by means of squatting on public land. In 1850—​and after decades of debate in Parliament—​the new Land Law was passed. However, its main objectives—​to provide for the legalization of land titles by way of proper demarcation and registration, and to prevent further encroachment on public land—​were largely frustrated (Bethell and Carvalho 1989).11 Nineteenth-​century Brazil was a capital-​scarce economy. Such local resources as existed were channeled to the most productive sectors, such as coffee, railways, the slave trade, sugar, and, later in the century, rubber. In the absence of even sparse data on domestic investment, one must rely on information on foreign investment to get an idea of trends in capital formation in Brazil. Foreign investment (both portfolio and direct) in 1825 amounted to just over £4.1 million; by 1895 this stock had increased to £79.6 million, roughly split between direct and private portfolio investment, on the one hand, and loans to the public sector, on the other. In 1930 the total stock stood at approximately £500 million, with public portfolio investments accounting for just over half of this amount (Abreu 1985).12 The stock of human capital in Brazil—​as proxied by the literacy rates of the population aged five years or more—​was very low throughout the nineteenth and early twentieth centuries. The 1872 census indicated a rate of 17.7%, which would increase modestly until 1920 (when it was at 28.8%), reaching 38.8% in 1940.13 From early on, the Brazilian economy was geared toward providing foreign markets with primary products. Estimates put Brazil’s share of world exports at 2.3% on average between 1830–​1850, inching up to a peak of 2.7% in 1850–​1870. From then on, this share would continuously drop and reached 1.7% on average between 1913 and 1929 (Federico and Tena-​Junguito 2016). As noted earlier, the imperial period marked the ascendancy of coffee as Brazil’s main export crop. In 1830 coffee accounted for just over 40% of total exports, amounting to £3.3 million; by 1889 coffee exports exceeded 60% of a total valued at £28.5 million.14 Over the course of the First Republic, the absolute value of merchandise exports increased further, exceeding £100 million in a couple of years. In the 1920s they averaged around £90 million per annum, largely due to high international coffee prices resulting

The Nineteenth and Early Twentieth Centuries    45 from the so-​called valorization schemes carried out by Brazil (IBGE 1990). Coffee’s share of total merchandise exports ranged from 50% to 70% in the early decades of the twentieth century. The remaining export items consisted of traditional staples such as sugar, cotton, tobacco, hides, and so on.15 Wild Amazonian rubber experienced a boom at the turn of the century, with its share of total exports peaking at nearly 40% in 1910. Competition from Southeast Asia on the eve of World War I would spell the end of this boom, however. During both the imperial period and the First Republic, the milréis (expressed as 1$000) was the official currency of Brazil. The exchange rate against sterling began the nineteenth century at between £0.29–​£0.35/​1$000, embarking on a downward trend from the late 1910s onward. In 1833 a (notional) gold parity of £0.18/​1$1000 was set, later (1846) reduced to £0.11/​1$000. Apart from the 1854–​1864 years, when gold-​backed note issues comprised a significant share of the money supply, for the remainder of the imperial period a fiduciary system prevailed. This arrangement also applied for most of the First Republic, barring the 1906–​1914 and 1927–​1930 intervals, during which convertible notes were issued by a currency board. Throughout this 100-​plus-​year period, the rate of exchange varied greatly, although its trend was unambiguously downward, reflecting persistent inflation in Brazil.16 The Brazilian banking system remained underdeveloped for most of the imperial period. The government-​controlled Bank of Brazil aside, few other banks operated outside the capital. Banking services were also, partially, performed by commercial houses and private banks. From mid-​century onward—​and on account of the growth of the export sector and economic activity at large—​several foreign banks set up operations in Brazil. Commercial banks other than the Bank of Brazil were also established, often with countrywide branches. Banking activity expanded in the early days of the First Republic, only to subside in the wake of the boom and bust episode known as the Encilhamento, which brought two successive waves of bank failures. Economic recovery at the start of the twentieth century was accompanied by renewed banking activity. From 1906 a refounded Bank of Brazil would gradually establish itself as the largest bank in the land, combining private banking operations with proto-​central bank functions until well into the 1980s.17 In line with the rest of the world, the size of the public sector in nineteenth-​and early-​ twentieth-​century Brazil was markedly smaller than it would become in the 1930s. Raymond Goldsmith puts central government revenues at around 10% of GDP during the imperial period, increasing to 12.5%–​16.4% during the First Republic (Goldsmith 1986). The share of central government receipts in the public sector total stood at just over 80% in 1856, and declined consistently afterward, reaching 54.2% in 1929. For most of the nineteenth century, import taxes comprised the single most important source of central government revenues, bringing in 50%–​65% of total tax receipts. This share declined during the early decades of the twentieth century, as taxes on domestic consumption and income gained importance amid the diversification of economic activity in general. A combination of note issues and debt (both domestic and foreign) helped finance recurrent budget shortfalls during the whole period under examination here. As

46   André Villela a result, foreign public debt, which stood at £5.3 million in 1830, reached £30.4 million in 1889 and £267 million in 1930 (Abreu and Lago 2014; Villela and Suzigan 2001). In what follows, I confine discussion to just two aspects of the Brazilian economy during the empire and the First Republic, namely its growth over time, and the resultant structural changes.18

3.2.  The Imperial Period: Falling Behind The publication, almost 20  years ago, of an influential collection of essays (Haber 1997) served to stimulate academic interest in the roots of Latin America’s relative economic backwardness.19 The qualifier “relative” in this sort of context usually refers to a comparison of growth rates (or per capita GDP levels) between different Latin American countries, on the one hand, and the United States. Against this yardstick, Brazil’s performance throughout the nineteenth century was underwhelming. It is estimated that US per capita GDP in 1800 was already twice as high as Brazil’s; by 1850 the ratio had increased to 3:1, and in 1890 it stood at just over 4:1.20 What accounts for this difference? Or, to put it differently, why did GDP per capita grow so slowly in nineteenth-​century Brazil? A  popular explanation would point to slavery as the reason for Brazil’s underperformance in the imperial period. After all, slavery has been often thought of as archaic and anti-​capitalistic, and this form of labor was all too pervasive in nineteenth-​century Brazil. However, this explanation is clearly insufficient, as it does not account for Cuba’s superior economic performance in the nineteenth century, despite the sharp decline of its terms of trade starting in the 1840s.21 Nor does it explain the marked contrast in economic performance of two contemporaneous slave-​based economies, the US South and Brazil’s coffee region, in favor of the former (Graham 1981). A different hypothesis for the contrasting fortunes of the North American and Latin American economies over the long run was put forward by Engerman and Sokoloff (1997) in their oft-​cited chapter.22 Essentially, these authors credit institutions such as slavery itself, schooling, access to land, and voting rights with the ultimate influence over what they perceive as Latin America’s (and Brazil’s) slower growth performance in the long run and in comparison with the rates of growth achieved by the United States and Canada. According to Engerman and Sokoloff, these institutional differences in turn stemmed from differences in the factor endowments (as captured, for instance, by climate, soil type, or the supply of indigenous labor) of Latin America vis-​à-​vis British North America.23 The Engerman and Sokoloff argument, identifying colonial institutions as the reason for slower long-​run growth rates in Latin America as a whole, has not gone unchallenged. In this regard at least two skeptical views stand out.

The Nineteenth and Early Twentieth Centuries    47 In the first of these, associated with the work of Leandro Prados de la Escosura, the very notion of Latin American economies “falling behind” sometime in the nineteenth century should be revised, on at least two counts. First of all, the benchmark normally used in these types of comparisons is the United States, admittedly the most successful instance of per capita GDP growth over the past two centuries (de la Escosura 2007). If instead of the United States a composite made up of rich (that is, current members of the OECD) countries were used as benchmark, Latin America’s growth record since the mid-​nineteenth century would not stand out as so disappointing. Indeed, from 1860 to 1938 the region managed to post growth rates on a par with the group of industrialized countries (de la Escosura 2007, 22).24 Second, the author shows that the severest drop in Latin America’s (and Brazil’s) per capita GDP vis-​à-​vis the Organisation for Economic Co-​operation and Development (OECD) countries took place between 1980 and 2000; therefore—​and contra Engerman and Sokoloff—​it would make more sense to search in the very recent past, rather than in the colonial era, for the factors that account for the continent’s relative economic underdevelopment. The second strand of criticism of the Engerman and Sokoloff view comes from Leticia Arroyo Abad. Contrary to those authors, whose argument endows factors occurring in the distant past with a great deal of persistence over centuries (path dependency), Arroyo Abad reminds us that endowments themselves are not fixed, but rather, change over time as countries are more or less open to the outside world. Indeed, as merchandise trade and flows of both capital and labor grew in Latin America (and the world as a whole) throughout the nineteenth century, the region’s endowments changed in tandem, resulting in varying rates of growth and distribution of income across the different countries (Abad 2013). The previously mentioned shortcomings of the Engerman and Sokoloff argument serve to reinforce the idea that there is no “silver bullet” or single factor when it comes to identifying the cause(s) of such a multivariate and complex phenomenon as economic growth. Institutions must certainly be included in any list of explanations, as must geography (or factor endowments). Trade—​both foreign and internal—​and economic policy are also likely candidates in any “holistic” explanation of why the nineteenth-​century Brazilian economy grew at a slow rate. One attempt at identifying the main factors accounting for such a trajectory was made by an often overlooked student of Brazilian economic history, Nathaniel Leff. In a number of articles published in the 1960s and 1970s in prestigious journals in the fields of economics and history, and subsequently revised and published in book form, Leff advanced an interpretation that involves three main points (Leff 1982; for a shortened version of the same arguments, see Leff 1997). The first point involves a composition effect, namely, the fact that the Northeast region as a whole—​which, at the start of the independent period (1822) was home to approximately half the Brazilian population—​probably experienced a decline in per capita GDP over the nineteenth century. This contrasted with the fortune of the Southeast, where economic growth attracted increasing numbers of workers, both free men and slaves, resulting in it becoming, by the end of the monarchic period, the demographic and

48   André Villela economic center of the country. These opposite trajectories, in turn, stemmed from the divergent paths taken by the main export staples in both regions. While Northeastern cotton and sugar experienced competition from more efficient suppliers in the United States and Cuba, respectively, in coffee—​the main export crop of the Southeast—​Brazil gradually established itself as the world’s leading producer and exporter, capable of exercising market power.25 (A similar point is also made by Furtado 1970, Chapter 18). The second—​and most original—​point emphasized by Leff concerns what appears to have been the occurrence of a peculiar form of the “Lewis model” in Brazil.26 For Leff, contrary to the original version of the model, the Brazilian historical experience consisted of a supply of labor that was for political reasons kept permanently “unlimited”—​at first, by importing of millions of African slaves and, from the last quarter of the nineteenth century onward, by subsidizing the entry of an almost equal number of European immigrants. Two outcomes of this deliberate policy may be discerned. On the one hand, there was little pressure on average wage rates in Brazil, with dire consequences for the welfare of the majority of the population. Consumer demand, as a result, was kept limited to simple wage goods. The other long-​term—​and negative—​ consequence of this excessive supply of labor was the limited incentive it provided for the substitution of capital for labor, and thus the increase in the marginal productivity of the latter. Leff ’s third and final reason for why the Brazilian economy grew so little in the nineteenth century has to do with the limited size attained by the modern—​in this case, export—​sector of the period. The idea here is that while the overall productivity of the Brazilian economy during the imperial period was low, the export sector—​which attracted more capital and entrepreneurial talent, and adopted the most productive technology—​displayed a relatively higher degree of efficiency. As a result, depending on the relative size of this sector (that is, its share of GDP) the average productivity of the economy will be higher or lower. Establishing the actual size of the export sector in nineteenth century Brazil is a difficult undertaking. As already noted, trade statistics for the period are far from reliable, although a recent revision has managed to correct some of the major errors in previous compilations (Absell and Tena-​Junguito 2016). Still, there remains the problem of constructing better estimates of the size of the Brazilian economy itself for the nineteenth century, which makes the export/​GDP ratios calculated by Absell and Tena-​ Junguito questionable. Based on the GDP estimates provided by Tombolo (2013), these authors claim that exports accounted for between 10% and 45% of Brazilian GDP in the nineteenth century, averaging about 30% over the whole period.27 This ratio appears too high, as it would imply an openness ratio (that is, imports plus exports as a percentage of GDP) of, roughly, twice that size (i.e., some 60%).28 If one accepts a more likely export ratio in the 15%–​20% range, then it becomes clear that the most productive sector in nineteenth-​century Brazil accounted for a small share of the economy as a whole. Furthermore, given the small size of both the manufacturing and service sectors, it remains that domestic agriculture employed the largest share of the workforce at the time. Ultimately, then, it was the absence of significant productivity

The Nineteenth and Early Twentieth Centuries    49 gains in the domestic agricultural sector that held back economic growth in the imperial period. Two interrelated factors explain the limited growth potential (and performance) of the domestic agricultural sector. First is the often rudimentary techniques that were employed (slash and burn technology with little recourse to metal implements such as hoes, scythes, and plows) in the production of foodstuffs (Linhares and Teixeira 1981). Second—​and, arguably, more important—​are the huge transport costs that characterized inland trade in Brazil. Geography, of course, played a part in this, as the almost 1,000-​mile-​long Serra do Mar mountain range, rising just a few miles in from the Southeast coast (where the largest urban settlements had been established since colonial times), posed a formidable obstacle to the movement of goods and people between this area and the hinterland. Moreover, the absence of a network of navigable rivers (let alone man-​made canals) near the coast, in addition to poor roads, made the mule train the main mode of transportation well into the second half of the nineteenth century. High transport costs, in turn, militated against specialization and accompanying gains from trade. As a result, commodity markets remained poorly integrated, and the incentives for the adoption of productivity-​enhancing technology were small.29 In the end, then, the more efficient export sector, comprising a relatively small share of the Brazilian economy, was unable to push the rest of the productive sector forward.30 This state of things would only begin to change with the advent of the railway. The first track was laid in 1854, but significant expansion did not occur until the 1880s.31 The potential impact of the new technology on overall efficiency (and, thus, economic growth) was tremendous, coming as it did to upgrade an antiquated transportation system. As a result, the volume of output produced for the market increased, as did the extent of interregional specialization (Leff 1997, 46).32 The tardiness and relatively modest scale of the introduction of such a growth-​enhancing transport technology should not be ascribed either to short-​sightedness on the part of the local elite or to acritical adherence to laissez-​faire ideology. Rather, it was the limited fiscal resources at the disposal of the government (provincial governments, especially) that prevented a more aggressive investment policy, either directly or through subsidies to the private sector (Leff 1997, 51–​54). One might speculate as to whether Brazil could have escaped this slow growth scenario that prevailed throughout the nineteenth century. However, returning to the major points highlighted in the preceding provides us with no indication of (historically meaningful) alternative scenarios that would have enabled the imperial economy to perform much better. Indeed, the factor of geography—​which locked the Northeastern economy into the production of sugar and cotton, with their declining terms of trade—​also figures in the way that mountainous topography and a dearth of navigable waterways both formed major obstacles to greater market integration and overall productivity gains. By the same token, to the landed elite’s unwillingness to provide tax resources, which could have enabled the state to fund the provision of public goods such as railroads or schools, must be added the possibility that the value of their taxable land was not that significant anyway, given their distance from markets. Finally, an official policy of restricting

50   André Villela the inflow of foreign labor (which would have pushed up local wages) was always going to have been highly unlikely, given the elite’s dependence on abundant and thus cheap hands to work the land (Leff 1997, 58–​59). Nevertheless, by the end of the imperial period things started to change, albeit very slowly. Railway construction proceeded apace, as did the abolition of slavery and the start of mass European immigration in the late 1880s, the latter providing the market with abundant and better educated labor.33 In addition, some of the institutional barriers that had hindered growth for most of the nineteenth century would come down, as in the cases of the ban on the free incorporation of joint-​stock companies (repealed in 1882) and the taxation of interprovincial trade (prohibited by the 1891 Constitution). Finally, the beginnings of modern industry in the late nineteenth century would introduce a further potential source of productivity gain to the Brazilian economy. It is to this slowly evolving economic scenario—​which coincides with the demise of the imperial regime and the onset of the Republic—​that I now turn.

3.3.  Modern Economic Growth (Slowly) Kicks In: The First Republic At the outset of the Republic, Brazil was still a very poor country, even by regional standards. It is estimated that Argentina’s per capita GDP in 1890 was almost three times as high, while Mexico’s was 30% higher. The difference in relation to the United States and the United Kingdom was even greater: 4:1 and 5:1, respectively. Other indicators attest to Brazil’s relative underdevelopment. The urbanization rate (measured by the percentage of the population living in cities of over 20,000 inhabitants) in 1890 was a mere 5.7%, as against 19.3% in Argentina and 14.8% in Chile. In schooling, Brazil displayed an enrollment rate of only 2.3% of the population, compared to 7% in Argentina in 1890; in the same year, 85% of the Brazilian population was illiterate (all figures in this paragraph in Franco and Lago 2012, 199). The early republican period was highly unstable, both politically and economically, thus contrasting with the relative stability of the previous regime (at least during most of the Second Reign).34 In the financial arena, major changes to monetary policy and capital market legislation helped fuel a boom in the Rio stock exchange in the 1890s, followed by an equally spectacular bust. The Encilhamento episode, with its aftermath of exchange-​rate collapse and two successive waves of bank failure, helped strengthen the position of monetary orthodoxy for the remainder of the First Republic. Still, there is no disputing the real effects of this early republican experiment in monetary profligacy in fostering—​if only unwittingly—​the first wave of import substitution industrialization in Brazil, in the cotton-​weaving industry (Fishlow 1972). Economic policy for the rest of the first republican decade was geared toward attempts at restoring exchange rate stability, which was attained in the early twentieth

The Nineteenth and Early Twentieth Centuries    51 century—​and only then after a bailout loan from British creditors, in tandem with the adoption of harsh deflationary policies.35 The resultant recovery of the exchange rate, in combination with coffee prices languishing at an all-​time low, led to the decision to intervene in both the commodity and money markets in 1906, which resulted, on the one hand, in the first so-​called coffee valorization scheme, involving the withdrawal from the market of part of Brazil’s ever-​increasing produce, and on the other, the setting up of the Conversion Office (Caixa de Conversão), marking Brazil’s embrace of contemporary gold standard orthodoxy.36 Economic growth in the decade before World War I stemmed, mostly, from the coffee and rubber booms and attendant infrastructure investments in railroads, ports, city improvements, and so on.37 Industrial production also benefited from a combination of growing markets, expanded infrastructure, and a stable currency (during 1906–​1914), which favored the importation of capital goods.38 Throughout, the Brazilian government adopted a mostly laissez-​faire approach in the economic sphere, except for crucial support (via special concessions or subsidies) to the railway and steel sectors and indirect backing (through loan guarantees) to the first coffee valorization scheme.39 The (admittedly incomplete) industrial survey carried out in 1907 provides a snapshot of the manufacturing sector in the early years of the twentieth century. In that year the number of industrial units reached 3,258, employing some 152,000 workers. The textile industry, individually, accounted for about a third of this workforce, with the bulk of the manufacturing sector consisting of the food-​processing and beverage industries. The survey also counted firms operating in other sectors of the light manufactured goods industry, such as shoes, matches, hats, tobacco, ceramics, and so on. As is to be expected, the metal and mechanical industries employed but a small fraction of total workers in the manufacturing sector in 1907 (Bonelli 2003). In spite of the performance of both the industrial and export sectors, overall economic growth in the first two decades of the republican period was disappointing, with per capita GDP barely budging between 1889 and 1913.40 World War I, rather paradoxically, helped to change this, as pent-​up demand for manufactured goods amid severe import restrictions helped fuel a second round of import-​substitution industrialization. In a repetition of the sequence observed in the early 1890s, a previous phase characterized by a stable (and relatively appreciated) rate of exchange facilitated the purchase of capital goods from overseas, to which was added local production of technologically simpler machinery (Franco and Lago 2012, 228). The increased productive capacity thus achieved was to be employed during the war years, allowing for an average growth of manufacturing output of 8.5% per annum during 1914–​1918 (Bonelli 1996, 84). By 1919, when the first full industrial census was undertaken, Brazil had achieved near self-​sufficiency in the production of light consumer goods, which accounted for some 80% of manufacturing value added. The intermediate goods industry, in turn, comprised 16.5% of manufacturing value added, with the remainder being made up of the more technologically complex capital goods and consumer durables industries. In those sectors, consumption was met mostly by imports (Fishlow 1972, 322–​323).

52   André Villela The 1920s mark the pinnacle of Brazil’s export-​led growth “model.” Indeed, with the end of the Amazonian rubber boom, coffee reigned supreme once again, making up 70% of total exports on average. Coffee exports doubled in value in the final decade of the First Republic, amounting to £57 million per annum, on average, against £29 million per annum in 1900–​1919. Most of this increase, in turn, is explained by the higher coffee prices that were ensured from 1922 on by the so-​called permanent policy of withdrawing from market excess output.41 The Brazilian economy grew at an average rate of 5.7% per annum between 1919 and 1929, while manufacturing output expanded at a slightly slower pace of 5.2% per annum. It was during this decade that the first large-​scale steel and cement plants were set up in Brazil. Together with investments in the chemical and mechanical industries, as well as the assembly lines for automobiles and light trucks built by Ford and General Motors, they testify to the increased sophistication of the industrial landscape.42 Three phases stand out in this decade: a boom in the immediate aftermath of World War I that extended until 1920; slowdown and recovery in 1921–​1914, followed by deflation in 1925–​1926; and a final boom in 1927–​1929. The first beginnings of this latter phase came in the early months of 1929, when difficulties in balance of payments set in motion a monetary contraction by the Stabilization Office. The resultant credit squeeze led to a drop in industrial output, but its greatest impact would be felt by the coffee valorization scheme. Two successive bumper crops, in 1927 and 1929, made further financing of coffee stocks untenable in the midst of the credit crunch, leading to massive shipments of coffee to the market and a price collapse (Bonelli 2003, 381–​382).43 Taking the 1900–​1930 period (for which there are reliable estimates of physical output) as a whole, it becomes clear that the escape from a regime of virtual stagnation of GDP per capita, which characterized the nineteenth century, began sometime in the 1900s—​although only slowly. Sustained growth of per capita GDP would continue, with a few bumps along the way, all the way to the Great Depression, and beyond.44 By 1930 the Brazilian economy was, certainly, a very different beast from the economy at the outset of the First Republic four decades earlier; it was bigger, more industrialized and urban, and average incomes stood some 30% higher. Nevertheless, in absolute terms Brazil was still a poor, mostly agrarian, country, its social indicators (literacy and urbanization rates, for instance) lagging behind those of the major Latin American economies at the time. Faster and sustained economic growth, coupled with equally faster social and political change, would be a feature of the post-​1930 era.

3.4.  Specialization and Long-​Term Growth Since the period discussed in this chapter broadly corresponds to that of so-​called classical globalization, that phenomenon serves as a good lens for taking stock of the

The Nineteenth and Early Twentieth Centuries    53 nature and pace of economic change in Brazil in the nineteenth and early twentieth centuries. The vast literature on the first great wave of globalization emphasizes two interrelated aspects of the period, namely, the so-​called Great Divergence and the Great Specialization.45 The former refers to the growing gap in economic growth between, on the one hand, a group of countries in Europe and their offshoots and, on the other, the vast majority of countries, which remained trapped in a low/​no growth regime. The Great Specialization, meanwhile, refers to the observed trend toward the specialization of the economies of different countries in the production and export of goods in which they possessed comparative advantages. Such specialization, in turn, was greatly aided by the significant reduction in transport costs (thanks, for instance, to railways and the steamship) over the nineteenth century, making the shipment of vast amounts of bulky merchandise over large distances feasible for the first time in human history. This chapter has discussed Brazil’s laggardly performance for most of the circa 1820–​ 1930 period, when per capita GDP increased by no more than 50%. This lackluster growth record places Brazil among the large group of countries that diverged from those few economies which managed to achieve high rates of economic growth in the period. One might speculate as to whether Brazil’s poor growth record was the to-​be-​ expected outcome of its specializing in the production and export of primary goods, in contrast to those economies that successfully industrialized and embarked on a British-​style mode of modern economic growth. However, the historical evidence does not back this up. In other words, it is not necessarily the case that the gradual escape from poverty achieved by a few countries over the course of Hobsbawm’s “long” nineteenth century was predicated on their experiencing some form of industrial revolution. Rather—​and as cogently argued by Harley (2014)—​modern economic growth in the period was driven, more than anything else, by the spread of capitalism and its attendant institutions, prominent among which were markets (for goods, services, and factors of production). The cases of economies that displayed a remarkable growth record before World War I  and where the manufacturing sector did not figure prominently are well known: Australia, New Zealand, Argentina, and Uruguay.46 What these countries have in common is their specialization in the export of temperate commodities, whereas Brazilian exports consisted of tropical goods. Perhaps it is the nature of the commodities being exported—​ rather than the specialization in the production and export of commodities as such—​that accounts for differing growth trajectories among these economies. In discussing the impacts of the first wave of globalization on the fortunes of developing countries, Jeffrey Williamson (2011) highlights three possible channels through which the growth prospects of Third World countries might have been hampered: (1) through a process of deindustrialization; (2) by means of the possibly perverse economic effects of disproportionate political power being conferred on a landowning elite; and (3) via exposure to price volatility, made worse by excessive concentration in the export of just a few commodities.47

54   André Villela The first factor does not apply to the Brazilian case, but the remaining two do. As already discussed, legal access to land in Brazil was restricted to a minority, who also exercised control of the political process. As a result, public policies that would have implied the provision of a larger amount of public goods (such as schools, hospitals, and investments in infrastructure in general) and, hence, economic growth were precluded. Also, Brazil’s specialization in the production and export of coffee—​itself a consequence, at least in part, of its geography—​may have meant a somewhat unfortunate draw in the so-​called commodity lottery, given the long-​term behavior of its international price. According to Williamson, the Brazilian economy displayed some of the highest degrees of price volatility among Third World countries between the 1860s and World War I. This was in part due to the high commodity export concentration ratio (compared to Argentina, for example). To make matters worse, Brazilian exports per capita were between one-​quarter and one-​half of the value exported by countries in the region, such as Argentina, Chile, Uruguay, and Cuba, all of which achieved much higher rates of growth in the period (Bulmer-​Thomas 2003, 68). Finally, compared to other commodities, coffee seems to have provided far fewer linkages to other domestic sectors, thus limiting even further its overall propensity to push the rest of the Brazilian economy forward.48 In sum, Brazil’s disappointing growth performance during the Great Specialization derived, to a large extent, from a combination of two factors: (1) a relatively small export sector (and, conversely, a large, unproductive domestic agricultural sector); and (2) its reliance on the sales of a commodity that was neither particularly successful in world markets—​in terms of both its prices (which were volatile) and overseas demand (which grew in line with incomes)—​nor did it have as widespread and significant an impact on domestic economic activity as other primary products had in various countries. We are back, then, to Nathaniel Leff ’s broad assessment of Brazil’s growth record in the nineteenth century, as discussed earlier. Ultimately, it appears that one could not reasonably have expected the Brazilian economy to have fared much better than it did at the time, given the several constraints under which it operated. Geography was one such constraint (another was institutions), with a major bearing on transport costs and on Brazil’s specialization in the production of coffee, a crop that failed to have as great—​and as positive—​an impact as temperate commodities had for frontier economies such as Argentina, Uruguay, Australia, the United States, and Canada. However, one should not conclude from the preceding discussion that some form of geographical determinism is being advocated. As Brazil’s post-​1930 history would go on to demonstrate abundantly, against the same geographical backdrop highly significant socioeconomic and political change could take place.

Notes 1. I  wish to thank the volume editors for their helpful comments. I  am also indebted to Marcelo de Paiva Abreu and Carlos Accioly for their careful reading of ad comments on an

The Nineteenth and Early Twentieth Centuries    55 earlier draft chapter. Finally, I would like to thank our dear friend, the late Werner Baer, for the invitation to contribute to this Handbook. 2. Rapid industrialization notwithstanding, commodities (with coffee looming large) would still comprise the bulk of Brazilian exports well into the 1960s. 3. This section follows a “canonical” division of topics aimed at providing a broad overview of the Brazilian economy from ca. 1800 to 1930. For a similar treatment (albeit confined to the imperial period), see Abreu and Lago (2014). 4. Aggregate rates of population growth in this period mask a distinct difference in the patterns observed depending on the racial/​ethnic groups concerned. Hence, the population of European descent grew at an average rate of 1.79% per annum between 1798 and 1872, while the free black population grew at 3.17% per annum. The slave population, meanwhile, actually dropped slightly, reflecting the harshness of their labor regime (and overall inferior living conditions) and the end of the slave trade in 1850. For a summary discussion of long-​term Brazilian population history, see Livi-​Bacci (2002). Due to political turmoil the population census was not carried out in 1930. 5. For the first half of the twentieth century, the best estimates are those by Haddad (1978). From 1947 onward, official GDP data calculated within a national accounts framework became available. 6. This view is shared by Furtado (1970, Chapter 25) and is, broadly speaking, the subject of consensus. In contrast, a highly implausible claim is made by Engerman and Sokoloff (1997) according to whom per capita GDP in Brazil grew at an average rate of 0.4% per annum between 1800 and 1850, only to decline 0.4% per annum between 1850 and 1913. For a critique of this contrarian claim, see Abreu and Lago (2014, 3). 7. As one approaches the end of the nineteenth century, though, export figures for items such as sugar, cotton, and tobacco lose their relevance as a proxy for output, as the local market becomes an important outlet for domestic producers (Abreu and Lago 2014, 4–​6). 8. The literature on the history of the coffee industry in Brazil is, of course, huge. A recent contribution that places this history within the broader international context is Marques and Tomich (2009). 9. Data in the Slave Voyages Database, http://​www.slavevoyages.org/​assessment/​estimates (accessed August 17, 2016). 10. Data on immigration in Merrick and Graham (1979, 37). The marked increase in the annual rate of immigrant arrivals that Brazil experienced at the end of the imperial period/​ early Republican period is partially explained by legislation enacted in 1884, which provided for the granting of government subsidies covering travel expenses for immigrant families coming into São Paulo. 11. On land policy in nineteenth-​century Brazil, see Dean (1971). The 1850 Law was instrumental, however, in the success of the official colonization schemes that were carried out in the southern provinces, and which provided for the parceling out of public land in smallholdings to European immigrants. 12. British capital dominated these flows for most of the period, although American investments made up an increasing share from World War I onward. 13. Brazilian literacy rates lagged behind not only those achieved by industrialized countries, but also most of the rest of Latin America (see Mariscal and Sokoloff 2000). Data on literacy rates in Brazil are in Ferraro (2002). On the relationship between low levels of human capital and Brazil’s poor economic record until the 1930s, see Barros (2016).

56   André Villela 14. Brazilian trade statistics for the early imperial period are notoriously inaccurate, as they rely on so-​called official (rather than market) values. For a recent attempt at revision, see Absell and Tena-​Junguito (2016). 15. For the breakdown of Brazilian exports, see Villela and Suzigan (2001, 63). 16. For example, the exchange rate, which in 1889 stood at close to parity (that is, at 27d/​ 1$000), would average a mere 4.9d/​1$000 in 1930. The history of the financial system during the empire is discussed in detail in Abreu and Lago (2001). 17. For the early history of banking in Brazil, see Cavalcanti (1893). The facts leading to the Encilhamento are discussed in Franco (1987) and Schulz (1986). For the history of banking during the First Republic, see Triner (2000). The development of the paulista capital market is the subject of Hanley (2005). On the connection between the coffee sector and capital markets in nineteenth-​century Brazil, see Sweigart (1987). 18. On the long-​run performance of the Brazilian economy, see Villela (2013). 19. Although the imperial period would only begin in 1822, following political independence from Portugal, since 1808 Brazil had become the administrative center of the Portuguese empire, which only served to reinforce its status as de facto economic center. In the following I treat the whole period extending from the early nineteenth century to 1890 as the “imperial” period. 20. If, alternatively, one were to choose as the baseline the group of Southern Cone countries, Brazil’s performance would not look so bad, yet the country would still stand out as a laggard: it has been estimated that Argentinean per capita GDP was 1.4 higher than Brazil’s in 1800, and this ratio increased to 3:1 by 1890; in the case of Chile, the ratios were 1:1 in 1800 and 2.5:1 in 1890; for Uruguay, 1.6:1 and 2.7:1 over the same period. All ratios calculated from data on GDP per capita (in 1990 international dollars) in the Maddison Project Database, http://​www.ggdc.net/​maddison/​maddison-​project/​home.htm, 2013 version (accessed August 20, 2016). 21. Cuba’s GDP per capita in 1800 was estimated by the Maddison Project as equivalent to just over $500 (in 1990 international dollars), against Brazil’s $683; by 1890, it had climbed to $1,510, while in the case of Brazil it had crawled to just $794. As for the so-​called net barter terms of trade (that is, the ratio of export prices to import prices—​a measure of the purchasing power of exports), these dropped by half in Cuba between the 1820s and 1880, while increasing by 70% over the same period in Brazil. This difference, essentially, reflects the contrasting fortunes of sugar and coffee, respectively, in the international market in this period. For the terms of trade, see de la Escosura (2009). 22. In later years these authors would refine their original arguments in a host of essays ultimately collected in Engerman and Sokoloff (2012). 23. It is worth noting that the argument linking factor endowments (“geography”) in the New World to different outcomes of European colonization is not new. In Brazilian historiography it was famously advanced in 1942 by Caio Prado, Jr. (Prado 1942; also published by the University of California Press in 1967 as The Colonial Background of Modern Brazil). In contrasting the fortunes of the tropical and temperate colonies in the Americas, Prado (1942)followed nineteenth-​century European historians, especially Paul Leroy-​Beaulieu (1882). For discussion see Leonidio (1999). 24. In this respect, it must be said that the Brazilian experience was somewhat different, as catching up with OECD countries would only start sometime around 1930, extending into the early 1980s.

The Nineteenth and Early Twentieth Centuries    57 25. One might wonder why, in light of the declining competitiveness of the sugar and cotton crops in the Northeast, resources were not redirected into coffee. This was no simple task, as Leff explains. Not only did the shallow capital markets of the day render this relocation more difficult, but climatic reasons made coffee less adapted to environmental conditions in the Northeast. To make matters worse, the author speculates that the Northeast may have also suffered from the so-​called Dutch disease, as the (nationwide) exchange rate, mostly determined by export earnings from the coffee sector, proved to be overvalued for the sake of the competitiveness of sugar and cotton exports (Leff 1982, Chapter 1). 26. The model was first presented by Lewis (1954). The Lewis model predicts that as countries develop, surplus labor is pulled in from the “subsistence” (agricultural) sector into the “capitalist” (industrial) one, attracted by the higher wages paid by the latter. At first—​and while the supply of labor is still unlimited—​the industrial sector manages to grow while preventing wage costs from denting its profits, which help finance capital investments in this sector. Over time the marginal productivity of workers in the industrial sector will increase with capital formation, while being driven down by the arrival of new workers. Eventually, the wage rates in both sectors will tend to equalize as workers leave the subsistence for the capitalist sector, increasing marginal productivity—​and hence wages—​in the former while driving down productivity (and wages) in the latter. 27. Exports/​GDP ratios in Absell and Tena-​Junguito (2016), derived from “eyeballing” Figure 3 in their article. Export ratios of 30% for Brazil, as posited by Absell and Tena-​ Junguito, are almost twice as high as the figures suggested by, among others, Leff (1982, 41) and Goldsmith (1986, 52). 28. Openness ratios of 60% in the nineteenth century contrast markedly (and implausibly) with 30% in Western Europe in 1870. For the latter, see O’Rourke, de la Escosura, and Daudin (2010, 106). 29. Long distances and high transport costs also reduced potential outlets for manufacturing goods produced in urban centers closer to the coast, thus thwarting productivity gains arising from economies of scale in production and distribution. 30. A simple exercise shows that with a population growth rate of 2% per annum, a share of the export sector of 20% of GDP and a domestic agricultural sector that expanded either at the same rate as total population (that is, displaying no productivity gains) or at a rate of 2.5% per annum (.5% annual productivity gain), Brazilian exports would have to grow, on average, between 7.5% and 9.5% per annum between 1850 and 1913 in order for GDP per capita to increase by 1.5% per annum. In practice, though, exports grew at just 2.8% per annum in this period. For this exercise, see Bulmer-​Thomas (2003, 61). Export growth rate in Absell and Tena-​Junguito (2016, 692). 31. Still, by 1890 the country would have only 9,973 kilometers of operating track. Total track of just over 26,000 in 1914 was equivalent to the amount achieved by the United States in the 1850s. See Leff (1997, 44–​45). 32. The social savings (that is, the difference between shipping a given amount of freight by railway and the cost of shipping the same amount over the same distance without the railway) accruing from the introduction of this novel transport technology has been estimated at 18% of Brazilian GDP in 1913. Comparable estimates for the antebellum United States put these gains at less than 4% of GDP, reflecting the country’s efficient transport system prior to the massive subsequent expansion of the American railway network. See Summerhill (2003, 97–​98).

58   André Villela 33. On the long and varied process of transition from slave to free labor in Brazil, see Lago (2014). For European immigrants’ higher levels of basic education when compared to native Brazilians, see Merrick and Graham (1979, 111) and Barros (2016, 164). 34. For the political history of the period, see Cardoso (1985)). 35. On economic policy during the first republican decade, see Franco (1987) and Fritsch (1988, Chapter 1). 36. It is worth noting that the Caixa de Conversão introduced convertibility to gold “at the margin” only, that is, strictly of those notes issued by this office, but not to the (much higher) volume of Treasury notes already in circulation. 37. In the wake of railway expansion and the massive inflow of European immigrants, the Brazilian coffee sector experienced a boom at the turn of the century, its output increasing from an average of 8 million 60-​kg bags in the 1890s to 14 million bags in 1900–​1913. By then São Paulo had taken over leadership of the industry from Rio de Janeiro, with the number of trees in the state increasing fourfold between 1886 and 1905 (from 150 million to 600 million) and stabilizing thereafter. See Holloway (1984, 264). Meanwhile, exports of wild rubber increased from an average of 20,000t in the 1890s to more than 35,000t between 1900 and 1913. See IBGE (1990, 347). 38. On the increased productive capacity prior to World War I, see Baer and Villela (1973, 221). For an excellent history of Brazilian industrialization, see Suzigan (2000). For the capital goods industry in particular, see do Lago, Almeida, and Lima (1979). Dean (1969) remains the classic treatment of the social origins of the Paulista industrial entrepreneurs. 39. Import tariffs throughout the First Republic were among the highest in the world. Still, their protective effect on industry was relatively minor compared to protection afforded by the exchange rate. For this point, see Fishlow (1972) and Villela (2000). 40. GDP estimates in constant “international” dollars from the Maddison Project Database, http://​www.ggdc.net/​maddison/​maddison-​project/​home.htm. Population, meanwhile, increased 2.4% per annum on average in the same period, from 14 million to 24 million. 41. From 1927 to 1930 coffee exports were also sheltered from exchange rate appreciation with the operation of a currency board, the Stabilization Office (Caixa de Estabilização), which fixed the value of the milréis at 5.9 pence. 42. However, light consumer goods would still dominate the industrial structure in the 1920s, with an estimated 80% of total value added. For industrialization in the 1920s, see Bonelli (1996, 83–​87) and Suzigan (2000, 90–​92). 43. For a detailed account of economic policy and its consequences during the final decade of the First Republic, see Frisch (1988). 44. As a result of their different rates of growth along this 30-​year period, agriculture lost 10 percentage points. of its share of Brazilian GDP, estimated at 35.8% in 1930, whereas industry increased its share of GDP from 11.6% to 14.8% over the same period, with the services sector comprising the largest part of the economy in 1930 at 49.4%. See Bonelli (2003, 374). 45. For an excellent treatment of the period, see Findlay and O’Rourke (2007, Chapter 7). 46. Although not a frontier economy, Denmark could be added to this list of fast-​growing economies where industrialization did not play a significant role. 47. The idea being that export price (or terms of trade) volatility hampers growth by “encouraging low-​risk and low-​returns investment projects as well as less investment” (Williamson 2011, 184). 48. On the limited linkage effects involving the coffee industry in Brazil, see Catão (1992).

The Nineteenth and Early Twentieth Centuries    59

References Abad, Leticia Arroyo. 2013. “Persistent Inequality? Trade, Factor Endowments, and Inequality in Republican Latin America.” Journal of Economic History 73 (1): 38–​78. Abreu, Marcelo de Paiva. 1985. “A dívida pública externa do Brasil, 1824–​1931.” Estudos Econômicos 15, 167–​189. Abreu, Marcelo de Paiva, and Luiz A. Corrêa do Lago. 2001. “Property Rights and the Fiscal and Financial Systems in Brazil: Colonial Heritage and the Imperial Period.” In Transferring Wealth and Power from the Old World to the New World: Monetary and Financial Institutions in the 17th through the 19th centuries, edited by Michael D. Bordo and Roberto Cortés-​ Conde, 327–​377. Cambridge: Cambridge University Press. Abreu, Marcelo de Paiva, and Luiz Aranha Correa do Lago. 2014. “A economia brasileira no império, 1822–​1889.” In A ordem do progresso: Dois séculos de política econômica no Brasil, 2nd edition, edited by Marcelo de Paiva Abreu, 1–​28. Rio de Janeiro: Elsevier., Chapter 1. Absell, Christopher D., and Antonio Tena-​Junguito. 2016. “Brazilian Export Growth and Divergence in the Tropics during the Nineteenth Century.” Journal of Latin American Studies 58: 677–​706. Bacha, Edmar, and Robert Greenhill. 1992. 150 Anos de café, 2nd edition. Rio de Janeiro: Salamandra. Baer, Werner, and Annibal V. Villela. 1973. “Industrial Growth and Industrialization: Revisions in the Stages of Brazil’s Economic Development.” Journal of Developing Areas 7: 217–​234. Barros, Alexandre Rands. 2016. Roots of Brazilian Relative Economic Backwardness. London: Academic Press. Bethell, Leslie, and José Murilo de Carvalho. 1989. “1822–​1850.” In Brazil: Empire and Republic, 1822–​1930, edited by Leslie Bethell, 45–​112. Cambridge: Cambridge University Press. Bonelli, Regis. 1996. Ensaios sobre política econômica e industrialização. Rio de Janeiro: Senai. Bonelli, Regis. 2003. “Nível de atividade e mudança estrutural.” In Estatísticas do século XX, 383–​426. Rio de Janeiro: IBGE. Bulmer-​Thomas, Victor. 2003. The Economic History of Latin America since Independence, 2nd edition. Cambridge: Cambridge University Press. Cardoso, Fernando Henrique. 1985. “Dos governos militares a Prudente-​Campos Sales.” In História geral da civilização brasileira, 6th edition, edited by Boris Fausto, 17–​57. São Paulo: Difel. Catão, Luis. 1992. “The Failure of Export-​Led Growth in Brazil and Mexico, c. 1870–​1930.” Research Papers Series. No. 31. Institute of Latin American Studies, University of London. Cavalcanti, Amaro. 1893. O meio circulante nacional. Rio de Janeiro: Imprensa Nacional. Contador, Claudio R., and Claudio L. Haddad. 1975. “Produto, moeda e preços: Brasil, 1861–​ 1970.” Revista Brasileira de Estatística 36, 407–​440. Dean, Warren. 1969. The Industrialization of São Paulo, 1880–​1945. Austin:  University of Texas Press. Dean, Warren. 1971. “Latifundia and Land Policy in Imperial Brazil.” The Hispanic American Historical Review 51, 606–​625. de la Escosura, Leandro Prados. 2007. “When Did Latin America Fall Behind?” In The Decline of Latin American Economies: Growth, Institutions, and Crises, edited by Sebastian Edwards et al., 1–​46. Chicago: University of Chicago Press/​NBER. de la Escosura, Leandro Prados. 2009. “Lost Decades? Economic Performance in Post-​ Independence Latin America.” Journal of Latin American Studies 41: 279–​307.

60   André Villela do Lago, Luiz Aranha Corrêa. 2014. Da escravidão ao trabalho livre:  Brasil, 1550–​1900. São Paulo: Companhia das Letras. do Lago, Luiz Aranha Corrêa, Fernando Lopes de Almeida, and Beatriz Mello Flores de Lima. 1979. A indústria brasileira de bens de capital: Origens, situação recente, perspectivas. Rio de Janeiro: Editoria FGV. Engermann, Stanley, and Kenneth Sokoloff. 1997. “Factor Endowments, Institutions, and Differential Paths of Growth among New World Economies:  A View from Economic Historians of the United States.” In How Latin America Fell Behind: Essays on the Economic Histories of Brazil and Mexico, 1800–​1914, edited by Stephen Haber, 260–​304. Stanford, CA: Stanford University Press. Engerman, Stanley L., and Kenneth L. Sokoloff. 2012. Economic Development in the Americas since 1500: Endowments and Institutions. Cambridge: Cambridge University Press. Federico, Giovanni, and Antonio Tena-​Junguito. 2016. “Lewis Revisited:  Tropical Polities Competing on the World Market 1830–​1938.” Working Papers in Economic History 16–​05, Instituto Figuerola/​Universidade Carlos III de Madrid, July 2016. Ferraro, Alceu R. 2002. “Analfabetismo e níveis de letramento no Brasil:  O que dizem os censos?” Educação e Sociedade 23 (81): 21–​47. Findlay, Ronald, and Kevin H. O’Rourke. 2007. Power and Plenty: Trade, War, and the World Economy in the Second Millennium. Princeton, NJ: Princeton University Press. Fishlow, Albert. 1972. “Origins and Consequences of Import Substitution in Brazil.” In International Economics and Development: Essays in Honor of Raúl Prebisch, edited by Luis E. di Marco, 311–​365. New York; London: Academic Press. Franco, Gustavo H.  Barroso. 1987. Reforma monetária e instabilidade durante a transição republicana. Rio de Janeiro: BNDES. Franco, Gustavo H. Barroso, and Luiz Aranha Corrêa do Lago. 2012. “O processo econômico.” In A abertura para o mundo 1889–​1930, edited by Lilia Moritz Schwarcz, 173–​237. Rio de Janeiro: Editora Objetiva. Fritsch, Winston. 1988. External Constraints on Economic Policy in Brazil, 1889–​ 1930. Houndsmill: Macmillan. Furtado, Celso. 1970. Formação econômica do Brasil, 10th edition. Rio de Janeiro: Imprensa Nacional. Goldsmith, Raymond W. 1986. Brasil, 1850–​1984: Desenvolvimento financeiro sob um século de inflação. São Paulo: Harper & Row do Brasil. Graham, Richard. 1981. “Slavery and Economic Development:  Brazil and the United States South in the Nineteenth Century.” Comparative Studies in Society and History 23 (4): 620–​655. Haber, Stephen, ed. 1997. How Latin America Fell Behind: Essays on the Economic Histories of Brazil and Mexico, 1800–​1914. Stanford, CA: Stanford University Press. Haddad, Claudio L. 1978. Crescimento do produto real no Brasil, 1900–​ 1947. Rio de Janeiro: Fundação Getulio Vargas. Hanley, Anne. 2005. Native Capital: Financial Institutions and Economic Development in São Paulo, Brazil, 1850–​1920. Stanford, CA: Stanford University Press Harley, Knick C. 2014. “British and European Industrialization.” In The Cambridge History of Capitalism, Vol. I, edited by Larry Neal and Jeffrey G. Williamson, 491–​532. Cambridge: Cambridge University Press. Holloway, Thomas H. 1984. Imigrantes para o café: Café e sociedade em São Paulo, 1886–​1934. Rio de Janeiro: Paz e Terra.

The Nineteenth and Early Twentieth Centuries    61 IBGE. 1990. Estatísticas históricas do Brasil, 2nd edition, Vol. 3. Rio de Janeiro: IBGE. Leff, Nathaniel. 1982. Underdevelopment and Development in Brazil, Vol. I: Economic Structure and Change, 1822–​1947; Vol. II: Reassessing the Obstacles to Economic Development. London: George Allen & Unwin. Leff, Nathaniel H. 1997. “Economic Development in Brazil, 1822–​1913.” In How Latin America Fell Behind:  Essays on the Economic Histories of Brazil and Mexico, 1800–​1914, edited by Stephen Haber. Stanford, CA: Stanford University Press. Leonidio, Adalmir. 1999. “Em torno das origens:  Leroy-​Beaulieu e o pensamento social brasileiro.” Estudos Sociedade e Agricultura 13: 119–​138. Leroy-​Beaulieu, Paul. 1882. De la colonisation chez les peuples modernes. Paris: Guillaumin. Lewis, W. Arthur. 1954. “Economic Development with Unlimited Supplies of Labor.” The Manchester School of Economic and Social Studies 22, 139–​191. Linhares, Maria Yedda, and Francisco Carlos Teixeira da Silva. 1981. História da agricultura brasileira: Combates e controvérsias. São Paulo: Brasiliense. Livi-​Bacci, Massimo. 2002. “500 Anos de demografia brasileira:  Uma resenha.” Revista Brasileira de Estudos de População 19, 141–​159. Mariscal, Elisa, and Kenneth L. Sokoloff. 2000. “Schooling, Suffrage, and the Persistence of Inequality in the Americas, 1800–​1945.” In Political Institutions and Economic Growth in Latin America: Essays in Policy, History and Political Economy, edited by Stephen Haber, 159–​ 218. Stanford, CA: Hoover Institution Press. Marques, Rafael, and Dale Tomich. 2009. “O vale do paraíba escravista e a formação do mercado mundial do café no século XIX.” In O Brasil Imperial, Vol. II: 1831–​1870, edited by Keila Grinberg and Ricardo Salles, 339–​383. Rio de Janeiro: Civilização Brasileira. Merrick, Thomas W., and Douglas H. Graham. 1979. Population and Economic Development in Brazil: 1800 to the Present. Baltimore, MD: Johns Hopkins University Press. Netto, Antônio Delfim. 2009. O problema do café no Brasil, 3rd edition. Campinas: Editora Unesp, Edições Facamp. O’Rourke, Kevin, Leandro Prados de la Escosura, and Guillaume Daudin. 2010. “Trade and Empire.” In The Cambridge Economic History of Modern Europe, Vol. I: 1700–​1870, edited by Stephen Broadberry and Kevin O’Rourke. Cambridge: Cambridge University Press. Prado, Caio, Jr. 1942. Formação do Brasil contemporâneo. São Paulo: Brasiliense. Schulz, John. 1986. A crise financeira da abolição. São Paulo:  Edusp/​Instituto Fernando Braudel. Summerhill, William R. 2003. Order against Progress: Government, Foreign Investment, and Railroads in Brazil, 1854–​1913. Stanford, CA: Stanford University Press. Suzigan, Wilson. 2000. Indústria brasileira: Origem e desenvolvimento, 2nd edition. São Paulo: Hucitec and Editora da Unicamp. Sweigart, Joseph. 1987. Coffee Factorage and the Emergence of a Brazilian Capital Market, 1850–​ 1888. New York: Garland. Tombolo, Guilherme Alexandre. 2013. O PIB brasileiro nos séculos XIX e XX: 200 anos de ciclos econômicos. Unpublished MsC dissertation, Universidade Federal do Paraná. Triner, Gail. 2000. Banking and Economic Development:  Brazil, 1889–​1930. Houndmills, UK: Palgrave. Villela, André. 2000. “Tarifas de importação e câmbio na gênese da indústria brasileira, 1901–​ 1928.” História Econômica e História de Empresas 3 (2): 27–​46.

62   André Villela Villela, André. 2013. “O desenvolvimento econômico no Brasil pré-​1945.” In Desenvolvimento econômico: Uma perspectiva brasileira, edited by Fabio Giambiagi, Fernando Veloso, Samuel Pessôa, and Pedro Cavalcanti Ferreira, 91–​128. Rio de Janeiro: Elsevier. Villela, Annibal V., and Wilson Suzigan. 2001. Política de governo e crescimento da economia brasileira, 1889–​1945, 3rd edition. Brasília: Ipea. Williamson, Jeffrey. 2011. Trade and Poverty: When the Third World Fell Behind. Cambridge, MA: MIT Press.

Chapter 4

B r azilian Stru c t u ra l i sm Joseph L. Love

4.1. Introduction Economic structuralism was probably more influential in Brazil than in any other Latin American country. This doctrine, often called “developmentalism” (desenvolvimentismo) in Brazil, characterized underdeveloped economies as distinguished by structural heterogeneity; that is, they had both heterogeneous technologies and production functions existing side by side in both industry and agriculture. The structuralist school of thought, associated with the UN Economic Commission for Latin America,1 was pioneered by the Argentinean Raúl Prebisch, who conceived of the international economy as a set of relations between an industrialized center and a periphery exporting foodstuffs and raw materials. Focusing on the problems of the periphery, the school emphasized structural unemployment,2 owing to the inability of traditional export industries to grow and therefore to absorb excess rural population; external disequilibrium, because of higher propensities to import industrial goods than to export traditional agricultural and mineral goods; and secularly deteriorating terms of trade—​all of which a properly implemented policy of industrialization could help eliminate.

4.2.  The Development of Structuralism in Brazil Celso Furtado was doubtless Brazil’s most important structuralist economist. In fact, if judged by the diffusion of his works, Furtado was probably the most influential Brazilian social scientist of the last century. In Latin America, where books are usually printed in editions of one to two thousand copies, Furtado’s works had sold some two hundred

64   Joseph L. Love thousand copies in Spanish and Portuguese by 1972. World sales of his works had reached a million copies by 1990, and half these books were published in Latin America (Furtado 1973). He was the most original and prolific of the structuralist writers in Brazil. Furthermore, he was one of the first of the dependency analysts in Latin America, and the first specifically to assert that development and underdevelopment were part of the same process of the expansion of the international capitalist economy. Yet Furtado did not introduce structuralism in Brazil, nor did Prebisch; that distinction belongs to François Perroux, the French economist who created the structuralist school of economics (Castro and Lessa 1973). Perroux defined structures as “the proportions and relationships that characterize an economic ensemble in space and time” (Perroux 1939, 194). From Portugal, where he had taught and published in the mid-​1930s, Perroux proceeded to São Paulo, Brazil, where he resided in 1936–​1937. At the time he was a member of the French mission of young but distinguished scholars who founded the University of São Paulo, among them Fernand Braudel and Claude Levi-​Strauss. And a more direct impact of French intellectual traditions on Brazilian economics occurred through their influence on Furtado when the latter took a PhD in the Faculté de Droit in Paris in 1948. Because so much of his early work was done in direct or indirect association with CEPAL and Raul Prebisch, Furtado’s opus forms the strongest link between the continent-​wide school of structuralism and the Brazilian national school, which he founded. Indeed, it is difficult to separate some of Furtado’s earliest contributions from those of Prebisch, and in the 1970s their views began to converge again, focusing on the consumption patterns of Latin America’s upper strata as the motor force in the allegedly non-​accumulating, dependent economies of the region. Throughout their careers, both men believed the state was the leading force in economic development, and thought it could provide the leadership that market signals, feeble or distorted by monopoly in backward economies, could not. Like Prebisch, Furtado was by profession a public servant, who, over the course of his career, was intermittently associated with both his national government and international organizations. And like Prebisch, Furtado was a “nonpartisan politician,” in the phrase of the Brazilian economist Francisco de Oliveira (1983, 14).3 Furtado, again like Prebisch, hailed from a remote region of his native country, a periphery of the periphery—​and for the Brazilian, his native region would remain a lifelong focus of writing and action. He lived in Brazil’s backward Northeast until age 20, and spent his childhood in contact with the arid and violent sertão of the small state of Paraíba, where his father was a judge. Like Prebisch, who was almost 20 years his senior, Furtado would pursue his university studies not in the provinces, but in the national capital. He arrived in Rio de Janeiro in 1940 to enroll at the University of Brazil. At the time, economics was not yet a permissible specialization there, and Furtado chose the traditional curriculum in law (Furtado 1973, 28–​32). However, the young man did have contact with the French economist Maurice Bye, a disciple of François Perroux; Bye had been teaching in Rio at the time of Germany’s defeat of France in June 1940, and remained in Brazil until 1942 (Furtado 1985, 18, 27). While at the university, Furtado

Brazilian Structuralism   65 switched from law to public administration, and at age 23 entered the Brazilian civil service. He soon joined Brazil’s expeditionary force in Europe, where he served as an officer in the Italian campaign (Furtado 1989, 14). After the war Furtado made his way to Paris, where he began his study of economics in 1946, working under Bye, his thesis advisor, and Perroux. At the time, Furtado was still an autodidact in economics (Furtado 1985, 14). He had come to that discipline by a circuitous intellectual route: from law to organization and administration, from organization to planning, and from planning to economics. He later wrote, I regarded planning as a social technique of the first importance, capable of increasing the degree of rationality of the decisions governing complex social processes by preventing the setting in motion of cumulative and irreversible processes in undesirable directions. (Furtado 1973, 32)

Given his penchant for planning and his apprenticeship under Bye and Perroux, Furtado came into contact with Perroux’s structuralism-​in-​formation (and a retreating corporatism4) before Prebisch had worked out his initial version of center-​periphery relations. But Furtado, more than Prebisch, was also interested in economic history. In 1948 Furtado presented a dissertation at the Faculté de Droit in Paris on the Brazilian economy during the colonial era (Furtado 1948). As Mauro Boianovsky shows, Furtado linked structure and history using economic models to interpret successive historical structures. In this regard he was influenced by Claude Levi-​Strauss, who had been a member of the Universidade de São Paulo (USP) mission and who adopted the Swiss linguist Ferdinand de Saussure’s structural distinction between synchrony and diachrony. Furtado held that Prebisch’s center-​periphery model was synchronic, whereas Furtado’s historical approach was diachronic, tying structure to historical change (Boianovsky 2015). In the second edition of Development and Underdevelopment (1971), Furtado added the subtitle A Structural View of the Problems of Developed and Underdeveloped Countries; this is one of the earliest instances of linking the CEPAL school with the term “structuralism.” Underdeveloped structures, wrote Furtado in 1961, are those in which full utilization of capital is not a sufficient condition for the complete absorption of the labor force at a level of productivity corresponding to the technology of the modern sector of the economy (Furtado 1964, 141). The main characteristic of underdevelopment is therefore technological heterogeneity of the various sectors of the economy, having its roots in economic history (Furtado 1964, 129). Furtado’s is probably the first definition of underdevelopment focused on heterogeneity. The well-​known essay by CEPAL economist Anibal Pinto—​defining underdevelopment in terms of the hybridity of technologies in Latin America—​only appeared in 1970.5 After completing his doctorate in 1948, Furtado returned to his native land, where he was employed by the Ministry of Finance to help produce Conjuntura Economica (Business Cycle), a new journal associated with the Fundação Getúlio Vargas,6 to which he had contributed while in Europe. Through an introduction provided by Otavio

66   Joseph L. Love Bulhões, a cofounder of the Fundação with Eugenio Gudin, Furtado later in 1948 joined the staff of the UN Economic Commission for Latin America, with which he would be associated for a decade. But also important in the development of Furtado’s economics was the work of John Maynard Keynes. Though Furtado had studied Keynes in France, the English economist had little influence on the Brazilian’s doctoral thesis, and it may be that Furtado’s major exposure to Keynes came through Prebisch. The latter arrived in Santiago in February 1949—​after Furtado—​but had published his Introduction to Keynes two years earlier. In any event, the influence of Keynes was obvious in Furtado’s first essay in economics, “General Characteristics of the Brazilian Economy,” written in 1949 and published the following year (Furtado 1950). Furtado worked under Prebisch, who published the Spanish version of his “manifesto”—​The Economic Development of Latin America and Its Principal Problems—​ in 1949, and the Brazilian soon became one of the older man’s leading collaborators. Furtado prepared the data, though he denies credit for the accompanying analysis,7 for the Brazilian section of the famous Economic Survey of Latin America for 1949. Prebisch and Furtado worked in tandem to marshal Brazil’s government behind CEPAL. They received critical support from Getúlio Vargas in 1951, his first year as a popularly elected president, to make CEPAL a permanent UN agency (Furtado 1985, 120–​122; Magarinos 1991, 140). The two economists also courted industrialists, participating in the debates of the National Confederation of Industries (CNI) in 1950. The organization and many individual manufacturers received Prebisch’s thesis warmly (Furtado 1985, 106). In the same year, Estudos Economicos (Economic Studies), the CNI journal, ran an article explaining and implicitly endorsing CEPAL’s position, and in 1953 the industrialists’ Confederation financially supported a regular CEPAL session in Brazil (Confederação Nacional das Industrias 1950; Sikkink 1988, 406; 1991, 155). A later CNI review, Desenvolvimento e Conjuntura (Development and the Business Cycle), founded in 1957, endorsed CEPAL’s interpretations and proposals in its first editorial.8 In general, industrial leaders in Furtado’s Brazil accepted state intervention and the “developmentalist” ideology associated with structuralism in the 1950s much more readily than did their counterparts in Prebisch’s Argentina (Sikkink 1991, 154–​157). President Getúlio Vargas saw in CEPAL’s program a key to greater national autonomy.9 CEPAL’s emphasis on the creation of infrastructure as the basis for industrialization was for him persuasive, and CEPAL helped train Brazilian economists and técnicos. Moreover, in the absence of PhD programs in economics before the mid-​1960s, CEPAL provided basic training in formal economics and planning (Haffner 2002, 44, 52; Ianni 1971, 131). That fact tended to signify a “depoliticization” of planning techniques, realized by technical experts, state economists, and by CEPAL itself. Thus, state participation in the economy in Brazil was increasingly seen as a guarantee, rather than a risk, for investors (Haffner 2002, 45). President Vargas’s government hosted the fifth plenary meeting of CEPAL in Petrópolis in 1953. On that occasion Celso Furtado presented a “Sketch of a Program of Development for Brazil,” a document presented by the joint team of CEPAL and the Banco Nacional de Desenvolvimento Econômico (Brazilian national development bank), known as the Grupo Misto CEPAL-​BNDE.

Brazilian Structuralism   67 Vargas’s enthusiasm for CEPAL’s ideas was matched by that of Brazilian President Juscelino Kubitschek (1956–​1961), but with a different emphasis—​putting growth ahead of independence from the international trading system. With his slogan “Fifty years of progress in five,” Kubitschek announced in his 1956 presidential address that CEPAL and the BNDE had devised a five-​year development plan. In carrying out the plan, Furtado—​an economist who had worked alternately in the Brazilian government and CEPAL—​figured importantly as head of the new regional development commission for Northeast Brazil (SUDENE).10 Kubitschek endorsed CEPAL’s analysis of deteriorating terms of trade and its interpretation of the persistent disequilibrium in Brazil’s balance of payments. The Brazilian president asserted the need for industrialization to absorb surplus labor in agriculture, and embraced CEPAL’s programming techniques.11 In the early phase of the Kubitschek government, Furtado took an academic sabbatical. Following an invitation from Nicholas Kaldor, in 1957 Furtado received a Rockefeller fellowship to work at King’s College, Cambridge, where Keynes once held court, and where other luminaries of macroeconomics, such as Kaldor, Richard Kahn (the inventor of the Keynesian multiplier), and Joan Robinson (the oligopoly theorist), then taught. Furtado left the United Nations after 10 years of service (much of it spent in Brazil) and at Cambridge wrote his best-​known book, The Economic Growth of Brazil (Furtado 1959a).12 The work would quickly take its place beside a handful of others as a classic study of Brazilian economy and society (see later discussion). Moreover, it opened a debate on the character and rhythm of Brazil’s industrialization that has yet to see a resolution. Returning home in 1958, Furtado became a director of the BNDE, with special responsibility for the poverty-​and drought-​stricken Northeast, his native region. At the outset of the new year, he was invited to a brainstorming session with President Juscelino Kubitschek, who was presiding over an unprecedented wave of economic expansion. Furtado persuaded the president to set up a special development program for the Northeast (Furtado 1989, 44–​45), and this endeavor led to the creation of a permanent regional autarky, the Superintendency for the Development of the Northeast (SUDENE). He also was among those who successfully urged Kubitschek not to introduce recessionary measures recommended by the International Monetary Fund to contain inflation (Furtado 1989, 70–​73).13 At the turn of the decade, the Northeast was the focus of political unrest and rural union organization, magnified in importance by international news media, which associated the Peasant League spokesman Francisco Julião with the Cuban Revolution. Furtado himself was frequently called a communist, and was investigated by Brazil’s National Security Council.14 After SUDENE was approved by congress at the end of 1959, Kubitschek named Furtado its first director; he was reappointed by Presidents Janio Quadros (1961) and Joao Goulart (1961–​1964). Under Quadros, Furtado’s position was raised to cabinet rank, and the younger man would soon have to do battle with the emerging development scheme of the Alliance for Progress, whose North American directors had a different set of priorities for the Northeast (Furtado 1989, 130). Despite his intense administrative and political activity, in 1961 Furtado published one of his

68   Joseph L. Love major works, Development and Underdevelopment, which would form the bridge between structuralism and dependency. By 1963 Furtado began to champion agrarian reform as necessary for the progress of the Northeast (Furtado 1989, 147). He found himself in an increasingly polarized situation, as a (perceived) radical. Furtado had also been appointed extraordinary minister of planning, and in 1962 he prepared a three-​year development plan that is better remembered for the orthodox short-​term anti-​inflation measures it contained than its long-​term structuralist features. When the coup d’état ousting Goulart came at the end of March 1964, Furtado was immediately dismissed from his post at SUDENE. Under the military dictatorship of Marshal Humberto Castelo Branco, he quickly lost his political rights as well. Furtado then began a long exile. In 1964 he briefly worked in Chile, where he contributed signally to the emergence of dependency analysis; after a year at Yale University, he returned to his alma mater, the University of Paris, as professor in 1965. He retained his post there, despite protracted stays in Brazil, beginning in 1975. He sophisticated his analysis of dependency and applied it in a region-​wide context in The Economic Development of Latin America (Furtado 1969). In 1985, as Brazil returned to a civilian and constitutional regime, President José Sarney named Furtado the Brazilian ambassador to the European Economic Community. He returned to Brazil to serve as minister of culture from 1986 to 1988. Throughout the 1980s Furtado continued to plead for public support and funds for the development of his beloved and scourge-​ridden Northeast. The preceding sketch of Furtado’s career should suffice to demonstrate that engagement in politics and policy formation did not preclude a fecund scholarly output, our main concern. All three major thrusts of Latin American structuralism—​the tendencies within the periphery toward unemployment, owing to structural (technological) heterogeneity; disequilibrium in the foreign sector; and deteriorating terms of trade—​were developed inventively in Furtado’s early work (Eatwell et al. 1987, 529). Following the publication of Prebisch’s The Economic Development of Latin America and Its Principal Problems (1949), which Furtado translated into Portuguese the same year, the Brazilian quickly drew further conclusions from Prebisch’s analysis of the business cycle and high import coefficients. Furtado argued that income tended to concentrate in Brazil during the upswing of the cycle, owing in part to a highly elastic labor supply that held down wages; in this manner, Furtado adumbrated by four years W. Arthur Lewis’s celebrated analysis of an infinitely elastic labor supply as the source of wage rigidity in underdeveloped countries. Further, the Brazilian hypothesized that much of the effect of the Keynesian multiplier “leaked” abroad, owing to the exporting groups’ high propensity to import (Furtado 1950, 11; Lewis 1954). Such analysis pointed again to the importance of an industrialization policy.15 The Brazilian economy during the Depression years, in Furtado’s view, made a decisive shift to growth based on industrialization for the domestic market, rather than on exports (Furtado 1985, 27–​28).16 Prebisch had noted earlier that the most economically advanced Latin American nations had made their largest strides toward

Brazilian Structuralism   69 industrialization when the international economy was in crisis, but Furtado supplied the term—​development by “external shocks.”17 For Furtado, a problem of equal importance with the small domestic market was the absence of an expanding foreign market—​an implication of the doctrine, or assumed fact, or deteriorating terms of trade for commodity exporters in world commerce (Furtado 1954a, 126). He wrote, “the aim of economic development must be to increase the physical productivity of labor,” and to do so beyond the export sector, where it was typically higher than elsewhere in the early stages of underdevelopment. Like other economists of the period, Furtado believed that labor’s physical productivity was “in the main” due to capital accumulation (Furtado 1954a, 127, 130). Such accumulation was difficult to achieve in the export sector, where the benefits of an increase in the physical productivity of labor might suddenly be transferred abroad by a fall in commodity prices. The export sector was also important on the demand side, especially when income was concentrated in a small group. This group emulated the consumption patterns of inhabitants of developed countries, and demand might not diversify, owing to a sharply skewed distribution of national income (Furtado 1954a, 132–​133). In fact, a similar point had been made by Ragnar Nurkse, the former League of Nations economist now at Columbia University, based on an extension of James Duesenberry’s interpretation of Keynes’s “consumption function”:  the Harvard economist had established that, from one period to another, the percentage of income consumed in the United States depended not on absolute income, but by consumers’ positions in the income pyramid; thus the consumption function was stable over time for a given stratum of income receivers. This fact was accounted for by a “demonstration effect” of the wealthiest groups on those of lesser affluence. Nurkse extended Duesenberry’s concept not to explain “size-​distribution” (roughly, class) behavior from one generation to the next in developed countries, but rather the emulation of consumption patterns of such nations by the upper classes in underdeveloped countries. Furtado endorsed this extension. He wrote that Brazil lacked “not incentives to invest, but incentives to save,” a problem exacerbated by “powerful stimuli to consume exercised by more advanced economies . . .” (Furtado 1954a, 134, 138, 144). The notion of “demand-​driven” and “demand-​distorted” patterns of growth would later gravitate to the center of Furtado’s vision of underdevelopment. At the time, Furtado believed in the power of fiscal policy to shape economic development, and for the former this included some (unspecified) form of compulsory savings. The state, Furtado contended, could more effectively ensure an adequate rate of savings and investment, rather than the conspicuous consumption-​oriented private sector (Furtado 1954a, 144; Bielschowsky 1985, 203). Furtado, a professional civil servant, saw in economic planning the advantages that had been proclaimed by Paul Rosenstein-​Rodan and Nurkse, namely, that because of indivisibilities in large-​scale investments and potential external economies not realizable by private firms, a state-​directed development effort was necessary:  Private investors would maximize the private, not the social, marginal net product. In underdeveloped countries, where markets did not function properly, the price mechanism was not a reliable guide for investment; in those areas, the utilization of factors of production

70   Joseph L. Love (as expressed, for example, by wages) varied widely across sectors. The rate of development could be accelerated, Furtado confidently believed, if investments were made “according to a co-​ordinated comprehensive plan” (Furtado 1954a, 139). Beginning in 1955, Furtado and Prebisch worked for official acceptance of “programming,” the CEPAL version of planning, in Brazil. CEPAL’s programming used techniques similar to Vassily Leontieff ’s input-​output analysis—​starting from a given rate of growth to be achieved by the economy, and estimating the necessary structural changes needed, as well as specific inputs to meet the target. Echoing Arthur Pigou and Rosenstein-​Rodan, but citing Keynes, Furtado reasoned that the interest of the entrepreneur and the collective interest do not always coincide. State investment had to focus on “strangulation points” (bottlenecks) and “germination points” (growth poles), which would diminish sectoral disequilibrium, a basic feature of underdevelopment (Furtado 1958a, 39, 42–​44).18 But planning was a matter of degree, and Furtado’s BNDE-​CEPAL commission advocated a comprehensive form in 1957; this proposal contrasted with the more limited and bottleneck-​oriented planning contemporaneously defended by Roberto de Oliveira Campos, head of the BNDE and a highly influential figure in the Kubitschek government (Bielschowsky 1985, 217). If Campos was more influential in formulating the administration’s Target Plan than Furtado, the president (1956–​1961) largely embraced the CEPAL analysis of underdevelopment, as shown earlier. Furtado later wrote that the government’s “Target Program” (Programa de Metas) was directly inspired by CEPAL, and a student of economic thought in Brazil credits Furtado himself with introducing programming in the country (Furtado 1970, 208; Bielschowsky 1985, 182). As indicated earlier, Furtado was the leader in the effort to historicize structuralism.19 True, Prebisch’s Economic Survey . . . 1949 had treated in brief the economic history of Latin America as a whole from the 1880s to the mid-​twentieth century, and had considered individually the four most industrialized nations—​Argentina, Brazil, Chile, and Mexico. In some respects, this volume was a model for the country case studies to be carried out between 1959 and 1963—​Furtado on Brazil, Anibal Pinto on Chile, Aldo Ferrer on Argentina, and later, Osvaldo Sunkel and Pedro Paz on the whole region, as well as Villareal on Mexico.20 But Prebisch’s interest was focused on the business cycle, not long-​term historical development. Furtado’s Economic Growth of Brazil derived principally from his pre-​CEPAL interests in defining the features of colonial Brazil. Although The Economy of Colonial Brazil, his pre-​CEPAL dissertation at the Faculté de Droit, University of Paris (1948), does not contain much formal economic analysis of any kind, The Brazilian Economy (1954b) and A Dependent Economy (1956), a briefer work, are structuralist treatments of Brazil’s economic history.21 These early historical essays offer evidence that Furtado’s contribution precedes Pinto’s, even though their “classic” studies both appeared in 1959—​Chile, A Case of Frustrated Development and Economic Growth of Brazil. The latter work covers the entire sweep of Brazilian history, and the colonial and nineteenth-​century sections compare and contrast the structures of the Brazilian and US economies, showing how Brazil’s monoculture and latifundia impeded the high savings and investment rates

Brazilian Structuralism   71 characteristic of the American economy. Focusing on the distribution of income and the size of the domestic market, Furtado provides one of the first uses of modern income analysis in a historical framework, and demonstrates the weak relationship between income and investment in an economy based on slavery (Baer 1974, 115). The work throughout is written from the point of view of a development economist, emphasizing the heterogeneity of technologies and production functions (including the vast subsistence sector) in the Brazilian economy. Turning to the problem of Brazil’s economic cycles, examined in earlier studies by Roberto Simonsen (1937) and João Normano (1935), Furtado saw in the weak monetization of the slave economy a kind of resilience, in that export stagnation or decline could be sustained as the free but plantation-​oriented population moved toward the backlands: the subsistence economy absorbed the excess labor supply after the exhaustion of successive export booms. In a slave-​based economy the response to depression is different from that of a fully capitalist economy; in the former, “entrepreneurs” have fixed costs (maintaining their slave populations) and are not in a position to contract their agricultural output. For example, when the sugar economy declined in the seventeenth century, the livestock economy became increasingly subsistence-​oriented, and average labor productivity, by inference, fell (Bielchowsky 1985, 243; Furtado 1963, 69–​7 1). This economic “involution,” as Furtado called it, was the opposite of development, since each historical export boom until coffee (brazilwood, sugar, gold, and—​ contemporaneous with coffee—​rubber) led to retrogression, not to sustained growth.22 The apparent aberrations in Brazilian financial policy in the period since independ­ ence could be explained in part by the fact that the structure of “dependent economies” was different from that of the industrial economies. In times of depression, the former suffered plummeting export prices, worsening terms of trade, a reduction in capital inflow, together with rigid requirements in foreign capital servicing. In trying to adhere to the gold standard, Brazilian statesmen had failed to understand the nature of their predicament and viewed their nation’s inability to keep to the standard as the result of bad management, rather than a problem with deeper causes (Furtado 1963, 174–​177). Differences in the growth and diversification of the production structure of the Brazilian and US economies in the first half of the nineteenth century were not accounted for by the greater degree of tariff protection in the United States, Furtado believed, but by the differences in social structure and income distribution, and therefore the size of the domestic market. In fact, Furtado estimated that Brazil’s continually falling exchange rate provided more protection for domestic industries than high tariffs would have (Furtado 1963, 107–​108). But more important, Brazil suffered from a small domestic market, lack of modern technology, entrepreneurship, and capital, and its small capacity to import (Bielschowsky 1985, 241). For Furtado, Brazil’s national market dated from the last quarter of the nineteenth century, when a modern working class came into existence. Beginning in the late 1880s, when wage labor replaced slave labor in São Paulo’s coffee fields, Brazil began to experience a significant home market. For Furtado, wages paid in the coffee sector permitted the “nucleus of a domestic market economy,” with the implication of an attendant multiplier effect (Furtado 1963, 167).

72   Joseph L. Love The big change in relative market size, however, occurred after the crisis of 1929, in which the coffee economy, which had risen to 70% of the value of national exports, collapsed. In Furtado’s estimation, the decisive shift toward an economy based on the stimulus of domestic demand took shape in the early 1930s. The American economist Werner Baer has noted that Furtado’s analysis of events in the Great Depression only accounts for 7% of the space in Economic Growth of Brazil, but it is the theme of the book that has generated by far the greatest amount of scholarly controversy (Baer 1974, 119). We have already seen that Furtado’s earliest treatment of the issue, the thesis in embryo, goes back to his “General Characteristics” in 1950.23 Furtado shared the view of Prebisch and CEPAL in the Economic Survey . . . 1949 that industrialization had historically occurred in periods of crisis in the larger Latin American economies. Elaborating on his analysis of 1950, Furtado pointed to Brazil’s rapid industrial growth during the Great Depression, partly due to the “socialization of losses” of coffee producers through exchange devaluation. This process helped maintain domestic demand by keeping up the employment level and purchasing power in the coffee sector, which permitted the rise of a significant domestic demand for industrial goods when foreign products were unavailable, owing to the absence of foreign exchange. The stockpiling and destruction of coffee in the face of grossly excess supply were financed through credit expansion, which in turn exacerbated the external disequilibrium and caused new exchange depreciation and a further socialization of losses (Furtado 1963, 205–​206). Furtado viewed the expansionary fiscal and monetary policies related to coffee as a form of unwitting Keynesianism, because the wealth destroyed in coffee beans was considerably less than that created by maintaining employment (Furtado 1963, 211). Furtado then noted that output of capital goods in Brazil by 1932 was 60% greater than in 1929. Furthermore, net investment in 1935, at constant prices, was greater than that in 1929, and the level of aggregate income of the latter year had been regained, despite the fact that the import of capital goods was only half of the 1929 figure (Furtado 1963, 218–​219). Therefore, the economy was undergoing profound structural change. Furtado’s views on Brazilian industrialization in the Depression touched off a long debate.24 In this exposition, it seems clear that Furtado was influenced by his Keynesian background, especially with regard to government intervention to sustain demand, and the significance of the domestic market in dynamizing production and income (Bielschowsky 1985, 191). Yet it now seems the case for Brazil, as for the other more industrialized countries in Latin America, that the world wars and the Depression were less important in producing “inward-​directed growth,” in Prebisch’s phrase, than was believed by some contemporaries to these events, and by CEPAL economists later.25 A now widely held view is that investment in industry (capacity) grew in line with export earnings for the period 1900–​1945, while output (but not capacity) tended to rise during the “shocks,” when imports had to be curtailed. Capacity could not grow appreciably during the Depression for lack of exchange credits to buy capital goods and inputs, nor during the world wars because of the unavailability of capital goods and fuels from the belligerent powers.26

Brazilian Structuralism   73 In addition to historicizing structuralism, Furtado explored the school’s potential in another direction, as did Hans W. Singer, who in 1950 had developed a model of the international trading process similar to Prebisch’s (Singer 1950, 473–​485). I refer to the problem now known as “internal colonialism.”27 Furtado and Singer independently built their analysis in the 1950s around perceived unequal exchange between industrial centers and agricultural peripheries. I will focus on the Furtado version, which was published first and in a fuller form, though Singer’s work was completed earlier.28 It was in the context of analyzing internal colonialism that Furtado first began to link development and underdevelopment as components of a single historical process. The model of the international trading process on which Singer and Furtado drew was that of CEPAL (formulated by Prebisch in 1949) and Singer’s very similar one, independently arrived at (Singer 1950). According to Prebisch and Singer, at the international level unequal exchange derived from differential productivities between industrial center and agricultural periphery in the world market, combined with different institutional arrangements in capital and labor markets. Technological progress in manufacturing, in any case, was shown in a rise in incomes in developed countries, while that in the production of food and raw materials in underdeveloped countries was expressed in a fall in prices relative to industrial goods. For Singer, the explanation of contrasting effects of technological progress was found in the disparate income elasticities of demand for primary and industrial goods. That is, the demand for agricultural and mineral goods rose less than proportionally to the rise in world income, and that for manufactured goods rose more. Furtado addressed the issue of internal colonialism in the late 1950s, as he became more deeply involved in the problems of his native Northeast. This agrarian and latifundium-​dominated region in 1956 had an annual per capita income of less than US $100, whereas the Center-​South enjoyed a level of income more than three times higher, because of the dynamic industrial economy organized around the cities of São Paulo and Rio de Janeiro. The gap between the Northeast and the Center-​South was larger than that between the per capita income of the latter region and those of Western Europe (Brasil 1959, 7, 14). Furtado estimated that the ratio between the growth rates of the lagging and leading regions was of the order of one to two for the decade after 1948 (Brasil 1959, 7). Moreover, the distribution of income within the Northeast was highly skewed, making the situation even more desperate for the masses. Like Prebisch, Furtado assumed the existence of market imperfections—​particularly the administered pricing of industrial goods—​and a virtually unlimited supply of labor in the backward region at the going wage in the industrial sector. But the Brazilian’s model was more complex than Prebisch’s international one, because it purported to measure the deterioration of terms of trade between the international price of agricultural goods sold abroad by Northeast Brazil against the domestic price of industrial goods, which the region had to buy from the Center-​South. Furtado analyzed the problem of the Northeast in terms of a triangular trade between the backward region, the foreign sector, and the developed area of Brazil (Brasil 1959, 22). Brazil’s Northeast had a surplus in its commercial balance abroad, but a deficit in its

74   Joseph L. Love balance of payments with its domestic trading partner, the Center-​South. The state was also an essential element in the trading process: in implementing its policy of import-​ substitution industrialization, the central government was subsidizing industrialists and penalizing agricultural exporters. This support took the form of differential exchange rates for importers of manufacturing-​related capital goods and importers who would use foreign exchange credits for other purposes.29 That the central government gave exporters poorer exchange rates than importers not only effected a sectoral transfer of income; the same action induced a regional transfer as well, because of the size of the export sector relative to real (national) income in the Northeast, compared to that of the Center-​South. Furthermore, the government stimulated industrial development by financing private enterprise, a process that principally aided the Center-​South. Finally, economies of scale and external economies in the industrial heartland of the Rio–​São Paulo area made the hitherto large industrial advantages of the region, relative to the Northeast, even greater as development proceeded. Therefore central government policies designed to stimulate industrialization had a major inequalizing effect on the regional distribution of income in the country. Furtado estimated that during the period 1948–​1956, the Northeast transferred US $24 million annually to the Center-​South, although a more accurate figure may be $15–​17 million yearly.30 Because of Brazil’s protectionist tariffs and related exchange policies, the Northeast was in no position to seek alternative supplies abroad for its manufacturing needs. It offered a captive market for the Center-​South, and its foreign exchange earnings gave it purchasing power in that region. But the relevant terms of trade now entered the picture: overall, prices of the South’s industrial goods rose more rapidly from 1948 to 1956 (the years studied by Furtado) than the exchange rate fell, that is, the rate at which Northeastern exporters gained more cruzeiros per unit of foreign currency (Brasil 1959, 27–​28). Furtado proposed industrialization as a solution to the Northeast’s economic problems (Brasil 1959, 49). He also stressed the need for agricultural development, implying the need for agrarian reform, because the cost of wage-​goods (i.e., foodstuffs) in the largest city of the Northeast, Recife, was rising faster than that of São Paulo. Consequently, if wage differentials were narrowing between São Paulo and Recife to meet rising costs of living in the latter, there would be little incentive for private capital to invest in the Northeast (Brasil 1959, 59; Furtado 1959b, 37).31 Large-​scale agrarian reform had yet to occur, however, and in the years following Furtado’s study, development strategies favoring the Center-​South continued. Despite efforts of the federal government to offset regional income concentration, Baer has concluded, the overall effect of development programs continued to favor the industrial Center-​South over the agrarian Northeast in the three and a half decades following Furtado’s analysis (Baer 1995, Chapter 12). In his two regional studies of 1959 surveyed in the preceding, Furtado had already perceived a relationship that he, Osvaldo Sunkel, Fernando Henrique Cardoso, and Andre Gunder Frank would develop in the mid-​1960s: that a structural and perverse relation existed between the growth of developed capitalist economies (and regions) and the growth of underdeveloped countries (and regions): “[There exists] . . . a tendency for

Brazilian Structuralism   75 industrial economies, as a result of their form of growth, to inhibit the growth of primary economies: This same phenomenon is occurring within our country” (Furtado 1959b, 13). It is notable for the history of dependency analysis that Furtado’s first published statement of the alleged causal relationship between development and underdevelopment appeared in the context of internal colonialism, rather than at the international level.32 The late 1950s and early 1960s were extremely productive for Furtado as a theorist of development. Not only did he write The Economic Growth of Brazil and develop a thesis on internal colonialism avant la lettre,33 but he wrote the essays that would constitute his book Development and Underdevelopment (Furtado 1961), which has probably been more influential than any of his other works, excepting Economic Growth of Brazil. It is notable that except for a year’s research at Cambridge in 1957–​1958, Furtado in this period was intensely involved in work for CEPAL in Brazil and other Latin American countries, followed by service in high positions in the Brazilian government from 1958 through 1964. In one essay in Development and Underdevelopment (a collection published in 1961), Furtado reviewed the major theories of growth, beginning with the classical British economists. He wanted to revive the use of the classical concept of economic surplus—​ that portion of the social product beyond the level required for the maintenance and reproduction of members of a society and their goods and services (i.e., total revenue less total costs).34 In the same essay, Furtado proceeded to critique neoclassical economics for its emphasis on equilibrium (as opposed to growth), while also highlighting the school’s ideological features. For instance, the idea of surplus had been discarded in the neoclassical synthesis between 1870 and 1890. For this school, “the social product was to be conceived in terms of ‘costs of factors’ [land, labor, and capital], savings being viewed no longer as a result of an existing surplus but as the result of an act of restraint or abstinence” (Furtado 1961, 79). In retrospect, it can be said that this work established Furtado as perhaps the first analyst of dependency. His passing reference to the problem in Operation Northeast (1959) had been explicit, but in Development and Underdevelopment he described how the European industrial economy by the nineteenth century had penetrated and transformed precapitalist economies. Underdeveloped economies were “hybrid structures,” and not simply undeveloped economies beginning to trace the path that Europe had already defined. Consequently, underdevelopment was a “discrete historical process through which economies that have already achieved a high level of development have not necessarily passed” (Furtado 1961, 129).

4.3.  Crisis in Import-​Substitution Industrialization The reformism of the 1960s was conditioned by and, for an increasing number of structuralists, made irrelevant by a long evolution of CEPAL’s views on its initial key

76   Joseph L. Love policy recommendation: import-​substitution industrialization (ISI). An ISI policy had seemed a brilliant success, especially in Brazil and Mexico, during the 1950s, but it was a success owed in part to unusually high commodity prices during the Korean War. In the latter 1950s, CEPAL began to consider the complexities of ISI. By 1957 the organization had distinguished between two types, which in the 1960s would be seen as phases, of import substitution. The first involved the relatively easy substitution of simple domestically produced consumer goods for previously imported items. The second, more difficult, type involved the production of intermediate goods and consumer durables, a shift from “horizontal” to “vertical” ISI—​so denominated because of the substitution of simple goods on a broad front in the first phase, and in the second, an integrated line of production of fewer final goods and their inputs. A third phase, the production of capital goods, would ensue at a later date (CEPAL 1957, 116; 1966, 19–​20; Rodriguez 1980, 202–​203). In 1956 CEPAL had still assumed the existence of a threshold in structural changes in the economy, beyond which “dependence on external contingencies” would diminish. Yet the following year the agency first suggested that dependence on “events overseas” might even increase as ISI advanced; all the same, it still held that “import substitution” consisted of lowering “the import content of supplies for the home market.”35 Argentina was Latin America’s most industrialized country, and despite its unique political phenomenon of peronismo, CEPAL tended to view it in 1957 as a trend-​setter for other Latin American nations. Argentina, CEPAL noted, had reduced its imports of finished goods to one-​third the total dollar amount. Yet its declining capacity to import had meant that reducing the importation of consumer goods was not sufficient to contain balance-​of-​payment difficulties; capital goods and fuels also had to be reduced, and this fact was reducing the rate of growth. Chile was seen as facing similar, though less dire, problems, and CEPAL seemed to wonder aloud whether the Argentine experience was the future of Latin America. Two conclusions followed: that primary exports and food production for domestic consumption had to be increased (the latter to relieve pressure on imports), and that a region-​wide common market must be developed to assure the future development of efficient manufacturing industries.36 Why should industrialization bring rising import requirements in its train? Using a simple two-​sector model, Furtado in 1958 explained the problem as one in which, by assumption, the advanced sector, A, had a larger import coefficient than the backward sector, B. As the economy developed, A’s coefficient grew ever larger as a share of the whole economy’s coefficient, and pari passu the average import coefficient tended to rise (Furtado 1958b, 406). If the terms of trade were deteriorating, the pressures on the balance of payments became even more acute. Thus, for CEPAL economists in the mid-​and late 1950s, the import requirements in the later stages of ISI, unless offset by capital inflows or rising exports, could cause “strangulation”—​a favorite CEPAL metaphor for stagnation caused by insufficient imports of capital goods and other industrial inputs. As a partial solution to stuttering ISI, CEPAL in 1957–​1958 formally appealed to its sponsoring states for a Latin American

Brazilian Structuralism   77 common market, which, CEPAL held, would provide incentives (through economies of scale) for the production of capital and intermediate goods (CEPAL 1958, 4).37 Yet in its early years the Latin American Free Trade Area, established in 1960, was only an expression of hope for alleviating the ills associated with ISI. Already in 1959 Prebisch had observed that the more economically advanced Latin American countries were becoming increasingly the hostages of external events, because they had compressed their imports to the absolute essentials for the maintenance of growth. Two years later he wrote, “It remains a paradox that industrialization, instead of helping greatly to soften the internal impact of external fluctuations, is bringing us a new and unknown type of external vulnerability” (Prebisch 1959, 268; 1961, 5). The agonizing reappraisal of ISI came in 1964. In that year a CEPAL study, though blaming Latin America’s declining rates of growth on deteriorating terms of trade in the 1950s, also noted that 80% of regional imports now consisted of fuels, intermediate goods, and capital equipment. Consequently, there was little left to “squeeze” in the region’s import profile to favor manufacturing (CEPAL 1964, 14, 21). Meanwhile, two monographs, highly critical of ISI, appeared in the agency’s Economic Bulletin—​one on the Brazilian experience in particular, and the other on Latin America in general.38 These articles pointed to problems that by the 1960s were beginning to affect other parts of the Third World as well. Examining the Brazilian case in the 1950s and early 1960s, the Brazilian structuralist Maria da Conceição Tavares argued that ISI had failed because of the lack of dynamism of the export sector, coupled with the fact that ISI had not diminished capital and fuel import requirements. Other problems were apparent ceilings on the domestic market, owing in part to highly skewed income distribution, which also determined the structure of demand; the constellation of productive resources—​for example, the lack of skilled labor; and the capital-​intensive nature of industrialization in more advanced phases of ISI, which implied little labor absorption. In the advanced stages of ISI, Tavares contended, the low labor absorption of manufacturing tended to exaggerate rather than to terminate the dualism of Brazil’s economy. Among other things, she argued that bottlenecks in the food supply, partly due to the antiquated agrarian structure, put unsustainable pressures on the import bill. Tavares recommended agrarian reform as a partial solution (Tavares 1964, 7–​8, 11, 12, 55). In the same issue of the Bulletin, Santiago Macario wrote a blistering critique of the way in which ISI had actually been practiced in Latin America, following up Prebisch’s observation the previous year (1963) that the region had the highest tariffs in the world. Macario observed that the governments of the four most industrialized countries—​ Argentina, Brazil, Mexico, and Chile—​had used ISI as a deliberate strategy to counteract a persistent lack of foreign exchange, and to create employment for expanding populations. But in those four countries, and in most of the others of the region, protectionism, primarily in the form of tariff and exchange policies, had been irrational, in that there was no consistent policy to develop the most viable and efficient manufacturing industries. On the contrary, the most inefficient industries had received the greatest protection; there had been over-​diversification of manufacturing in small markets in the

78   Joseph L. Love “horizontal” phase; and these factors had contributed, in some instances, to real dis-​ savings (Macario 1964, 65–​67, 77, 81). Nor did Latin American manufactures hold their own in international markets, continued Macario, at a time when exchange earnings had become critical for the future of industrialization. On the positive side, there were tendencies in the early 1960s to abolish exchange controls, quantitative restrictions, and multiple exchange rates, and a related tendency to begin tariff reduction; yet Macario asserted that Latin American tariffs were still being built on a makeshift basis, resulting in a gross misallocation of scarce resources (Macario 1964, 67, 78, 81). Rational criteria were needed to develop industries—​such as the use of factors in greatest abundance (e.g., labor), or the promotion of industries that could earn foreign exchange. Equally important, thought Macario, was the establishment for each country of a “uniform level of net protection” (1964, 84). Overall, his thesis was less that CEPAL’s policy prescriptions had initially been wrong—​which Tavares’s analysis in some ways suggested—​than that the region’s governments had flagrantly ignored CEPAL’s technical advice, pursuing, in Macario’s (1964, 67) words, “import substitution at any cost.” Though Hirschman suggested four years later that Tavares and Macario had issued the death certificate of ISI somewhat prematurely, other scholars soon added new charges, such as ISI’s having increased the concentration of income with regard both to social class and to region (within countries).39 CEPAL answered this challenge in a several studies and programs, culminating in 1990. The chief articulator of CEPAL’s new theoretical and programmatic thrust was a Chilean economist named Fernando Fajnzylber. He had been head of the foreign trade department in the Allende government and had held posts in several UN agencies, notably the UN Industrial Development Organization (UNIDO) in Vienna. In The Incomplete Industrialization of Latin America (1983), Fajnzylber had compared Latin American industrialization with that of the newly industrialized East Asian Tigers, but had focused on Japan. He had done a study tour there as a Chilean engineering student in 1962, just as Japan was moving into its leading position in industrial exports. In Incomplete Industrialization, Fajnzylber described Japan’s export-​driven industrialization, a model that other East Asian countries were following. It called for selective import-​substitution industrialization, focusing on industries that could open up export markets using advanced technologies. Japan and its followers, Fajnzylber wrote, especially Korea, had learned to adapt, innovate, and compete. In Latin America, he argued that protection should be provided for industries that required apprenticeship in new technologies for both workers and managers—​this, in contrast to the region’s historical pattern of protection, an ad hoc affair that Fajnzylber termed “frivolous protection” (Fajnzylber 1983, 116). In Japan the state was heavily involved in development through the Ministry of International Trade and Industry (MITI), and the government offered subsidies and protection for groups with cutting-​edge technologies that could penetrate foreign markets. Fajnzylber wanted Latin American governments and firms to emulate Japan, combining planning and market-​based export strategies. This would require alliances of Latin America’s still relatively weak industrial bourgeoisie with other actors.40

Brazilian Structuralism   79 In 1990 CEPAL regrouped under the banner of “neostructuralism,” articulated by Fajnzylber in 1990 and enthusiastically endorsed by cepalinos Osvaldo Sunkel and Gustavo Zuleta in the same year. Like the neoliberals, Fajnzylber condemned the excesses of protectionism in Changing Production Patterns with Social Equity (CEPAL 1990)—​a complaint that Prebisch, as noted earlier, had made as early as 1963 in Towards a Dynamic Development Policy—​but Fajnzylber further argued that growth could only be sustained by progressively introducing high-​technology, high-​productivity goods for the international market. The goal was to endogenize technological innovation by establishing research and development traditions in Latin America. But he also called for a greater degree of equity in the distribution of income, consistent with rising levels of productivity in agriculture as well as industry. The two goals of new forms of production and equity . . . must be achieved within the context of greater international competitiveness, based more on the deliberate and systematic absorption of technical progress by the production process (with corresponding rises in productivity) than on the maintenance of low real wages. . . . Emphasis must be placed on the systemic nature of competitiveness. Sustained growth based on competitiveness is incompatible with the continued existence of lags as regards equity. (CEPAL 1990, 14)

Thus, for Fajnzylber, competitiveness resulting from the acceptance of lower salaries was a “spurious” competitiveness. An “authentic” form would be based on apprenticeship in the use of more advanced technology, which, in turn, results in higher wages. Policies that simultaneously promote competitiveness, apprenticeship, and a more equal distribution of income, he contended, are often complementary.41 New technology, of course, is not entirely a public good, because of patent rights and royalties; moreover, the introduction of new technologies—​almost by definition focusing on saving labor—​also results in job losses. Therefore the state, as in Europe and the United States, should subsidize technological research, including the introduction of new technologies in agriculture as well as industry, to lessen technological heterogeneity throughout the economy. But the state must also participate in “human capital” formation through education and basic scientific research; both will have positive externalities, that is, spillover effects. Because this is a long-​term project, private firms are much less likely to make these investments (Rodriguez 2006, 377–​398). Changing Patterns called attention to the fact that Latin American growth in the 1980s was a negative 8.3%—​thus the phrase “the lost decade.” Accompanying this shrinking product were persistent macroeconomic imbalances, a regressive adjustment in income distribution, a weakening of the public sector, and a decline of capital formation. In addition, Latin America’s share of world exports fell from 7.7% in 1960 to 3.9% in 1988. One of the few bright spots of the period was the rapid growth of exports of manufactures based on research and development over the years 1962–​1985 at an annual average of 8.1% (CEPAL 1990, 20, 21, 40).42 In neostructuralism the state was to serve as regulator, and structural heterogeneity would be replaced by “inclusive growth,”

80   Joseph L. Love as opposed to relying on cheap labor and currency devaluations to advance in international markets. Unfortunately, Fajnzylber did not live to see the maturation of neostructuralism, dying in 1991 at the age of 51. Other Brazilian structuralists contributed to the development and diffusion of structuralism in their country, and they also played important roles in government, especially in the twenty-​first century. Two who collaborated on a widely read textbook on structuralist economics were Antonio Barros de Castro and Carlos Francisco Lessa. Their Introduction to Economics had gone through 38 editions in Portuguese by 2011, and 56 editions in Spanish by 2004. Barros de Castro earned his doctorate at the Universidade Estadual de Campinas (UNICAMP) and taught economics at the Federal University of Rio de Janeiro; he was also a visiting professor at the University of Chile and Cambridge University. President of the National Development Bank (BNDES) under President Itamar Franco for six months in 1992–​1993—​after that of China, the largest development bank in the world—​ Barros served as director of planning for the BNDES in 2005–​2007.43 He wrote a two-​ volume analysis of national economic problems called Seven Essays on the Brazilian Economy, which focused on uneven regional development in agriculture and industry (Castro 1969–​1971). As a planner he was especially concerned with protecting sectors of the national economy under conditions of globalization. Carlos Lessa, who also took his doctorate at UNICAMP and later taught there, directed the division of social programs at the BNDES in the latter 1980s and taught at CEPAL in Chile. He served President Lula da Silva as head of the BNDES in 2003–​2004. As a disciple of Furtado, Lessa defended an active role for government in directing and planning the Brazilian economy. A third structuralist, Guido Mantega, took his doctorate in sociology at the University of São Paulo and taught at the Pontifical Catholic University in that state. He was an advisor to President Luiz Inacio Lula da Silva and helped design the economic program of Lula’s Partido dos Trabalhadores (Workers’ Party). When Lula first took office in January 2003, Mantega became minister of planning, budgeting and management, with a mandate to reduce the bureaucracy. Although structuralism (developmentalism) is now largely a historical phenomenon, Fundação Getúlio Vargas (FGV) professor Luiz Carlos Bresser-​Pereira has sought to revive it in a revised form, bringing structuralism closer to orthodox economics. Bresser faults President Lula da Silva for letting the real’s exchange rate rise from 7 to the dollar in 2003 to 2.20 in 2010. This policy allowed importers to capture the gains of exchange rate policy. He favors a “once and for-​all-​devaluation” for the reason that “markets are excellent coordinators of the economy because they guarantee . . . that prices are right in the competitive microeconomic sector.” This devaluation could be effected by placing an export tax on commodities proportional to their international prices (Bresser-​Pereira 2016, 243, 249, 251). Bresser believes that in developing countries there is a cyclical and chronic overvaluation of exchange rates (the Dutch disease). Like the neostructuralists of CEPAL after 1990, Bresser defines economic development as a process of accumulating capital through the incorporation of technological

Brazilian Structuralism   81 progress that raises labor productivity, salaries, and the standard of living. This policy still requires industrialization as the centerpiece of moving labor from lower to higher value added per capita, but it should focus on the export sector, in line with the neostructuralism of CEPAL. Other economists who updated structuralism became influential formulators of economic policy under Presidents Lula and Rousseff in the twenty-​first century—​Luciano Coutinho, Fabio Erber, and Joao Carlos Ferraz. All played important roles in the BNDES, holding top positions there in the new century. At the institutional level, structuralism remains one of the important schools in Brazilian departments of economics. USP professor Eleuterio Prado noted in 2001 that orthodox economics in Brazil had not achieved “a position of uncontested supremacy in Brazil” (Prado 2001, 18). Ramon Garcia Fernandez and Carlos Eduardo Suprinyak surveyed Brazilian economics departments in 2016, and found that while “orthodox” (neoclassical) departments are most numerous, “heterodox” and “pluralist” departments together are larger in number. When the Brazilian Association of Graduate Programs in Economics (ANPEC) was organized in 1973, the FGV, a center of orthodoxy, withdrew from the association because non-​neoclassical departments were admitted. But, partly because the Ford Foundation, which supported ANPEC, favored pluralism, the FGV came back into the association. Non-​orthodoxy remains most conspicuous at UNICAMP, the São Paulo state university at Campinas, where the structuralists Maria Conceição Tavares, Wilson Cano, Carlos Lessa, and Luciano Coutinho held professorships (Fernandez and Suprinyak 2016).

4.4. Conclusions This chapter has dealt with the Brazilian adaptation of the principal tenets of Latin American structuralism, first promulgated in 1949 by the Argentine economist and central banker Raúl Prebisch. Structuralists held that Latin American economies were characterized by heterogeneous technologies, structural unemployment, external disequilibrium, and secularly deteriorating terms of trade. Celso Furtado was the major Brazilian contributor to structuralist thought, and pointed to the importance of economic planning to correct structural imbalances. He and Prebisch influenced the governments of Getúlio Vargas and Juscelino Kubitschek, and Furtado played a key role in establishing the national development bank (BNDES) and the Northeast development agency, SUDENE. Furtado “historicized” CEPAL structuralism and showed how losses in the coffee sector were spread across the whole economy in the 1930s. He furthermore developed a model of internal colonialism and arguably was the first dependency theorist. The crisis of structuralism in the mid-​1960s ultimately resulted in neostructuralism in 1990, a reformed version of the doctrine that emphasized the export market, technological change, and continual “learning by doing.” That so many structuralists held top

82   Joseph L. Love posts in the BNDES in the new century shows that structuralism in Brazil seems alive and well, with a firm base at UNICAMP.

Notes 1. Later renamed the Economic Commission for Latin America and the Caribbean. I use “CEPAL,” the Spanish and Portuguese acronym, throughout. 2. Note that this is different from Keynesian unemployment, which increases cyclically with recession. 3. Furtado did not belong to any political party, according to his Fantasia desfeita (Furtado 1989, 96). 4. See Perroux’s (1950) apologetic defense of corporatism. 5. Although this was a much fuller treatment of the matter than Furtado’s. See Pinto (1970). 6. Along with the Revista Brasileira de Economia, the professional economics journal. 7. Furtado, interview by the author, 1990. 8. See Desenvolvimento e Conjuntura 1 (1) (July 1957): 5–​15 (including CEPAL’s deteriorating terms-​of-​trade argument, the structuralist thesis that inflation is principally caused by bottlenecks, and the need for government planning or programming). Later numbers in the period examined (through 1960) were generally favorable to CEPAL. 9. See Vargas’s annual messages to the Brazilian congress in 1951 and 1954, cited in Haffner (2002, 45). 10. In the government of President Joao Goulart (1961–​1964), Furtado became minister of planning. 11. Kubitschek (1956, 47–​48, 54, 275, 278, 362). 12. Published in English as The Economic Growth of Brazil: A Survey from Colonial to Modern Times, translated by Ricardo W. de Aguiar and Eric Charles Drysdale (Berkeley: University of California Press, 1963). The English title is slightly misleading, since formação indicates qualitative aspects of development as well as quantitative growth. 13. For a discussion of other individuals and groups seeking to influence Kubitschek, see Skidmore (1967, 178–​182). 14. He was labeled a communist, for example, by the renowned though now reactionary social essayist Gilberto Freyre, and by Carlos Lacerda, governor of Guanabara (the former Federal District, containing the city of Rio de Janeiro and replaced by Brasilia as the national capital in 1960) (Furtado 1989, 68–​69, 133). 15. In this essay, Furtado also employed the terms of trade in gauging the country’s total buying power abroad, a notion which CEPAL was now calling the “capacity to import” (unit price of exports times quantities sold), a key indicator of the state of the economy, and the numerator of the coefficient later known as the “income terms of trade.” (Income terms of trade is defined as the unit price of a given good times the number of units exported divided by the unit price of imports.) In A Fantasia Organizada, Furtado (1985, 70) takes credit for pioneering the use of the idea of capacity to import; but it is also found in United Nations: Economic Commission for Latin America, Economic Survey of Latin America: 1949 (New York, 1951), a study written by Prébisch and published in Spanish the same year as Furtado’s “Características.” Presumably Furtado was a major contributor to the CEPAL product.

Brazilian Structuralism   83 16. Furtado probably drew his inspiration from Prebisch in United Nations, Economic Survey of Latin America: 1949, 171–​172, in which the latter wrote that Argentina maintained aggregate demand in the agricultural sector during the Depression by government crop purchases. 17. Furtado (1950, 28) wrote, “The shock caused by the external crisis (1929) gave [. . .] the Brazilian economy the opportunity to develop its internal market.” 18. Cf. the similar and contemporaneous emphasis by Hirschman (1958). 19. Furtado saw this as one of his major contributions to structuralism. Furtado, letter to author, Paris, December 22, 1982. 20. Furtado, Formação Econômica do Brasil (1959a); Anibal Pinto Santa Cruz, Chile, Un caso de desarrollo frustrado (Santiago de Chile: Editorial Universitaria, 1959); Aldo Ferrer, La economía Argentina: Las etapas de su desarrollo y problemas actuales (México: Fondo de Cultura Económica, 1963); Osvaldo Sunkel and Pedro Paz, El subdesarrollo latinoamericano y la teoría del desarrollo (Madrid: Siglo Veintiuno de España, 1970). Subsequently a more specialized structuralist work appeared on Rene Villareal, El desequilibrio externo en la industrialización de México (1929–​75):  Un enfoque estructuralista (México:  Fondo de Cultura Económica, 1976). Villareal argues, however, that structuralism accounts more adequately for Mexico’s external disequilibrium in the period 1939–​1958 than in 1959–​1970. 21. Furtado, “L’economie coloniale bresilienne”; Furtado, A economia brasileira (Rio de Janeiro:  Editora A  Noite, 1954); Furtado, Uma economia dependente (Rio de Janeiro: Ministério da Educação e Cultura, Serviço de Documentação, 1956). 22. On “involution” see Furtado (1963, 71). 23. The thesis also appears in later approximations and sketches of the work that would become Furtado, The Economic Growth [Furtado, Formação Econômica do Brasil] in 1959, namely, Furtado, A economia brasileira (1954b, 132–​143), and Furtado, Uma economia dependente (1956, 32, 57–​66). 24. For the best summary of Furtado’s arguments and the subsequent debate in Brazil, see Suzigan (1986, 21–​73). 25. For case studies of Latin American countries, including Brazil, see essays in Rosemary Thorp, ed. (1984). 26. During World War II, Brazil’s growth was perhaps less hampered because of the existence of a small capital goods sector. For discussion of the revisionist literature on Brazil and other Latin American countries, see Suzigan (1986); Love (1980, 57–​59); Cardoso and Brignoli (1979, 197) (summarizing a literature on Mexico, Argentina, Brazil, and Chile), and 199. See also the essays in Thorp, ed. (1984). 27. By my definition, internal colonialism is a process of unequal exchange, occurring within a given state, characteristic of industrial or industrializing economies, capitalist or socialist. As the economy becomes more differentiated with regard to region, factors and income flow from one or more geographically definable area to another, based primarily on price mechanisms, and secondarily (or not at all) on fiscal transfers; the state may nonetheless play a decisive role in setting price ratios, and differential regional effects of foreign trade are relevant. At the minimum, the process involves a structural relationship between leading and lagging regions (or city and hinterland) of a territorial state, based on monopolized or oligopolized markets, in which growth is progressively “inequalizing” between populations of these constituent geographic elements, rather than “equalizing.” Internal colonialism is distinct from colonialism per se, in which an alien state enforces

84   Joseph L. Love monopsony in labor markets, or even prescribes wage levels and labor drafts, such as the repartimiento of the Spanish American empire or the corvée of French colonial Africa. The definition by itself does not, of course, establish that the phenomenon exists. 28. See Singer (1964, 262–​267). For Furtado’s work, see Brasil (1959). That Furtado was the principal author of this statement is indicated in Furtado (1959, 35). 29. A  related policy, the confisco cambial, also adversely affected the Northeast:  the government “confiscated” a share of the earnings of traditional exporters (sugar and cocoa planters in the Northeast, coffee and cotton growers in the South) by maintaining an overvalued exchange rate—​in effect, collecting a tax. This was a way of getting around the federal constitution, which prohibited the central government from levying any tax on exports (Furtado 1959b, 49). 30. For the estimation procedures, see Love (1996, 280). 31. In 1957, Furtado noted, the absolute cost of food in Recife was a quarter more than that of São Paulo, and much of the food was imported from the South (Brasil 1959, 60). 32. In the previous year, Furtado had made the connection between development and underdevelopment at the international level in an academic thesis, but his Desenvolvimento e subdesenvolvimento was not published until 1961. 33. The term “internal colonialism” became widely known after the mid-​1960s, when it was introduced in the Latin American literature by the Mexican sociologist Pablo Gonzalez Casanova. For more on the history of the concept and the phrase, see Love (1989). 34. Note that this surplus is not Marx’s surplus value. Its source is not the exploitation of labor, but technological progress, as for Prebisch. Furtado (1961, 78–​79). 35. United Nations, Economic Commission for Latin America, “The Situation in Argentina and the New Economic Policy,” Economic Bulletin for Latin America 1 (1) (Jan. 1956): 30; United Nations, Economic Commission for Latin America, “Preliminary Study of the Effects of Postwar Industrialization on Import Structures and External Vulnerability in Latin America,” Economic Survey of Latin America, q.v. (1956): 115. 36. United Nations, Economic Commission, “Preliminary Study,” 128, 150, 151. 37. The seventh session of CEPAL in 1957 adopted a resolution calling for steps toward the creation of a region-​wide common market. 38. Tavares (1964); and Macario (1964). 39. Inter alia, Hirschman argued that the failure of ISI was not inevitable, but depended on the interaction of social and political factors with economic elements. See Hirschman (1971, 85–​123, esp. 103). On ISI failures, see the discussion of Sunkel and Furtado in the following section, and the literature surveyed in Baer (1972, esp. 107). 40. On the weak industrial bourgeoisie, see F.  H. Cardoso, Empresario industrial e Desenvolvimento Economico no Brasil (Saã Paulo: Difusão Européia do Livro, 1964), and his other studies of that group in Brazil and Argentina cited in Love (1996, 287–​288). 41. For support of this claim, see Gabriel Porcile, “La teoria estructuralista del desarrollo,” in Ricardo Infante, ed., El desarrollo inclusive en America Latina y el Caribe: Ensayos sobre politicas de convergencia productiva para la igualdad (Santiago: CEPAL, 2011), 43, 61. 42. An outstanding success in this area was the Brazilian firm EMBRAER, which eventually became a world leader in producing jet airplanes for short-​term hauls. See Werner Baer and Joseph L. Love, “Brazil’s EMBRAER: Institutional Entrepreneurship” in forthcoming collection edited by Jerry Davila. 43. On the importance of the BNDES, see Musacchio and Lazzarini (2014).

Brazilian Structuralism   85

References Baer, Werner. 1972. “Import Substitution and Industrialization in Latin America.” Latin American Research Review 7 (1): 95–​122. Baer, Werner. 1974. “Furtado Revisted.” Luso-​Brazilian Review 2 (1): 114–​121. Baer, Werner. 1995. The Brazilian Economy. 4th ed. New York: Praeger. Bielschowsky, Ricardo. 1985. “Brazilian Economic Thought in the Ideological Cycle of Developmentalism (1930–​64).” PhD dissertation, Leicester University. Boianovsky, Mauro. 2015. “Between Lévi-​Strauss and Braudel:  Furtado and the Historical-​ Structural Method in Latin American Political Economy.” Journal of Economic Methodology, 22 (4): 413–​438. Bresser-​ Pereira, Luiz Carlos. 2016. “Reflexões sobre o novo desenvolvimentismo e o desenvolvimentismo clássico.” Revista de Economia Política 36 (2):  (143) (abril–​junho): 237–​265. Brasil. 1959. Uma política de desenvolvimento econômico para o Nordeste. Rio de Janeiro: Conselho de Desenvolvimento, Grupo de Trabalho para o Desenvolvimento do Nordeste. Cardoso, Ciro F. S., and Hector Perez Brignoli. 1979. Historia económica de América Latina, Vol 2: Economías de exportación y desarrollo capitalista. Barcelona: Editorial Crítica. Castro, Antonio Barros de. 1969–​1971. Sete ensaios sôbre a economia brasileira. 2 vols. Rio de Janeiro: Forense. Castro, Antonio Barros de, and Carlos F. Lessa. 1967. Introdução à economia: Uma abordagem estruturalista. Rio de Janeiro: Forense. Castro, Antonio Barros de, and Carlos F. Lessa. 1973. Introducción a la economía: Un enfoque estructuralista, 11th edition. México: Siglo Veintiuno. CEPAL. 1951. Economic Survey of Latin America: 1949. New York. CEPAL. 1956a. “Preliminary Study of the Effects of Postwar Industrialization on Import Structures and External Vulnerability in Latin America.” Economic Survey of Latin America, q.v.: 115–​163. CEPAL. 1956b. “The Situation in Argentina and the New Economic Policy.” Economic Bulletin for Latin America 1, no. 1: 26–​45. CEPAL. 1957. Economic Survey of Latin America: 1956. New York. CEPAL. 1958. “Bases for the Formation of the Latin American Regional Market.” Economic Bulletin for Latin America 3 (March). no. 1, 4. CEPAL. 1964. The Economic Development of Latin America in the Postwar Period. New York. CEPAL. 1966. The Process of Industrial Development in Latin America. New York. CEPAL. 1990. Changing Production Patterns with Social Equity. Confederação Nacional das Industrias. 1950. “Interpretação do processo de desenvolvimento econômico da América Latina.” Estudos Econômicos 1 (3–​4): 271–​306. Fajnzylber, Fernando. 1983. La industrialización trunca de América Latina. Buenos Aires: Centro de Economía Transnacional. Ferrer, Aldo. 1963. La economía Argentina: Las etapas de su desarrollo y problemas actuales. México: Fondo de Cultura Económica. Ferrer, Aldo. 1967. The Argentine Economy. Translated by Marjorie Urquidi. Berkeley: University of California Press. Fernandez, Ramon Garcia, and Carlos Eduardo Suprinyak. 2016. “Manufacturing Pluralism in Brazilian Economics: The Role of ANPEC as Institutional Mediator and Stabilizer.” Paper

86   Joseph L. Love presented at the History of Economics Society Conference, Duke University, Durham, NC, June. Furtado, Celso. 1948. “L’economie coloniale bresilienne (XVIe et XVIIe siecles):  Elements d’histoire economique appliques.” PhD dissertation, Faculté de Droit, Université de Paris. Furtado, Celso. 1950. “Características gerais da economia brasileira.” Revista Brasileira de Economia 4 (1): 1–​37. Furtado, Celso. 1954a. “Capital Formation and Economic Development.” International Economic Papers 4: 124–​144. Furtado, Celso. 1954b. A economia brasileira. Rio de Janeiro: Editora A Noite. Furtado, Celso. 1956. Uma economia dependente. Rio de Janeiro:  Ministério da Educação e Cultura, Serviço de Documentação. Furtado, Celso. 1958a. “Fundamentos da programação econômica.” Econômica Brasileira 4 (1–​2): 39–​44. Furtado, Celso. 1958b. “External Disequilibrium in the Underdeveloped Economies.” Indian Journal of Economics 38: 403–​410. Furtado, Celso. 1959a. Formação econômica do Brasil. Rio de Janeiro: Editôra Fundo de Cultura. Furtado, Celso. 1959b. A operação Nordeste. Rio de Janeiro: Ministério da Educação e Cultura, Instituto Superior de Estudos Brasileiros. Furtado, Celso. 1961. Desenvolvimento e subdesenvolvimento. Rio de Janeiro: Editôra Fundo de Cultura. Furtado, Celso. 1963. The Economic Growth of Brazil: A Survey from Colonial to Modern Times. Translated by Ricardo W. de Aguiar and Eric Charles Drysdale. Berkeley:  University of California Press. Furtado, Celso. 1964. Development and Underdevelopment. Translated by Ricardo W. de Aguiar and Eric Charles Drysdale. Berkeley: University of California Press. Furtado, Celso. 1969. Formação econômica da América Latina. Rio de Janeiro: Lia. Furtado, Celso. 1970. Economic Development of Latin America:  Historical Background and Contemporary Problems. Translated by Suzette Macedo. Cambridge:  Cambridge University Press. Furtado, Celso. 1973. “Adventures of a Brazilian Economist.” International Social Science Journal [UNESCO] 25 (1–​2): 28–​38. Furtado, Celso. 1982. Letter to author, Paris, December 22. Furtado, Celso. 1985. A fantasia organizada. Rio de Janeiro: Paz e Terra. Furtado, Celso. 1989. A fantasia desfeita. São Paulo: Paz e Terra. Furtado, Celso. 1990. Interview by the author, Joseph L. Love. Rio de Janeiro, May 31. Garcia Fernandez, Ramon, and Carlos Eduardo Suprinyak. 2016. “Manufacturing Pluralism in Brazilian Economics: The Role of ANPEC as Institutional Mediator and Stabilizer.” Paper presented at the History of Economics Society Conference, Duke University, Durham, NC, June. Haffner, Jacqueline Angélica Hernández. 2002. A CEPAL e a industrialização brasileira, 1950–​ 1961. Porto Alegre: EDIPUCRS. Hirschman, Albert O. 1971. A Bias for Hope: Essays on Development and Latin America. New Haven, CT: Yale University Press. Hirschman, Albert O. 1968. “The Political Economy of Import-​Substituting Industrialization in Latin America.” The Quarterly Journal of Economics 82 (1): 1–​32. Hirschman, Albert O. 1958. The Strategy of Economic Development. New Haven, CT:  Yale University Press.

Brazilian Structuralism   87 Ianni, Octavio. 1971. Estado e planejamento econômico no Brasil (1930–​ 1970). Rio de Janeiro: Civilização Brasileira. Kubitschek, Juscelino de Oliveira. 1956. Mensagem ao Congresso Nacional: 1956. Rio de Janeiro: Imprensa Nacional. Lewis, W. Arthur. 1954. “Economic Development with Unlimited Supplies of Labour.” The Manchester School of Economic and Social Studies 22 (2): 139–​191. Love, Joseph L. 1980. São Paulo in the Brazilian Federation, 1889–​1937. Stanford, CA: Stanford University Press. Love, Joseph L. 1989. “Modeling Internal Colonialism:  History and Prospect.” World Development 1 (6): 905–​922. Love, Joseph L. 1996. Crafting the Third World: Theorizing Underdevelopment in Rumania and Brazil. Stanford, CA: Stanford University Press. Macario, Santiago. 1964. “Protectionism and Industrialization in Latin America.” Economic Bulletin for Latin America 9 (1): 61–​101. Magarinos, Mateo. 1991. Diálogos con Raúl Prébisch. México: Banco Nacional de Comercio Exterior. Musacchio, Aldo, and Sergio G. Lazzarini. 2014. Reinventing State Capitalism: Leviathan in Business, Brazil and Beyond. Cambridge, MA: Harvard University Press. Normano, João Frederico. 1935. Brazil: A Study of Economic Types. Chapel Hill: University of North Carolina Press. Oliveira, Francisco de, ed. 1983. Celso Furtado: Economia. São Paulo: Atica. Perroux, François. 1939. “Capitalismo e corporativismo:  Socialização do produto—​ corporativismo e salário.” Revista Forense 77 (36): 402–​405. Perroux, François. 1950. “The Domination Effect and Modern Economic Theory.” Social Research 17 (2): 188–​206. Pinto, Anibal. 1970. “Naturaleza e implicaciones de la ‘heterogeneidad estructural’ de la América Latina.” El trimestre económico 37(145):  83–​ 100. México:  Fondo de Cultura Económica. Pinto Santa Cruz, Anibal. 1959. Chile, Un caso de desarrollo frustrado. Santiago de Chile: Editorial Universitaria. Prado, Eleutério F. S. 2001. “A ortodoxia neoclássica.” USP: Estudos Avançados 15 (41): 9–​20. Prébisch, Raúl. 1959. “[International Trade and Payments in an Era of Coexistence:] Commercial Policy in Underdeveloped Countries.” American Economic Review 49 (2): 251–​273. Prébisch, Raúl. 1961. “Economic Development or Monetary Stability:  A False Dilemma.” Economic Bulletin for Latin America 6 (1): 1–​25. Rodriguez, Octavio. 1980. La teoría del subdesarrollo de la CEPAL. México: Siglo Veintiuno Editores. Rodriguez, Octavio. 2006. El estructuralismo latinoamericano. México:  Siglo XXI and CEPAL. Simonsen, Roberto. 1937. História econômica do Brasil (1500–​1820). São Paulo:  Companhia Editora Nacional. Singer, Hans W. 1950. “The Distribution of Gains between Investing and Borrowing Countries.” American Economic Review: Papers and Proceedings 40 (2): 473–​485. Singer, Hans W. 1964. “Trade and Fiscal Problems of the Brazilian Northeast.” In International Development: Growth and Change, edited by H. Singer, 262–​267. New York: McGraw-​Hill.

88   Joseph L. Love Sikkink, Kathryn A. 1988. “Developmentalism and Democracy:  Ideas, Institutions, and Economic Policy Making in Brazil and Argentina (1955–​1962).” PhD dissertation, Columbia University. Sikkink, Kathryn A. 1991. Ideas and Institutions: Developmentalism in Brazil and Argentina. Ithaca, NY: Cornell University Press. Skidmore, Thomas E. 1967. Politics in Brazil, 1930–​ 1964:  An Experiment in Democracy. New York: Oxford University Press. Sunkel, Osvaldo, and Pedro Paz. 1970. El subdesarrollo latinoamericano y la teoría del desarrollo. Madrid: Siglo Veintiuno de España. Suzigan, Wilson. 1986. Industria brasileira: Origem e desenvolvimento. São Paulo: Brasiliense. Tavares, Maria Conceição. 1964. “The Growth and Decline of Import Substitution in Brazil.” Economic Bulletin for Latin America 9 (1): 1–​59. Thorp, Rosemary, ed. 1984. Latin America in the 1930s: The Role of the Periphery in World Crisis. London: Macmillan in Association with St Antony’s College, Oxford. Villareal, Rene. 1976. El desequilibrio externo en la industrialización de México (1929–​75): Un enfoque estructuralista. México: Fondo de Cultura Económica.

Chapter 5

Brazil’s Imp ort -​ Su bstitu t i on Indu striali z at i on Werner Baer

5.1.  Early Industrial Growth Substantial industrial growth in Brazil began in the late nineteenth century. It was associated with the large wave of European immigration that followed the abolition of slavery in 1888. Free immigrant labor represented a large enough market to lead many importers of manufactured goods (mainly textile and food products) to establish local production facilities. By 1907 there existed close to 3,000 manufacturing establishments, employing about 136,000 workers in such sectors as textiles, food processing, and beverages. Capital goods production did not exist, except for some primitive ironworks, naval yards, and repair shops for the transportation sector (Dean 1969, Chapter 6; Villela 2011, 41). World War I caused a shortage of many types of manufactured goods, as industrial countries switched from civilian to war goods production. The shortage of manufactured imports did not result in investments in new capacity, as Brazil did not have a capital goods industry. The shortages were partially met by increased use of existing production capacity (Baer 2014). The dynamism of Brazil’s economy in the 1920s was based on a booming coffee sector, as its share of exports rose from 56% in 1919 to 75% in 1924. A slight appreciation of the exchange rate and rising internal prices decreased the protection of domestic industry from foreign competition, and throughout most of the 1920s industry grew at a very slow pace. The average yearly growth rate of industrial output fell from 4.6% in 1911–​1920 to 3% in 1920–​1929. Since the textiles sector constituted the principal industrial sector at the time, its stagnation explains the overall weak performance of industry. A closer

90   Werner Baer examination, however, reveals much faster growth in other industrial subsectors, such as metallurgical products, cement, iron and steel, and paper products (Baer 2014, 34–​35; Villela 2011, 42). Despite the growth of industries since the 1890s, one cannot define Brazil’s industrial growth as “industrialization.” That term would only apply where industry becomes the leading growth sector of the economy.

5.2.  Import Substitution of the 1930s The depression of the 1930s had a severe negative impact on Brazil’s exports, whose value fell from US$445.9 million in 1929 to US$180.6 million in 1932. In addition to the decline in export receipts, the inflow of foreign investments had come almost to a complete halt. This situation became critical as the foreign exchange needed to finance the country’s external debt (which amounted to US$1.3 billion in 1931), not counting the remittances of profits by private entities, forced the government to take drastic actions (Abreu 1990, Chapter 3, for further details). The government suspended part of its foreign debt services and introduced exchange and other direct controls, combined with a devaluation of the currency. The latter increased the price of imports. The combination of these measures resulted in a decline of imports from US$416.6 million in 1929 to US$108.1 million in 1932. Additionally, with the steep decline in world coffee prices, the federal government took over the coffee-​support program from individual coffee-​producing states (mainly São Paulo). A National Coffee Council was founded, which bought all coffee production, destroying a large quantity that could not be sold or stored (for further details, see Baer 2014, 38; Dean 1969, Chapter 7; Furtado 1972). The curtailment of imports and the continued domestic demand resulting from the income generated by the coffee support program caused shortages of many manufactured goods and a consequent rise in their relative prices. This acted as a catalyst for a burst of domestic industrial production. By 1931 industrial production had fully recovered from a decline that had begun in 1928, and it doubled in the following eight years. Especially noteworthy was the growth of textile production (147%), metal products (three times larger in the late 1930s than in 1929), and paper products (almost seven times larger).1 Unlike during World War I, industrial growth was based not only on the more intensive use of existing production capacity, but also, especially in the second half of the 1930s, on the creation of new capacity. One could characterize the rapid industrial growth of the 1930s as Brazil’s first experience with import-​substitution industrialization, that is, the industrial sector became the Brazilian economy’s leading sector, whereas in earlier years industrial growth had accompanied the expansion of the primary export sector.

Brazil’s Import-Substitution Industrialization    91

5.3.  Import-​Substitution Industrialization as a Deliberate Development Policy During World War II there was substantial industrial growth, but with little expansion of productive capacity. Industrial production grew at an annual rate of 5.4% in the period 1939–​1945. Noteworthy in this period were average yearly growth rates of metal products (9.1%), textiles (6.2%), shoes (7.8%), and beverages and tobacco (7.6%), which were industries whose imports were drastically curtailed. Investment activities fell at first, but rose again in 1945. This was mainly due to the capital equipment Brazil was allowed to import during the war to construct its first large integrated steel mill at Volta Redonda (Baer 1969). The establishment of the Volta Redonda plant is especially significant in that it marked the most direct participation of the state thus far in productive activities. The construction and operation of the plant occurred under the auspices of a state-​owned enterprise, the Companhia Siderurgica Nacional (CSN; National Steel Company). The expanding role of the state in the industrialization process and, indeed, with the process of development more generally formed part of the corporatist ethos associated with President Vargas’s Estado Novo (New State). This marked a radical change from the less interventionist approach associated with successive administrations during the nineteenth and early twentieth centuries. The drastic decline of imports during World War II and the boom of exports resulted in a substantial increase in the country’s foreign exchange reserves, from US$71 million prior to the start of the war to US$708 million in 1945. In February 1945, the government established a foreign exchange regime without restrictions, except for some limitation on the remittance of profits. Brazil’s currency, the cruzeiro, was kept at its prewar value of CR$18.50 per dollar and did not change until 1953, although prices rose 285% from 1945 to 1953. The combination of an overvalued currency and the elimination of most import restrictions resulted in a spurt of imports, which eliminated the accumulated wartime foreign exchange reserves within one year. This forced the government to reintroduce foreign exchange controls, which were also supplemented by import controls through a system of import licensing (Baer 2014, 54–​61). The post–​World War II industrialization drive was initially the consequence of measures taken in order to cope with balance of payments difficulties. Only gradually, predominantly in the 1950s, did various efforts at balance of payments protection become part of a deliberate policy to promote industrialization through import substitution. The protection of domestic industry occurred through different systems of exchange controls and in some years through a multiple exchange rate system (Baer 2014, 54–​61).

92   Werner Baer The government also made use of protective tariffs and of SUMOC2 Instruction 113, which was designed to attract foreign investments by enabling firms to import capital equipment without the need for exchange cover. The Tariff Law of 1957 expanded and solidified protection of domestic industry. In many cases tariffs were as high as 60%, 80%, and 150%. The creation in December 1950 of the Joint Brazil–​ United States Economic Development Commission provided an important vision to policymakers about infrastructural bottlenecks and the potential for structural changes in the economy. The Commission also recommended the creation of a development bank, which was established in 1952 as the BNDE (National Bank for Economic Development).3 It was to become the basic financier of state and private enterprises. Its role became especially significant in financing heavy capital investment across industry as the industrialization process intensified into the 1960s and the 1970s. Import-​substitution industrialization (ISI) resulting from these measures had a notable growth impact in the first half of the 1950s (with industrial production growing at 9% per year), but this expansion would be even more pronounced in the second half of the 1950s under the presidency of Juscelino Kubitschek, who established a Targets Plan, which made the state even more active in speeding up ISI. Under this plan, specific programs were established to promote such industries as automobile and utility vehicle construction, shipbuilding, and heavy machinery. The programs were organized through the BNDE, and the favored industries were given special treatment for importing capital goods, raw materials, components, and the like for specific periods of time (Baer 2014; Villela 2011). Brazil’s ISI program did not rely solely on foreign private investments; it was also complemented by a substantial growth in state enterprises. The establishment of large integrated steel mills was deemed essential to the deepening of the industrial sector, and since there was little interest from the foreign private sector and the domestic private sector did not have the technical and financial capabilities to enter this field, state firms were established (Baer 1969). Additionally, the expansion of public utility services, such as power generation and distribution, was increasingly provided by public firms (both at the federal and state levels), as private domestic and foreign firms showed little interest in an increasingly regulated sector.

5.4.  The Impact of Import-​Substitution Industrialization Policies The ISI policies of the 1950s resulted in high rates of economic growth. The average yearly real growth rate of gross domestic product (GDP) in the period 1947–​1962 was over 6%, and in the years 1956–​1962 it was 7.8%. From 1947 to 1961 the overall real product increased by 128%. The real agricultural product rose by only 87%, while the

Brazil’s Import-Substitution Industrialization    93 Table 5.1 Brazil: Sectoral Distribution of GDP (Percentages) 1950

1960

1970

1980

1990

2010

Agriculture

13.3

10.2

4.6

5.0

5.8

6.2

Industry

28.7

32.2

33.0

29.9

32.4

30.8

Services

58 100

57.6 100

62.4 100

65.0 100

61.8 100

63.1 100

Source: Conjuntura Econômica; IBGE, Sistema de Contas Nacionais.

Table 5.2 Brazil: Sectoral Distribution of Labor (Percentages)

Agriculture

1950

1960

1980

1990

2000

2010

64.4

58.9

38.1

26.4

22.3

16.7

Industry

16.5

16.8

22.8

22.8

19.5

20.3

(Manufacturing)

(11.5)

(11.8)

(12.7)

(14.7)

(12.0)

(12.1)

Services

19.1

24.3

39.1

50.8

58.2

63.0

100.0

100.0

100.0

100.0

100.0

100.0

Source: Conjuntura Econômica; IBGE, Anuário Estatístico do Brasil, various years.

industrial product increased by 262%. For the absolute increase of GDP between 1947 and 1961, agriculture was responsible for only 18%, whereas the nonagricultural sector contributed the rest. From the 1930s on, a characteristic of Brazil was pronounced rural-​urban migration. In the early post–​World War II years, the rate of the country’s urban population growth was almost twice that of the general population. Unfortunately, as is clear from the data in Tables 5.1 and 5.2, the industrial sector absorbed only a small proportion of migrants to the cities. Most survived in the service sector, which at the time consisted to a large extent of marginal activities, including informal street vending. The lack of labor absorption by the most dynamic sector was a puzzle that generated a fairly large literature trying to explain the phenomenon. Some claimed that there was a distortion in factor prices (industrial wages made artificially higher through social legislation than would be warranted, given the huge supply of labor), while capital was made artificially cheap by the development bank and other incentives to invest in import-​ substitution industries. Assuming different production techniques to choose from, this distortion of factor prices led to a tendency to choose the more capital-​intensive ones. However, although this may offer a logical explanation, the real world was rather more complicated. Many multinationals that invested in Brazil chose to ship older capital equipment to their Brazilian plants (which was possible through special Brazilian

94   Werner Baer import regulations); thus Brazil’s ISI in the 1950s and 1960s utilized comparatively labor-​ intensive equipment in relation to the technological state of the art at the time. In some other industries, such as special types of steel, production technology was rigid and was not influenced by relative factor prices. The capital-​intensive nature of Brazil’s new industries contributed to a continued concentration of income. In other words, as the capital/​labor ratio of the dynamic sectors of the economy (i.e., industry) was greater than the capital/​labor ratio of the traditional sectors, the net result (other things being equal) was an increase in the concentration of income in Brazil.

5.5.  The Import Coefficient Problem Brazil’s government promoted ISI across the board. Its major goal was to lower the import coefficient that is, the lowering of imports as a share of GDP. The value of imports of goods and services as a proportion of GDP declined from 9% in 1949 to 5% in 1960.4 During this period of import substitution, exports were neglected and little attempt was made to diversify them. In 1955, coffee still accounted for 59% of exports, and by 1964, 53%; other primary products made up 38% in 1955 and 45% in 1964. However, by the 1960s, imports as a share of GDP began to rise again, reflecting a rise of imported inputs into industry (Baer 2014, 193–​194). It was at this stage that policymakers realized that there was a limit to ISI and that export diversification was crucial in order to assure future foreign exchange earnings to pay for imported inputs into Brazil’s industries. The problem, however, was that most ISI industries were established behind high tariff walls, with high costs and many firms producing goods of low quality.5 It would take special incentives (such as lower taxes for export earnings and subsidized credit for export) to begin a process of export diversification.

5.6.  Regional Inequalities Inequalities in the geographic distribution of income and growth have been a characteristic of the Brazilian economy since colonial times. By the time of ISI, Brazil’s major export product was coffee, and most of its production was located in the Southeast of the country, while only 43% of the population lived in that region. Most of Brazil’s new industries were located in the Southeast, and by the early 1960s this region accounted for more than 75% of the country’s industrial product (Baer 2014, Chapter 12). Most new industries were located in the states of São Paulo, Rio de Janeiro, and Minas Gerais. The explanation for this regional concentration was that these states had the highest per capita income and thus represented the greatest source of demand for industrial products. They also had the best infrastructure and most of the skilled labor.

Brazil’s Import-Substitution Industrialization    95 The concentration of ISI in the Southeast of the country resulted in a dynamic that worsened the situation of the poorer regions, especially Brazil’s Northeast, which at the time had more than a third of the country’s population. The Northeast continued to produce primary export products (such as sugar, cotton, and cacao), but with ISI was forced to purchase its manufactured products from Brazil’s Southeast, whose prices were much higher than imported products. Thus the terms of trade of the Northeast declined, which meant that the poorest region of Brazil was transferring resources to the most prosperous, the Southeast. In other words, the analysis of the center-​periphery relationship developed by the original theorists of ISI—​Raúl Prebisch and his Economic Commission for Latin America (ECLA) group—​and which was used to justify ISI, now repeated itself within Brazil (Baer 2014, 254–​262).

5.7.  Neglect of Agriculture Brazil’s investment in agriculture during the intensive ISI period was quite small, especially in agricultural products for the domestic market. Although the growth of the population was smaller than the growth of the food supply, there was another factor that cast a shadow on this favorable picture. While the intensive rural-​urban migration resulted in an urban population growth of about 5.4% per year in the 1950s, most of the increase of food production occurred in new lands placed under cultivation, rather than increased productivity in existing lands. Since the rapidly rising demand for food in urban centers had to be supplied from increasingly distant places, there was an increasing strain on the country’s precarious rural-​urban transportation network and on the agricultural marketing system. It was generally recognized by the 1960s that further industrial growth would be severely hampered if advances were not made in agricultural productivity near the main consuming centers. If these trends continued, the rise of relative food prices would not only increase inflationary pressures, but also lead to rising social tensions.

5.8.  The 1960s: Import-​Substitution Industrialization in Crisis By the early 1960s the Brazilian economy had lost its dynamism. After GDP growth reached a peak of 10.3% in 1961, it fell to 5.1%, 1.5%, and then 2.4% in 1962, 1963, and 1964, respectively. The immediate cause of the stagnation that set in after 1961 was the continuing political crisis that the country experienced after the resignation of Jânio Quadros from the presidency in August 1961. The turbulent years that followed until the overthrow of the Goulart government in April 1964 were devoid of any consistent economic policy.6

96   Werner Baer After the coup of 1964, the new military regime concluded that the path to economic recovery lay in remedying a number of problems that had emerged during the ISI period. The primary concerns at first were with bringing inflation and its price distortions under control, modernizing capital markets, indexing controlled prices, and using tax incentives to influence the allocation of resources in desired directions (such as investments in neglected regions) (Baer 2014, 73–​74). Foreign trade policy was considered of central importance by the post-​1964 military governments. The rapid growth and diversification of exports was deemed essential to the long-​term health of the economy. In order to achieve this diversity, state export taxes were abolished, administrative procedures for exporters were simplified, and export tax incentives and subsidized credit were instituted. In 1968 a crawling-​peg exchange rate policy was introduced to avoid overvaluation of the currency. It consisted of frequent (but unpredictable) small devaluations of the cruzeiro. This would keep the currency from becoming overvalued as long as some inflation was still present, while keeping speculation against the currency at a minimum and keeping the exchange rate from becoming a political issue. The outward orientation of policies on the import side consisted mainly of a tariff reform in 1966, which resulted in the lowering of nominal tariffs from an average of 54% in 1964–​1966 to 39% in 1967. Subsequent changes again led to a rise in rates, but not to pre-​reform levels. Real protection was also reduced in the late 1960s and early 1970s by the fact that the rate of devaluation of the cruzeiro was smaller than the rate of inflation. Collectively, these policies, which pushed the industrialization strategy in a more outward-​oriented direction, have become known as post-​ISI. For a while this policy set proved highly effective, giving rise to a period from the late 1960s to the early 1970s known as the “Brazilian Economic Miracle.” After a period of stagnation (1962–​1967), when annual real GDP growth was only 3.7%, Brazil experienced a period of high growth that lasted from 1968 to 1973, when the annual growth rate of GDP averaged 11.3%, which was generally attributed to the reforms instituted by the military regime. During those years, industry was again the leading sector, expanding at yearly rates of 12.6%. Within manufacturing, the highest growth rates were achieved in transport equipment, machinery, and electrical equipment, while traditional sectors, like textiles, clothing, and food products, experienced much slower growth rates. External trade grew at rates substantially higher than the economy. In the years 1970–​1973, the average yearly growth rate of exports was 14.7% and of imports 21%. The resulting trade deficit was also accompanied by a rising deficit in the service balance. Until 1974, however, this was more than covered by a massive inflow of official and private capital. In these years Brazil succeeded in diversifying its commodity export structure, and on the import side there was a notable increase in capital goods. Coffee, which in 1964 amounted to 53% of exports, fell to 13% in 1974; manufactured goods rose from 5% to 36%; and capital and intermediate goods rose from 60% of imports in the early 1960s to 85% in 1972. The post-​1964 policies clearly opened the economy

Brazil’s Import-Substitution Industrialization    97 to foreign trade. Whereas the ISI policies of the 1950s decreased the import coefficient (import/​GDP ratio) from 16% (1947–​1949) to 5.4% in 1964, it rose again to 14% in 1974.

5.9.  The Oil Shock of 1973 and Its Consequences The oil shock of November 1973 quadrupled the price of petroleum. At the time Brazil was relying on imports for over 80% of its oil consumption and thus its import bill rose from US$6.2 billion in 1973 to US$12.6 billion in 1974, and the current account from a deficit of US$1.7 billion to US$7.1 billion. At the time Brazil had two options for reacting to the oil shock: it could either substantially reduce growth in order to diminish its non-​oil import bill, or it could opt for continued relatively high growth rates. The latter would cause a substantial decline in the country’s foreign exchange reserves and/​or a substantial increase in its foreign debt. It opted for the latter. In 1975 the Second National Development Plan (PND II) was introduced. It consisted of a huge investment program, with the following two goals: (1) import substitution of basic industrial products (such as steel, aluminum, copper, fertilizers, and petrochemicals), and capital goods;7 and (2) the rapid expansion of economic infrastructure (hydro and nuclear power, alcohol production, transportation, and communications). Many of these investments were undertaken by state enterprises (in energy and steel), whereas others (especially capital goods) were carried out by the private sector, with massive financial support from the BNDES. The goals of these programs were (1) to act as a strong countercyclical policy vis-​à-​vis the impact of the oil crisis and maintain a reasonable rate of growth; (2) to change the structure of the economy through a new round of import substitution and export diversification and expansion; and (3) to encourage international lenders to finance the current account deficit and to postpone external adjustment. It has also been claimed that the basic ideas behind PND II were to increase the country’s self-​sufficiency in sectors such as energy and to develop new types of comparative advantages. Large sums of state investments that occurred at the time were justified on the grounds that in the short run the returns on investment in infrastructure and heavy industry would be too low to attract private capital. The growth impact of the PND II was quite positive, as the yearly real GDP expansion was 7% (led by industry, which grew at yearly rates of 7.5%). The sectors that experienced exceptional growth rates during this period were metal products, machinery, electrical machinery, paper products, and chemicals (Baer 2014, 78). Taking the ratio of imports to domestic production as a measure of ISI, there occurred a notable decline between 1976 and 1981 in following sectors listed in Table 5.3.

98   Werner Baer Table 5.3 Import Coefficient, Selected Sectors 1976 and 1981 1976

1981

Paper

0.13

0.08

Cellulose

0.05

0.01

Polyethylene

0.72

0.02

Plastic tubes

0.45

0.03

Steel

0.15

0.05

Fertilizers

1.34

0.85

Aluminum

0.58

0.14

Capital goods

0.64

0.40

Brazil’s option of growth through external debt was justified on the grounds that future savings of foreign exchange resulting from the investment programs—​due to import substitution and to the development of new export capacity—​would ultimately bring about a situation in which Brazil could produce trade surpluses large enough to service and repay its international debt.

5.10.  The Debt Crisis and Its Aftermath The debt crisis, which exploded in the early 1980s and resulted in the (economically) “lost decade” of the 1980s, led the country to push hard to promote nontraditional exports and to decrease imports. Throughout the 1980s, Brazil had to negotiate with its foreign creditors and the International Monetary Fund (IMF) to turn over and extend the servicing of its debt. The conditions for various adjustments of the foreign debt gradually forced Brazil into a full-​scale acceptance of what become known as neoliberal policies (for a discussion of neoliberalism in Brazil, see Amann and Baer 2002). Tariffs were gradually lowered, the market reserves of various products (such as computers) were eliminated, many state-​owned industries were privatized, and various artificial stimuli for exports were removed (Silber 2011). The first round of cautious trade reforms took place under the civilian administration of President Sarney in 1987. Then, in 1990, the directly elected president Fernando Collor de Melo instituted a rolling four-​ year program of trade liberalization that became known as the Abertura Comercial (Trade Opening). This comprised phased tariff reductions and, more importantly, a comprehensive rolling back of nontariff barriers. The latter, including outright import prohibitions (the so-​called Anexo C) had comprised the main protectionist mechanisms of the ISI era. Besides Brazil’s unilateral trade reform efforts, the early 1990s saw the creation of MERCOSUL, a customs union embracing Brazil, Argentina, Paraguay, and Uruguay.

Brazil’s Import-Substitution Industrialization    99 Suddenly, regional trade in industrial products became significantly freer. The partial exception to this was the still politically sensitive automobile sector. In the latter, a managed trade regime was established between MERCOSUL’s two major automotive producers, Brazil and Argentina. In addition, various measures were instituted to facilitate foreign investments. These centered on privatization of public utilities and state-​owned enterprises, among them the poster child of the early ISI era, CSN. Another icon of Brazil’s industrialization drive, Petrobrás, saw its monopoly in domestic onshore and offshore oil exploration activities broken as the result of a 1995 constitutional amendment. Collectively, the intent of these policies was to increase efficiency through foreign competition and to increase the inflow of foreign direct investment. Economic liberalization gained momentum throughout most of the 1990s, and, as Villela (2011, 53)  observes, “[.  .  .] stabilization, trade liberalization and privatization combined to usher in a new development model, leaving behind 60 years or so of ISI. Globalization had finally caught up with Brazil, which embraced market-​friendly reforms more out of necessity than from actual conviction as to their merits.”

5.11.  Inflation Targeting, Interest, and the Exchange Rate The success of Brazil’s Plano Real (Real Plan) stabilization program (introduced in 1994) was in part connected with the appreciation of Brazil’s currency, which was the result of the high interest rates that were important stabilization instruments used at the time. These high base interest rates (varying between 7% and 10% in real terms) attracted a substantial amount of capital, which appreciated the currency and thus facilitated stabilization efforts. For a while this came to be known as a policy of controlling inflation via an exchange rate anchor. Although this policy was discontinued at the turn of the century and a policy of inflation targeting was introduced, there has been a continued emphasis on the use of high interest rates to underpin price stability. As during the first decade of the twenty-​first century Brazil’s interest rates were among the highest in the world and attracted substantial portfolio inflows, Brazil’s real came to be one of the most appreciated currencies in the world. The strong real made it very difficult for many of the country’s industries to achieve international competitiveness. However, this did not result in balance of payments problems since at the same time there occurred for over a decade a world commodity boom, originating in the high growth rates of a number of Asian countries, led by China. Thus over the second half of the first decade of the 2000s commodities rose to over 60% of total exports, while manufactured goods declined to 36%. This is ironic, since what originally led Brazil and similar countries to industrialize was too great a dependence on the export of commodities. By the second decade of the twenty-​first century, however, commodity prices declined and contributed to a period of stagnation.

100   Werner Baer It is not clear to what extent Brazil’s appreciated exchange rate was responsible for the decline of manufactured exports. It could be that Brazil’s use of tax incentives and subsidized credit to stimulate nontraditional exports led some of its trade partners to interpret these policies as amounting to dumping and thus to take retaliatory action.

5.12.  Import-​Substitution Industrialization Redux? Industrial Policy under Presidents Lula and Rousseff As a partial response to the deindustrialization process that had been unfolding since the early 1990s, the PT (Workers’ Party) administrations of President Lula and President Rousseff (2002–​2016) gradually adopted a more interventionist industrial policy stance. To some extent, this development can also be seen as a product of the growing influence of the structuralist school of economic thought within the administration. One aspect of the resurgence of interventionism centered on an expansion of BNDES lending to strategic or technologically dynamic sectors, notably biotechnology, aerospace, and automotive products. This development can, in many ways, be seen as merely an intensification of a long-​ established facet of public policy in Brazil; even the more free-​market-​oriented administration of Fernando Henrique Cardoso (1995–​2002) had engaged in reasonably activist industrial and technology policy. What made the policy stance of his two successors more eye-​catching was their willingness to revert to protectionist measures when deemed necessary. Thus in the rapidly expanding oil and gas sector, the leading national oil major, Petrobrás, was obliged to purchase equipment from domestic producers. As a result, there was a rapid expansion in the shipbuilding and offshore engineering sector. The continued growth of this sector looked set to be all but guaranteed by the program to develop newly discovered “pre-​salt” offshore oil deposits. While the ISI era might not be returning, a sense that the cause of trade liberalization was, at least, in retreat was further evidenced by the lack of progress in resuming tariff reductions or signing fresh free trade agreements. By the end of the Rousseff administration in 2016, tariffs on industrial products in Brazil remained very high by international standards. In the automotive sector, especially, the trade regime made it all but commercially impossible to import assembled vehicles. Thus, as during the 1960s and 1970s, most domestic demand was met through Brazil-​based assembly plants owned by foreign multinationals. Looking to the future, what prospects remain for the ISI-​style mechanisms adopted by Lula and Rousseff? Following the arrival of a more centrist administration in mid-​ 2016, the indications so far point toward the pursuit of a more pro-​free-​trade agenda, in

Brazil’s Import-Substitution Industrialization    101 particular involving the signing of trade accords with regional and extra-​regional partners (including, possibly, the post-​Brexit UK). At the same time, the role of the BNDES has already been scaled down and lines of credit cut back. Perhaps most significant in political terms, the entire framework surrounding the Petrobrás procurement process is the subject of a wide-​ranging criminal investigation. This investigation—​known as Lava Jato—​has already highlighted the means by which an activist trade and industrial policy offered an avenue for the corrupt financing of political parties. It also provided one of the central de facto motives for the impeachment of President Rousseff. With Lava Jato ongoing, and oil prices well below their 2008 peak, investment in the oil sector has sharply contracted. This has in turn resulted in a financial crisis in Rio de Janeiro state, home to most of the country’s oil services and equipment activities. Thus, as Brazil approaches the third decade of the twenty-​first century, it is fair to say that the ISI policy agenda is well and truly again in retreat.

5.13.  The Infant That Didn’t Grow? Many apologists for ISI in Brazil (and other countries that followed its model) have claimed that ISI was an application of the traditional “infant industry” argument for protection, that is, that in a closed economy many new industrial sectors would have the opportunity and time to “learn by doing,” and after a while many protected industries would achieve productivity levels that would make them internationally competitive. Unfortunately, extensive studies carried out under the sponsorship of IPEA (the research center of Brazil’s Planning Ministry) found that many sectors of Brazil’s industry were substantially lagging behind in productivity (De Negri and Cavalcante 2014). It was noted that by the end of the first decade of the twenty-​first century, Brazil’s labor productivity was about 25% of productivity in rich countries (De Negri and Cavalcante 2014, 37–​38). It was also shown that in 1995 the productivity of the United States was 6.6 times greater than that of Brazil, and by 2009 had grown to 7.1 times Brazil’s productivity. The IPEA productivity studies place the blame of Brazil’s low productivity on the country’s low human capital, low research and development (R&D) as a proportion of GDP, and decades of low investment in the country’s infrastructure.

5.14.  The End of Import-​Substitution Industrialization and Deindustrialization In the last decade, many studies of emerging countries have noted the phenomenon of deindustrialization, that is, shifting toward services at an earlier stage of development

102   Werner Baer than was the case in older industrial economies.8 This has also been the case in Brazil (see Table 5.1). Between 1950 and 1980, agriculture’s share declined from 13.3% to 4.6%, industry’s rose from 28.7% to 33%, while services’s share changed very little, from 58% to 62.4%. In roughly the same period, from 1950 to 1980, one can observe a change in the sectoral distribution of employment (Table 5.2); agriculture declined from 64.4% to 38.1%, industry rose from 16.5% to 22.8%, and services from 19.2% to 39%. By 1980 the share of industry in GDP reached its peak of 33% (with manufacturing reaching 21.1%), and similarly, employment in industry reached 22.8% (with manufacturing reaching 12.7%). Bonelli et al. (2013, 69) analyzing Brazil’s industrialization via shares of GDP in constant prices, find that this process began in the mid-​ 1970s and continued steadily until the second decade of the twenty-​first century. Within manufacturing they note a decline of the shares of textiles and related products, chemical products, and plastics, while there was a notable growth in the shares of pharmaceutical products, machinery, electrical equipment, and transport equipment. In the mid-​1990s, as already noted, Brazil introduced a successful stabilization program while the country followed policies of trade and market liberalization. Using interest rates as one of its principal policy instruments and adopting a floating exchange rate, there occurred a long period of currency appreciation. This has had a negative effect on Brazilian industry, making it less internationally competitive and contributing to a decline in its participation in GDP, falling to 18% by 2010. Traditionally, industries have been viewed as the main driver of economic development, while services have been associated with low-​productivity activities (Baumol 1967). The best-​known exposition of this is found in the writings of Kaldor (e.g., 1966), who viewed manufacturing as the leading “engine of growth.” However, recent experiences, such as that of Brazil, present a slightly different picture. Growth has gone hand in hand with an increasing share of the tertiary sector in total GDP and employment. This shift has been called “tertiarization,” and seems to be occurring at an earlier stage of economic development than was the case with older industrial economies. It has in turn prompted fears of “premature deindustrialization” (McMillan and Rodrik 2011). The latter has to take into account that services do not automatically close avenues to industrial development. There has been an increasing trend in the twenty-​first century for industry to outsource tasks that were previously performed internally. Thus the growing role of services does not automatically imply deindustrialization. Also to be taken into account is that “modern” services can create positive spillovers into the entire economy. Many modern services have become increasingly tradable, highly productive, and technologically intensive (Rowthorn and Ramaswamy 1999). As far as Brazil is concerned, evidence suggests that structural change has favored low-​ productivity activities within services and that structural change has not been growth-​ enhancing (Cruz et al. 2008; McMillan and Rodrik 2011). However, it must be noted that the quality of employment has increased since 2000. Within services, labor has shifted from informal to formal activities. The share of labor in formal services has increased from 48% to 55% between 2000 and 2009 (Aldrighi and Colistete 2012). When viewed

Brazil’s Import-Substitution Industrialization    103 from this perspective, Brazilian services may not have performed as badly as has been suggested.

Notes 1. The link of Brazil’s coffee support program to the import substitution of the 1930s was first analyzed by Celso Furtado in his classic work Formação econômica do Brasil (11th edition, 1972). 2. SUMOC was the forerunner of Brazil’s Central Bank. 3. Which would later add the word “social,” becoming the current BNDES. 4. Imports of capital goods as a percentage of total supply declined from 59% in 1949 to 12.9% in 1962; for intermediate goods it fell from 25.9% to 8.9%; and for consumer goods it fell from 10% to 1.1%. 5. Critics of ISI pointed out that this dilemma was the result of across-​the-​board ISI, without any consideration of being more selective by choosing only sectors with a potential comparative advantage. The problem, however, was that many advanced industrial countries were protective of their older industries, such as textiles and shoes. 6. For details on the political situation of the period, see Skidmore (1967, Chapter 6) and Roett (1984). 7. For individual industry studies, see Baer (1969) and Amann (2000). 8. Some parts of this section are based on the article “Industrialization and De-​ Industrialization in Emerging Economies: The Cases of Brazil and India” by Werner Baer and Rahul A. Sirohi (2016).

References Abreu, Marcelo de Paiva. 1990. A ordem do progresso:  Cem anos de política econômica republicana, 1889–​1989. Rio de Janeiro: Editora Campus. Amann, Edmund. 2000. Economic Liberalization and Industrial Performance in Brazil. Oxford: Oxford University Press. Amann, Edmund, and Werner Baer. 2002. “Neoliberalism and Its Consequences in Brazil.” Journal of Latin American Studies 34 (4): 945–​959. Baer, Werner. 1969. The Development of the Brazilian Steel Industry. Nashville, TN: Vanderbilt University Press. Baer, Werner. 2014. The Brazilian Economy: Growth and Development, 7th edition. Boulder, CO: Lynne Rienner. Baumol, William J. 1967. “Macroeconomics of Unbalanced Growth: the Anatomy of Urban Crisis.” American Economic Review 57 (3): 415–​426. Bonelli, Regis, Samuel Pessoa, and Silvi Matos. 2013. “Desindustrialização no Brasil: Fatos e interpretação.” In O futuro da indústria no Brasil, edited by Edmar Bacha and Monica Baumgarten de Bolle, 45–​79. Rio de Janeiro: Civilização Brasileira. da Cruz, Marcio José Vargas, Gabriel Porcile, Luciano Nakabashi, and Fábio Dória Scatolin. 2008. “Structural Change and the Service Sector in Brazil.” Working Paper No. 95. Universidade Federal do Paraná, Departamento de Economia.

104   Werner Baer De Negri, Fernanda, and Luiz Ricardo Cavalcante, eds. 2014. Produtividade no Brasil: Desempenho e determinantes. Brasília: IPEA and ABDI. Dean, Warren. 1969. The Industrialization of São Paulo, 1880–​1945. Austin:  University of Texas Press. Furtado, Celso. 1972. Formação econômica do Brasil, 11th edition. São Paulo:  Companhia Editora Nacional. Kaldor, Nicholas. 1966. Causes of the Slow Rate of Economic Growth of the United Kingdom. Cambridge: Cambridge University Press. McMillan, Margaret S., and Dani Rodrik. 2011. “Globalization, Structural Change and Productivity Growth.” National Bureau of Economic Research, No. w17143. Roett, Riordan. 1984. Brazil: Politics in a Patrimonial Society, 3rd edition. New York: Praeger. Rowthorn, Robert, and Ramana Ramaswamy. 1999. “Growth, Trade, and Deindustrialization.” IMF Staff Papers 46 (1): 18–​41. Silber, Simão David. 2011. “Foreign Trade and Foreign Investments: The Brazilian Experience in the Last Two Decades.” In The Economies of Argentina and Brazil: A Comparative Perspective, edited by Werner Baer and David Fleischer, 441–​467. Cheltenham, UK: Edward Elgar. Skidmore, Thomas E. 1967. Politics in Brazil, 1930–​ 64:  An Experiment in Democracy. New York: Oxford University Press. Villela, André. 2011. “A Bird’s Eye View of Brazilian Industrialization.” In The Economies of Argentina and Brazil: A Comparative Perspective, edited by Werner Baer and David Fleischer, 38–​65. Cheltenham, UK: Edward Elgar.

Chapter 6

Ex periences of I nfl at i on and Stabili z at i on, 196 0–​1 9 9 0 Fernando de Holanda Barbosa

6.1. Introduction This chapter surveys the Brazilian experience of inflation and stabilization during the roughly three decades from the 1960s to the start of the 1990s. It is worth remembering that until the Real Plan in 1994, inflation was so entrenched in the daily life of Brazilian society that some analysts claimed that it was simply a national-​cultural phenomenon on a par with football, carnival, samba, and black beans and rice. The early 1960s saw stagflation and an inflation rate that rose until it reached 90% in 1964. The import-​substitution industrialization (ISI) model was facing several problems that had to be tackled. The stabilization plan put into effect in 1964 succeeded in reducing inflation to a level of 20% by the end of 1967, which endured until 1973. However, in 1974, the rate of inflation doubled and remained near 40% until 1979. Some analysts single out the first of the oil shocks that occurred in 1974 as the source of the inflation rate in that year. However, that cannot have been the case because the increase in the oil price did not transmit into fuel products. In fact, the real reason for the leap in the rate of inflation was the increase of the money supply in 1973, the final year of Delfim Netto’s tenure as finance minister of the Garrastazú Médici government. By the end of the 1970s, the ISI model had reached its end, and the Brazilian economy entered a new phase with a rising inflation rate and a decreasing growth rate. The rate of inflation increased to a new level by nearly 100% in each of the following years: 1979, 1980, 1981, and 1982. Public finances were in disarray, the external debt unbearable. In 1983, the rate of inflation resumed its increasing trend without reaching a new steady state. In 1986, due to the Cruzado Plan’s price freezing, the rate of inflation was lower than in 1985, but over the subsequent three years, 1987,

106    Fernando de Holanda Barbosa 1988, and 1989, inflation accelerated, reaching almost 2,000% by 1989. In 1990, the inflation rate did not increase, due to the Collor Plan, but nevertheless was close to 1,600%. At around this point we can say that Brazil ceased to be a chronic inflation country and instead entered the pathology of hyperinflation. It remained there until the Real Plan in 1994. The choice of a monetary regime in which the central bank works as if it were a department of the internal revenue service, with the task of issuing money to finance government expenditures, characterizes chronic inflation and hyperinflation, as shown in section 6.2. When the central bank uses a nominal anchor to pin down the price level, it can choose one of three options, namely, (1) exchange rate targeting, (2) monetary targeting, or (3) inflation targeting. Thus, there are several monetary regimes to choose from. The choice of a particular monetary regime is not due to Brazilian culture but rather is the end result of a political game played by different groups within Brazilian society, each seeking to protect their own interests. A useful way to analyze inflation, as well as other macroeconomic variables from a theoretical point of view, is in terms of impulses and propagation mechanisms (Slutsky 1937). This framework provides a taxonomy of the different types of stabilization: orthodox, heterodox, and neo-​orthodox.1 In an orthodox stabilization, the plan deals with the impulse mechanism that yields the trend rate of inflation by using fiscal and monetary tools. In a heterodox stabilization, the plan tries to stop the propagation mechanism, the inertial component of inflation, by using incomes policy, such as prices and wages controls, or even freezing the price system. In a neo-​orthodox stabilization, the plan takes into account the impulse as well as the propagation mechanism. Indeed, neo-​ orthodox stabilization is a combination of the orthodox and the heterodox approaches to combat inflation. Heterodoxy by itself cannot get rid of inflation. One way to analyze inflation is to look at it as a tax rate on money. The inflation tax is a source of revenue for the government. To cut the inflation revenue tax, it is necessary to use one of four options: (1) increase other taxes; (2) cut expenditures; (3) borrow issuing public debt; or (4) combine the three previous tools. Thus, to decrease the trend rate of inflation, it is necessary to use fiscal policy. This chapter analyzes five stabilization plans. The first is the PAEG (Government Economic Action) Plan, a neo-​orthodox stabilization plan. The other four are heterodox plans—​Cruzado, Bresser, Summer, and Collor—​implemented during the so-​called lost decade of the 1980s. All of them failed to stabilize inflation. No stabilization plan was attempted between PAEG and the Cruzado Plan. The PAEG Stabilization Plan began as a fully orthodox plan, but during its implementation the social cost imposed by the inertial component of inflation became apparent. Simonsen (1970), who was the mastermind of this plan, devised an indexation mechanism with a forward component in wages readjustment, decreasing the backward component. This mechanism, which intended to preserve the average worker’s real wage, became a standard tool of stabilization plans in Brazil, despite the criticism it received at that time. The PAEG Plan gave up orthodoxy, and it became the first neo-​orthodox stabilization plan because it discovered the role played by the

Experiences of Inflation and Stabilization, 1960–1990    107 inertial component of the inflation process and designed mechanisms to deal with it. This inertial component will play the most important role in the four heterodox plans to be analyzed in this chapter.

6.2.  Chronic Inflation, Hyperinflation, and Stabilization Chronic inflation is the phenomenon of high inflation sustained over a long period. This definition describes its chief characteristic but does not provide empirical criteria to identify episodes of chronic inflation. The threshold of 10% is one candidate for a benchmark; inflation rates above 10% for a long period would then amount to chronic inflation. The inspiration for this definition probably comes from Cagan’s (1956) definition of hyperinflation, namely that hyperinflation begins in a month when the price level increases at least by 50% and ends when the price level drops below 50% and stays there for at least one year. This was Cagan’s empirical criteria to identify episodes of hyperinflation; there is no theoretical justification for choosing these arbitrary thresholds to define either chronic inflation or hyperinflation. We will use the type of impulse mechanism to define both chronic inflation and hyperinflation. Chronic inflation is the result of a regime where money issues finance the public deficit on a permanent basis. In this environment, there is no nominal anchor.2 The central bank cannot control the exchange rate, the money supply, or the nominal interest rate. The fiscal deficit financed by money is the impulse for both chronic inflation and hyperinflation models. In countries with chronic inflation, there are lagged indexation mechanisms for both prices and wages, yielding inertial inflation, because economic agents have learned to live with permanent inflation. Inertial inflation is the propagation mechanism, but it is not the inflation source. In chronic inflation processes, the indexation intervals shrink due to the acceleration of inflation. The change of the average period of the lagged indexation may cause important changes in the dynamics of inflation, despite the fact that it does not affect the steady states of the model. The dynamics of such an economy can be very unstable under parameter changes or shocks that hit the economy. Furthermore, the inertial component of inflation would impose a very high social cost to reduce inflation. Hyperinflation is a pathology that arises when the price of money goes to zero in finite time. Why would people be willing to get rid of money? One possibility is a bubble caused by a self-​fulfilling prophecy: everybody believes that tomorrow money will be worthless, and so rid themselves of it today. A  second reason for money to become worthless grounded on fundamentals lies in a fiscal crisis. In the monetary fiscal regime in which the central bank finances the treasury, we can think of the central bank as a money rental business, just like a car rental business. The central bank owns the real

108    Fernando de Holanda Barbosa quantity of money, as the car rental company owns the car. The central bank’s income stems from renting the money to society. The rental price charged by the central bank is equal to the value of the services provided by money, which is equal to the nominal interest rate times the real quantity of money. The central bank uses this income to provide the funds needed by the treasury to pay government expenditures. The residue is the cash flow for the central bank. The value of money, as any other asset, is the discounted value of its cash flow. Thus, this regime works as long as the rental price received by the central bank is larger than the resources transferred to the treasury (i.e., as long as its cash flow is positive) (Barbosa 2016). When this cash flow becomes negative, hyperinflation would be the outcome of an unsustainable regime. The present value of the cash flow is the intertemporal budget constraint of the government in a fiscal monetary regime that finances the government issuing money. We define hyperinflation as beginning in the month when the intertemporal budget constraint is not sustainable, conditional on no change in the economic policy regime, and ending in the month when the constraint is satisfied. Some stabilization programs succeeded in reducing the inflation rate from a high to a low rate (PAEG of 1964–​1967; see Israel 1985). This type of stabilization plan reduces inflation without a change in the economic policy regime. The other type of stabilization changes the economic policy regime itself. We define a regime as the rules of the game, encompassing goals and instruments of policy. The end of hyperinflation occurs with a change in regime. There is some controversy about this notion of regime change. According to Sargent (1982), hyperinflation ends with a change in both fiscal and monetary regimes. Dornbusch (1987) disagrees with this verdict and claims that hyperinflation ends with the change in the monetary regime; the change in the fiscal regime occurs after the stabilization of the price level.3 To test these two competing hypotheses, we must analyze the end of each hyperinflation.

6.3.  PAEG (March 1964 to December 1967) Roberto de Oliveira Campos, as minister of planning, and Octávio Gouvêa de Bulhões, as minister of the treasury, carried out the economic policy of the first government of the military regime, under President Castello Branco (1964–​1967). Mario Henrique Simonsen, an informal government advisor, designed the PAEG Plan. The plan was very similar to the Triennial Plan of Celso Furtado, prepared for the João Goulart government in 1962. Both intended to combat inflation using a combination of monetary and fiscal policy, and both used the Harrod/​Domar model as a theoretical framework to formulate economic growth policy. As PAEG’s full title (Government Economic Action Plan) suggests, it was not just a stabilization plan. The lack of macroeconomic discipline of the Kubitschek government

Experiences of Inflation and Stabilization, 1960–1990    109 (1956–​1960) led the Brazilian economy into stagnation and rising inflation in the beginning of the 1960s. The PAEG Plan had two main goals: stabilization, and resuming growth. The PAEG Plan carried out a fiscal policy in two stages. In the first stage (1965–​1966), it increased consumer tax rates, enlarged the income tax base, broadened the system of withholding tax at source, and indexed the tax at the time the tax was due. This eliminated the Tanzi effect, the loss of real value of tax revenue due to the lag in collecting the tax. This tax indexation became widespread among states and municipalities setting aside the Tanzi effect during the Brazilian inflation experience analyzed in this chapter. Inflation also affects the government budget through government expenditures (Barbosa 1987). Unexpected inflation decreases real government expenditures. Thus, if stabilization is not fully expected, real government expenditures could increase, deepening the fiscal deficit. The second stage of the fiscal policy was a fiscal reform that came into effect in 1967. The fiscal reform abolished inefficient taxes, such as a sales tax with a cascading effect. It introduced value-​added taxes, one for the federal government (IPI) and another for the states (ICM). These two value-​added taxes were a compromise solution because as part of Brazil’s organization as a federal system, income tax belongs to the federal government and property tax to the municipalities. Import and export taxes pertain to the federal government. The PAEG Plan implemented two other reforms: (1) social security, and (2) banking and financial. The social security reform extinguished several social security institutions and established just one, a national institute. The usury law in place at that time, with a ceiling of 12% per year for the nominal rate of interest, prevented the development of the financial sector. The first step in the financial reform was to abolish this law. The second step was to create a legal framework for its institutions. The banking reform adopted the American system of segmented markets with commercial banks, investment banks, consumer credit banks, real estate banks, and brokers. Only commercial banks could issue demand deposits, and the other banks could issue different types of time deposits. The integration of the big financial institutions, which had previously all operated in segmented markets, became a reality. This system evolved, and a decision enacted by the National Monetary Council in 1988 allowed the functioning of universal banks. The banking reform also created the Central Bank of Brazil, as an independent central bank like the Federal Reserve in the United States. However, it differed from the institutional organization of its United States counterpart in that a council was established, the National Monetary Council, in charge of monetary, credit, and financial regulation and policymaking. This council was presided over by the minister of finance, but the executive branch of government would have just three votes among nine members. Six members had a fixed mandate. The rationale for this setup was the need to coordinate monetary and fiscal policy. To develop the mortgage market, the real estate banks issued time deposits with an indexation mechanism. To finance the public deficit, the federal government issued

110    Fernando de Holanda Barbosa indexed bonds (ORTN, National Treasury Indexed Liabilities) with the same indexation mechanism of the savings deposits. One of the most important tools of the PAEG Plan was to increase the savings ratio through forced savings. Until 1964, an employee after working 10 years at the same firm would acquire tenure because the cost of firing him would have been incredibly high. (In fact, it was common for firms to fire workers before they completed 10 years.) The PAEG Plan changed this law and introduced a fund, called FGTS (Fundo de Garantia do Tempo de Serviço; Length of Service Guarantee Fund) that every worker would be obliged to contribute to with 8% of his or her salary. The workers could draw from this fund either when fired or to buy a house, under certain conditions. The other component of forced savings came from three taxes, on fuels, electricity, and telecommunication services. These taxes financed investment, and they could not be used in consumption. The PAEG plan intended to be an orthodox plan using only monetary and fiscal tools to combat inflation. Those in charge of implementing it recognized, after a short period, that the inflation rate was not falling as expected. The government tasked Simonsen with finding out what was happening. He discovered the inflation feedback component of inflation, or “inertial inflation,” as it is commonly known today in all macroeconomics textbooks. The inflation literature at that time did not address this issue. Inflation feedback exists because prices and wages adjust based on past inflation. To cut this link between the past and the present, it is necessary to change the indexation mechanism. A new law made it mandatory for wage adjustments to use a formula that combined past inflation and expected inflation in such way that it preserved the average real wage of the worker. However, since this expected rate was underestimated, there was a sharp drop in real wages. Subsequently the government recognized the problem and corrected the wage indexation formula to take into account any discrepancy between actual and expected inflation. Because of this new formula, the inflation rate declined without causing a recession. The PAEG Plan succeed in bringing the trend rate of inflation to less than 2% per month by the end of 1967, as shown in Figure 6.1. This rate was much lower than the monthly rates of inflation during the first quarter of 1964. In hindsight, we can say that the PAEG Plan committed at least two mistakes. The first was to allow the Banco do Brasil to continue to issue money as if it were a central bank. This arrangement gave rise to a new government budget besides the fiscal budget, namely the monetary budget resulting from the consolidation of Central Bank and Banco do Brasil accounts. This institutional arrangement produced quasi-​fiscal deficits and a lack of transparency in Brazilian public finances for a long time. The second criticism comes from the growth strategy. Growth theory in the 1960s did not include human capital in its models, as can be seen in the well-​known Harrod-​Domar model. Thus, the PAEG Plan did not include education among the reforms that changed a number of Brazilian institutions in the 1960s.

Experiences of Inflation and Stabilization, 1960–1990    111 12

Monthly inflation rate

10 8 6 4 2

Ja

n.

1 Ap 964 r. 1 Ju 964 l. 1 Oc 964 t. 1 Ja 964 n. 1 Ap 965 r. 1 Ju 965 l. 1 Oc 965 t. 1 Ja 965 n. 1 Ap 966 r. 1 Ju 966 l. 1 Oc 966 t. 1 Ja 966 n. 1 Ap 967 r. 1 Ju 967 l. 1 Oc 967 t. 19 67

0

Inflation

Figure 6.1. PAEG Plan. Source: FGV Revista Conjuntura Econômica.

6.4.  Cruzado Plan (February 1986 to December 1986) The hypothesis behind the Cruzado Plan was that Brazilian inflation was 100% inertial (Lopes 1986). The source of inflation was inflation itself, due to the use of widespread indexation mechanisms. Thus, the strategy of this plan was (1) to cut the link between the past and the present, like breaking a turnstile, and (2) to forbid any type of indexation shorter than one year. The Cruzado Plan introduced a new currency. The cruzado replaced the cruzeiro as the Brazilian currency, cutting three zeros. All prices were frozen for an indeterminate period. The economy was de-​indexed, and no indexation mechanism was allowed for short periods. The plan extinguished the National Treasury Indexed Liabilities (ORTN) created by PAEG. The National Treasury Liability (OTN) replaced it, and its price remained fixed for one year. Wages, rents, school fees, and mortgage installments were converted from cruzeiro to cruzado by using the average value over the six months prior. Minimum wages received an 8% bonus above their average value. The Cruzado Plan also introduced a trigger point to readjust wages, namely every time inflation reached 20%.

112    Fernando de Holanda Barbosa A tablita, a device imported from stabilization plans in Spanish-​speaking Latin American countries, converted contractual liabilities, to avoid income transfers between creditors and debtors. The exchange rate was fixed at 13.80 cruzados per dollar, but the central bank could change this at its own discretion. At first, the Cruzado Plan was a great success. The inflation rate dropped below 1% per month from March to July 1986. However, the Brazilian public did not trust inflation to remain at the same rate in the near future, and went on a shopping spree, overheating the economy and causing a shortage of goods. Due to price controls, supply bottlenecks became common. Price freezing became unsustainable. In an open economy, one way to solve this problem is to increase imports and to decrease exports; but in this case the result was to aggravate the balance of payments problem. By the end of 1986, Brazilian international reserves were depleted and the country was ready to default. Monetary policy, conducted on the assumption that expected inflation was near zero, was expansionist. The central bank was pumping money into the economy in a clear inconsistency with the goal of having a low inflation rate. The Brazilian public was getting rid of this money buying goods. The only way to stop this behavior was for the central bank to realize that it would have to raise interest rates. However, the Brazilian central bank faced a problem when it decided to increase the interest rate. Most financial institutions that were carrying treasury bills had a very high leverage ratio. Once the central bank raised the interest rate, these financial institutions would go broke. To solve this problem the central bank created a new bill, today called treasury financial bills (LFTs), indexed daily to the interbank market interest rate.4 The duration of this bill is zero, meaning that the interest rate does not affect its price. This bill allowed the Brazilian economy to avoid dollarizing during hyperinflation, as other Latin American countries like Argentina had done. This bill backed the indexed money issued by commercial banks with daily indexation. Money has three functions: unit of account, means of payment, and reserve of value. In the Brazilian inflation experience, domestic currency was the means of payment, and indexed money was a reserve of value. People with access to banks used indexed money because this provided a hedge against inflation. The only way poor people could protect from inflation was to buy and store goods as soon as they received their salary. As for the unit of account, it would depend on the value of the good and the kind of business. For example, big supermarket chains used their own unit of account to mark their prices, and then converted the prices into domestic currency at the cashier. Before the creation of the Brazilian Central Bank in 1964, Banco do Brasil, a commercial bank owned by the government, was the Brazilian central bank. Banco do Brasil, created in 1808 when the Portuguese court came to Brazil, was a copy of the Bank of England, at that time both a commercial and a central bank. However, while the Bank of England changed along the way and became a central bank, Banco do Brasil did not change and had a powerful lobby defending its status quo. When the Brazilian Central Bank began operations in 1965, there was an agreement between the two institutions to create a special account facility (the Conta Movimento) that would allow Banco do Brasil to draw money from the central bank. During the Cruzado Plan, this overdraft

Experiences of Inflation and Stabilization, 1960–1990    113 30

92.00 90.00 88.00

20

86.00 84.00

15

82.00 10

Output

Monthly inflation rate (%)

25

80.00 78.00

5

76.00

Inflation

May. 1987

Apr. 1987

Mar. 1987

Feb. 1987

Jan. 1987

Dec. 1986

Nov. 1986

Oct. 1986

Sep. 1986

Aug. 1986

Jul. 1986

Jun. 1986

May. 1986

Apr. 1986

Mar. 1986

Feb. 1986

–5

Jan. 1986

0

74.00 72.00

Output

Figure 6.2. Cruzado Plan. Sources: FIBGE and FGV.

facility was closed. The Brazilian Central Bank became a central bank, and Banco do Brasil just a commercial bank. Toward the end of 1986, the Sarney government enacted an economic policy package known as Cruzado II. It increased indirect taxes, and readjusted the prices of public utilities and fuel products. Besides being a very weak package, it was too late to change the course of events, and the inflation rate in January 1987 was nearly 10%. The Cruzado Plan caused a transitory increase in the level of economic activity, measured by industrial product index, as shown in Figure 6.2.5 In the last quarter of 1986, with the international reserves depleted, the Central Bank had to stop servicing the foreign debt. In the first quarter of 1987, the minister of finance, Dilson Funaro, resigned. The verdict of the facts, shown in Figure 6.2, rejected the hypoth­esis that Brazilian inflation was purely inertial. Inflation had rejoined the hyperinflation path before the Cruzado Plan.

6.5.  Bresser Plan (June 1986 to December 1986) In May 1987, Bresser Pereira became minister of finance, replacing Funaro, who had served as minister during the Cruzado Plan. In June 1987, the Bresser Stabilization Plan

114    Fernando de Holanda Barbosa

40

90.00

35

89.00 88.00

30

86.00 20 85.00 15

Output

87.00

25

84.00

10

83.00

Inflation

Jan. 1989

Dec. 1988

Oct. 1988

Nov. 1988

Sep. 1988

Jul. 1988

Aug. 1988

Jun. 1988

Apr. 1988

May. 1988

Mar. 1988

Jan. 1988

Feb. 1988

Dec. 1987

Nov. 1987

Oct. 1987

Sep. 1987

Jul. 1987

81.00 Aug. 1987

0 Jun. 1987

82.00 Apr. 1987

5

May. 1987

Monthly inflation rate (%)

was set forth. This plan involved three stages. The first stage froze prices and wages for a period of 90 days. In the second stage, after the frozen period, prices and wages could be readjusted monthly to avoid the problems of shortage of goods faced by the Cruzado Plan. Unlike the previous plan, there was no trigger point to readjust salaries. A new index, called the Price Reference Unit (URP), corrected wages. This index was the geometric average of the monthly Consumer Price Index for the three previous quarters. This index provided the limit allowed for increasing prices. By September, the inflation rate was rising, as shown in Figure 6.3, and the economy was back on the hyperinflation path. The level of activity, as measured by the industrial production index, decreased during the first months of the plan. By the third quarter of 1987, industrial production began to increase, reaching its peak by the first quarter of 1988. The Bresser Plan did not intend to be a fully heterodox stabilization plan; the intention was to use fiscal and monetary policy to control demand. Thus, the third stage of the Bresser Plan, to be implemented in 1988, would take care of the public deficit problem. At the end of 1987, Bresser presented a series of measures to reduce the deficit, but his own government rejected them. The only option for him was to present his letter of resignation.

Output

Figure 6.3.  Inflation, April 1987–​January 1989. Sources: FIBGE and FGV.

Experiences of Inflation and Stabilization, 1960–1990    115

6.6.  Summer Plan (January 1989 to December 1989) Bresser Pereira’s deputy Mailson da Nóbrega became the new finance minister. His strategy was to try put in place what we could call a “black beans and rice” economic policy—​not a gourmet recipe, but a simple and efficient staple. The goal was to stabilize the rate of inflation to around 15% per month and to reduce the public deficit. However, by December 1988 the inflation rate was at 29%. The gradualist approach to combating hyperinflation was always bound to fail. The pathology of hyperinflation requires not feijão but cold turkey. In January 1989, in the Brazilian summer, a new stabilization plan applied a shock treatment to Brazilian hyperinflation. This time the plan was called the Summer Plan. It used the standard heterodox procedure of price freezing and de-​indexation. Prices were frozen for an indeterminate period, with the intention that it would not last more than 90 days. This concern reflected the experience of the Cruzado Plan, with scarcity and bottlenecks in the supply of goods whose prices were out of equilibrium. The de-​ indexation prohibited monetary correction clauses in contracts spanning a period of less than 90 days. The Summer Plan cut three zeros from the cruzado and named the new currency cruzado novo. After a devaluation of 17% on January 15, the exchange rate was fixed to the new currency at the price of one cruzado per US dollar. As with the Cruzado Plan, a tablita converted the values of contracts drawn between January 1, 1989, and January 15, 1989. The factor used per day assumed a monthly inflation rate of 13.56%. The Summer Plan extinguished OTN (the National Treasury Liability), replacing it with the National Treasury Bond (BTN). The average value in OTNs during the period from January to December 1988 converted wages to cruzados novos. These values were computed to February 1 by the change in value of the Price of Reference Unit index, a change of 26.05%. The Summer Plan intended to carry out a fiscal adjustment cutting public spending, privatizing state-​owned enterprises, and firing public employees. Law 7,730, approved by Congress on January 31, authorized the issuance of public debt just to pay the interest rate and to roll over the principal. The primary deficit was supposed to be zero. Congress vetoed the fiscal measures, and the Summer Plan became another heterodox plan to fail. In July 1989, inflation returned to the same rate as before the plan, as shown in Figure 6.4, and the Brazilian economy was back on the hyperinflation path. As in the Cruzado Plan, the level of activity, measured by the industrial production index, had a transitory movement, increasing at the beginning of the Summer Plan but after some months returning to its former level.

40

90.00

35

89.00 88.00

30

86.00 20 85.00 15

Output

87.00

25

84.00

10

83.00

Jan. 1989

Dec. 1988

Oct. 1988

Nov. 1988

Sep. 1988

Jul. 1988

Aug. 1988

Jun. 1988

Apr. 1988

May. 1988

Feb. 1988

Inflation

Mar. 1988

Jan. 1988

Dec. 1987

Oct. 1987

Nov. 1987

Sep. 1987

Jul. 1987

81.00

Aug. 1987

0

Jun. 1987

82.00 Apr. 1987

5

May. 1987

Monthly inflation rate (%)

116    Fernando de Holanda Barbosa

Output

Figure 6.4. Summer Plan. Sources: FIBGE and FGV.

6.7.  Collor Plan (March 1990 to December 1990) The Collor Plan of March 1990 had, like the PAEG Plan, two goals: stabilization and growth.6 The ISI strategy had ended during the Geisel government (1974–​1979), with the substitution of capital goods and basic inputs. The Brazilian economy was essentially an Asian Tiger during the period 1920–​1980, growing at an average rate of 7% per year. However, in the so-​called lost decade of the 1980s, growth came to a halt and inflation accelerated. The Brazilian economy faced two crises. First, the ISI strategy was over and the domestic market was overprotected from international competition, with both tariff and nontariff barriers. Second, the country’s public finances were in disarray, with the federal government issuing money to pay its bills and defaulting on its foreign debt. There was no money left for the government to finance investment in infrastructure and to supply capital to its state-​owned enterprises (SOE), which covered sectors such as banking, electricity, mining, oil, steel, and telecommunications. The Collor Plan opened up the economy overnight, using a cold turkey strategy, arguing that private interests, vested or not, would not allow a gradualist approach. The Collor Plan also began privatization of some of its SOEs. The first was the Brazilian

Experiences of Inflation and Stabilization, 1960–1990    117 state’s crown jewel, the National Steel Company (CSN), created during the Vargas government (1930–​1945) and perceived by many Brazilians as a symbol of the industrialization process.7 The Collor Plan carried out a state reform, closing several state agencies, such as the Brazilian Institute of Coffee (IBC) and the Alcohol and Sugar Institute (IAA), agencies in charge of state intervention in those sectors. The plan also reduced the number of ministries to 12, attempting to bring some order to the chaotic federal public organization. The Collor Plan followed the standard procedure of heterodox plans by introducing a new currency, the cruzeiro, freezing prices for a short period, converting wages by the average of the previous year, and using a tablita to convert assets and liabilities. The plan had two innovations. First, the exchange rate was flexible, meaning that money was the anchor of the plan. The second innovation was a blocking and partially confiscatory monetary reform; 70% of all financial assets included in the broad definition of money M4 (currency + demand deposits + savings deposits + public securities held by the public + time deposits + exchange bills), except currency, were blocked and transformed in time deposits, with a rate of interest of 6% per year plus monetary correction based on the fiscal BTN. The plan blocked those assets for 18  months, with no secondary market allowed and the understanding that the owners would have their assets returned later in 12 monthly installments. One rationale behind the blocking and partial confiscation was the hypothesis of monetary overhang. Monetary overhang occurred in some European countries after World War II, such as in Germany in 1948, due to monetized public deficits and controlled prices. This phenomenon implies that the real quantity of money that individuals are holding is larger than the quantity of money they are willing to hold. To eliminate this overhang, either the central bank reduces the quantity of nominal money, or it allows the price level to rise. The option of increasing real income is not feasible due to resource constraints. Policymakers do not like to use the option of price level increases because an increase in the price level will amount in the short run to an increase in the rate of inflation. They fear that this transitory increase in the rate of inflation will become permanent since inflation may feed itself. At the time of the Collor Plan, did the Brazilian economy have a monetary overhang? Despite the hyperinflation, it is very likely that this was not the case because the financial markets were working without any restrictions and prices were free to adjust. Thus, the money market broadly defined was in equilibrium, without any discrepancy between desired and existing real cash balance. As such, we can conclude that a monetary overhang was not a good rationale for blocking financial assets. A second rationale for the blocking and partial confiscation of financial assets given by some economists was that it would make it impossible for individuals to bet against the stabilization plan by exchanging goods for their financial assets. According to this argument, previous heterodox plans had failed because individuals had “attacked” by buying goods and, therefore, increasing prices and yielding inflation. Thus, the argument goes, the best strategy is to remove their ammunition from the battle for such time as is necessary to stabilize the economy. However, the weak point of this argument is

118    Fernando de Holanda Barbosa that it ignores the fact that individuals will bet against a stabilization plan when they believe that it will not succeed. A simple way to measure the credibility that a stabilization plan has managed to attain is through gauging the gap between the level of inflation expected by the public and the level of inflation targeted by the policymaker. Blocking and confiscating financial assets is like trying to measure a patient’s fever by throwing away the thermometer. In such an environment, the policymaker lacks the proper information to conduct monetary policy—​for instance, to raise the interest rate if price freezing produces an overheating of the economy. A stylized fact of hyperinflation experiences is currency substitution. The use of a foreign currency as a medium of exchange is known as dollarization. In the Brazilian experience, there was a different type of currency substitution. Instead of dollarization, the Brazilian economy created “indexed money,” issued by commercial banks and backed by daily indexed government bonds. The demand-​deposit banking account was linked to the indexed money account, a kind of money market fund, and each time a check was drawn the money would be automatically transferred from the indexed money account to the demand account. As a result, there was a surge in the size of the financial sector, as shown in Table 6.1. The value added of the financial sector was 8% of gross domestic product (GDP) in 1980. By the end of the 1980s, it reached 21%. In 1995, the year following the end of hyperinflation, the value added of the financial sector was back to its normal level, 7% of GDP. A third rationale for the blocking and partial confiscation of financial assets, according to Carvalho (1990), was the monetary reform proposed by Belluzzo and Almeida (1990) in a paper that circulated in 1989. The basic idea behind the monetary reform was to get rid of the dual monetary regime system, with both a conventional money and an indexed money. The origin of Brazilian hyperinflation, in the view of this

Table 6.1 Brazilian Financial Sector: Value Added (% GDP) Year

Value Added (%)

Year

Value Added (%)

1980

8

1988

14

1981

10

1989

21

1982

10

1990

12

1983

12

1991

10

1984

12

1992

12

1985

12

1993

14

1986

8

1994

12

1987

14

1995

7

Source: IBGE. Brazilian Geographical and Statisical Institute.

Experiences of Inflation and Stabilization, 1960–1990    119 paper, was not a fiscal problem, but rather was linked to the cumbersome monetary regime in place. The monetary reform had as its main goal the recovery of the instruments of monetary policy. Blocking financial assets was necessary in order to control the remonetization of the economy. The partial confiscation of financial assets was a one-​time capital levy to promote social justice. However, the monetary reform proposed by Belluzzo and Almeida was based on an incorrect diagnosis of Brazilian hyperinflation. Indexed money, like other indexation mechanisms, was not the source of this pathology; it was just part of the hyperinflation propagation mechanism. Optimal monetary policy theory supplies the fourth rationalization for the central bank’s starting of a stabilization program reducing the stock of money. In order to minimize the social cost of a stabilization anchored by money, in an inertial inflation environment the first stage is to reduce the stock of money, which raises the interest rate, causing a recession and reducing the rate of inflation. Then, at the second stage, once the public’s expectation of the rate of inflation changes, the central bank increases the money supply to remonetize the economy due to an increase in the real quantity of money that individuals are willing to hold. This monetization is made at a variable speed, with a declining rate of growth, until it reaches the rate of growth of money consistent with the targeted rate of inflation. For the central bank to be able to choose the rate of growth of money, it is necessary to change the economic policy regime. The Collor Plan addressed the fiscal problem by cutting government expenditures, eliminating subsidies, and increasing tax revenue, but there was no change in the economic policy regime because most of the fiscal adjustment was transitory rather than permanent. Besides blocking financial assets, there was a capital levy of 8% on those assets made by using the tax on financial transactions (IOF). The government intended to combat tax evasion, but there was a gap between what was desired and achieved. The state reform stalled because it was unpopular. It faced great resistance from public employees. The Brazilian Constitution enacted in October 1988 is considered a “Citizen Constitution.” It states that it is the obligation of the state to provide education, health care, and social security for all citizens. It did not specify how to fund these obligations. These proposed measures would increase the overall tax burden. The new Constitution also transferred some tax revenues from the federal government to states and municipalities. The fiscal situation of the federal government was in disarray before the new constitution. Stabilizing the economy demanded getting rid of the inflation tax, which was in the range of 3%–​5% of gross domestic output. After the enactment of the new constitution, it became worse because it was necessary to increase taxes to finance the new obligations. Thus, the necessary fiscal consolidation was of a magnitude that was very difficult to carry out in the short run without great political support. The Collor Plan did not change the economic policy regime, and inflation went back to its hyperinflation path, as shown in Figure 6.5. Like the other heterodox stabilization

120    Fernando de Holanda Barbosa 90

88.00

80

87.00 86.00

60

85.00

50 84.00 40

Output

Monthly inflation rate

70

83.00

30

82.00

20

81.00

0

80.00

Fe

Ja

n.

19 89 b. 19 8 M ar 9 .1 9 Ap 89 r. 19 8 M ay 9 .1 9 Ju 89 n. 19 89 Ju l. 1 9 Au 89 g. 19 Se 89 p. 19 Oc 89 t. 19 No 89 v. 19 De 89 c. 19 Ja 89 n. 19 Fe 90 b. 19 9 M ar 0 .1 99 0

10

Inflation

Output

Figure 6.5. Collor Plan. Sources: FIBGE and FGV.

plans, it lasted less than a year. The overall level of activity in the economy decreased in the first months after the plan’s announcement, but recovered its loss by the end of the second quarter of 1990, and then remained flat for the rest of the year. In January 1990, the Collor government introduced a new stabilization plan, the Collor Plan II, which used the standard procedure of heterodox plans: (1) price freezing, (2) conversion of wages and rents, (3) a tablita, and (4) de-​indexation. Before freezing, the government increased the prices of wheat, electricity, fuel products, and telecommunication services. It extinguished the National Treasury Bonds (BTN) and created the Benchmark Interest Rate (TR, Taxa Referencial). It prohibited overnight deposits for individuals and nonfinancial corporations and created Financial Investment Funds (FAF in the Portuguese acronym). This time the plan did not even last one quarter because the economic team had eroded its credibility. A small political party had elected President Collor and he did not have a significant political support base in Congress, facing strong opposition both on the right and the left of Brazilian society. The industrial sector was against him because he opened up the economy. The financial sector opposed him because he blocked financial assets. The Marxist-​inspired left did not like him because he broached the antithesis of their ideology, privatization. Public employees hated him because of his state reform plan. The poor people who elected him were very unhappy with the fiasco of the Collor Plan stabilization. This trajectory culminated in Collor’s impeachment in 1992.

Experiences of Inflation and Stabilization, 1960–1990    121

6.8. Conclusion PAEG was the only stabilization plan that succeeded in bringing inflation down during the period 1960–​1990. The neo-​orthodox plan combined monetary fiscal policies with incomes policy to deal with the inertial component of inflation. In spite of its success, it did not change the economic policy regime. Brazil remained a chronic inflation country, issuing money to finance public deficits. PAEG also succeeded in building the foundations for resumed growth. Besides introducing several reforms, such as in social security, banking, the financial sector, and the tax system, it increased the savings ratio through a policy of forced savings. All of the heterodox plans reported in this chapter—​Cruzado, Bresser, Summer, and Collor—​failed to stabilize the economy. With the exception of the Collor Plan and the Bresser Plan, they produced a transitory increase in production. These heterodox plans focused on the inertial component of inflation. In the short run, these plans eliminated this component. However, the source of the inflation process, a fiscal problem that requires a permanent increase in tax revenue and/​or a cut in expenditures, remained unsolved. Thus, inertial inflation returned. It is interesting to note that following the first heterodox experiment, the Cruzado Plan, the economy returned each time to its hyperinflation trajectory in a progressively shorter time than after the previous experiment. The Collor Plan, like the PAEG plan, was not only a stabilization plan. It established an agenda that has been followed since then by Brazilian economic policymakers. This agenda includes opening up the economy, privatization, and state reform. Brazil’s inflation was finally stabilized with the Real Plan in 1994. This plan, besides getting rid of the inertial component of inflation, initially changed the monetary regime, and later, in 1998, the fiscal regime. The change in the monetary regime eliminated the inflation tax as a permanent source of finance for government expenditures. The main goal of the Central Bank was to control inflation, first by using an administered exchange rate system and then, from mid-​1999 onward, through an inflation target system. The fiscal regime changed in 1998 when the government established a primary fiscal surplus for each fiscal year that would make the public debt sustainable.

Notes 1. I borrow this classification from Nakano, cited in Faro, ed. (1990, 138). 2. This section, based on Barbosa (2017), is a verbal description of the mathematical representation of the economic policy regime in which the central bank finances government expenditures. 3. Franco (1990), who was among the policymakers behind the Real Plan, held the same position as Dornbusch on this issue. 4. For more details about this bill and its effects on monetary policy, see Barbosa (2006).

122    Fernando de Holanda Barbosa 5. There are no monthly estimates of GDP in Brazil. As a proxy for this variable, we use the industrial production index, seasonally adjusted. 6. The government called the plan the “New Brazil Plan,” but this name did not become widely used by the Brazilian public. 7. For a detailed analysis of the Brazilian steel industry, see Baer (1969).

References Arida, Persio, ed. 1986. Inflação zero: Brasil, Argentina, Israel. Rio de Janeiro: Paz e Terra. Baer, Werner. 1969. The Development of the Brazilian Steel Industry. Nashville, TN: Vanderbilt University Press. Barbosa, Fernando de Holanda. 1983. A inflação brasileira no pós-​guerra: Monetarismo versus estruturalismo. Rio de Janeiro: IPEA. Barbosa, Fernando de Holanda. 1987. “Inflação, indexação e orçamento do governo.” Revista Brasileira de Economia 41: 251–​273. Barbosa, Fernando de Holanda. 2006. “The Contagion Effect of Public Debt on Monetary Policy.” Revista de Economia Política 26: 231–​238. Barbosa, Fernando de Holanda. 2017. Exploring the Mechanics of Chronic Inflation and Hyperinflation. Cham, Switzerland: Springer. Barbosa, Fernando de Holanda, Rudiger Dornbusch, and Mario Henrique Simonsen, eds. 1992. De la estabilizacion al crecimeiento en America Latina. Chile: Cinde. Barbosa, Fernando de Holanda, and Mario Henrique Simonsen, eds. 1989. Plano Cruzado: Inércia x inépcia. Rio de Janeiro: Editora Globo. Belluzzo, Luis G. M., and Julio G. S. Almeida. 1990. “Crise e reforma no Brasil.” São Paulo em Perspectiva 4: 63–​75. Bresser Pereira, Luiz Carlos, and Yoshiaki Nakano. 1987. The Theory of Inertial Inflation: The Foundations of Economic Reform in Brazil and Argentina. Boulder, CO: Lynne Rienner. Cagan, Phillip. 1956. “The Monetary Dynamics of Hyperinflation.” In Studies in Quantity Theory of Money, edited by Milton Friedman, 25–​117. Chicago: University of Chicago Press. Carvalho, Carlos Eduardo. 2003. “O fracasso do Plano Collor:  Erros de execução ou de concepção?” Economia 4: 283–​331. Dornbusch, Rudiger. 1987. “Lessons from the German Inflation Experience of the 1920s.” In Macroeconomics and Finance, Essays in Honor of Franco Modigliani, edited by Rudiger Dornbusch and Stanley Fischer, 337–​366. Cambridge, MA: MIT Press. Dornbusch, Rudiger, and Holger Wolf. 1990. “Monetary Overhang and Reforms in the 1940s.” NBER Working Paper No. 3456. Faro, Clovis de, ed. 1990. Plano Collor, avaliação e perspectivas. Rio de Janeiro: Livros Técnicos e Científicos Editora. Faro, Clovis de, ed. 1991. Plano Collor II. Rio de Janeiro: Livros Técnicos e Científicos Editora. Folha de São Paulo. 1987. A tragédia do Cruzado, do congelamento à ameaça de hiperinflação, dia após dia, nas páginas da Folha de São Paulo. São Paulo: Folha de São Paulo. Franco, Gustavo H.  B. 1990. “Fiscal Reforms and Stabilization:  Four Hyperinflations Examined.” The Economic Journal 100: 176–​187. Lopes, Francisco. 1986. O choque heterodoxo: Combate à inflação e reforma monetária. Rio de Janeiro: Campus. Pazos, Felipe. 1972. Chronic Inflation in Latin America. New York: Praeger.

Experiences of Inflation and Stabilization, 1960–1990    123 Plano Collor. 1990. Revista de Economia Política 10: 114–​120. Sargent, Thomas. 1982. “The Ends of Four Big Inflations.” In Inflation, edited by R. E. Hall, 41–​ 98. Chicago: University of Chicago Press. Simonsen, Mario Henrique. 1966. A experiência inflacionária no Brasil. Rio de Janeiro: Instituto de Pesquisas e Estudos Sociais, IPES. Simonsen, Mario Henrique. 1970. Inflação:  Gradualismo x tratamento de choque. Rio de Janeiro: APEC. Slutsky, Eugen. 1937. “The Summation of Random Causes as the Source of Cycle Processes.” Econometrica 5, 105–​146. Williamson, John, ed. 1985. Inflation and Indexation. Washington, DC:  Institute for International Economics.

Chapter 7

L eviathan C a p t u re d Neoliberalism as Solution and Problem in Brazil Philippe Faucher

One of the great mistakes is to judge policies and programs by their intentions rather than their results. —​Milton Friedman É preciso fazer o bolo crescer, para depois dividi-​lo. —​Antonio Delfim Netto Ha varias maneiras de definir o neoliberalismo, mas talvez a maneira mais simples é dizer que é a ideologia da luta de classes ao inverso. —​Luiz Carlos Bresser Pereira

7.1. Introduction This may come as a surprise for those who have been following Brazil’s economic development over the years: apparently the country, in a worldwide comparison, was a champion of neoliberalism during the period 1980–​2002. According to a 2009 International Monetary Fund (IMF) study, Brazil has adopted most types of reforms intended to increase domestic competition and open its economy to foreign capital (Ostry, Loungani, and Furceri 2016). With its oversized bureaucracy, impenetrable fiscal code, and extended public sector, Brazil does not invite obvious comparisons with a capitalist champion such

Leviathan Captured   125 as Singapore, or with any of the top 10 countries of the World Bank’s Doing Business ranking (Brazil is currently in 116th position).1 This is the first indication that, whatever neoliberalism stands for, it is not necessarily associated with performance. These observations lead to two lines of thought. The first is that the idea that neoliberalism promotes a more efficient allocation of resources, greater wealth, and faster growth should be rejected as far as Latin America is concerned.2 The second is that neoliberal economic reforms in Brazil did not result in an obvious reduction of the role of the state. Political resistance and economic necessity managed to preserve (or rebuild) the interventionist state. This is the scenario we explore in this chapter. Brazil’s economic development has followed a singular path. Historically a resource economy—​like most of its Latin America neighbors—​its large domestic market allowed for late industrialization behind high protectionist barriers, an extended network of state-​owned enterprises, and regulations meant to channel foreign investment in selected sectors—​mostly consumer goods and in particular the car industry. Following the political turmoil that culminated in the military coup of 1964, this state-​ led development model, labeled “dependent-​development” (Evans 1979), was credited with producing a “miracle” (1968–​1973), when the growth rate averaged 11%. Evidently, however, something went wrong, as subsequently the Brazilian economy struggled for more than 20  years (1975–​1995) along an uncomfortable path of moderate booms and busts (nicknamed the “chicken flight”) marked by large fiscal deficits and record high inflation. After a slow recovery, a short boom (2007–​2010) occurred, sparked by a combination of restored consumer confidence, social transfers, and exceptional export revenues pulled by high commodities prices on international markets. Since then, the performance of the economy has moved from disappointing to catastrophic, with the country entering a major recession from 2014. This analysis begins with a short recap of what neoliberalism stands for and how it was an inspiration, with mixed results, in several countries of the region. We then look at the situation in Brazil at the end of the military regime, when the first debate over the economic role of the state took place under the leadership of some prominent entrepreneurs.3 Policies such as privatization and trade liberalization were introduced shortly after a civilian, Fernando Collor de Melo (1990–​1992), was elected president. But the neoliberal agenda was relegated to second place, as all energy was turned toward the fight against inflation and the debt crisis. The “dragon,” as inflation was nicknamed, was eventually defeated and price stability was restored—​after at least a half-​dozen unsuccessful attempts—​ with the application of the “Real Plan” under President Cardoso (1995–​2002). At this time, a number of “market-​oriented” reforms were introduced into the development model, with the intention that the market would gain in flexibility and efficiency. Such reforms are clearly part of a neoliberal agenda—​in particular, financial markets liberalization.

126   Philippe Faucher Efforts to reduce the role of the state are most likely to take place under pressure within a short window of opportunity, and structural reforms are at all times subject to reversals, as groups will beg for exemptions, or will fight to regain lost advantages or to obtain new privileges. Questions have been raised about the long-​term sustainability of reforms. Taylor (2015) has documented in detail how Brazil’s state capitalism was reborn as a consequence of continuous political competition for control over public regulation and the appropriation of collective resources. The presidential victory of the Workers’ Party (PT) in 2002 did not initiate a break with the macroeconomic orthodoxy that had been dutifully followed by preceding governments. However, besides ending inflation—​a major victory—​little had been achieved in previous administrations to address the income distribution problem. Under President Lula, social programs were improved, transfers increased, and the minimum wage was raised substantially. Millions of Brazilians were pulled out of absolute poverty, and an important expansion of demand resulted. It was also a period during which the interventionist state was rebuilt with the intention of stimulating growth and generating the surplus required for sustainable social transfers. Large (quasi) state-​ owned enterprises such as Petrobrás, Embraer, and Electrobras are major investors, and the public development bank (BNDES, National Bank for Economic and Social Development) benefited from an important increase in the amounts of credit available. Meanwhile protectionism and price controls—​classic instruments of the early import-​ substitution period—​were reintroduced.

7.2.  Neoliberalism in Perspective 7.2.1. What Do We Mean by Neoliberalism? In its economic incarnation, neoliberalism is a rather simple set of ideas centering on the expected benefits of allowing market forces to determine resource allocation, with limited interference from public authorities. This reduced role of the state came as a reaction against state planning and regulation at a time when postwar growth in the industrialized world was facing increasing difficulties. It has involved deregulation, trade and financial liberalization, and privatization, and commands an overall orthodoxy in monetary policy. Neoliberalism opened the way to the second phase of globalization (the first having ended in 1930), characterized by an increase in international trade, accompanied by plant relocations, capital mobility, and higher exchange-​rate volatility. As an inspiration and prescription for economic policy, neoliberalism as theory, in its most recent incarnation, was the work of a select group of academics. The Austrian School, represented by Friedrich von Hayek, and the Chicago School of Milton Friedman are the most commonly cited origins. These Nobel Prize winners (1974 and 1976, respectively) found their inspiration in developing neoliberal ideas to apply to

Leviathan Captured   127 economic theory.4 This relatively limited academic notoriety was shattered when neoliberalism erupted onto the political scene with the first economic team—​labeled the “Chicago boys”—​of General Augusto Pinochet, who headed the Chilean military coup in 1973 and ruled Chile from 1974 to 1990. The association with right-​wing conservative politics was confirmed when British Prime Minister Margaret Thatcher (1979–​1990) and US President Ronald Reagan (1981–​ 1989) actively supported competition, deregulation, and privatization, advocated a smaller role for the state, and engaged in historical confrontations with trade unions. Neoliberalism has been associated with dictatorship, US imperialism, multinational enterprises, and, later, globalization. As expected, in Latin America it was seldom endorsed, and mostly criticized and denounced, by academics and politicians alike. Soon the label entered the political debate. Any indication in favor of a reduction of state intervention, in particular when social transfers were targeted, was labeled neoliberal and denounced by opponents. Thus, in order to avoid the neoliberal label, economic reforms that implied a reduction in the role of the state would be called “market oriented,” and more substantial changes, such as privatizations, were referred to as “structural reforms.” But, as governments from left to right have experienced, macroeconomic orthodoxy, given high levels of international capital mobility, remains irremediably neoliberal.5 A notorious example is indeed the PT’s President Lula, who was branded a neoliberal as he pledged fidelity to financial orthodoxy to enhance his electoral chances (Moraes and Saad-​Filho 2005).

7.2.2. Neoliberalism as Political Proposition Opposing markets and states as competing institutions in resource allocation is an ideological proposition of limited value. Recent studies acknowledge the diversity of productive institutions, presenting a more nuanced view of the “varieties of capitalism” (e.g., Boschi 2011; Hall 2001; Schneider 2013). Ha-​Joon Chang demonstrates that the fundamental components of the neoliberal framework need to be carefully assessed with due historical perspective. For instance, the bans on child labor and on slavery are the results of the historical evolution of human rights; given that any debate on these issues is in many countries today considered politically unacceptable, very few would consider these bans to be manifestations of state intervention upon the labor market. More recently, the right to a clean environment limits access to technologies that are considered harmful. As such, a free market is a political proposition in the sense that “[i]‌t is only because some state regulation [. . .] can be so totally accepted [. . .] that some markets appear to have no ‘intervention’ at all and therefore appear to be free” (Chang 2001, 6). The determination of what constitutes “market failure” is also based on what it is believed a normal market is expected to do. Monopolies, oligopolies, and income inequality are normal outcomes of a performing capitalist economy, and failures from the perspective of the market. These are manifestations of the power of firms as “institutions

128   Philippe Faucher of production.” The emergence of markets is engineered, and the organizations of production are regulated by the state—​such as through the legal determination of property rights—​according to the power structure in place. In short, debate about the role of the state needs to be placed in the appropriate political context; the market cannot be depoliticized. In terms of a more pragmatic set of coherent policies, probably the most fully formed expression of neoliberal orthodoxy—​and one that has generated numerous comments and debates in Latin America—​is the list of 10 propositions outlined by John Williamson in 1989 and known as the “Washington Consensus.”6 The policy shift associated with market reforms involved all actors of society: the bureaucracy, at the time referred to as a “state bourgeoisie” due to the sense that this group behaved as if they owned the public enterprises they were responsible for; the political elite, who used and abused power and influence; and business and economic interest groups, who invested in rent-​seeking along with most other political, economic, and social institutions. Thus the policy “outcomes” typically mandated by reform in a neoliberal direction are the following: (a) Reduce state economic intervention; (b) Increase competition and thus the probability that efficiency will be rewarded and inefficiency punished; (c) Create incentives for economic agents to comply with market forces rather than, for example, seeking political protection from them; (d) Integrate domestic and international markets; (e) Shift responsibility over investment and growth to the private sector; (f) Create pressures for public agencies better to enforce regulations without discrimination (Armijo and Faucher 2002, 2).

7.2.3. Brazil in Perspective Other Latin American countries underwent market reforms in a process that started in the 1970s and was well underway by the end of the century.7 Comparing Brazil with the major economies of the region, namely Argentina, Chile, and Mexico, we observe a strong policy convergence. Using an index for economic liberalization developed by Morley, Machado, and Pettinato (1999), we can see that “although the timing of the introduction of reforms differed significantly among countries, the difference in degree of liberalization had practically disappeared by 1995” (Armijo and Faucher 2002, 6). This view tallies with our own qualitative assessments and additional observations. Essentially, a similar reform agenda was implemented, with some chronological differences—​Brazil being the laggard. This could be interpreted in different ways. Either neoliberal market reforms were the only policy option and Brazil introduced such

Leviathan Captured   129 reform when domestic conditions and/​or economic exigencies made it necessary; or, after considering directions taken in neighboring countries, Brazilian authorities decided to follow the same path. In either case, the advent of reforms was the result of both foreign influence and domestic choice. However, one major divergence is that Brazil, unlike other countries of the region, did not engage significantly before 2000 in any substantial reform of social security (Armijo and Faucher 2002, 10).

7.2.4. Political Support for Market Reforms Economic neoliberalism refers to a large set of policies implemented with the aims of introducing or restoring competition in markets, reducing the role of the state, and fostering fiscal responsibility, all in the hope of bringing stability and promoting growth. Income redistribution is not a concern, but should happen, so it is claimed, as the result of democratic political battles over the use of the available surplus. From a political perspective, these are issues over which revolutions are fought. Neoliberalism, at least in Brazil, amounts in theory to a full frontal attack on the vested interests of major groups in society. Domestic firms should resist trade liberalization, unions oppose pressure for more flexibility in the labor market, public entrepreneurs resist privatization, and bureaucrats fight to preserve their control over agencies and their lifelong benefits. It is widely believed that for reforms to be introduced, there must be exceptional circumstances. For a shift in policy to occur, representing a potential challenge to vested interests, a window of opportunity is required.8 In Brazil’s case, following democratization in 1985, and as major instability, market disorganization, low growth, and record high inflation evaded initial stabilization attempts, a political consensus developed as to the need to act resolutely. This particular “decisive moment” created a specific configuration of conditions. The gains that could be extracted by some sectors—​such as private banks, who thrived from unrestricted inflation—​were declining. Meanwhile, the collective costs imposed and the subsidies made attainable by preserving the status quo were no match, as perceived by a majority of actors, for the benefits associated with stability and growth.9 An “economic crisis” is possibly a necessary condition for structural reforms, but it is not a sufficient condition, as political legitimacy is also required (Armijo and Faucher 2002, 11). Recent work has emphasized the powers given to the Brazilian president by the Constitution of 1988 (Alston et al 2006). This is plausible, as institutions can provide a leader with insulation, allowing him or her the authority to initiate structural reforms even when influential interest groups and a majority may oppose them. State autonomy,10 combined with the technocratic insulation that characterized the early development model, has been credited with the robust average growth of 7% from early 1950 to late 1970. But again, this is not a sufficient condition, as our comparison with other countries of the region shows a large variation in the measure of the degree of “political

130   Philippe Faucher insulation” of the executive, even though the reform agenda has, as mentioned, been convergent.11 Since political leadership is required, market reforms are needed to garner significant political support. All such reform is the result of some amount of political bargaining given that a support coalition is required—​in particular, although not exclusively, under democratic rules. For example, trade liberalization allowed large firms, foreign and national, and exporters to negotiate for partial exemptions and delays in tariff reductions along with tax rebates (Armijo and Faucher 2002, 25). Additionally, the rules that applied at different stages of the privatization process, the type of currencies allowed, and the limits imposed on foreign bidders were designed to generate political support among potential domestic buyers. Finally, popular support for market reforms should not be disregarded. Of course, numerous considerations, besides a neoliberal agenda, will come into the decision to vote for a specific candidate. It remains the case that at least four presidential elections in Brazil were won by candidates openly promising to introduce and/​or pursue a neoliberal agenda (Collor in 1989, Cardoso in 1994, Cardoso in 1998, and Lula in 2002). It could be argued that voters were manipulated by neo-​populist leaders and that patronage prevailed; but the seeming paradox also presented itself in Chile (Aylwin in 1989, Frei in 1993, and Lagos in 1999), in Mexico (Zedillo in 1994 and Fox in 2000), and in Argentina (Menem in 1995 and de la Rua in 1999), to mention only these four countries. The urgency of ending inflation and the promise of macroeconomic stability are the most plausible explanations for this apparent electoral support for leaders associated with market-​oriented reforms.

7.3.  The Collapse of the Early Developmental State: Neoliberalism as Solution Broadly speaking, neoliberalism-​inspired reforms were enacted in Brazil between 1985—​the end of authoritarian rule—​and 2003, the start of Lula’s first mandate. (This is something of a simplification; the pace, scope, and depth varied from one reform to the next.) The earlier Brazilian developmental state had emerged as a compromise between factions of the traditional elite, the emerging urban industrial entrepreneurs, and the working class. This arrangement was by nature unstable. It required a strong state, using corporatist mechanisms to promote development and coordinate interventions. Created during Vargas’s Estado Nôvo (1937–​1945), the model was threatened by political competition among elites and opportunistic popular mobilization. The subsequent short democratic interval (1946–​1964) came to an abrupt end with the military coup of

Leviathan Captured   131 1964. With order restored through repression, the developmental strategy was revived and the public sector expanded, boosting economic growth, with public agencies enjoying considerable autonomy (Faucher 1979, 1981).12 Weyland (1998) accurately characterizes the process as a manifestation of decline, through which public agencies multiplied and were progressively captured by interest groups; autonomy and coordination, as the state became larger, got lost in the process. The developmental coalition collapsed as the state lacked the tax revenues for its investment projects. Entrepreneurs, supported by politicians and gorging on protectionism, tax incentives, and subsidies, successfully lobbied for tax exemptions, and repeatedly resisted any fiscal adjustment. The decline in tax revenues, accounting for a large share drained by privileged treatment for business (and tax evasion), was momentarily compensated by foreign loans. But borrowing came to an abrupt halt as the debt service dangerously expanded with the 1980s crisis, and inflation exploded. The two oil shocks, followed by the debt crisis of the early 1980s, seriously weakened the supporting coalition of the military government. Out of frustration with the weak performance of the economy, influential business leaders launched a public campaign against what they considered excessive intervention and the discretionary power of the technocracy (Diniz 2010; Kingstone 1999). Facing opposition from all sectors of society, it took nearly 10 years for the military to negotiate a viable and “honorable” exit (essentially making General Golbery do Couto e Silva responsible for the liberalization policy of distensão while chief of staff of President Geisel and President Figueiredo). From the perspective of restoring stability and confidence, extensive market reforms, including trade liberalization, privatization, and deregulation, were in order. Although the business community adhered to neoliberalism, firms had different views, and remained divided on the range, pace, and depth of reforms that were required. This was the result in part of the divide-​and-​rule structure of rent-​seeking that had ruled state-​business relations. As Weyland (1998, 55) argues, “[b]‌usiness firms were subject to numerous controls and eligible for many benefits, from subsidized loans to protection from competition. It was in their interest to advance individual requests for special favors, rather than banding together and pressing broad-​based demands on the state, which could provoke retaliation.” The various stances among Brazilian entrepreneurs, and the bargaining games these led to, are seen in two of the main reforms involved in the dismantling of the import substitution industrialization model: trade liberalization and privatization.

7.3.1. Trade Liberalization Trade liberalization began in 1988. The number of goods subject to the “similarity test” was reduced.13 Simultaneously, and facing uncharacteristically little opposition, the average nominal tariff was reduced from 55% in 1987 to 27% by 1990, and to around 10%

132   Philippe Faucher by 1994. The average effective tariff declined from 68% to around 12% over the same period, and a large number of non-​tariff barriers were eliminated (Amann and Baer 2008; Ter-​Minassian 2012, 27). As a result, the ratio of trade to gross domestic product (GDP) increased, a positive sign of integration. Liberalization slowed almost to a halt around 2000. Foreign exchange crises, including with Mercosul partners, faltering competitiveness, and the subsequent appreciation of the real were to blame. It should be noted that Brazil, in its assertive new role on the international scene, adopted with other emerging countries an attitude of challenge toward the multilateral trade agenda, which contributed to the failure of the Doha round. It was the most vocal opponent to the failed US-​sponsored “Free Trade Area of the Americas.” Although there have been calls for further liberalization, mostly from economists, is it not a proposition often publicly supported by economic and political leaders. The import-​substitution argument—​inspired by the infant industry theory—​remains the main justification (i.e., Brazil has a sufficiently large domestic market to sustain competitive production lines in goods and services, operated by national and international firms). With a trade to GDP ratio of 21% (2015), Brazilians compare themselves with the United States (21%) and not South Korea (70%). Barriers are still considered to be incentives for investment, rather than what they are: obstacles to competitiveness.

7.3.2. Privatizations During the military government (1964–​1985)—​highly nationalistic but also committed to the promotion of a strong domestic capitalist class—​sectors that were not considered “strategic” (e.g., energy) were not intended to remain under public ownership. It was clear from the mid-​1970s that state-​owned enterprises (SOEs) were to be transferred to “selected private groups” with the active participation of BNDE.14 It took more than 10 years to overcome bureaucratic resistance to the privatization program being given a place on the agenda. Early privatization, between 1980 and 1990, was limited to 38 enterprises, mostly small firms that had been taken over by the federal government because they were experiencing difficulties. Receipts amounted to US$723 million. The  BNDES was later officially selected by President Collor (1990–​1992) as the “operational agent” of the National Privatization Program. It remained in that role under President Cardoso (1995–​2002). During the Collor government, 15 SOEs were privatized, for total proceeds of US$3.5 billion. This was an important step, as the scope of privatization was expanded to include firms formerly considered “strategic,” such as Usiminas (in 1991), the large steel mills founded in 1956, which were a symbol of Brazil’s industrialization (Schneider 1992). As stabilization became the main priority, considerations for cash—​initially mostly in the form of debt certificates (84%), and then, from 1995, in currency—​were given more emphasis. The Franco government privatized 18 enterprises for revenues estimated

Leviathan Captured   133 around US$5 billion. Embraer, the aircraft manufacturer, was sold during that period. In this early phase, proceeds from foreign investors represented only 5% of total sale revenues (BNDES 2002). Changes in the privatization program were also introduced during Cardoso’s presidency. Foreign investors were invited to bid, but more important, all sectors could potentially be privatized at least in part; transport, highways, electric power generation and distribution, petrochemicals, sanitation, rail transport, ports, telecommunications. This required, in some instances, a constitutional amendment. Unions’ support was bought by introducing a degree of “popular capitalism” with workers offered share ownership, following the precedent set in the 1980s by the Thatcher government in the United Kingdom. In the privatization of Usiminas, employees bought shares for a total of US$167 million at the auction, plus US$428 million from the Time and Service Guarantee Fund (FGTS) (BNDES 2002). BNDES not only defined a minimum auction price, but also participated in the identification of potential acquirers, providing loans and equity using its own funds and government-​controlled pension funds (Musacchio and Lazzarini 2014).The mining consortium Vale do Rio Doce (Vale), privatized in 1997, offers an example of such a creative financial package: “[. . .] Vale is directly controlled by Valepar, which is a consortium of owners including BNDES, Japanese group Mitsui, Brazilian banking group Bradesco, and a host of pension funds of SOEs such as Previ (from Banco do Brasil) and Petros (from Petrobras)” (Musacchio and Lazzarini 2014). In all, BNDES oversaw sales of US$105.3 billion (BNDES 2002, 2). As an expected complement to privatizations, regulatory agencies were created to supervise and enforce post-​privatization rules and to introduce competition in natural monopolies such as electrical distribution, fixed telecommunication networks, and transportation (Trubek et al. 2013). With Cardoso’s election in 1995, neoliberalism was again on the agenda. Market reforms gained new momentum with all energies mobilized and policies made to converge in a renewed attempt to end hyperinflation and rein in what had become a chaos in public finance.

7.4.  Neoliberal Pragmatism and the Consolidation of Market-​Oriented Reforms It is difficult to identify a consistent neoliberal agenda given that, as recent works demonstrate, reforms were introduced in isolation with limited institutional complementarities (Angel-​Tapias 2016). The policy coordination of market reforms that allowed for the success of Cardoso’s Real Plan stands out as an exceptional moment. “By reducing interventionism, Cardoso also seeks to mitigate some of the root causes of

134   Philippe Faucher state weakness. Slimming down the state apparatus through administrative reform and privatization can facilitate central coordination and control and cut many of the particularistic connections between public agencies and business people” (Weyland 1998, 66). At this stage—​following several failed attempts—​fighting inflation required more than a balanced budget. It involved fiscal adjustment, de-​indexation, and a structural reform of the financial sector. Restoring confidence in the new currency required that the Central Bank be given enough resources and authority to manage the exchange rate and to allow capital to flow freely. An orthodox macro-​economic policy, combined with a commitment to a primary surplus and strict limitations on the borrowing capacity of public banks and state governments, were all essential elements of the Real Plan. It was all very delicate, and never more so than when, in January 1999, the exchange rate anchor was abandoned, resulting in a sharp devaluation against the US dollar (Amann and Baer 2003). Neoliberal reforms were in most cases partial in their application and limited in time. With inflation under control and financial stability restored (a primary surplus was achieved only in 1999), combined with a strong export performance, mostly in agro-​ business and minerals, and an inflow of foreign investment (in part tied to the privatization process) (Amann and Baer 2002, 949), the need to push for a neoliberal agenda vanished. But on monetary matters, capital mobility, a primary surplus, a floating exchange, and high lending rates became the exclusive policy prescription (Coes 2009, 184). To prevent the outflow of capital, the real interest rate was raised on several occasions, increasing borrowing costs and thus weakening the fiscal position of the government. The Law of Fiscal Responsibility was introduced.15 Lula, through both mandates and to ensure price stability, maintained a larger than required primary surplus (it reached 4.7% of GDP in 2004) combined with very high interest rates (as high as 19% in 2005). These, combined with large trade surpluses, pushed the value of the real, while reserves rose. Of course, these are costly policies with a negative impact on growth, and also contributed to further income concentration. Both Lula’s PT and his political base more broadly were disappointed by and frequently denounced this continuous endorsement of orthodox neoliberal macroeconomic policy. In particular, the combination of high interest rates with an overvalued exchange reduced private investment, created pressure on industry, and generated unemployment (Bresser-​Pereira 2014; Morais and Saad-​Filho 2005).

7.5.  Leviathan Captured: Neoliberalism as Problem 7.5.1. What Was Achieved? It is remarkable that none of the expected results of neoliberalism, as listed earlier, was achieved. In order:

Leviathan Captured   135 (a) Public spending decreased between 1993 and 2000 and has been increasing steadily since. (b) Concentration increased, because (1)  privatization resulted in the creation of private monopolies, poorly regulated (Amann and Baer 2008); (2) by lowering imports of manufactured goods, exchange valuation contributed to deindustrialization, particularly in the capital intensive and high value-​added sectors. (c) Selective protection using subsidized credit and protective regulation were used to shelter noncompetitive industries from foreign competition. (d) Integration in the global market was pulled by international price increases and was limited to primary goods and commodities. (e) As growth showed signs of weakness, BNDES was provided with ample resources to compensate for the lack of private investment. (f) Finally, regulation was not adequately enforced, and politically inspired price controls were applied. From Lula’s second term, state-​led development returned. There were three reasons for this. First, an economic surplus was indispensable for the left-​inclined PT to apply its redistribution policy. Second, encouraging private investment required that entrepreneurs participated in consultation through a newly created corporatist structure. Third, rapid growth could only be sustained if infrastructures were improved, while slow growth required added public incentives. In all, because the owners of capital did not share the political agenda of the left, not only did the government have to increase its subsidies to private investment, it also had to share its decision-​making authority over investment.

7.5.2. Leviathan Captured The “happy globalization” period of the early 2000s (Valladão 2013, 2), resulting from the combined effect of the international commodity boom, important social transfers, and the steady flow of foreign investment, came to a halt with the global crisis of 2008. In reaction, the government took actions meant to stimulate trade, protect domestic industries, promote investment with subsidized credit, and control energy prices, and it launched several ambitious infrastructure programs. In many respects these interventions are reincarnations of the development model of the early phase of industrialization. However, both the economic environment and the distribution of political power among actors are radically different this time. The following passage summarizes the path that was followed in this most recent period. Economic theory provides numerous justifications for state intervention. The more benign “market failure” idea suggests that public intervention will compensate for insufficient investment in some critical sectors. A more interventionist perspective is provided by Schumpeterian theory; state intervention is required to provide for research and development in order to support innovation and to socialize part of the risk in the hope that it will result in the creation of comparative advantage for

136   Philippe Faucher local firms. Finally, inspired by Keynesian theory, is the idea that the state is required to adjust its investments and stimulate consumption, providing countercyclical support in periods of economic slowdown.16 These justifications can be combined at will. The capacity to coordinate a diversity of agencies and remain autonomous from business is associated with a strong state. In such a situation, public intervention claims not to be influenced by politics, using objective technocratic rules to provide preferential treatment for firms and to determine investment decisions.17 Political science offers a distinct point of view. State autonomy and technocratic neutrality are viewed with suspicion, as conceivable only in exceptionally stable democratic institutional settings with multiple checks and balances and a general requirement for transparency. The reality is believed more often to amount to rent-​seeking, clientelism, patronage, patrimonialism, or plain corruption, resulting in situations where state agencies are captured by interest groups seeking special favors. Any suggestion of political discrimination using collective resources constitutes a political liability in most countries. Brazil’s early developmental state provided inspiration in numerous studies (e.g., Evans 1995; Faucher 1993; Woo-​ Cummings 1999)  showing that public investment, combined with technocratic expertise and strong market understanding, could result in the implantation of robust industrial activity. The collapse of the developmental state is associated with lost state capacity, proliferation of agencies, absence of coordination, fragmentation, and capture. In such a setting, state capitalism is no more than the selective appropriation of collective resources to stimulate growth and private accumulation. Instead of promoting competition, public intervention, by providing support for a selected number of firms, contributes to concentration. Increased public intervention when institutional capacity is lacking opens the way to clientelistic discrimination, the antithesis of neoliberal goals. An illustration of such a contra-​neoliberal phenomenon is the creation, during Lula’s first term, of a neo-​corporatist structure, the Council for Economic and Social Development (Doctor 2007).

7.5.3. The Return of Industrial Policy Budget rules and fiscal regulations as determined by the Constitution of 1988 operate as strict constraints. To be maintained and increased, social transfers require that a surplus be produced from growth, or through an increase in taxes rates.18 Because tax revenues are a function of economic activity, the government has an interest, to preserve its political margin, in stimulating growth through various programs and subsidized credit. We consider briefly the most recent (2016) policy shifts in trade, privatization and regulation, investment through BNDES, and price controls in order to illustrate the fact that neoliberalism was sacrificed to the benefit of growth promotion (Musacchio and Lazzarini 2016).

Leviathan Captured   137

7.5.3.1. Trade Despite calls for liberalization, trade policy became more protectionist as foreign competition, stimulated by an overvalued exchange rate, increased. Exports have been systematically financed using BNDES-​subsidized credit (rising from 4% of total BNDES loan disbursements in 1996 to 11.5% in 2009). The practice has been exposed by US Eximbank as not conforming to World Trade Organization (WTO) regulations (Armijo, 2017). In comparison with neighboring countries, the average tariff remains high and an index of nontariff barriers reaches a significantly higher level (20%) than the country’s peers (Taylor 2015). After an initial liberalization drive, overall trade openness has made little progress in the past 25 years. Using its hegemonic position in Mercosur, Brazil has adopted a position in international trade negotiations intended to “control the pace and process of trade liberalization along ‘neo-​mercantilist’ lines” (Taylor 2015, 8). An Organisation for Economic Co-​operation and Development (OECD) comparative study of the G20 countries shows that Brazil has implemented a number of trade restrictions since 2008. It is the fourth most protectionist member of the G20. The most important measures implemented concern government procurement and what foreign partners consider “discriminatory taxation” (Barone and Bendini 2015, 19).

7.5.3.2. Privatization and Regulation Despite significant privatization, the Brazilian government maintains significant interests as majority or minority shareholder in a number of leading firms. 19 As mentioned earlier, there were some “symbolic” privatizations. The government holds a minority position but retains effective control through “golden shares,” or using some state-​controlled financial instruments such as BNDES (or the subsidiary BNDESPar), SOEs’ pension funds—​such as Previ and Fincef—​or reverting to cross-​shareholdings by other SOEs (Taylor 2015, 15).20 Following the privatization of SOEs, agencies were created to regulate monopolies in essential services (communications, transport), basic inputs (oil, gas, electricity, water) and infrastructure (roads, trains). These agencies were meant to protect firms from the influence of political interests with the intention of providing stimulative tariffs and long-​term guarantees to reassure investors, while monitoring firms’ activities to avoid monopoly pricing (Amann and Baer 2005). Initial independence, provided by using fixed terms, congressional review, and financial autonomy, was compromised through budgetary inducements and partisan appointments. Bureaucratic competition between agencies and dedicated ministries also challenged the intended regulatory autonomy. As a result, the government retained much of its influence, and has the capacity to manipulate prices at will in the newly privatized sectors (Taylor 2015, 18–​20). Tariff adjustments have proven to be a source of tensions, as in the case of Electropaulo, the São Paulo electricity distributor acquired by the US group AES(Applied Energy Services) in 1998 (Amann and Baer 2005). The downward pressure on tariffs led some concessionaires to consider abandoning their concession, which would have resulted in renationalization of the concerned utilities; clearly the government has not been able to

138   Philippe Faucher subject the functioning of the economy to market forces (Amann and Baer 2005). On the other hand, measures to give back to the state control of the regulated sectors were discussed by the cabinet, but ultimately were abandoned (Mueller 2009, 159–​160).

7.5.3.3. Investment BNDES, created in 1952, is the most important institution of Brazilian state-​led development policy. Its mandate and resources have constantly grown. Through allocation of low-​cost credit to firms, BNDES has nurtured the development of all major sectors of industry and, more recently, has contributed to their international expansion. By all accounts, BNDES is a highly performing public bank, as it is no longer used as a tool to rescue failed enterprises. A first concern arises, as with all types of public intervention, when the institution, which has access to practically unlimited amounts of public funds, lends at subsidized rates to hand-​picked national champions. The concern reaches another level when evidence indicates that politically connected firms have advantage in obtaining loans (Musacchio and Lazzarini 2012, 26). The bank accounted in 2011 for 70% of long-​term bank lending in the country and supplied 30% of investment in industry and infrastructure after retained earnings (Colby 2012, 6; Taylor 2015). BNDES is relying on forced savings coming from FAT (Fundo de Amparo ao Trabalhador [Worker Support Fund]) paying a preferential interest Taxa de Juros de Longo Prazo (TJLP [Long-​term Interest Rate]), below the Central Bank benchmark Sistema Especial de Liquidação e Custodia (SELIC). The implicit subsidy has been evaluated at 7.6% for each dollar loaned (Musacchio and Lazzarini 2014). Both the importance and the low cost of credit provided raise issues that are usefully synthesized in Armijo’s “trilemma,” in a similar fashion to the Mundell-​Fleming “impossible trinity.” In her words, The trilemma encapsulates the larger conflict inherent in state-​led capitalist development: that of balancing autonomous, technocratic decision-​making (“Expertise”), and the immediate demands of ordinary citizens for a better-​life, as expressed through their political representatives (“Democracy”), with the delicate challenge of using state monies and regulatory authority to stimulate voluntary, decentralized, competitive, and efficient private investment (“Markets”). (Armijo 2017)

Of course, “market” represents the neoliberal option. But political science informs us of a fourth possibility: that of businesses privatizing “expertise” for their own benefit. Serious evidence points in this direction. An example is provided by BNDES financing the international expansion of selected “national champions.” Friboi, a giant meatpacker, was provided with credit for mergers and acquisitions of competitors; the merger of wood producer Aracruz with Votorantin was arranged by the bank, as were the consulting and construction activities of the engineering firm Odebrecht, in Angola (Armijo 2017. Concentration of loans is high, with

Leviathan Captured   139 six firms, all private, cumulating from 2008 to 2012 more than 50% of the total loan portfolio. That “BNDES acts as a full-​service merchant bank to Brazil’s largest transnational firms, financing almost any activity that forms part of one of these firms’ corporate strategic vision” (Armijo 2017) raises little opposition, as it conveys to the underlying nationalism that neoliberalism is the strongest antidote and best ally of state capitalism. Industrial policy, as expected, has also used tax exemptions both at state and federal levels of government, to discriminatively support productive activities, particularly in manufacturing. Finally, domestic content requirements are widespread in sectors where government procurement is an important segment of demand, such as pharmaceuticals and medical equipment, energy (wind, solar, gas), and information technology.

7.5.3.4. Price Controls SOE-​controlled “strategic inputs” such as petroleum, electricity, and, in a not-​so-​distant past, steel and transport, were sold at subsidized prices to private firms for them to become competitive. Managing prices has generally been the rule. As soon as producing alcohol from sugarcane as a substitute for imported gasoline became a priority, prices were administered to remain below that of gas as an incentive for consumers and to save on foreign currencies. Most obviously, price controls were also enforced as a stopgap measure to combat inflation. These practices were meant to end once control of inflation was achieved. More recently, there are numerous instances of government interference through price control. The price of gasoline was kept low in 2012 when import prices were rising. This manipulation was in part responsible for significant losses to Petrobrás in 2013. Electricity is another instance often referred to. In both cases the regulating agencies (ANP and ANEEL, respectively) could not resist the government’s pressure (Musacchio and Lazzarini 2014).

7.6.  Conclusion: The Game Remains “Fazer o bolo crescer” Neoliberalism means a general orientation in government economic policy toward trade liberalization, favorable views on globalization and capital market opening, and a concern to reduce the size of the state through deregulation and self-​imposed limits on fiscal deficits and debt. Currently it is difficult to find a government that doesn’t share some of these views and try to act upon them—​not that they succeed. Neoliberalism is of limited relevance for developing economies. Brazil provides a useful illustration of this. State intervention is a function of market failures manifested as suboptimal allocation of resources, which in turn is often politically motivated. Market flexibility is a function of competition. The experience in Brazil’s economy,

140   Philippe Faucher dominated by collusion and capture, points resolutely in the opposite direction. These are the factors that have led to the development of state intervention and ultimately much privatization of collective resources. What has been achieved in Brazil? Although never acknowledged publicly—​with the exception of economist and politician Roberto Campos21—​neoliberalism as a pro-​market inclination was an inspiration in Brazil more than 30 years ago when the country undertook, of necessity, a shift in its development model away from import-​ substitution industrialization and state capitalism. Because of raging inflation and severely imbalanced markets, liberalization became the only game that was supported by international finance, multilateral organizations, and the local banking sector. At the time there was no viable policy alternative to market-​oriented reforms, that is, promoting competition and factor mobility. What matter are the causes and the results of such reforms. As the old developmental Brazilian state collapsed under its own weight, it was necessary to introduce reforms to enhance state capacity. Once stabilization was achieved, market-​oriented reforms were introduced in conjunction with an increase in fiscal extraction. The strengthening of state capacity was a prerequisite for a renewed developmental coalition. A number of observations can be submitted after a brief survey of the economic policies of the past 30 years. There has never been a clear neoliberal agenda of “rolling back the state.” Early privatization was undertaken, with little political support, out of frustration by BNDES technocrats with recurrent losses of irrecoverable public firms in their portfolio that undermined the overall performance of the bank. Privatization was conducted behind technocratic insulation. Trade liberalization was swiftly introduced; tariffs were lowered progressively, as other protective barriers were dismantled. But overall Brazil remains a protected market. Cardoso found support within his governing coalition to adopt constitutional amendments that allowed the extension of the privatization agenda into strategic sectors. To prevent a return of inflation, he implemented a macroeconomic policy locked into the dominant financial orthodoxy, including high primary surpluses and record high interest rates. Lula managed, benefiting from a favorable political and economic conjuncture, to reconcile for a while growth and redistribution. The experiment in top-​down social democracy was short-​lived. Brazil confirms that the liberalization of capital flows results in more vulnerability. The 2008–​2009 financial crisis mobilized important resources, and the government pledged substantial investments in infrastructure to push the economy out of an anticipated recession. The pull provided by favorable commodity prices stopped as prices fell. The growth in primary exports pushed the value of the currency to a point where domestic manufactured products were no longer competitive. Under the circumstances, BNDES had little choice but to cajole local businesses with subsidized credit in order for them to invest. This is a repeat of the capture of state capacity. But equally important has been the ending of the governing political coalition.

Leviathan Captured   141 The return of state capitalism has been documented, looking at several policies. Most market reforms have been “rolled back” and the developmental state was reborn. Most, if not all, countries do intervene to stimulate economic activities, promote national champions, and subsidize research and development. There are plenty of excellent reasons to do so. Whatever the government, state intervention has been prevalent in Brazil since the start of its industrialization process, and it will remain so. State capitalism, with all its limitations, combines:  a strong expression of nationalism (of the “new-​world” type22), the urgency of the development of a large and populated country, and necessary limits to the pressures imposed by globalization. All of these factors are an unequivocal antidote for any manifestation of market fundamentalism and act as source of inspiration for a “new developmentalism” (Bresser Pereira 2014, 2016). Lula’s democratic popularity and international praise for a redistributive agenda raised hopes that a new pro-​poor development model was emerging (Amann and Barrientos 2014). The severe recession of recent years, the consecutive corruption scandals, and the traumatic political crisis that ended in the impeachment of President Dilma Rousseff (2011–​2016) are demonstrations that the “social contract” supporting an agenda of redistribution and equity is weak and can be reversed. Emerging economies are basically unfair societies, with a minority oligarchy enjoying the concentration of wealth and power. Brazil is no exception. To stimulate growth and engineer political support, industrial policy returned in full strength during the Lula years. An elaborate business-​state conciliation structure was created, reminiscent of corporatist arrangements. Various agendas were simultaneously pursued: to help national champions in becoming world leaders, to contribute to the expansion of the agro-​ business industry, to promote trade, to use procurement to stimulate the creation of high-​tech suppliers in offshore petroleum production and transport. Overall, the main goal has been to stimulate local production and employment using all means available, while preventing the return of inflation. It all proved unrealistic and undermined the government’s legitimacy. Why would a democratically elected left-​wing government use forced savings to subsidize private investment? The answer is because growth became a political necessity. In the absence of a strong popular mandate, without the means to challenge the economic elites, social transfers without sacrifice are only possible when a surplus is available. The fazer o bolo crescer (make the cake grow) non-​theory of development is not so much an economic necessity as a political reality that the left, no matter how moderate it is, has not been able to escape. In the future, adjustments reminiscent of a neoliberal agenda will most likely be introduced episodically and accompanied, as is to be expected, with market rhetoric and the usual oppositions. Brazil’s pension system is in dire need of a major reform. Subsidized prices on energy are not sustainable and BNDES’s unrivaled autonomy is regularly challenged. It is clear that reforms are badly needed. But this should not be interpreted as a manifestation of neoliberalism. These reforms, and others likely to occur, belong to the necessary changes in which all democratically elected governments

142   Philippe Faucher are engaged, inspired by a concern for more effective spending of public funds and a performing economy.

Notes World Bank, Doing Business, http://​www.doingbusiness.org/​, consulted on June 10, 2016. This view is empirically supported for Latin America (Serra and Stiglitz, 2008, 4). This episode has been thoroughly studied by Peter Kingstone (1999). The publications that most contributed to the dissemination of these authors ideas are Von Hayek’s The Road to Serfdom (1944) and Friedman’s Capitalism and Freedom (1962). 5. There is no escaping the dictatorship of the financial market, as argued by radical economist Chossudovsky (2003). There are few attempts. As a symbolic—​if costly—​gesture, Brazil made in 2006 an early payment of US$16 billion to the IMF in an assertion of its restored financial health. Brazil’s announcement was shortly followed by Argentina for an amount just short of US$10 billion. 6. “Williamson’s Washington Consensus centred on ten reforms: (i) fiscal discipline in order to eliminate public deficits; (ii) a change in the priorities of public spending: withdrawal of subsidies and increased spending in health and education; (iii) tax reform: broadening tax bases and reducing tax rates; (iv) positive real interest rates, determined by the market; (v) exchange rates determined by the market, which must guarantee its competitiveness; (vi) liberalization of trade and opening of the economy (Williamson did not attach any priority to the liberalization of capital flows); (vii) no restrictions on foreign direct investment; (viii) privatization of public enterprises; (ix) deregulation of economic activity; (x) a solid guarantee of property rights” (Serra, Spiegel, and Stiglitz, 2008, 4). In their opening chapter, Serra, Spiegel, and Stiglitz (2008, 3) remind us that it is unfair to associate Williamson with “market fundamentalism.” 7. This and the following section use material previously published in Armijo and Faucher (2002). 8. “Windows of opportunity result when:  1) the rental streams fall short of expectations to members of the configuration of power; 2) a new member or organization enters the configuration of power because of an unanticipated economic and political shock; and 3) the beliefs of some members in the configuration of power change either because of the change in the economic and political outcomes or because of an exogenous event” (Alston et al. 2013, 14). 9. In another words, hurdles to change—​understood as arising when “new institutions have redistributive consequences that cannot be renegotiated due to transaction costs and commitment problems, such that those in power prefer to block the change and retain a larger expected share of a smaller pie” (Alston et al. 2016)—​can, in special circumstances, temporarily be overcome. 10. State autonomy refers to the ability of public institutions to take initiative and develop policies and implementation strategies (Przeworski 1990). 11. For a detailed account of the indicators used for this estimation, consult Armijo and Faucher (2002, 13–​19). 12. The state provided major impetus for the rapid growth of the economy from 1940 to 1980 (7.1% per year) and the country’s development from an agrarian to an industrial economy (Weyland 1998, 51).

1. 2. 3. 4.

Leviathan Captured   143 13. The “similarity test” meant that import licenses were granted under the condition that no other firm was producing similar goods domestically. This highly protectionist requirement was administered by CACEX (Carteira de Comercio Exterior), created in 1953 and abolished in 1990 by the Collor government. 14. In a 1976 note, Marcus Vianna, than head of BNDE (BNDE became BNDES in 1982), expressed the view that state capitalism was not the desire of the government (Vianna, quoted in Musacchio and Lazzarini 2014). 15. The Law of Fiscal Responsibility imposed spending limits for all levels of government, including state and municipal. 16. This presentation is transposed from Colby (2012, 10). 17. Neutrality is the norm with one exception. Discrimination is tolerated in the case of very large firms and financial institutions on the understanding that they are “too big to fail.” 18. This avenue is unlikely as “total tax revenue has been rising steadily, from 17% of GDP at the outset of the military regime in 1964 to nearly 36% in 2013, in what may have been the world’s greatest peacetime expansion in taxation” (Taylor 2015, 9). 19. “The total number of companies wholly owned by the Brazilian federal government has fallen, from 251 in 1979 to a total of 141 of much smaller size in 2014” (Taylor 2015, 15). 20. As mentioned earlier in the case of Vale, SOEs’ pension funds have been required to participate in the financial packages organized under BNDES leadership: “Because the government has voice in the management of state-​controlled SOEs and these companies have a voice in the management of their pension funds, governments in Brazil (from distinct political factions) have been able to strategically use pension funds in their favour” (Musacchio and Lazzarini 2016, 125). 21. Roberto Campos was nicknamed “Bobby Fields” (a literal translation of his name) in reference to his assumed neoliberal and pro-​US views. 22. In Brazil the ideas of nation and identity were elaborated out of a determination to gain independence from the metropolis and create a new and better society.

References Alston, Lee J., Marcus André Melo, Bernardo Mueller, and Carlos Pereira. 2006. “Political Institutions, Policymaking Processes and Policy Outcomes in Brazil.” Washington, DC: Inter-​American Development Bank. Latin American Research Network Working Paper #R-​509. Alston, Lee J., Marcus André Melo, Bernardo Mueller, and Carlos Pereira. 2013. “Beliefs, Leadership and Economic Transitions; Brazil 1964–​2012.” Mimeo. Available at http://​extranet.sioe.org/​uploads/​isnie2013/​alston_​melo_​mueller_​pereira.pdf, accessed July 2016. Alston, Lee J., Marcus André Melo, Bernardo Mueller, and Carlos Pereira. 2016. Brazil in Transition:  Beliefs, Leadership, and Institutional Change. Princeton, NJ:  Princeton University Press. Amann, Edmund, and Werner Baer. 2002. “Neoliberalism and Its Consequences in Brazil.” Journal of Latin American Studies 34 (4): 945–​959. Amann, Edmund, and Werner Baer. 2003. “Anchor Away: The Cost and Benefits of Brazil’s Devaluation.” World Development 31 (6):1033–​1046.

144   Philippe Faucher Amann, Edmund, and Werner Baer. 2005. “From the Developmental to the Regulatory State:  The Transformation of the Government’s Impact on the Brazilian Economy.” The Quarterly Review of Economics and Finance 45: 412–​431. Amann, Edmund, and Werner Baer. 2008. “Neo-​liberalism and Market Concentration in Brazil.” Quarterly Review of Economics and Finance 48: 252–​262. Amann, Edmund, and Armando Barrientos. 2014. “Is There a Brazilian Model of Development?” United Nations University World Institute for Development Economics Research. WIDER Working Paper 2014/​134. Angel Tapias, Alejandro. 2016. “Never Again. The Fear Factor in Policy Complementarity: Brazil, Chile and Mexico.” Doctoral thesis, Départment de science politique, Université de Montréal. Armijo, Leslie Elliott, and Philippe Faucher. 2002. “We Have a Consensus: Explaining Political Support for Market Reforms in Latin America.” Latin American Politics and Society 44 (2): 1–​40. Armijo, Leslie Elliott. 2017. “The Public Bank Trilemma:  The BNDES and Brazil’s New Developmentalism.” In Democratic Brazil Divided, edited by Peter Kingstone and Timothy Power, 230-​–​248. Pittsburgh: Pittsburgh University Press. Banco Nacional de Desenvolvimento Econômico e Social (BNDES). 2002. “Privatization in Brazil.” Powerpoint presentation, September. Available at www.bndes.gov.br, accessed July 2016. Barone, Barbara, and Roberto Bendini. 2015. Protectionism in the G20 (2015). Paris: OECD. DG EXPO/​B/​PolDep/​Note/​2015_​136. Boschi, Renato, ed. 2011. Variedades de capitalismo:  Politica e desenvolvimento na América Latina. Belo Horizonte: Editora UFMG. Bresser-​Pereira, Luiz Carlos. 2014. A construçao politica do Brasil. São Paulo: Editora 34. Bresser-​Pereira, Luiz Carlos. 2016. “New Developmentalism and Developmental Economics.” Mimeo. Available at www.bresserpereira.org.br, accessed September 2016. Chang, Ha-​Joon. 2001. “Breaking the Mould: An Institutionalist Political Economy Alternative to the Neoliberal Theory of the Market and the State.” United Nations Research Institute for Social Development. Social Policy and Development Program Paper Number 6. Chossudovsky, Michel. 2003. “Capitalism with a ‘Human Face.’” Centre for Research on Globalization. Available at http://​www.globalresearch.ca/​articles/​CHO303C.html, accessed July 2016. Coes, Donald V. 2009. “Exchange Rate Policy, Perceptions of Risk and External Constraints under Lula.” In Brazil under Lula: Economy, Politics and Society under the Worker-​President, edited by Joseph Love and Werner Baer, 115–​133. New York: Palgrave Macmillan. Colby, Seth. 2012. “Explaining the BNDES: What It Is, What It Does and How It Works.” CEBRI Artigos 3, 31. Diniz, Eli. 2010. “Democracy, State and Industry: Continuity and Change between the Cardoso and Lula Administrations.” Latin American Perspectives 38 (3): 1–​19. Doctor, Mahrukh. 2007. “Lula’s Development Council: Neo-​Corporatism and Policy Reform in Brazil.” Latin American Perspectives 34 (6): 131–​148. Evans, Peter B. 1979. Dependent Development: The Alliance of Multinational, State and Local Capital in Brazil. Princeton, NJ: Princeton University Press Evans, Peter B. 1995. Embedded Autonomy: States and Industrial Transformation. Princeton, NJ: Princeton University Press

Leviathan Captured   145 Faucher, Philippe. 1979. “Croissance et répression: La double logique de l’État dépendant, le cas du Brésil.” Revue canadienne de science politique 12 (4): 747–​774. Faucher, Philippe. 1981. “La crise de croissance du régime autoritaire brésilien.” Nord-​ Sud: Revue canadienne des études latino-​américaine 6 (12): 27–​51. Faucher, Philippe. 1993. “Politicas de ajuste o erosao do Estado no Brasil?” Dados 36 (3): 393–​418. Friedman, Milton. 1962. Capitalism and Freedom. Chicago: University of Chicago Press. Hall, Peter, and D. Soskice. 2001. Varieties of Capitalism:  The Institutional Foundations of Comparative Advantage. Oxford: Oxford University Press. Kingstone, Peter R. 1999. Crafting Coalitions for Reform:  Business Preferences, Political Institutions, and Neoliberal Reform in Brazil. University Park:  Pennsylvania State University Press. Morais, Lecio, and Alfredo Saad-​Filho. 2005. “Lula and the Continuity of Neoliberalism in Brazil:  Strategic Choice, Economic Imperative or Political Schizophrenia?” Historical Materialism 13 (1): 3–​32. Morley, Samuel A., Roberto Machado, and Stefano Pettinato. 1999. “Indexes of Structural Reform in Latin America.” Eclac serie reformas economicas 12. Santiago:  United Nations Economic Commission on Latin America and the Caribbean (ECLAC). Mourougane, Annabelle. 2011. “Refining Macroeconomic Policies to Sustain Growth in Brazil.” Paris: OECD Economic Department Working Papers 899. Mueller, Bernardo, and André Rossi de Oliveira. 2009. “Regulation during the Lula Government.” In Brazil under Lula:  Economy, Politics and Society under the Worker-​ President, edited by Joseph Love and Werner Baer, 93–​114. New York: Palgrave Macmillan. Musacchio, Aldo, and Sergio G. Lazzarini. 2012. “Leviathan in Business:  Varieties of State Capitalism and their Implications for Economic Performance.” Harvard Business School Working Paper No. 12–​108, June. Musacchio, Aldo, and Sergio G. Lazzarini. 2014. “State-​Owned Enterprises in Brazil: History and Lessons.” Paris:  OECD. Workshop on State-​Owned Enterprises in the Development Process. Available at www.oecd.org. Musacchio, Aldo, and Sergio G. Lazzarini. 2016. “The Reinvention of State Capitalism in Brazil, 1970–​2012.” In New Order and Progress: Development and Democracy in Brazil, edited by Ben Ross Schneider, 107–​133. New York: Oxford University Press. Ostry, Jonathan D., Prakash Loungani, and Davide Furceri. 2016. “Neoliberalism: Oversold.” Finance and Development (June):  53.2, 38–​ 41. Washington, DC:  International Monetary Fund. Power, Timothy P. 1998. “Brazilian Politicians and Neoliberalism:  Mapping Support for the Cardoso Reforms 1995–​1997.” Journal of Interamerican Studies and World Affairs 40 (4): 51–​72. Przeworski, Adam. 1990. The State and the Economy under Capitalism. Chur:  Harwood Academic Press. Serra, Narcis, and Joseph E. Stiglitz, eds. 2008. The Washington Consensus Reconsidered. New York: Oxford University Press. Schneider, Ben Ross. 1992. “Privatization in the Collor Government: Triumph of Liberalism or Collapse of the Development State?” In The Right and Democracy in Latin America, edited by Douglas A. Chalmers, Maria do Carmo Campello de Souza, and Atilio Boron, 225–​238. New York: Praeger.

146   Philippe Faucher Schneider, Ben Ross. 2013. Hierarchical Capitalism in Latin America. New York: Cambridge University Press. Taylor, Matthew M. 2015. “The Unchanging Core of Brazilian State Capitalism, 1985–​2015.” Washington, DC: American University, School of International Service. Working Paper No. 2015–​8. Ter-​Minassian, Teresa. 2012. “Structural Reforms in Brazil: Progress and Unfinished Agenda.” Washington, DC: Inter-​American Development Bank. Policy Brief No. IDB-​PB-​158. Trebat, Thomas. 2007. Brazil’s State-​Owned Enterprises. Cambridge:  Cambridge University Press. Trubek, David M., Diogo R. Coutinho, and Mario G. Schapiro. 2013. “Towards a New Law and Development: New State Activism in Brazil and the Challenge for Law Institutions.” University of Wisconsin Law School. Legal Studies Research Paper No. 1207. Valladão, Alfredo G. A. 2013. “Emergent Brazil and the Curse of the ‘Hen’s Flight.’ ” CEPS Working Document 379. Von Hayek, Friedrich A. 1994. The Road to Serfdom. Chicago: University of Chicago Press. Weyland, Kurt. 1998. “From Leviathan to Gulliver? The Decline of the Development State in Brazil.” Governance 11 (1): 51–​75. Williamson, John. 2008. “A Short History of the Washington Consensus.” In The Washington Consensus Reconsidered:  Towards a new Global Governance, edited by Narcis Serra and Joseph E. Stiglitz, 14–​30. Oxford: Oxford University Press. Woo-​ Cummings, Meredith, ed. 1999. The Developmental State. Ithaca, NY:  Cornell University Press. World Bank. Doing Business Index. Available at http://​www.doingbusiness.org.

Chapter 8

Grow th Vol at i l i t y and Ec onom i c G row t h in Bra z i l Jorge Arbache and Sarquis J. B. Sarquis

8.1. Introduction One of the most puzzling aspects of Brazil’s economic performance is that, despite the country’s ability to grow at high rates, its long-​term growth has been disappointing.1 Between 1960 and 2015, per capita income grew by 2.25% per year, a relatively low figure for a major emerging economy. In 2015, the country’s per capita income was 29% of the average of high-​income countries, about the same level registered in 1960. As such, despite possessing considerable natural resources and having achieved formidable growth in the postwar period (Growth Commission 2008), Brazil has still yet to realize its full potential. Brazil has faced significant challenges to its economic and social development and has not yet established a robust path of high-​potential growth. In fact, it is still in the processes of overcoming the “middle-​income trap” (Im and Rosenblatt 2013)  and catching up with high-​income countries. It inevitably has to continue to defy its low rates of capital and productivity growth. At the same time, it has to progressively deal with new challenges such as globalization, the digital economy, and a rapidly aging population. Behind Brazil’s deficient long-​term trajectory lies high growth volatility. It has experienced substantial growth booms and busts as measured by international standards. Even the spectacular “economic miracle” of 1967–​1974 was followed by a long recession in the 1980s, which rendered the economy more prone to macroeconomic instability. Growth acceleration and collapses have constrained overall long-​term performance and have impaired the country’s route of convergence to the income frontier. Recently,

148    Jorge Arbache and Sarquis J. B. Sarquis again, a significant boom-​bust episode has occurred, and the economy is currently in a long process of economic recovery. These recurrent episodes suggest that the persistent nonrealization of Brazil’s long-​ term growth potential is associated with the high volatility that the economy has exhibited over decades—​that this is the mechanism by which growth has been systematically constrained. As the cross-​country empirical evidence suggests, high growth volatility is not necessarily neutral, and may be associated with low average growth rates (Arbache and Page 2007; Dabusinkas et al. 2012; Easterly et al. 2000; Ramey and Ramey 1995). Both microeconomic and macroeconomic reasons for that abound, including shortsighted behavior, uncertainty, risk aversion, investment discouragement, financial restrictions, and fiscal and current account imbalances. Such factors can have direct negative effects on growth and its volatility. They also can have protracted adverse effects, including through asymmetric impacts on economic and social variables over booms and bursts. As such, growth volatility could be a critical aspect behind Brazil’s poor growth performance. The purpose of this chapter is to provide an analysis of growth volatility and its impact on Brazil’s growth performance and long-​term trajectory. As far as we are aware, this topic has not yet been addressed in the literature. Instead, most systematic analyses have concentrated on the determinants of growth over time using standard models and growth accounting techniques.2 In this chapter, we discuss some of the critical features of Brazil’s economic growth that are associated with growth volatility. Particular attention is given to long-​term underperformance, macroeconomic regularities, and the possible causes and implications of Brazil’s limited growth potential and degree of international convergence. The discussion takes both a historical and an international perspective, as the case of Brazil is not only significant per se, but also representative of many emerging economies, particularly in Latin America. The chapter proceeds using two data sets for Brazil’s gross domestic product (GDP) per capita: the World Bank’s World Development Indicators (WDI) from 1960 to 2015, and the authors’ calculations based on the Brazilian Institute of Geography and Statistics (IBGE) series for GDP and population from 1900 to 2015.3 The former indicators allow international analysis of Brazil’s convergence and growth volatility, while the latter enables a longer-​term analysis of Brazil’s growth trajectory and of empirical features with regard to growth volatility. Section 8.2 of the chapter presents Brazil’s underperformance in terms of convergence with the international frontier, with very limited “catching up” observed from 1960 to 2015, as well as a comparative deterioration in relation to upper middle-​income countries. Section 8.3 discusses the country’s declining growth potential from a longer-​ term perspective using data from 1901 to 2015. Section 8.4 focuses on the inter-​temporal negative correlation between growth rates and growth volatilities over Brazil’s centennial path. Moreover, it compares these Brazilian statistics internationally, taking into consideration particularly the G20 advanced and emerging economies. Section 8.5 illustrates the potential long-​term effects of volatility, by means of alternative

Growth Volatility and Economic Growth in Brazil    149 simulations that take into account excessive annual growth deviations from the actual trend. Section 8.6 concentrates on the accumulative and asymmetric effects of volatility on GDP per capita. It identifies growth acceleration and collapse episodes and how the latter can further aggravate the long-​term growth trajectory. Section 8.7 considers the performance of economic variables over these two kinds of episodes, showing their respective asymmetric impacts and providing additional evidence of the non-​neutrality of growth volatility. Section 8.8 discusses the economic channels and mechanisms through which growth and volatility interplay over Brazil’s long-​term path. Section 8.9 concludes, emphasizing the need to promote a set of macroeconomic and structural policies to foster growth dynamics that are not only stronger, but also more persistent and less volatile.

8.2.  Is Brazil Catching Up? A key question to be posed is whether Brazil has managed to converge or is in the process of converging with developed countries in terms of per capita income. As economic theory indicates (Barro and Sala-​i-​Martin 2003; Solow 1956), not only does per capita income tend to grow over time thanks to technological progress, less developed countries are also expected to catch up with countries that are at—​or closer to—​the world’s frontiers of knowledge and technology. Less developed countries can grow faster than more advanced economies while displaying higher returns of capital. They tend to converge by attracting international flows of capital and benefiting from international knowledge diffusion and technology spillovers. However, as many countries’ experiences have shown, “catching up” can be a hard task. For several reasons, the process of converging with leading economies can be partial or conditional on a set of complex requirements, as extensively discussed in the empirical literature (Barro 2015). Many underlying economic and non-​economic factors that distinguish countries’ institutions, structures, and economic dynamism, in addition to a set of initial conditions, affect countries’ economic growth and convergence rates (Acemoglu et al. 2005; Lucas 1990). Therefore, the expectation of international convergence in per capita income is not necessarily warranted; in fact, this measure has become increasingly spread and unequal across countries (Milanovic 2016). Despite cases of countries that exhibit considerable degrees of convergence, such as Japan and the so-​called Asian Tigers (Hong Kong, Singapore, South Korea, and Taiwan), most developing and emerging countries in Asia, Africa, and Latin America have not yet displayed significant changes in the direction of higher degrees of convergence. A few cases of fast-​growing economies in recent history, notably China,4 remain promising, but must still overcome the test of the middle-​income trap (Arias and Wen 2015). In fact, some developing and emerging economies that experienced important episodes of rapid economic growth in their history have not yet moved toward the convergence path. To eventually achieve it, they would need to move closer to a trajectory of sustained growth

150    Jorge Arbache and Sarquis J. B. Sarquis 0.4

0.4

0.35

0.35

0.3

0.3

0.25

0.25

0.2

0.2

0.15

0.15

0.1

0.1

0.05

0.05 0 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

0

Brazil

Latin America

Figure  8.1.  Ratios of Brazil’s and Latin America’s per capita income to that of high-​income countries. Data source: WDI; authors’ calculations.

and persistence in both capital accumulation and technological development. This is Brazil’s situation. As Figure 8.1 shows, any potential convergence of Brazil’s GDP per capita with that of the group of high-​income countries has not been sustained from 1960 to 2015.5 In 1960, Brazil’s GDP per capita was equivalent to 31% of the group’s level. After a decline in the early 1960s, Brazil succeeded in accelerating convergence from the mid-​1960s until the late 1970s. The peak was reached with a per capita income ratio of around 36% over the second half of the 1970s. Since then, the ratio has exhibited declining or stagnant trends, falling to 25% in the transition from the twentieth century to the twenty-​first, and, more recently, recovering to around the same level as at the beginning of the 1960s. Despite significant high growth over recent decades, Brazil has not consolidated a long-​term convergence to effectively climb the economic ladder. The Brazilian stagnant or non-​convergent trend indicates the presence of underlying economic constraints upon the country’s potential to grow faster than and thus catch up with the frontier. Furthermore, over the same period, as depicted in Figure 8.2, Brazil has also exhibited a declining trend in relation to peers among upper-​middle-​income countries. The average per capita income of this group has considerably expanded in the last decades. The group’s average has caught up with Brazil, from a ratio close to 3 at the beginning of the 1980s, to around 1.5 in 2015. Such a move largely derives from the high and sustained growth rates found in China and other Asian middle-​income countries over recent decades. Some of these countries have found at least conditional convergence with Brazil, which can be regarded as a leader within the group, and are set to challenge the middle-​income trap. Other Latin American countries share Brazil’s non-​convergent pattern to a considerable extent. The region’s average ratio of per capita income over time also exhibits

Growth Volatility and Economic Growth in Brazil    151 4

4 3.5

3.5

3

3

2.5

2.5

2

2

1.5

1.5

1

1

0.5

0.5 0 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

0

Brazil

Latin America

Figure 8.2.  Ratios of Brazil`s and Latin America’s per capita income to that of upper-​middle-​ income countries. Data source: WDI; authors’ calculations.

declining trends vis-​à-​vis both frontier and peer countries. The negative trend appears to be stronger than in Brazil, and the implications are at least as worrisome for many countries in the region. Certainly, the term “middle-​income trap” can be subject to criticism on both theoretical and empirical grounds, since it is not clear whether there are definitions of “trap” and of “middle-​income” that remain consistent and robust over time. However, as argued by Im and Rosenblatt (2013), who discuss such issues, it is undeniable that Brazil and many middle-​income countries, not only in Latin America, but also in Asia and Africa, “have been substantially off-​track in recent decades in terms of catching up to either a US standard or a high-​income OECD standard” (Im and Rosenblatt 2013, 7). As they show, among 13 major developing countries of these regions, only China exhibited an average growth rate over the 1980–​2011 period higher than the growth rate necessary to catch up with the United States or the Organisation for Economic Co-​operation and Development (OECD) in 50 years.

8.3.  Low and Declining Growth Potential In order to explain Brazil’s challenges in converging toward the highest living standards, we first examine the country’s growth performance over time. Today’s living standards are the result of successive annual rates of per capita income growth. As stated by Barro

152    Jorge Arbache and Sarquis J. B. Sarquis and Sala-​i-​Martin (2003, 6), “if we want to understand why countries differ dramatically in standards of living [. . .], we have to understand why countries experience such sharp divergences in long-​term growth rates.” For a long time, Brazil has been viewed as a country of promising progress with potentially high rates of returns and growth (Growth Commission 2008). At the beginning of the twentieth century, it was one of the countries in the New World that, like the United States, managed to attract considerable flows of immigrants who came in search of jobs and opportunities and therefore helped to build the country’s potential. Thus, Brazil eventually moved into the initial stages of industrialization in the 1930s. This process intensified in the 1950s and was characterized by the accumulation of technologies and physical and human capital (Baer 2010). Through this industrialization phase, growth accelerated, showing an average upward trend from the 1920s to the 1970s (Table 8.1). In the transition to the following decades, the economy experienced a succession of foreign shocks in the 1970s and early 1980s, such as the sharp rise in oil prices and in international interest rates. Thsey caused considerable harm to the balance of payments and also exposed domestic economic vulnerabilities. These vulnerabilities were characterized by poor monetary policy and fiscal management, especially during the 1980s and early 1990s, and by procrastination over badly needed reforms (Arbache 2004). In sharp contrast to the most recent decades, Brazil underwent a long period of recession, and the economy grew on average a mere 0.29% in 1981–​2000.

Table 8.1 Brazil Per Capita Income Growth (%), 1901–​2015 Period

Average

Standard Deviation

Coefficient of Variation

1901–​2015

2.53

4.14

1.64

1901–​1940

2.24

4.92

2.20

1941–​1980

4.27

3.31

0.78

1981–​2015

1.03

3.27

3.16

1901–​1920

1.47

5.24

3.57

1921–​1940

3.01

4.57

1.52

1941–​1960

3.89

3.21

0.83

1961–​1980

4.64

3.45

0.74

1981–​2000

0.29

3.51

12.10

2001–​2015

1.64

2.84

1.73

Data source: IBGE and Instituto de Pesquisa Econômica Aplicada (IPEA); authors’ calculations. Averages are arithmetic means of annual growth rates.

Growth Volatility and Economic Growth in Brazil    153 Indeed, it was in the 1980s that Brazil’s trajectory of high growth and international convergence eventually faltered. Since then, growth has been low or constrained, and the economy has remained distant from its previous convergence path. Some growth conditions were improved through the 1990s thanks to greater openness to trade, other economic reforms, and fiscal adjustment initiated in the middle of the decade. Building on some positive effects of these measures and benefiting from high commodity prices, growth recovered in the early twenty-​first century, but at a limited pace of 1.64% over the 2001–​2015 period. The average annual growth per capita rose from 1.47% in 1901–​1920 to 3.01% in 1921–​ 1940. Brazil managed to expand at high rates of growth, especially in the three decades following World War II. The average rate reached 3.89% in 1941–​1960, and peaked at 4.64% in 1961–​1980, this last phase including the so-​called economic miracle that provided the country with a solid manufacturing base (Baer 2010). As Figure 8.3 indicates, over that period, besides attaining high levels, growth also became substantially more persistent and less volatile. However, such recent aspects of Brazil’s economic performance can be deceiving; the country’s more long-​term levels of overall growth, which averaged 2.53% over the 1901–​ 2015 period, would not be sufficient to bring it back onto a more promising trajectory of convergence with leading economies over the next decades. Taking into account some realistic degree of convergence, Johansson et al. (2013) project that Brazil could grow on average at a rate of 3.4% annually from 2011 to 2030 and eventually at a rate of 2.1% annually by 2060. Accordingly, Brazil’s GDP per capita, as a percentage of that of the United States, would increase from about 22% in 2011 to 40% by 2060. Important challenges to closing this gap will therefore remain, and considerable growth acceleration might be required to produce a faster and more effective process of convergence. 12 11 10 9 8 7 6 5 4 3 2 1 0 –1 –2 –3 –4 –5 –6 –7 1901 1904 1907 1910 1913 1916 1919 1922 1925 1928 1931 1934 1937 1940 1943 1946 1949 1952 1955 1958 1961 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 2015

12 11 10 9 8 7 6 5 4 3 2 1 0 –1 –2 –3 –4 –5 –6 –7

Annual growth

Trend of annual growth

Figure 8.3.  Brazil’s annual growth of per capita income and its linear trend over time (%). Data source: IBGE and Instituto de Pesquisa Econômica Aplicada (IPEA) and IMF projections for 2016 and 2017; authors’ calculations based on GDP and population series.

154    Jorge Arbache and Sarquis J. B. Sarquis Despite its acceleration from the 1920s to 1970s, Brazil’s potential growth rate shows a declining trend over the long term, as depicted in Figure 8.3. This statistical result is largely due to the growth brake of the 1980s that interrupted the fast expansion of the 1960s and 1970s. The emergence of the 1980s crisis and its negative effects, some of which still persist today, are not an isolated fact in Brazil’s economic growth history. On the contrary, they are part of Brazil’s intertemporal dynamics, reflecting macroeconomic imbalances and structural weaknesses that manifest in the empirical regularities of Brazil’s long-​term growth performance. Overall, Brazil’s long-​term growth trajectory displays features that contrast with the expected trajectories of advanced economies and of countries that have seemingly been more successful in overcoming the middle-​income trap. Indeed, Brazil has experienced extraordinary boom cycles that could have eventually resulted in a successful trend of rising per capita income. However, most of the advancement in terms of grater convergence brought about by boom cycles was later undermined by downturns. Despite its contribution to Brazil’s growth, the high and strong growth persistence of the economic miracle remains a outlier in the country’s economic history. Over the long term, as registered more recently, Brazil has undergone sufficiently severe bust cycles to jeopardize the cumulative process of growth.

8.4.  High Growth Volatility A key feature of Brazil’s growth trajectory is the excessive and continuous volatility of its growth rates. In this section, we study this feature intertemporally, examining the long-​term dynamic relationship between growth and volatility, and internationally, comparing Brazil with advanced and emerging economies in this regard. The volatility of growth can be measured by the standard deviation of the annual per capita income growth rate over a period. It can also be defined in relative terms by the coefficient of variation of the growth rate. These measures are shown in Table 8.1 for Brazil over different periods. Furthermore, Figure 8.4 gives a dynamic picture of these moments, depicting the 10-​year moving average of growth and the 10-​year standard deviation of growth from 1901 to 2015. Figure 8.4 clearly shows the intertemporal negative correlation between growth and volatility in Brazil. Brazil managed to accelerate its economic growth from the 1920s to 1970s, while also reducing the standard deviation of growth. The correlations between low volatility and higher growth are also clear in the period since the economic miracle. High standard deviations of growth coexist with low growth rates, particularly during the 1980s, while the growth recovery from the 1990s to the first decade of the 2000s was accompanied by a lowering of volatility. Finally, volatility has recently risen, while 10-​ year growth has dramatically shifted downward. Such a negative correlation renders stronger results, too, if volatility is measured in relative terms, as by the coefficient of variation in Table 8.1.

Growth Volatility and Economic Growth in Brazil    155 8

7

7

6

6

5

5

4

4

3

3

2

2

1

1

0

0

–1

–1

1903 1907 1911 1915 1919 1923 1927 1931 1935 1939 1943 1947 1951 1955 1959 1963 1967 1971 1975 1979 1983 1987 1991 1995 1999 2003 2007 2011 2015

8

Growth

Standard deviation

Figure 8.4.  Brazil’s backward 10-​year moving average growth and standard deviation (%). Data source: Instituto de Pesquisa Econômica Aplicada (IPEA); authors’ calculations.

The economic miracle culminated in both the highest rates of economic growth and the lowest levels of growth volatility in relative terms, as shown in Table 8.1 for the two-​decennial periods. In Figure 8.4, the 10-​year growth peak in 1976 coincides with one of the troughs in the 10-​year standard deviation. This was just slightly higher than the troughs registered in the 1950s and the beginning of the twenty-​first century. As mentioned before, the economic miracle was special, as it featured a strong persistence of economic activity, not only during the upward trend to 1976, but also through the downward trend since then. Such a degree of persistence has not yet been repeated, despite the greater stability brought about since the mid-​1990s that helped to build the moderate growth recovery over the first decade of the 2000s. In order to allow for international comparisons, Table 8.2 presents the average annual per capita income growth rates and volatility measures for G20 countries from 1961 to 2015. The G20 countries constitute a suitable benchmark for Brazil, in that this group comprises economies of comparable sizes, but at various stages of development and in various regions.6 The table also displays average estimates for the group and for 24 Latin American countries. Brazil’s average rate of growth is higher than Latin America’s mean rate, but lower than that of the the G20. At the same time, the volatility of growth in Brazil is lower than in Latin America on average, both in terms of standard deviation and of the coefficient of variation. Despite displaying lower average growth than the G20, Brazil has higher growth variability as measured by the standard deviation.7 The contrasts between Brazil and other Latin American countries (inside or outside the G20), on the one hand, and Asian emerging economies, on the other hand, are also consistent with the suggested negative correlation between growth volatility and performance. China, India, Indonesia, and Korea have all registered higher growth rates and lower volatility than Brazil, and therefore much lower coefficients of variation.

156    Jorge Arbache and Sarquis J. B. Sarquis Table 8.2 Annual Per Capita Income Growth: G20 Countries’ Descriptive Statistics, 1961–​2015 Average

Standard Deviation

Coefficient of Variation

Argentina

1.51

5.61

3.71

Australia

1.93

1.73

0.90

Brazil

2.25

3.80

1.69

Canada

1.93

2.09

1.08

China

6.86

6.85

1.00

France

2.14

1.93

0.90

Germany

1.89

2.01

1.06

India

3.26

3.20

0.98

Indonesia

3.57

3.59

1.01

Italy

2.11

2.68

1.27

Japan

3.15

3.66

1.16

Korea, Rep.

5.91

3.83

0.65

Mexico

1.79

3.23

1.81

Russian Federation

0.67

6.95

10.38

Saudi Arabia

1.07

6.61

6.16

South Africa

1.01

2.50

2.48

Turkey

2.53

3.82

1.51

United Kingdom

2.01

2.08

1.04

United States

2.05

2.02

0.98

G20 (exc. Brazil)

2.52

3.58

2.11

G20

2.51

3.59

2.09

Latin America

1.78

4.11

4.02

Data source: WDI; authors’ calculations. Note: G20 here comprises its 19 member countries, not accounting therefore for the EU. The G20 statistics refer to the average statistics for these countries, while the line above excludes Brazil.

Indeed, one of the most remarkable features of Brazilian growth has been its economic instability. Due to various factors and channels, economic instability generates growth volatility. Such volatility results from and aggravates the variability of aggregates, notably consumption, savings, investments, and exports. The high level of variability of GDP growth and its components are a common feature of developing and emerging countries in general (Agenor et  al. 1999), but in particular of Latin America among middle-​income countries.8

Growth Volatility and Economic Growth in Brazil    157 Table 8.3 Annual Per Capita Income Growth: Brazil and G20 Descriptive Statistics, 1961–​2015 Mean

Standard Deviation

Coefficient of Variation

Brazil

G20

Brazil

G20

Brazil

G20

1961–​2015

2.25

2.51

3.80

3.59

1.69

2.09

1961–​1980

4.59

3.43

3.66

3.55

0.80

1.15

1981–​2000

0.34

1.91

3.48

3.21

10.13

2.23

2001–​2015

1.69

2.30

2.80

2.44

1.66

1.22

Data source: WDI; authors’ calculations.

Brazil’s growth underperformance, which has been especially severe since the 1980s, has contrasted with relatively greater moderation of volatility among a number of G20 countries, as suggested by Table 8.3. In comparison to other G20 countries, high volatility appears to be one determining factor behind the very low growth experienced since the 1980s in Brazil. Despite the significant improvement brought about by stabilization programs and other reforms since the 1990s, Brazil’s performance has remained below the G20’s average and is characterized by excessive growth variability.

8.5.  Volatility and Growth Simulations In order to explore further the impacts of volatility on growth performance, we simulate what would have happened to Brazil’s trajectory under alternative growth scenarios. The scenarios rely on assumptions that the economy could have grown in a manner closer to its long-​term trend, avoiding excessive high and/​or low rates of growth in some years. Table 8.4 presents the simulations using both data series employed in earlier sections, covering the periods 1901–​2015 (IBGE) and 1961–​2015 (WDI). In simulation 1, we substitute the average long-​term growth rate (average rate) for years where the actual growth rate was below the average rate minus one standard deviation (i.e., we replaced the worst growth years with the long-​term growth rate). As a result, the long-​term growth rate would have been about 1 percentage point higher. In simulation 2, we substitute the average long-​term growth rate for years when the actual growth rate was above the average rate plus one standard deviation (i.e., we replaced the best growth years with the long-​term growth rate). As a result, the long-​ term growth rate would have been close to 1 percentage point lower.

158    Jorge Arbache and Sarquis J. B. Sarquis Table 8.4 Simulating the Average Long-​Term Growth Rate under Different Conditions (%), 1901–​2015 and 1960–​2015 Simulated Average Growth Rates (%) Simulations

1901–​2015

1960–​2015

No simulation. Actual growth rates

2.53 (4.14) ((0.00))

2.25 (3.80) ((0.0))

Simulation 1. Substitute long-​term growth rate for years where actual growth rate was below (average rate—​one standard deviation)

3.62 (2.98) ((1.09))

3.12 (2.83) ((0.87))

Simulation 2. Substitute long-​term growth rate for years where actual growth rate was above (average rate + one standard deviation)

1.45 (3.02) ((1.08))

1.24 (2.69) ((1.01))

Simulation 3. Substitute long-​term growth rate for years where actual growth rate was below (average rate—​one standard deviation) or above (average rate + one standard deviation)

2.54 (1.81) ((0.01))

2.10 (1.60) ((0.15))

Notes: Standard deviations in parentheses (); and differences to actual average rates in (()) Source: Authors’ calculations.

In simulation 3, we substitute the average long-​term growth rate for years when the actual growth rate was either below or above the average rate, plus/​minus one standard deviation (i.e., we replaced the worst/​best growth years with the long-​term growth rate). As a result, the long-​term growth rate would have been close to the actual average rate. These exercises demonstrate that the excessive volatility induced by both high and low growth rates does indeed have a very considerable impact on growth performance. It suffices to note that the differences of average rates between simulations 1 and 2 are about 2 percentage points, which is close to the actual long-​term average rates for the two series used. Moreover, we can begin to explore the asymmetric effects of volatility on GDP per capita and therefore on international convergence. Despite a rough equivalence of average growth rates between simulation 3 and the actual growth trajectory,9 especially for the 1901–​2015 period, the higher the volatility and the longer its duration, the greater the negative accumulative effects on per capita income. These arguments are more clearly summarized in Figure 8.5, which depicts the overall impacts of the simulations in Table 8.4 on per capita income over the long term. Had the economy avoided low growth rates over the period 1961–​2015, income would have been 61.7% higher in 2015. Over the longer period, starting in 1901, per capita income would have grown to a level 254% higher in 2015. This would have led Brazil into the group of high-​income countries, at a level just 4% below the group’s average.

Growth Volatility and Economic Growth in Brazil    159 300 254.4

250 200 150 100

61.7

50

9.7

0

–4.9

–50 –100

–41.2 Simulation 1

–69 Simulation 2

1961–2015

Simulation 3

1901–2015

Figure  8.5.  Per capita GDP in 2015 under different growth simulations from 1961 and from 1901: Gap with actual data (%). Note: Growth conditions as in Table 8.4; each simulation reflects results using the 1961–​2015 and 1901–​2015 series in order. Source: Authors’ calculations.

Figure 8.5 also shows that, had the economy avoided high growth rates from 1961, income would have been 41.2% lower than the actual level observed in 2015. Accounting for the cumulative impact since 1901, the level would have been 69% lower in 2015. Finally, had the economy avoided both extreme rates, per capita income would have been closer to the real figure. The discrepancies depend, ultimately, on the sample period. However, overall, given the size of the non-​realized income in simulation 1, the harm to the economy of inferior rates of growth substantially outweighs the benefits of superior rates of growth. Figure 8.6 compares per capita income under the actual growth trajectory—​the baseline—​and alternative per capita income trajectories, as given by simulations 1–​3. The income gap in 2015 between the baseline and simulations 1 and 2 can mostly be accounted for by the growth acceleration that took place until 1980, especially during the economic miracle years, and the collapse of growth rates since then. That explains the similarity between the baseline and simulation 1 trajectories, as well as the discrepancy between the baseline and simulation 2 trajectories, at least up to 1980; and it explains the greater divergence between the baseline and simulation 1 trajectories, as well as the greater convergence between the baseline and simulation 2 trajectories since 1980. While the incidence of growth expansions was greater before 1980, the incidence of growth contractions grew substantially after 1980. Figure 8.7 shows the decadal incidence of years where actual annual growth was above long-​term growth. The upward trend, which showed persistency up to the 1970s, halted in the 1980s. Since then, the incidence has remained at much lower levels, suggesting again that potential growth and growth persistency have receded in the most

160    Jorge Arbache and Sarquis J. B. Sarquis 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

0

Actual Series

Simulation 1

Simulation 2

Simulation 3

Figure 8.6.  Simulated trajectories of per capita income under different growth scenarios ($), 1961–​2015. Note: Simulations as in Table 8.4; comparison with the actual per capita income trajectory. Source: Authors’ calculations.

10 9 8 7 6 5 4 3 2 1 15 20 11 -20

10 -20 01 20

-20 00 91 19

90 -19 81 19

-19 80 71 19

-19 70 61

0

No trend

19

96 –1 51 19

0

19 41 –1 95 0

94 –1 31 19

19 21 –1 93 0

19 11 –1 92 0

19 01 –1 91 0

0

With trend

Figure 8.7.  Incidence of years with actual growth above long-​term growth, 1901–​2015 (%). Source: Authors’ calculations. Note: The incidence of years with trend is based on a linear trend, in place of average growth, for the long-​term growth.

recent decades, despite some improvement from the 1990s to the 2000s. Once more, in this section, our findings suggest that high volatility can have a detrimental impact on growth. It renders growth weaker and less persistent, with considerable losses of income over the long term.

Growth Volatility and Economic Growth in Brazil    161

8.6.  Growth Accelerations and Collapses In this section, we focus on specific episodes of growth acceleration and growth collapse. While volatility constrains long-​term growth, including through the asymmetric effects of expansions and contractions, acceleration and collapse episodes can change the growth trajectory even more dramatically, inflicting persistent shifts over and beyond typical or regular cycles. These episodes magnify high growth volatility, which has adverse consequences on parameter stability in regression models and economic analysis, especially in low-​growth periods (Arbache and Page 2007; Hausmann, Pritchett, and Rodrik 2005; Johnson, Ostry, and Subramanian 2007;Pritchett 2000).10 Growth episode analyses in those circumstances are advisable in order to identify substantially different behavior beyond the expected trend and cyclical regularities. In order to examine growth accelerations and collapses, we employ a methodology to identify these two kinds of episodes, similar to that developed by Arbache and Page (2007). The episodes of acceleration and collapse are defined in two steps: the reference and the remaining years. The reference years are at least three years in a row in which two conditions are observed for every year: (a) annual growth is higher than the long-​term average growth; and (b) the forward four-​year moving average growth rate is higher than the backward four-​year moving average. The remaining years are those years in a row, immediately after the reference years, in which condition (a) still holds but condition (b) does not.11 Applying this methodology for the 1965–​2015 series, we find only one acceleration episode for Brazil. It consists of seven consecutive years, from 1968 to 1974 (reference years 1968–​1971).12 Following that, there are three collapse episodes, comprising a total of 14 years: 1987–​1992 (reference 1987–​1989), 1996–​1999 (reference 1996–​1998), and 2012–​2015 (reference 2012–​2015). When using the 1901–​2015 series, the same results are found for the period starting in 1961.13 Additionally, the 1996–​1999 collapse turns to be two years longer, thus encompassing the period from 1996 to 2001. For the years prior to 1961, we identify two growth acceleration episodes in total, encompassing nine years: from 1932 to 1934 (reference 1932–​1934), and from 1957 to 1962 (reference 1957–​1959). Overall, over 115 years since 1901, Brazil has experienced three growth acceleration episodes totaling 16 years, and three growth collapses, also totaling 16 years. Figure 8.8 compares Brazil’s average growth rates during these two kinds of episodes. The acceleration phases considerably exceed average growth by 4 to 6 percentage points, while collapses were 3 percentage points below the long-​term mean. Even more noteworthy, the gaps between acceleration and collapse episodes range from 6 to 9 percentage points. For 1901–​2015, the excessive growth of the acceleration episodes vis-​à-​vis the long-​ term trend represents around 47% gain in GDP per capita. The three collapse episodes,

162    Jorge Arbache and Sarquis J. B. Sarquis 10

2 0 –2

6.85

6.32

6 4

8.75

8.16

8

2.25 2.53 –0.58 –0.53 Long-term

Accelerations 1961–2015

Collapses 1901–2015

Gap between accelerations and collapses

Figure  8.8.  Average growth rates during episodes of growth acceleration and collapse (%), 1961–​2015 and 1901–​2015. Note: Growth conditions as in Table 8.4; each simulation reflects results using the 1961–​2015 and 1901–​2015 series in order. Data source: WDI; authors’ calculations based on the methodology of Arbache and Page (2007).

however, amount to a loss of 206% in GDP per capita relative to the long-​term trend up to 2015. The various acceleration and collapse episodes further expose the economy to a very bumpy, low persistent-​growth trajectory. Collapses curtail aggregate demand, and particularly capital accumulation and technological progress, in a more severe way than regular business cycle downturns. They generate deeper and longer recessions, augment (long-​term) unemployment through hysteresis effects, and erode the hard-​to-​sustain conditions that stimulate innovation and the development of skilled labor. As collapses are more likely or effectively become more frequent, growth volatility is eventually exacerbated and growth performance tends to be further restricted. Growth persistence over and beyond business cycles deteriorates and is further dependent on the moderation of growth volatility.

8.7.  Macroeconomic Volatility and Asymmetric Responses in Accelerations and Collapses The non-​neutrality of growth volatility manifests through different economic and financial variables and channels. Table 8.5 presents the actual figures of a selection of key indicators during growth accelerations and growth collapses. Considering their contrasting responses to these two types of episode—​responses that are sometimes highly asymmetric—​helps to understand how volatility can by different means affect economic growth.14 Total factor productivity (TFP) has on average grown at low rates in Brazil and constitutes perhaps one of the sources of the GDP per capita gap between Brazil and

Growth Volatility and Economic Growth in Brazil    163 Table 8.5 Means of Selected Indicators, 1961–​2015 Means over Available Sample

Mean

Growth Accelerations

Growth Collapses

Total factor productivity, growth rate (%)

1961–​2014

0.39

3.28

–​1.89

Investment, growth rate (%)

1971–​2015

3.72

16.58

–​1.63

Capactity utilization (%)

1970–​2014

81.76

87.60

79.94(*)

Current account (% GDP)

1960–​2015

–​1.95

–​2.87(**)

–​1.90(**)

Current account, difference from previous period (% GDP)

1961–​2015

–​0.01

–​0.85(*)

0.04

Domestic credit provided by the financial sector (% GDP)

1960–​2015

70.23

39.34

105.44

Domestic credit provided by the financial sector, trend adjusted (% GDP)

1960–​2015

–​0.02

–​6.34(*)

19.81

Indicator

Source: IPEA, Penn Tables, and WDI; authors’ calculations. Notes: Unilateral T-​tests suggest that the aacceleration and collapse means above either are unequal or differ from the overall mean, always at the 5% significance level, except in some cases, as indicated, at 10% (*) or 20% (**) significance levels.

OECD countries (Johansson et al. 2013). It expands at higher rates during acceleration periods than it falls during collapses. The same applies to investments, which respond even more strongly over acceleration episodes. In fact, there seem to be asymmetric reactions of TFP and investments to accelerations and collapses. While the latter induce the contraction of these variables, growth accelerations amplify their expansion. However, on average, from a long-​term perspective, their growth is limited and insufficient to spur convergence. The responsiveness of TFP and investments to accelerations and their poor performance on average, further impaired by collapses, indicate the existence of important obstacles to high and persistent rates of growth. Another evidence of such constrained growth and investment dynamics is the higher level of capacity utilization in the acceleration period than that registered on average or during collapses. As discussed by Bacha and Bonelli (2016), the variations, overexpansions and contractions, in capacity utilization as well as in the price and productivity of capital, reflect such restricted dynamics. Investment booms during expansions are not sufficient to induce sustained expansions of capital formation. Consistent with such limitations, the current account deficit is higher through growth acceleration as consumption and especially investment expand at much higher rates. In contrast, during collapses the current account balance tends to shift upward as part of the exchange rate and balance of payment adjustments. Typically, under these

164    Jorge Arbache and Sarquis J. B. Sarquis conditions, the exchange rate also displays severe volatility, with excessive local currency appreciation during booms that can hamper the development of competitive manufacturing exports. These restrictions and adjustments seem to be associated with the growth cycles of an economy dependent on external financing. Consistent with that view, business cycles and growth dynamics reflect features of an internationally credit-​constrained economy (Sarquis 2009, 2011). The stock of domestic credit as a ratio to GDP tends to grow during collapses, as a sign of excessive debt leverage and domestic consumption smoothing. Certainly, however, in recent decades Brazil has undergone considerable financial deepening, besides some degree of greater trade and financial openness. Financial deepening can partially explain a higher credit ratio to GDP in collapses taking place after a long acceleration period. Nevertheless, the results are robust for the consideration of trends in financial deepening. As discussed by Easterly et al. (2000) and Carvalho and Gabaix (2013), financial deepening might have ambiguous or nonlinear effects on growth volatility. It might, on the one hand, strengthen the ability to respond to shocks, while, on the other hand, exacerbating exposure to financial shocks. A similar analysis applies to trade and financial openness, which might contribute to risk diversification, while rendering the economy more volatile as it becomes more responsive to foreign demand and credit shocks. In fact, debt accumulation in foreign and domestic markets is an issue that pertains both to openness and to financial deepening. Unemployment, poverty, and income inequality also show unequal and asymmetric responses during acceleration and collapse episodes. They tend to deteriorate disproportionately more during collapse episodes than they improve during accelerations. Thus, avoiding growth collapses seems to be far more effective for poverty alleviation than supporting growth bursts. Of course, there are other specific reasons explaining the responses of each indicator during different growth circumstances. Whatever the reasons, the analysis tends to suggest that growth volatility is indeed non-​neutral in relation to other key variables that might determine and restrict growth performance, as well as its social imbalances. Contractions and collapses can have persistently adverse effects, rendering their overall impact particularly asymmetric and detrimental to long-​term GDP, despite the mostly temporary benefits of expansions and accelerations.

8.8.  Additional Discussions of Volatility and Growth in Brazil Ever since Ramey and Ramey (1995) proposed a causal relationship between growth and its volatility, the literature has found both negative and positive effects of volatility on growth (see Dopke 2004 for a summary). When analyzed either over its own

Growth Volatility and Economic Growth in Brazil    165 long-​term trajectory or in contrast to advanced countries and emerging and developing economies, namely in Asia, the case of Brazil tends to show more evidence in favor of adverse effects of volatility on growth. The workings of these phenomena and their causes can be viewed from both microeconomic and macroeconomic perspectives. The Brazilian experience confirms that volatility reduces economic predictability, generates ambiguous signals, and therefore reduces the appetite for long-​term return investments. It precipitates shortsighted behavior and excessive debt leverage, and thus increases uncertainties, risk aversion, and financing costs. All these factors are detrimental to sustained investments and to long-​term projects, particularly those intensive in innovation and human capital (Aghion et al. 2010). Volatility therefore lowers growth and reduces the persistency of growth rates over medium-​to long-​term horizons (Pritchett and Summers 2014). As suggested empirically in the previous section, high volatility curtails TFP and investment dynamics, while also affecting savings and further exacerbating the obstacles to high and sustained growth. It also brings instability to monetary and fiscal policy, as well as to financial markets. This manifests in pressures over domestic interest rates, and a higher premium of longer-​term credit and of foreign credit risks. All these financial costs, including term spreads and risk premia, respond intensively to domestic and international shocks and through financial accelerator mechanisms render investment and economic activity overall more volatile, as discussed in the case of Brazil by Sarquis (2009). Moreover, shifts in this risk perception are associated with exchange rate movements and with drastic changes in the current account. Through all of these mechanisms, the economy becomes sensitive to, for instance, cyclical gains of exports of commodities and semi-​manufactured goods, or to excessive capital flows, via the concentration of foreign investments in mergers and acquisitions, portfolio investments, and public services such as energy and telecommunications. Furthermore, as observed in the Brazilian case, political economy mechanisms come into play. Volatility encourages firms and individuals to turn to the public sector for security, including the search for public jobs, tax breaks, and other demands, such as public services and public goods, which result in bigger government, increased government interventions, and a higher public fiscal burden (Jetter et al. 2013; Rodrik 1998). As more resources are withdrawn from the private sector, this can in turn lower growth rates, at least in the short run (Afonso and Castro 2016; Afonso and Furceri 2010; Barro 2001; Barro and Lee 1994).15 Brazil has been trapped in a growth dynamic that has become both weaker and more volatile over time. It has suffered from the interplay of low savings and investment rates as much as from low productivity and high costs of financing and production factors. As Bacha and Bonelli (2004, 2016) document, low rates of investment in Brazil result not only from low savings, but also from high costs of capital. The latter was historically induced by the import-​substitution industrialization strategy and by a lack of trade openness, and is affected by high financial costs, insufficient long-​term financing, and excessive exchange rate volatility.

166    Jorge Arbache and Sarquis J. B. Sarquis Growth volatility, as manifested in acute crises of various sorts (such as balance of payment, fiscal, currency, or monetary crises), is not merely a symptom of an unbalanced growth trajectory. It reflects underlying macroeconomic and structural weaknesses that severely restrict growth, impeding the sustained moderate to high rates that would assure convergence. As with many developing and emerging economies, Brazil is confronted with different macroeconomic and structural weaknesses. A good number of these pertain to domestic and foreign credit constraints (Cabalero and Krishnamurthy 2001), which restrict investments, rendering them and economic activity in general more volatile, and eventually curbing mean growth (Aghion and Banerjee 2005; Aghion et al. 2010). The major triggers of the tightening of domestic constraints in Brazil reside in fiscal imbalances.16 By aggravating uncertainties and financing costs, fiscal imbalances augment volatility and restrict and even halt public and private investments. For instance, Adrogué et al. (2006) emphasize the workings of high levels of government consumption that impose high real interest rates, lowering investment and growth. Such fiscal imbalances and the volatility they generate also interfere in debt maturity and the term structure not only of the public debt, but also in the formation of long-​term financial and credit markets that are key to sustained growth and development. As Blanchard (2005) argues, fiscal dominance can be so entrenched that, within the inflation target regime, the necessary tightening of interest rates in view of inflation pressures can lead to adverse effects in terms of both debt sustainability and real depreciation. This situation obtains especially for high levels of debt, of foreign-​currency-​ denominated debt, and of risk premia, and monetary policy responses further amplify volatility and its negative effects on investment dynamics. In open economies, pressures of this kind can be particularly severe. International financial openness, as much as trade openness, might in principle provide a mechanism by which a country could diversify risks and smooth shocks. However, at the same time, openness could expose it to greater volatility, as exogenous shifts in the terms of trade and in capital flows can, through foreign (credit) and domestic (monetary and fiscal) constraints, further disrupt economic activity (Easterly et al. 2000). The patterns of trade and financial integration matter, and therefore countries respond differently to more or less integration. Trade openness can also be an important trigger of TFP improvement and a balanced growth strategy. Some empirical studies document the positive effects of trade liberalization on productivity growth in Brazil (e.g., Ferreira and Rossi 2003; Sarquis 2011). Sarquis (2011), for example, gathers evidence of causal relationships running from imports, particularly of machines and equipment, to TFP and to exports of manufacturing goods. Sarquis (2011) also suggests that the sequencing of trade and financial integration helps to explain the different performances of some Asian and Latin American countries, as they oriented their strategies differently in this regard. While Asian countries have tended to focus more on trade-​led growth strategies, Latin American countries have concentrated on strategies that were unavoidably dependent on international

Growth Volatility and Economic Growth in Brazil    167 finance (Agenor et  al. 1999). Despite remaining less open to trade and international finance than advanced economies, Asian and Latin American countries developed different sequencings of openness. The former have focused on trade from the start, and have managed to develop an export-​led growth strategy that implies savings generation. Latin American economies, on the contrary, have concentrated on gaining access to international finance, and through an import-​substitution strategy have become more dependent on external savings and more exposed to growth volatility. In the case of Brazil, which could be regarded as one of constrained capital import-​ led growth (Sarquis 2011), severe foreign financial constraints are still reflected in often-​ persistent current account deficits and high country-​risk premia. Additionally, these constraints interact with domestic constraints, which manifest in low savings, unsettled public debt dynamics, and high real interest rates. Brazil only managed to temporarily avert some of these negative features, for instance exhibiting surpluses in the current account due to favorable commodity prices and terms of trade—​but only for a couple of years. As discussed in the previous section, factors such as interest rates and terms of trade associated with commodity booms can lead to excessive appreciation of the local currency during growth expansions. This exacerbates exchange rate volatility and also restricts the long-​term development of exports, particularly of a competitive manufacturing sector. Therefore, high volatility can induce considerable changes in the economic structure that might further magnify volatility pressures. The case of Brazil is illustrative of these mechanisms, to the extent that the economy has relied not only on low investments but also and increasingly on investments in more oligopolized or monopolized businesses, and in commodities and semi-​manufactured sectors that are excessively volatile and highly subject to public and even private interventions. Alberola and Benigno (2017) discuss how financial openness interacts with a commodity-​exporting economy, and show mechanisms by which the commodity boom crowds out more dynamic tradable sectors and leads to a growth trap, delaying convergence with the world technology frontier, and therefore suggest the adoption of capital account management policies. Empirical evidence shows that countries that, like Brazil, are dependent on commodities exports have a slower long-​term economic growth rate compared with countries with more diversified exports.17 Export diversification can lead to more persistent trade gains, while also minimizing adverse impacts arising from the volatility of commodity prices and exports (Cavalcanti, Mohaddes, and Raissi 2014; Loayza, Servén, and Ventura 2007).18 As shown by Lederman and Maloney (2007, 2008), it is not the exporting of primary goods itself, but rather the low degree of export diversification that is associated with a low growth rate. Arbache and Page (2007) stress that the countries that are most dependent on commodities exports, despite being able to accelerate growth for a while, eventually experience strong collapses, meaning that in the long run average growth is low. They also show that terms of trade are among the main causes of these acceleration and collapse episodes. Cardoso and Teles (2010) document how the fluctuations of Brazilian

168    Jorge Arbache and Sarquis J. B. Sarquis products around the potential GDP growth between 1900 and 2008 are strongly associated with shocks in the terms of trade. Commodity exports to some extent serve to alleviate the external financing constraint by improving the trade balance and moderating current account deficits. However, it must be noted that besides adding additional instability due to its excessively volatile nature in international markets in terms of both quantity (demand) and prices, the recent commodity expansion in Brazil has been accompanied by premature deindustrialization and “servicification.” The latter is expressed by the fall of the share of manufacturing in GDP from around 34% in the mid-​1980s to 10% in 2015, and by the very high and increasing share of services in GDP, around 74%. These processes have geared the economy toward a concentration in mostly low-​tech, low-​productivity, final consumption services, and suggest that Brazil is undertaking a kind of “reverse structural change” (Arbache 2016). This helps to explain the stagnated productivity, the relatively low and decreasing industrial density (Arbache 2012, 2016), the low indices of intra-​industry trade with some of Brazil’s most important partners (Sarquis 2011), and the poor participation of Brazil in global value chains (UNCTAD 2013). In many ways, industrialization, trade openness, and trade composition can have important implications for moderating Brazil’s volatility and enhancing its long-​term growth strategy. Sarquis (2011) presents evidence that, in the case of Brazil, the expansion of manufacturing exports, especially interacting exports and imports of machines and equipment, have been more significant in inducing economic growth than increases in commodity exports. Manufacturing export growth contributes more directly to capital accumulation and to other growth factors, such as technological adaptation and diffusion. It also tends to be more persistent and less volatile.

8.9. Conclusions This chapter shows that growth is extremely volatile in Brazil, and that this growth volatility has had considerable impacts on the country’s growth potential and performance over the last century. High volatility above and beyond ordinary business cycles constrains long-​term growth, with particularly adverse and asymmetric cumulative effects of growth acceleration and collapse episodes on GDP. Growth volatility and collapses are both harmful to long-​term growth prospects, likely shifting potential growth downward on a persistent basis. They help explain the underperformance of Brazil’s growth and its poor convergence with advanced economies. Such an explanation could be extended to other developing and emerging economies that are still viewed as being in the middle-​income trap. In view of these considerations, the main economic challenge facing Brazil is not attaining high rates of growth, but rather growing in a more stable and sustained manner. All else being equal, merely attaining high rates of growth can be shortsighted

Growth Volatility and Economic Growth in Brazil    169 and misleading, while also inducing economic distortions, unbalanced responses, and, ultimately, new sources of volatility. Economic policies should be geared toward breaking the high volatility/​low growth vicious circle. Preemptively reducing the risks of growth collapses, moderating contractions, and strengthening persistency are key policy objectives that can per se contribute to enhancing growth in the long-​term. Such objectives should be part of a policy framework aiming to render the economy more dynamic on a sound, balanced, and sustained basis. This framework, designed to address the main sources of growth volatility and to promote growth persistence, should contemplate and reconcile both macroeconomic and structural policies, addressing the supply and demand sides of the economy. The way forward requires the establishment of sound fiscal policies at all levels of government. Fiscal discipline should also be made as consistent as possible with a strategy of public investment for development and with the ability to respond responsibly to adverse shocks to aggregate demand. Monetary and financial policies are key to assuring macroeconomic stability, while also determining increases in the propensity to save and to deepen capital and credit markets, in particular for pension funds and long-​term investment financing. At the same time, they should better hedge against risks, including of sudden outflows of capital, and improve the competitiveness and inclusiveness of the financial system, with cheaper and more accessible financial services. Market reforms and trade and investment policies should seek to strengthen and further integrate the Brazilian economy into the world economy, while aiming to diversify the economic structure, domestic markets, and international linkages. Internationalization of Brazilian companies, participation in global value chains, and the attraction of efficiency-​seeking foreign direct investments should be coupled with the development of activities of a higher value-​added and more knowledge-​ dependent nature. In many ways, this agenda can be pushed forward in areas in which Brazil benefits from static and dynamic comparative advantages, such as agribusiness, forests, biodiversity, renewable energy, health, aerospace, mining, and ultra-​deep oil exploration.

Notes 1. This chapter was especially prepared for The Oxford Handbook of the Brazilian Economy. We would like to thank, without implicating, Fabiano Bastos, Renato Baumann, Regis Bonelli, and Martin Raiser for their comments on earlier versions of this chapter. The views expressed herein are solely those of the authors and do not necessarily reflect the views of the institutions with which they are affiliated with. 2. See, for instance, Adrogué et al. (2006) and Bacha and Bonelli (2004, 2016). 3. Both GDP series are measured in constant prices. 4. Among G20 countries, India and Indonesia are two other significant cases, having exhibited over 1961 to 2015 average rates of growth lower than China’s but significantly higher than Brazil’s and the G20’s average, as shown in Table 8.2 in Section 8.4.

170    Jorge Arbache and Sarquis J. B. Sarquis 5. If, instead of constant prices, GDP per capita is measured in PPP, some international comparisons in this section would change. However, we stick to constant series not only for the sake of consistency with the subsequent sections—​where the same series are used—​but also because our main interest is the analysis of real GDP per capita growth and its cumulative effect on GDP per capita levels. 6. Of course, the relationship between growth volatility and long-​term growth is to be expected to vary from country to country according to their institutional and economic development specificities and to their exposure to changes in global economic conditions. Output fluctuations and therefore growth volatility relate to shocks as much as to the manner in which each economy copes with those shocks. They are determined by the extent to which the individually rational actions of firms and households, and the policy interventions of governments, add up to collective behavior that either brings the economy closer to full employment and efficient resource utilization or does not. 7. The G20 average statistics are substantially affected by China and Russia, which present respectively high growth and low volatility and low growth and high volatility, relative to the other countries. The correlation coefficient between the statistics in columns 1 and 2 returns positive but not significantly different from 0. Once we remove Russia, it increases. Alternatively, once we remove China, it becomes negative. 8. Along with Latin America, this feature has been particularly characteristic of African countries, as discussed by Arbache and Page (2007, 2015) and Arbache et al. (2010). 9. A feature that is due to growth rates being normally distributed, especially for longer series, whose trend can be accounted for. 10. Growth rate volatility is hardly a significant predictor of growth in a single equation framework, even after including the usual control variables. That is due to the potential endogeneities that characterize the relationship between these variables, as well as to nonlinearities. As both good and bad times push in opposite directions, estimating growth in a standard single equation framework can lead to misleading conclusions. The empirical evidence suggesting no link between output volatility and growth (Dawson and Stephenson 1997; Posch and Walde 2011; Solow 1997) is at least partly explained by this factor. 11. As in Arbache and Page (2007), there are two merits in this methodology. First, it identifies both growth accelerations and decelerations, thus allowing for a much wider investigation of growth volatility. Second, it endogenizes the country’s economic conditions into the method, such that there is no need to impose any parameter to identify growth acceleration and deceleration. In other words, it does not use an exogenous threshold rate to identify growth episodes. Instead, it defines acceleration and deceleration relative to the country’s long-​run economic performance. 12. Despite the growth rates for 1975 and 1976 being higher than the historical average, we do not include them due to condition (b), which holds up to 1971. 13. Both the 1961–​2015 and 1901–​2015 series reveal the same episodes, namely the same reference years, since 1961. These results are also robust for the use of a linear trend, instead of a historical average, and for the change in the window of the moving average as defined in condition (b). We also tested the window for three-​year and five-​year, in place of four-​year, moving averages. 14. Arbache and Page (2007), Conceição and Kim (2014), and Bedir (2015) show that this asymmetric relationship is a pervasive, cross-​country phenomenon, valid for both developing and developed countries. The higher the growth volatility, though, the higher the

Growth Volatility and Economic Growth in Brazil    171 asymmetric response of indicators during growth episodes. As such, the asymmetric relationship is especially relevant for developing countries. 15. Jetter (2013) finds that after addressing the underlying endogeneity of government size in a simultaneous estimation framework, both the positive and the negative effects receive strong statistical support. 16. From a political economy perspective, fiscal issues in Brazil are even more complex, as imbalances arise at three different levels (federal, state, and municipal). 17. Empirical evidence has suggested that the production and export of commodities is not a problem in itself, inasmuch as some countries have become rich through exports of basic goods. Representative examples include Canada, Australia, and Norway. However, a comparison between the Brazil of today and the experiences of these countries has to be treated with caution. First, Brazil is seemingly going in this direction via deindustrialization, or at least after a significant industrialization process. Second, there are important differences in the timing of Brazil’s experience, the size of its population, and the institutions and policies that are in place to cope with fluctuations in the terms of trade. 18. There is also evidence that misaligned and overvalued foreign exchange rates also have negative effects on economic growth (Berg and Miao 2010).

References Acemoglu, D., S. Johnson, and J. A. Robinson. 2005. “Institutions as a Fundamental Cause of Long-​Run Growth.” In Handbook of Economic Growth, edited by P. Aghion and S. N. Durlauf, 386–​472. Philadelphia: Elsevier. Adrogué, R., M. Cerisola, and G. Gelos. 2006. “Brazil’s Long-​Term Growth Performance—​ Trying to Explain the Puzzle.” IMF Working Paper No. 06/​282. Available at http://​www.imf. org/​external/​pubs/​ft/​wp/​2006/​wp06282.pdf. Afonso, A., and D. Furceri. 2010. “Government Size, Composition, Volatility and Economic Growth.” European Journal of Political Economy 26: 517–​532. Afonso, J. R., and K. P. Castro. 2016. “Carga tributária no Brasil: Redimensionada e repensada.” Revista de Admnistración Tributaria 40. CIAT/​AEAT/​IEF, 1–​16. Agenor, P. R., C. J. McDermott, and E. S. Prasad. 1999. “Macroeconomic Fluctuations in Developing Countries:  Some Stylized Facts.” IMF Working Paper No. 9935. Available at https://​www.imf.org/​external/​pubs/​ft/​wp/​1999/​wp9935.pdf. Aghion, P., and A. V. Banerjee. 2005. Volatility and Growth. Oxford: Oxford University Press. Aghion, P., G. M. Angeletos, A. Banerjee, and K. Manova. 2010. “Volatility and Growth: Credit Constraints and the Composition of Investment.” Journal of Monetary Economics 57 (3): 246–​265. Alberola, E., and G. Benigno. 2017. “Revisiting the Commodity Curse: A Financial Perspective.” BIS Working Papers No. 609. Arbache, J. 2004. “Do Structural Reforms Always Succeed?” WIDER Research Paper No. 2004/​58. Available at https://​core.ac.uk/​download/​pdf/​6379841.pdf. Arbache, J. 2012. “Is Brazilian Manufacturing Losing Its Drive?” Department of Economics, University of Brasília. Available at http://​papers.ssrn.com/​sol3/​papers.cfm?abstract_​ id=2150684. Arbache, J. 2016. “The Contribution of Services to Manufacturing Competitiveness in Brazil.” In Innovation and Internationalization of Latin American Services, edited by R. A.

172    Jorge Arbache and Sarquis J. B. Sarquis Hernandez, A. Hualde, N. Mulder, and P. Sauvé. Santiago, 65–​98: ECLAC and El Colegio de la Frontera Norte. Arbache, J., D. Go, and J. Page. 2010. “Is Africa at a Turning Point?” In Is Africa at a Turning Point? Growth, Aid and External Shocks, edited by D. Go and J. Page. Washington, DC: World Bank, 13–​86. Arbache, J. and J. Page. 2007. “More Growth or Fewer Collapses: A New Look at Long Run Growth in Sub-​Saharan Africa.” World Bank Policy Research Working Paper Series No. 4384. Available at https://​ssrn.com/​abstract=1029499/​ Arbache, J., and J. Page. 2015. “Hunting for Leopards: Long-​Run Country Income Dynamics in Africa.” In Growth and Institutions in African Development, edited by A. K. Fosu, 159–​196. London: Routledge. Arias, M. A., and Y. Wen. 2015. “Trapped: Few Developing Countries Can Climb the Economic Ladder or Stay There.” The Regional Economist (October 2015). Federal Reserve Bank of St. Louis. 5–​9. Bacha, E. L., and R. Bonelli. 2004. “Accounting for Brazil’s Growth Experience—​1940–​2002.” IPEA Texto para Discussão No. 1018. Available at http://​www.ipea.gov.br/​portal/​index. php?option=com_​content&view=article&id=4598 Bacha, E., and R. Bonelli. 2016. “Accounting for the Rise and Fall of Brazil’s Growth after World War II.” In Macroeconomics and Development: Roberto Frenkel and the Economics of Latin America, edited by M. Damill, M. Rapetti, and G. Rozenwurcel, 188–210. New York: Columbia University Press. Baer, W. 2010. Economia brasileira. São Paulo: Editora Nobel. Barro, R. J. 2001. “Human Capital and Growth.” American Economic Review 91 (2): 12–​17. Barro, R. J. 2015. “Convergence and Modernization.” Economic Journal 125 (585): 911–​942. Barro, R. J., and J. W. Lee. 1994. “Sources of Economic Growth.” Carnegie-​Rochester Conference Series on Public Policy 40: 1–​46. North-​Holland. Barro, R. J., and X. Sala-​i-​Martin. 2003. Economic Growth. Cambridge, MA: MIT Press. Bedir, S. 2015. “The Asymmetric Impact of Growth Fluctuation on Human Development.” Eurasian Journal of Social Sciences 3 (2): 24–​34. Berg, A., and Y. Miao. 2010. “The Real Exchange Rate and Growth Revisited: The Washington Consensus Strikes Back?” IMF Working Paper No. 10/​58. Available at https://​www.imf.org/​ external/​pubs/​cat/​longres.aspx?sk=23684.0. Blanchard, O. 2005. “Fiscal Dominance and Inflation Targeting:  Lessons from Brazil.” In Inflation Targeting, Debt, and the Brazilian Experience, 1999 to 2003, edited by F. Giavezzi, I. Goldfajn, and S. Herrera, 49–​80. Cambridge, MA: MIT Press. Caballero, R. J., and A. Krishnamurthy. 2001. “International and Domestic Collateral Constraints in a Model of Emerging Market Crises.” Journal of Monetary Economics 48 (3): 513–​548. Cardoso, E., and V. Teles. 2010. “A Brief History of Brazil’s Growth.” FGV Discussion Paper No. 241. Available at http://​bibliotecadigital.fgv.br/​dspace/​handle/​10438/​4146. Carvalho, V., and X. Gabaix. 2013. “The Great Diversification and Its Undoing.” American Economic Review 103 (5): 1697–​1727. Cavalcanti, T., K. Mohaddes, and M. Raissi. 2014. “Commodity Prices Volatility and the Sources of Growth.” Journal of Applied Econometrics 30 (6): 857–​873. Conceição, P., and N. Kim. 2014. “The Asymmetric Impact of Growth Fluctuation on Human Development:  Evidence from Correlates of Growth Decelerations and Accelerations.” Journal of Development Areas 48 (3): 31–​45.

Growth Volatility and Economic Growth in Brazil    173 Dabusinskas, A., D. Kulikov, and M. Randveer. 2012. “The Impact of Volatility on Economic Growth.” Eesti Pank WPS 7/​2012. Available at www.bankofestonia.ee/​pub/​en/​dokumendid/​ publikatsioonid/​seeriad/​uuringud/​. Dawson, J. W., and E. F. Stephenson. 1997. “The Link between Volatility and Growth: Evidence from the States.” Economics Letters 55 (3): 365–​369. Dopke, J. 2004. “How Robust Is the Empirical Link between Business-​Cycle Volatility and Long-​Run Growth in OECD Countries?” International Review of Applied Economics 18 (1): 103–​121. Easterly, W., R. Islam, and J. Stiglitz. 2000. “Explaining Growth Volatility.” Washington, DC:  World Bank. Available at http://​documents.worldbank.org/​curated/​en/​ 447071468761424497/​Explaining-​growth-​volatility. Ferreira, P., and J. Rossi. 2003. “New Evidence from Brazil on Trade Liberalization and Productivity Growth.” International Economic Review 44 (4): 1383–​1407. Growth Commission—​ Commission on Growth and Development. 2008. The Growth Report:  Strategies for Sustained Growth and Inclusive Development. Washington, DC: World Bank. Hausmann, R., L. Pritchett, and D. Rodrik. 2005. “Growth Accelerations.” Journal of Economic Growth 10 (4): 303–​329. Im, F. G., and D. Rosenblatt. 2013. “Middle-​ Income Traps:  A Conceptual and Empirical Survey.” World Bank Policy Research Working Paper No. WPS 6594. Available at http://​documents.worldbank.org/​curated/​en/​969991468339571076/​ Middle-​income-​traps-​a-​conceptual-​and-​empirical-​survey. Jetter, M. 2013. “Volatility and Growth: Governments Are Key.” IZA DP No. 7826. Available at http://​ftp.iza.org/​dp7826.pdf. Jetter, M., A. Nikolsko-​Rzhevskyy, and W. T. Smith. 2013. “The Effects of Wage Volatility on Growth.” Journal of Macroeconomics 37: 93–​109. Johansson, A., Y. Guillemetti, F. Murtin, D. Turney, C. de la Maisonneuve, P. Bagnoli, G. Bousquet, and F. Spinelli. 2013. “Long-​Term Growth Scenarios.” Economics Department Working Papers, No. 1000. OECD. Johnson, S., J. D. Ostry, and A. Subramanian. 2007. “The Prospects for Sustained Growth in Africa: Benchmarking the Constraints.” NBER Working Paper No. 13120. Available at http://​ www.nber.org/​papers/​w13120. Lederman, D., and W. Maloney, eds. 2007. Natural Resources:  Neither Curse nor Destiny. Washington, DC: World Bank and Stanford University Press. Lederman, D., and W. Maloney. 2008. “In Search of the Missing Resource Curse.” Economia: Journal of the Latin American and Caribbean Economic Association 9 (1): 1–​39. Loayza, N. V., R. Rancière, L. Servén, and J. Ventura. 2007. “Macroeconomic Volatility and Welfare in Developing Countries: An Introduction.” The World Bank Economic Review 21 (3): 343–​357. Lucas, R. E., Jr. 1990. “Why Doesn’t Capital Flow from Rich to Poor Countries?” American Economic Review 80 (2): 92–​96. Milanovic, B. 2016. Global Inequality: A New Approach for the Age of Globalization. Cambridge, MA: Harvard University Press. Posch, O., and K. Walde. 2011. “On the Link between Volatility and Growth.” Journal of Economic Growth 16 (4): 285–​308. Pritchett, L. 2000. “Understanding Patterns of Economic Growth:  Searching for Hills and Plateaus, Mountains and Plains.” World Bank Economic Review 14 (2): 221–​250.

174    Jorge Arbache and Sarquis J. B. Sarquis Pritchett, L., and L. H. Summers. 2014. “Asiaphoria Meets Regression to the Mean.” NBER Working Paper No. 20573. Ramey, G., and V. A. Ramey. 1995. “Cross-​Country Evidence on the Link between Volatility and Growth.” American Economic Review 85 (5): 1138–​1151. Rodrik, D. 1998. “Why Do More Open Economies Have Bigger Governments?” Journal of Political Economy 106 (5): 997–​1032. Sarquis, S. J. B. 2009. “Business Cycles in a Credit Constrained Small Open Economy.” PhD dissertation, London School of Economics and Political Science. Sarquis, S. J.  B. 2011. Comércio internacional e crescimento econômico no Brasil. Brasília: Fundação Alexandre de Gusmão. Solow, R. M. 1956. “A Contribution to the Theory of Economic Growth.” Quarterly Journal of Economics 70 (1): 65–​94. Solow, R. M. 1997. “Is There a Core of Usable Macroeconomics We Should All Believe In?” American Economic Review 87 (2): 230–​232. UNCTAD. 2013. Global Value Chains and Development: Investment and Value Added Trade in the Global Economy. Geneva: UNCTAD.

Pa rt  I I

M AC ROE C ON OM IC P OL IC Y A N D I N ST I T U T ION S

Chapter 9

The Brazi l ia n Devel opm e nt  Ba nk Luiz Ricardo Cavalcante

9.1. Introduction Created in the early 1950s, the Brazilian Development Bank (BNDES)1 has played a major role in the national economy ever since. The Bank has historically been the main provider of long-​term credit to both productive and infrastructure investments in Brazil (Baer 2002, 293; Bonelli and Pinheiro 1994, 26).2,3 The role of BNDES and of its subsidiaries (Finame, which focuses on capital goods, and BNDESPAR, which carries out capitalization operations) in the Brazilian economy become clearer as one examines the capital and credit markets in the country, which have traditionally been marked by restrictions. As a result, public banks financed relevant investments, and the industrialization process in the country was strongly associated with the role played by BNDES. The bank has been one of the most important—​if not the most important—​instruments of the industrial policies adopted in Brazil since the mid-​twentieth century. In fact, BNDES played a central role during the import-​substitution industrialization (ISI) process described in Chapter 5 of this Handbook, and it is hard to point out a major investment in Brazil that was not financed by BNDES. In addition, the bank played a major role in financing infrastructure projects, and coordinated several government programs over time, as in the case of the privatization process that took place in the 1990s. In 2015, the total assets of the bank reached BRL 930.6 billion (US$238.3 billion)4 and, according to estimates of BNDES, its disbursements represented 15% of the gross fixed capital formation in Brazil in 2014 (Coutinho 2015). According to Ferraz et al. (2014, 302), the bank holds two-​thirds of the credit with more than five years’ maturity in the country. Nevertheless, most analyses—​which tend to highlight the benefits of having a bank aimed at reducing credit market failures—​do not mention the costs associated with BNDES. However, it is readily apparent that any public bank that provides subsidized credit creates a fiscal cost for the government. Given the gap between the interest rates

178   Luiz Ricardo Cavalcante of the bank and the cost of capital for the Brazilian government, these costs may be very high. According to estimates of the Brazilian National Treasury, the subsidies alone provided by that institution to BNDES reached BRL 28.5 billion in 2015.5 As shown further in this chapter, the critiques of BNDES are not limited to the fiscal costs, as some authors argue that this kind of bank may have crowding-​out effects, as well as negative impacts, on the allocation of resources or on inflation control, for example. Adjudicating the debate about the costs and benefits of BNDES is not easy. Of course, the provision of subsidized credits generates new investments and creates jobs. These investments may also create positive (but hard to estimate) externalities. However, the resources to supply these credits must be extracted from somewhere, and a bank such as BNDES creates fiscal costs for the Brazilian government. Although estimating these fiscal costs is a relatively simple task, some other negative impacts mentioned in the previous paragraph are not easily measurable. The aim of this chapter is to assess the role played by BNDES based upon a survey of its costs and benefits as reported in the literature. We believe that no single formal model can provide a conclusive answer to the question since it involves different views on the role of the state in the economy. In fact, formal models for evaluating this kind of initiative strongly rely on the assumptions of the model and focus on specific questions and time spans (and not on a broad view of the bank in the long run). As such, the same data may lead to opposing conclusions under different assumptions. However, a survey of the results of formal models, combined with an analysis of the qualitative results reported in the recently published literature, may help to understand the impacts of BNDES on the Brazilian economy. In section 9.2, we provide some theoretical background to the creation and the existence of development banks. We use this background to support a brief discussion about the long-​term context that marked the bank’s evolution. Based upon the theoretical background and on the historical perspective, we discuss, in section 9.4, the contemporary issues concerning the role of BNDES in the Brazilian economy, stressing its costs and benefits. Finally, we present the concluding remarks in section 9.5.

9.2.  Theoretical Background on Development Banks The expression “development bank” is, unsurprisingly, associated with financial institutions that somehow aim to provide funding to economic development. This broad definition encompasses institutions as diverse as multilateral development banks, which focus mainly on credits to governments, such as the International Bank for Reconstruction and Development (IBRD) and the Inter-​American Development Bank (IADB); alternative financial institutions, which focus on microcredit, for example; and national development banks. The last category is formed by state-​owned (or at least government-​sponsored) financial institutions that aim to provide funding for industrialization and infrastructure (see Aghion 1999, 83).6

The Brazilian Development Bank    179 The rationale for national development banks has been the subject of fierce arguments in the economic literature concerning the association between financial intermediation activities and economic development. Far from being a merely academic debate, the disagreements about the role of the state in the financial system may lead to different and frequently antagonistic policies. In this section, we briefly discuss the different views on the association between financial intermediation activities and economic development in order to provide some theoretical background for the discussion of the cost and benefits of establishing national development banks. This background will be useful to discuss BNDES in historical perspective (section 9.3) and the contemporary issues concerning the role of BNDES in the Brazilian economy (section 9.4). We consider three main views concerning the association between financial intermediation activities and economic development: the neoclassical approach, the Keynesian view (both new-​Keynesian and post-​Keynesian) and the historical perspective. Each of these approaches is discussed in sequence. The neoclassical approach is based on empirical evidence of the association between the development of the financial system and the growth of per capita gross domestic product (GDP). In the late 1960s, Goldsmith (1969) collected data about these variables and argued that banks and financial intermediaries provided a more efficient allocation of factors of production. In the early 1970s, McKinnon (1973) and Shaw (1973), whose work laid the basis for the financial liberalization school, alleged that any measure to promote economic activities through loans at artificially low costs—​a phenomenon they label “financial repression”—​would result in the de-​stimulation of savings and thus investment. They advocated financial liberalization, which would lead to higher levels of savings, investments, and economic growth. Thus, this approach essentially rejects the creation of development banks. A series of econometric studies were published in the 1990s in order to investigate the causal relations between financial development and economic growth. Using cross-​ country regressions, Levine (1997, 707) argues that “the initial level of financial development is a good predictor of subsequent rates of economic growth, physical capital accumulation, and economic efficiency improvements over the next 30 years even after controlling for income, education, political stability, and measures of monetary, trade and fiscal policy.” However, no clear policy implications about development banks emerge from this kind of analysis. The Keynesian view can be divided into new-​and post-​Keynesian approaches. The new-​Keynesians point to information asymmetries among economic agents involved in credit operations. Stiglitz and Weiss (1981), for example, argue that information is asymmetrically distributed among economic agents, while expected returns of banks in their credit operations are not a rising function of interest rates, due to the adverse selection phenomenon. This occurs when interest rates are very high and many credit-​ worthy borrowers leave the market. This results in a concentration of loans to projects of greater risk, raising the probability of default, thus diminishing the expected returns to the bank. As a result, in certain circumstances, even though there are borrowers who are willing to pay higher interest rates, most financial institutions will not offer

180   Luiz Ricardo Cavalcante credit facilities due to the default risk, which leads to the phenomenon known as credit rationing. It is in accordance with this interpretation that Stiglitz (1994) justifies the direct intervention of the state in the financial system. The proposed interventions are not only limited to so-​called prudential regulation, but also include direct credit, financial repression, and the regulation of competition among banks (Stiglitz 1994, 42). In addition, the proposal includes public credit institutions that measure the social return of projects and thus can finance projects that the private sector, taking into account only private returns, would avoid. Post-​Keynesian authors question the prominence of savings over investments, and therefore the emphasis on financial liberalization. In fact, post-​Keynesians look upon state intervention in the financial system as a way to guarantee low interest rates, which are crucial for the industrialization process of developing countries. Arestis (1997, 152)  and Studart (1995) argued that the state should not only regulate the financial system to minimize credit rationing, but also provide credit to sectors lacking adequate credit (especially for long-​term projects) and publicize sectors with good perspectives to other financial intermediaries and savers. Finally, located within the historical approach, authors such as Gerschenkron (1962) and Cameron (1967, 1972) argued that late-​industrializing countries needed long-​term credits from universal banks and from governments to finance the development process. Gerschenkron (1962, 14), in particular, points out that the gradual characteristic of the industrialization process in England did not require the development of specific institutions to provide long-​term capital, as it underwent a “natural” process of accumulation. According to Gerschenkron (1962, 14), The industrialization of England proceeded without any substantial utilization of banking for long-​term investment purposes. The more gradual character of the industrialization process and the more considerable accumulation of capital, first from earnings in trade and modernized agriculture and later from industry itself, obviated the pressure for developing any special institutional devices for provision of long-​ term capital to industry.

On the other hand, late-​industrializing countries required specific instruments to finance their development because (1)  capital was scarce and diffused; (2)  there was general distrust of industrial activities; (3) larger industrial enterprises were required to catch up; (4) industrialization took place in capital-​intensive sectors; (5) there was a dearth of entrepreneurial talent. Although these analyses were based on the industrialization experiences of countries in continental Europe in the nineteenth century, they may be also applied to the ISI model of Brazil and provide the grounds for the creation of an institution like BNDES. Zysman (1983) expanded the dichotomy involving capital-​market and credit-​based financial systems by analyzing and comparing the patterns of industrial development in five advanced nations (France, Great Britain, Japan, West Germany, and the United States) and proposed three distinct types of financial systems:  (1) capital-​market; (2) public credit-​based; and (3) private credit-​based financial systems.

The Brazilian Development Bank    181 The neoclassical critique of the Keynesian and historical views has typically been that state intervention would lead to distortions in the allocation of resources, as well as corruption. Such criticism was especially directed to the new-​Keynesian school by authors such as Jaramillo-​Vallejo (1994, 54), who claimed that the type of intervention advocated by Stiglitz (1994) “has led everywhere to a burst of corruption and other undesirable effects.” He also claimed that “government failures” could be bigger than the market failures that had motivated intervention in the first place. Although the post-​ Keynesian and historical visions did not use the concept of market failure, the criticism about the intervention of the state in the financial system can clearly be extended to the solutions offered by these schools. Similarly, Lazzarini et al. (2015, 238) contrast a positive view about development banks (which they label the “industrial policy view”) with a negative view labeled the “political view” which “sees lending by development banks causing multiple sources of credit misallocation.” Despite these criticisms, development banks have actively participated in the credit market in various developed and developing countries. A benchmark study prepared by the Business Development Bank of Canada (BDC) refers to a database that includes 235 development institutions from 92 countries (BDC 2009). Musacchio and Lazzarini (2015, 274) register the existence of 288 development banks globally in 2011. The basic argument that sustains the creation of such institutions is that they contribute to reducing the rationing of long-​term credit. In the circumstances where there are subsidies involved, these banks may create incentives for projects that present positive externalities. Besides, development banks have frequently acted as coordinators of the industrialization processes in developing countries by influencing the sectoral and regional allocation of capital. Another possible reason why national development banks are frequently well accepted is that while their benefits are relatively concentrated, their cost is diffuse. As a result, lobbies against these kinds of institutions tend to be weaker than the lobbies in favor of them.

9.3.  BNDES in Historical Perspective: A Brief Review of the Long-​Term Context Following a suggestion of the Brazil–​United States Joint Commission (CMBEU, in the Portuguese acronym), BNDES was founded in 1952 “as a government agency with administrative autonomy and registered as a company, initially under the jurisdiction of the Finance Ministry” (BNDES 2012, 21).7 Typologies such as those proposed by Gerschenkron (1962) and Zysman (1983) are helpful in understanding the reasons that led Brazil to create the BNDES in the early 1950s; for instance, Baer (2002, 293) argues that Gerschenkron’s (1962) arguments are highly applicable to the Brazilian case of the 1950s and 1960s. Brazil had no well-​developed capital market at the time, which made it

182   Luiz Ricardo Cavalcante 250.00

Picking winners and investment maintenance

200.00

150.00

100.00

50.00

Early ISI period. Relatively low disbursements. Focus on infrasrtucture. Steel industry and power generation.

Financer of the private sector. Increasing disbursements. II PND: petrochemicals and capital goods

Crisis of the 1980s (and early 1990s): resuce operations. No clear sectoral priorities.

Privatization (telecomunications and energy)

52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 12 14 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20

BRL billion

US$ billion

Figure 9.1.  BNDES disbursements, 1952–​2015 (2015 BRL billion and USD billion). Source: Elaborated by the author from data published by BNDES (http://​www.bndes.gov.br).

difficult for both the private domestic sector and state enterprises to finance long-​term investments.8 During its existence of more than 60  years, BNDES has faced starkly different environments. Significant structural change has taken place in the Brazilian economy since the early 1950s, and the Brazilian financial system, as well as its regulation patterns, has undergone profound change. During this period, the bank “consistently responded to government policy preferences” (Doctor 2015, 197). In this section, we briefly review the different phases of BNDES and the role played by the bank in the Brazilian economy.9 In order to indicate the significance of the bank during each phase, Figure 9.1 shows the evolution of BNDES disbursements between 1952 and 2014 in constant 2015 BRL billion and in US$ billion. Different periods in BNDES’s role are also schematically indicated on the graph; the following sections briefly describe these periods in further detail.

9.3.1. BNDES in the Early Import-​Substitution Industrialization Period In the early years of its existence, BNDES played an important role as an instrument to support public investments in infrastructure and basic industries. However, according to Baer and Villela (1980, 425), “the scarcity of funds kept the BNDE’s share of gross

The Brazilian Development Bank    183 capital formation below 3% prior to 1956, and only the projects having top priority were funded.” During the four first years of operation, the bank focused mainly on the transport sector, while in the late 1950s and early 1960s most resources were directed to the power and industrial sectors. As the bank’s initial financing activities were small, its staff spent a large amount of time researching macro and micro aspects of Brazil´s economy. That made BNDES “an influential institution in the planning of future governments” (Baer and Villela 1980, 426). In fact, even today BNDES “provides technical services and policy advice to both government and enterprises to enhance the economy’s development outcomes” (Doctor 2015, 210). It might seem odd, at first glance, that a public bank was used to finance the public sector itself. However, as pointed out by Cavalcante and Uderman (2011), due to its “institutional nature” (a state bank endowed with more autonomy and readiness than the government itself), BNDES not only coordinated actions, but also made possible investments in infrastructure and the influx of international capital directed to these projects. This proposition is consistent with a point of view stressed even in the late 1960s, when Pinto (1969, 34, quoted in Baer and Villela 1980, 425)  argued that “the BNDE would enjoy a substantial amount of insulation from the inflexible administrative practices, for which the Brazilian Federal bureaucracy is known.” The shift to industry in the late 1950s was also accompanied by the buying of equity in a selected number of firms.

9.3.2. BNDES as Financier of the Private Sector The importance of BNDES grew considerably during the government of President Juscelino Kubitschek (1956–​1960), whose Targets Program (Programa de Metas) was formulated by the bank’s staff. In fact, it became the bank’s task to decide which types of enterprises were eligible for government support. Pinto (1969, 41, quoted in Baer and Villela 1980, 426) noted that at the time the bank was “by its legal nature, the controller of decisions, and not just a manipulator of incentives. This subtle but significant clash between the bank’s roles may have been responsible for a major change in the bank’s loan policies, which later came to favor manufacturing industries to the detriment of infrastructure projects.” Despite the “liberalizing rhetoric” of the banking reform that took place between 1964 and 1965, in practice there was a significant increase in BNDES disbursements. The higher levels of indexation of the economy (the correção monetária, monetary correction) legitimated the mandatory savings used to fund BNDES and created more room for the increase of its disbursements. With the availability of new funds, the bank’s activities began to grow and change, and its share of gross capital formation rose from 2.7% in 1964 to 6.6% in the following year. During this period, the share of the public sector in the disbursements of BNDES (which floated around 90% between 1952 and 1964) started to decline after 1964 and, in the 1970s, was around 20% (Cavalcante 2004).

184   Luiz Ricardo Cavalcante

9.3.3. BNDES during the Crisis of the 1980s The 1980s in Brazil are commonly referred to as a “lost decade,” in which rates of economic growth and investment slowed markedly. However, as shown in Figure 9.1, BNDES disbursements did not reduce significantly until the end of the decade. During that period, BNDES was used by the government as an instrument to bail out firms in several sectors. This involved, occasionally, the take-​over of a number of firms by the state, with the role of the bank being to strengthen their finances and general operations. BNDES also privatized some companies whose control had been taken over by the bank due to previous defaults (Pinheiro and Giambiagi 2000, 20). As Pinheiro and Giambiagi (2000, 21) point out, the experience of BNDES during the 1980s provided the institutional foundations for the privatization process of the 1990s. Thus, the accumulated experience of this period enabled BNDES to coordinate the privatization process that would take place in the Brazilian economy during the 1990s, as explained in the next subsection. Besides these “bail-​out” operations, the bank during the 1980s seemed to have neither a clear plan of expansion, nor sectoral priorities.

9.3.4. BNDES as the Coordinator of the Privatization Program Although some public companies had been sold in the 1980s, it was from the mid-​1990s onward that the Brazilian privatization program expanded rapidly, including “public utilities, such as telecommunications, electricity distribution, railroads, ports and some of the major highways” (Amann and Baer 2005, 424). During this period, BNDES became the major federal institution directly involved in the privatization process and came to be known, according to Francisco Gros, president of the bank from 2000 to 2002, as the “privatization bank” (BNDES 2002). The bank was responsible for the executive coordination of auctions, and a large proportion of BNDES disbursements in that period were linked to the acquisition of state firms by the domestic and foreign private sectors, since multinational companies were allowed from 1995 onward to take loans from BNDES to take part in the privatization process. This included privatization through concession contracts for various types of infrastructure investments (Baer 2008, 231). Most of the loans at the time were not directed to capital formation, and the disbursements to infrastructure, which corresponded to 25.8% of the total in the period between 1990 and 1993, rose to 43.7% in 1998. The bank financed bids by private firms to acquire privatized companies, and the BNDESPAR took part in several acquisitions. According to data published by the bank, its total applications between 1995 and 2002 reached around US$96 billion. This value is similar to the total revenues from the privatization process (US$93 billion) in the same period. Of course, only a share of BNDES’s total applications were used to finance the bids of the privatization process. Even so, its prominent role in funding privatization is undeniable.

The Brazilian Development Bank    185 Again, it may seem odd that a state-​owned bank was used to finance the selling of state-​owned companies; besides, multinational companies frequently had access to low interest rate credit operations abroad. However, the perceived risk of a devaluation of the Brazilian currency at the time created an incentive to avoid credit operations in foreign currencies (as the privatized companies would generate revenues in BRL). Thus, the credits provided by BNDES worked as a kind of “exchange rate hedge” to cover the accounting risk assumed by the companies. In fact, writing in the late 1990s, Pinheiro and Giambiagi (1999, 18, quoted in Baer 2002, 308) stressed that as Brazil was still considered too risky for investments and a devaluation of the local currency was thought probable, loans in foreign markets were only a partial solution. As a result, the government intervened and financed the lenders either directly (by allowing deferred payments) or through BNDES. Another reason why BNDES played an important role in the financing of the privatization process is pointed to by Cavalcante and Uderman (2011), discussing the role played by BNDES in funding electricity investments in Brazil. They argue that to cope with the problems associated with the perception of “weak regulation” and the uncertainty regarding energy prices, BNDES was also used by companies that took part in the bids as a “regulation hedge,” that is, as a way of sharing the risk of changes to the rules established by the Brazilian government.

9.3.5. BNDES from 2003 Onward In the first decade of the 2000s, Brazil had increasing access to international capital markets, and the development of its internal capital market enabled many firms to obtain financing from nongovernmental sources. Credit markets also benefited from a set of reforms, which included a new bankruptcy law, the spread of payroll loans, and the spread of fiduciary alienation contracts.10 However, as seen in Figure 9.1, BNDES disbursements increased significantly during this decade. Privatization came to be considered a “neoliberal issue” and there was a revival of the “developmentalist” strategy as well as of explicit industrial policies. Since 2004, a series of three different industrial policies have been put in place in Brazil (Ferraz et al. 2014, 297). Based upon the reported South Korean experience (Amsdem 1989), BNDES tried to promote mergers and acquisitions in order to strengthen Brazilian companies and to transform them into global players. Amsdem (1989) argues that countries like South Korea implemented successful industrial policies because they managed to create large-​scale national leading companies that were able to compete internationally. Government subsidies were used to induce these large firms to move to technology-​intensive activities. As a result, these policies managed to promote a structural change in the local economy. It is quite clear that reports like this inspired BNDES, particularly from 2007 onward, as the bank began to focus on the empowerment of Brazilian companies in order to make them able to compete in global markets.

186   Luiz Ricardo Cavalcante However, Almeida (2009) argues that, in practice, the focus does not seem to have been placed on structural changes, but rather on scale (for which no subsidies would be required, as private investment banks could have financed these operations). The lack of a focus on technology-​intensive sectors was also questioned by Antônio Barros de Castro, who declared that there was “an option for the strengthening of the industry of the past,” and by Carlos Lessa (president of BNDES between 2003 and 2004), who associated this movement with a return of the Brazilian economy to the condition of commodities exporter. Luciano Coutinho (president of the bank between 2007 and 2016), however, argued that sectors where Brazil was more competitive (agroindustry and commodities) had been prioritized.11 The focus was placed on sectors where, according to the official industrial policy adopted in Brazil at the time (the “Productive Development Policy”), Brazil needed to consolidate and expand its leadership.12 After the 2008 financial crisis, BNDES launched the “investment maintenance program” (PSI in the Portuguese acronym) as a countercyclical measure. Essentially the idea was to provide low-​cost credits in order to sustain investment levels in spite of the adverse international scenario. In an interview in 2012, the president of BNDES, Luciano Coutinho, mentioned the need to “counter-​balance the enormous credit crunch” that followed the financial crisis of 2008 (BNDES 2012, 179). As a result, BNDES disbursements rose significantly from 2008 onward, not only to finance long-​term investments, but also to support countercyclical policies (Ferraz 2012). Since additional funds were required (either to increase the disbursements or to cover for interest rate differentials), the National Treasury had to transfer resources to BNDES. That increased its share in the capital structure of the bank from 11.2% in 2005 to 57.2% in mid-​2015.13 Given the interest rate differentials, these transfers created a large cost for the National Treasury (see section 9.4.4 in this chapter). Initially, the program—​as well as the Brazilian economy—​managed to sustain investments. In 2010, for example, GDP growth reached 7.6%, while most countries still faced the effects of the financial crisis. On the other hand, however, the high fiscal costs of the program contributed to the deterioration of public accounts and led to the confidence crisis that became evident in 2015, when GDP decreased by almost 4%.

9.4.  Contemporary Issues Concerning BNDES After more than 60  years of history, and in spite of its significance in the Brazilian economy, BNDES is, more than ever, a controversial topic among economists and policymakers. In this section we present the terms of the debate about the impacts of BNDES on the Brazilian economy, and systematize the main arguments about its role in investments, structural change, firm performance, and fiscal costs. Each of these issues is treated in the following subsections.

The Brazilian Development Bank    187

9.4.1. Investments As discussed in the previous section, the main function of BNDES is to provide credit on more favorable terms in order to facilitate investment. Therefore, several authors highlight the contribution of the bank in increasing capital formation in Brazil. Documents published by BNDES or official presentations by the bank’s staff usually stress the high ratio between BNDES disbursements and gross capital formation in the country.14 This proposition may rely both on deductive arguments and on empirical evidence. The deductive arguments rely on a simple idea: lower interest rates reduce the weighted average capital cost of the projects, which become, as a result, more attractive. Meanwhile, several authors have assembled empirical evidence that firms that would not have invested in the absence of BNDES credits decided to invest following access to its loans. Additionally, based on cross-​national statistical evidence and on a case study of BNDES, Krieckhaus (2002) argues that public savings mobilized through development banks substantially influence economic growth rates. Microeconomic evidence has also been obtained by Ottaviano and Sousa (2014). Using microdata, their analysis concludes that “government support of the type provided by BNDES can indeed help relax credit constraints that prevent constrained firms from performing as otherwise identical unconstrained ones” (2014, 3). Support for exports is also a way of creating incentives to invest. Between 2000 and 2015, the annual average disbursements in constant 2015 BLR were 8.31 billion. Around 30% of this total was directed to “other transport equipment” (which includes aircraft manufacturing).15 Of course, this kind of support helps Embraer—​a local manufacturer of aircraft—​to keep its presence in the international market and to increase its investments in Brazil. However, support for engineering services exports is far from universal. Newspaper articles have criticized the funding of infrastructure investments abroad.16 Data available from the BNDES website indicate that between 2002 and 2015 more than US$14 billion was contracted to finance Brazilian contractors abroad.17 Most credits were directed to Construtora Norberto Odebrecht S.A. (63.50%) and to Construtora Andrade Gutierrez S.A. (20.00%).18 The main target counties were Angola (28.33%), Venezuela (22.94%), Dominican Republic (18.44%), Argentina (14.52%) and Cuba (6.01%). Despite the argument given by BNDES that this helps to export engineering services (Coutinho 2013), there seems to be no conclusive evidence that the externalities from these investments contributing to the Brazilian economy would be higher than those that would be associated with infrastructure in the country itself (see Mendes 2014a). Another critique of the role played by BNDES in relation to investment levels in Brazil has to do with “crowding-​out” effects. As stated by the Organisation for Economic Co-​ operation and Development (OECD 2013, 10), “further development of long-​term credit markets is hampered by a lack of private participation, owing to an uneven playing field caused by strong financial support to the national development bank which dominates long-​term lending.”

188   Luiz Ricardo Cavalcante Finally, Bolle (2015, 1)  argues that to maintain price stability, the Central Bank of Brazil “is forced to raise interest rates more than it might do otherwise in the absence of BNDES lending.” This “second-​order effect” has negative impacts on investment. Similarly, one could argue that resources from the National Treasury and from mandatory savings are ultimately transferred to the firms supported by BNDES. However, on the other hand, these costs create additional pressure on public accounts and, consequently, on interest rates, causing a negative impact on investment. This perception is essentially similar to that stated by the president of the Central Bank of Brazil between 2003 and 2011, Henrique Meirelles. In an interview with the newspaper Valor Econômico in 2010, he argued that “to the extent that there is an interest rate fixed outside the monetary policy, it is plausible to think, in spite of the lack of empirical evidence [. . .], that it may require a higher basic interest rate than otherwise.” In short, Meirelles recognized that BNDES credits “reduce the efficiency of the monetary policy.”19 There seems to be no consensus in the literature about these contradictory effects. Although Bolle (2015) estimated the impacts of BNDES on interest rates, no conclusive analysis of the net effects has yet been published. However, assuming a neutral effect on overall investments, one could argue that the transfers by BNDES may allow large-​ scale investments that would not take place in the absence of support from the government. Some of these investments may have positive externalities, as we discuss in the next subsection.

9.4.2. Structural Change A historical analysis of BNDES shows that the bank, on several occasions, defined “sectoral priorities” for investments or, at least, was itself the instrument for implementing priorities defined by the government. Therefore, especially through support for large-​ scale or technology-​intensive investments that would not otherwise have taken place, BNDES may have contributed to structural changes in the Brazilian economy. These changes resulted either from the creation of new firms or from investments of existing firms. In the case of the second National Development Plan (II PND in the Portuguese acronym), for example, BNDES supported investments in the petrochemical sector made by existing firms that had theretofore focused on traditional sectors (construction, for example). Besides, by allocating resources—​which are ultimately public ones in a given sector—​BNDES signals to private agents the priority sectors for investment. This kind of coordination of expectations may contribute to additional private investments linked to the supported ones in a “cumulative causation” process as described by Myrdal (1957). The debate about the role of development banks in promoting structural changes generates two levels of reactions. Some authors, such as Almeida (2009), do not contest the theoretical potential of development banks to promote structural changes, but argue that the recent “picking winners” strategy adopted by BNDES has reinforced low-​ technology sectors and has not supported a structural change process (in contrast to the

The Brazilian Development Bank    189 South Korean experience). Other authors, however, would question the capacity per se of governments to pick strategic sectors and would leave this decision to the market. In this view, the argument is that public banks would misallocate public resources according to politicians’ economic or political interests. That may, as a consequence, lead to higher levels of corruption.20 Although the issue of structural change is pervasive in the debate about the bank, it might be argued that the use of workers’ mandatory savings to create new jobs—​ regardless of their impact on the sectoral structure of the economy—​should be a priority for BNDES. Besides, it seems that incentives directed to employees (like profit sharing) rely on financial sustainability indicators (which involve administrative expenditures, default rates, and return on assets), which have little to do with the sectors benefited by the credits provided by the bank.

9.4.3. Firm Performance There are several empirical studies that support the claim that access to BNDES loans improves firm performance (as measured by productivity or exports indicators, for example) or firm growth (as measured by the number of employees). The basic argument is that firms that face credit constraints tend to invest less than would be necessary to improve their performance. The evidence of the impacts of the loans are based on propensity score matching methods (or similar) applied to microdata and on comparisons of firms that are as similar as possible, except regarding access to BNDES support. Ottaviano and Sousa (2014) survey these works and note that several papers tackle the issue of whether firm productivity is related to BNDES loans. Some papers focus on the impacts of BNDES loans on labor productivity and on total factor productivity (TFP) (see, for example, Coelho and De Negri, 2010). In particular, Ribeiro and De Negri (2009) focus on loans specifically aimed at the acquisition of domestic capital goods, which are supposed to have a more direct impact on firm productivity. Ottaviano and Sousa (2008) use similar methods and conclude that firms that have access to BNDES loans demonstrate better performance than those that are not supported by the bank. However, they also conclude that the impact depends on firm size; BNDES loans have a positive impact on the performance of larger firms, but a negative impact on the performance of smaller ones. Finally, Ottaviano and Sousa’s (2014, 14) conclusions reinforce the view that BNDES loans should focus on higher technology projects. In spite of some mixed results, the conclusions are, in general, broadly intuitive: if BNDES support is a sort of subsidy (as the credits typically involve lower interest rates than the market interest rates), firms that have access to these loans should present better performance indicators than identical firms with higher capital costs. Yet, none of the papers focuses on the impacts of the subsidies on the results of the firms. This point, however, is addressed by Lazzarini et al., who used econometric techniques on a smaller database (286 publicly listed companies in the São Paulo Stock Exchange) to conclude that “BNDES’s loans and equity allocations do not appear to have a consistent effect on

190   Luiz Ricardo Cavalcante the performance and investment decisions of the firms in our sample, other than a reduction in financial expenses due to the effect of governmental subsidies” (2015, 250; emphasis added). The impacts of BNDES disbursements on the performance of the public sector, meanwhile, have not yet been widely evaluated. However, Bast (2015) attempts to assess the effects of BNDES loans to municipal governments. In particular, the author uses a matching differences-​in-​differences method and concludes that the disbursements to the Modernization Program for Public Finance Management (PMAT in the Portuguese acronym) had no effects on tax revenues of municipalities.

9.4.4. Fiscal Costs Although the impacts of BNDES on investments, on structural changes, and on firm performance are not universally agreed on (as shown in the previous subsections), they are mostly associated with benefits brought about by the bank. However, these benefits are also associated with fiscal costs that allow BNDES to provide credit with more favorable conditions. It is very hard to estimate the overall costs because they rely on assumptions about interest rate differentials. However, recently published figures give a partial idea of the fiscal costs of BNDES. According to a law issued in 2009, the National Treasury has to estimate the fiscal impact of its credits to BNDES taking into account the capital cost for the Union and the remuneration paid by the bank. Essentially, these estimates consider the opportunity cost for the National Treasury of allocation of capital to BNDES at lower interest rates (the so-​called credit subsidies or implicit subsidies), as well as direct costs generated by the interest rate equalizations associated with the PSI (the so-​called financial subsides or explicit subsidies), especially from 2008 onward. Figure 9.2 notes the evolution of these subsidies between 2008 and 2015. As shown in Figure 9.2, the total subsidies of the National Treasury reached BRL 28.5 billion in 2015. This value is essentially the fiscal cost associated with the expansion of BNDES disbursements that followed the financial crisis of 2008. This value is higher than the fiscal cost of the cash-​transfer program adopted in Brazil to reduce poverty (Bolsa Família), which required resources of around BRL 27.1 billion in 2015, according to official data.21 As the costs to the National Treasury are associated with long-​term contracts, even if no additional credits are given, they will keep on existing (though decreasing) up to 2060. However, these costs are only a part of BNDES’s total costs. Any other transfer of the Union to the bank at interest rates below the opportunity cost to the National Treasury may be considered a cost associated with BNDES. Besides, a significant source of funding for the bank is workers’ mandatory savings at low interest rates. Although these costs are not directly associated with the National Treasury, there is, in this case, a clear transfer of resources from workers to the bank. It is very hard to estimate the total costs associated with BNDES given the large variety of different conditions that regulate its contracts.

The Brazilian Development Bank    191 35.00 28.53

R$ bn (i.e Billions of Reais)

30.00 25.00

18.97

20.00 15.00

9.56

10.00 5.00 2008

2009

2010

2011

Financial subsidies

2012

2013

Credit subsidies

2014

2015

Total

Figure 9.2.  Subsidies of the National Treasury to BNDES, 2008–​2015. Source: National Treasury, available at http://​goo.gl/​aXSUgI, accessed April 25, 2016.

However, if one considers that the total assets of the bank reached BRL 930.6 billion in 2015 (according to the bank’s financial statements22), a difference of 1% in the remuneration of the financers (be they the National Treasury or workers), for example, would result in a yearly cost of BRL 9.3 billion. If we assume that the total assets remuneration corresponds, on average, to the long-​term interest rate (TJLP in the Portuguese acronym) and that the gap between this rate and the opportunity cost for the government set by the Special Custodial and Clearing System (SELIC in the Portuguese acronym) was around 7 percentage points in 2015, a total credit (“implicit”) subsidy of around BRL 65.1 billion can be roughly estimated.23 If the financial (“explicit”) subsidies of almost BRL 19.0 billion (Figure 9.1) are considered as well, total costs may reach BRL 84.1 billion. Official BNDES documents contest numbers like these, based upon the argument that they depend on assumptions about the trajectories of the TJLP and of SELIC. If these rates converge, the costs tend to reduce (Coutinho 2015). This is true, but, in this case, the attractiveness of BNDES credits tends to shrink, as well as its capacity to create incentives for new investments. Another argument used by BNDES is that the bank generates profits and, since it belongs exclusively to the Union, these profits return to the Brazilian government anyway.24 In short, given the way that the money transits inside BNDES, the resources transferred by the government to the bank partially return in the form of profits. If the total profits in 2015 (BRL 6.2 billion, according to the bank’s financial statements) are considered, the 84.1 billion estimate of costs reduces to BRL 77.9 billion. If taxes paid by BNDES are also considered, around BRL 3.3 billion is also to be subtracted from the total cost. Yet that still leaves around BRL 74.6 billion.

192   Luiz Ricardo Cavalcante These estimates are approximate, and different assumptions may lead to different figures. However, it is evident that the costs of BNDES is not low. These costs are to be weighed against the potential benefits, mentioned in the previous subsections, of having a national development bank, in order to decide to what extent BNDES may be used to promote the development of the Brazilian economy.

9.5.  Concluding Remarks The main purpose of this chapter has been to analyze different views on the costs and benefits of the role played by BNDES in the Brazilian economy, and to leave it to the reader to adjudicate over the question of the ongoing role of the bank, according to the reader’s own beliefs and specific circumstances. In order to support this discussion, we have provided some theoretical background about the creation and the existence of development banks (section 9.2); a brief review of the history of BNDES throughout its more than 60 years of existence (section 9/​3); and a review of the contemporary issues (investments, structural change, firm performance, fiscal costs) concerning the role of the bank in the Brazilian economy (section 9.4). As stressed in the introduction, the debate about the costs and benefits of BNDES is complicated and hard to judge. On the one hand, a national development bank such as BNDES contributes to increasing capital formation, as it provides credit on more favorable conditions to selected projects. However, it is also argued that the presence of BNDES loans forces the Central Bank to raise interest rates to a level that would otherwise be lower. These opposing views somehow blur BNDES’s net effects on aggregate investment levels. At any rate, the coordination of BNDES can make possible large-​ scale and technology-​intensive investments that would not take place in the absence of subsidies and of a guideline from the government. In fact, the industrialization of “backward” countries frequently involves a kind of net transfer of resources from society to a number of selected sectors. That recalls Gerschenkron’s (1962) argument and appears to fit the Brazilian case in the 1970s, for instance. However, the transfer of resources to projects that are not associated with a “development project,” or that show no positive externalities, seems to exceed the role to be played by a development bank like BNDES. Therefore, the challenge would be to select investments that lead to structural changes or that present positive externalities, as the strengthening of existing activities could be financed directly by the private banking system. Besides the analysis of the impact of the supported projects, the fiscal costs generated by the bank have to be taken into account as well. As shown in section 9.4, an aggregate number is not available. In spite of that, estimates generated to fulfill legal constraints indicate that just the subsidies provided by the National Treasury to BNDES reached BRL 28.5 billion in 2015. An overall number that takes into account interest rate differentials, as well as profits and taxes generated by BNDES, may reach figures as high as, for example, 74.6 billion in 2015. Although these estimates are rough and the figures may vary

The Brazilian Development Bank    193 according to different conditions and assumptions, it is clear that the costs of the BNDES are not low and cannot be disregarded. All of the aspects mentioned in this chapter may vary over time. In fact, not only the costs, but also the benefits associated with BNDES will rely on specific circumstances. Besides, different views on the role of the state in the economy may lead to different conclusions. Nevertheless, a transparent and focused discussion, supported by accurate statistics and well-​built models, can help policymakers to determine the ongoing role to be played by BNDES in the Brazilian economy.

Notes 1. In Portuguese, Banco Nacional de Desenvolvimento Econômico e Social (National Bank for Economic and Social Development). 2. The author would like to thank Carlos Azzoni, David Kupfer, Edmund Amann, Eduardo Viotti, João De Negri, Mansueto Almeida, Marcos Mendes, and Simone Uderman for their comments and suggestions. 3. The author is Legislative Advisor at the Brazilian Federal Senate. 4. Data refer to December 31, 2015. Conversion to USD based upon the exchange rate of 3.9045 BRL/​USD reported by the Brazilian Central Bank for the same date. In the previous year, the total assets of BNDES were BRL 877.2 billion. As the exchange rate was 2.6562 BRL/​USD at that date, the total assets in 2014 were US$330.2 billion. These amounts make BNDES one of the largest development banks in the world, after the China Development Bank, which boasted, in 2014, assets of around RMB 10.32 trillion or US$1.66 trillion (available at http://​goo.gl/​vinYHx; accessed August 4, 2016), and the German KfW, with assets of US$602.1 billion (available at https://​goo.gl/​1UeSxe; accessed April 14, 2016). BNDES is larger than the Japan Finance Corporation, which held total assets of US$167.4 billion in 2014 (available at https://​goo.gl/​aSORvO; accessed April 14, 2016) and much larger than National Financiera (Nafinsa), which held assets of MXN 366.2 billion or US$22.2 billion the same year (available at http://​goo.gl/​2hrMS7; accessed August 4, 2016). 5. Source: Brazilian National Treasury. Available at http://​goo.gl/​f2Pc32; accessed April 14, 2016. 6. In the Brazilian case, besides BNDES, there are also regional development banks aimed at providing credit at more favorable conditions for specific regions of country, such as the Banco do Nordeste and the Banco da Amazônia. These banks are controlled by the federal government. Some states also have their own development banks and development agencies (agências de fomento), and the three states of the southern region of the country have a shared development bank (Banco Regional de Desenvolvimento do Extremo Sul). 7. The original acronym was BNDE (National Bank of Economic Development). In 1982, when it began to manage resources of the Social Investment Fund (Finsocial), it was renamed National Bank of Economic Development (Cavalcante 2004). 8. For details, see Baer (2008, Chapter 10) and Baer and Villela (1980). 9. For brevity, only a few stylized facts about BNDES are presented. Detailed histories of BNDES are available in several of the references cited in this chapter. 10. In this case, the assets financed by the loans are pledged in guarantee and maintained under lien until the final settlement of the agreements. This reduces credit risks and interest rates as a result.

194   Luiz Ricardo Cavalcante 11. All of these statements extracted from an article by Consuelo Dieguez about Luciano Coutinho published in the magazine Piauí (available at http://​goo.gl/​VSbV5T; accessed April 27, 2016). 12. A series of slides describing the “Productive Development Policy” are available at http://​ goo.gl/​aJT1h9; accessed June 6, 2016. 13. Data available at http://​goo.gl/​NJvLHn; accessed August 4, 2016. 14. BNDES disbursements represented 15% of the gross fixed capital formation in Brazil in 2014 (Coutinho 2015). If the regional development banks controlled by the federal government are included, this number may reach around 17.5%. 15. Data used to calculate these numbers are available at http://​goo.gl/​EduTgn; accessed August 5, 2016. 16. According to Folha de São Paulo, interest rates for the contracts to build the Mariel Port in Cuba, for example, varied between 4.44% and 6.91%. According to the newspaper, these figures are consistent with the size and complexity of the project, but are “atypical” considering the Cuban risk. Available at http://​goo.gl/​WKn8GM; accessed April 25, 2016. 17. Available at http://​goo.gl/​xG3FgF; accessed May 3, 2016. 18. These numbers do not consider contracts with subsidiaries (as in the case of the Companhia de Obras e Infraestrutura, which is a subsidiary of Odebrecht). 19. These statements by Meirelles extracted from an interview published in the newspaper Valor Econômico on July 5, 2010, and translated into English by the author of this chapter. 20. See, for example, Jaramillo-​Vallejo (1994). Also, Lazzarini et al. (2015, 238) note that “the general conclusion [of the literature on lending behavior of state-​owned banks] is that state owned banks may misallocate credit by targeting firms based on political criteria rather than on the merits of specific projects.” These authors tested this hypothesis for the case of BNDES and concluded (239) that “firms that donate to winning candidates are more likely to receive funding in the form of loans from BNDES.” However, they also conclude (250) that “BNDES is not generally picking bad projects with negative implications for its own financial health.” 21. Available at http://​goo.gl/​y9zcBM; accessed April 14, 2016. 22. Available at http://​goo.gl/​9QhJiD; accessed April 25, 2016. 23. This is, of course, a very rough estimate, as different sources may have different capital costs to BNDES. A more detailed estimate of these costs is presented by Mussachio and Lazzarini (2015, 294–​299). These authors estimate a gap of 7.5 percentage points (ibid., 296). Besides, we did not consider that several contracts involving the National Treasury establish a ceiling of 6% for the TJLP (Mendes 2014b). In these cases, the estimated costs may be higher. 24. Even if BNDES is (at least theoretically) intended to provide funding to projects that are not attractive to private banking, the bank yearly generates significant profits. Again, this is a consequence of the gap between the cost of its sources of capital and market costs.

References Aghion, B. A. 1999. “Development Banking.” Journal of Development Economics 58: 83–​100. Amann, Edmund, and Werner Baer. 2005. “From the Developmental to the Regulatory State: The Transformation of the Government’s Impact on the Brazilian Economy.” Quarterly Review of Economics and Finance 45 (2–​3): 421–​431.

The Brazilian Development Bank    195 Amann, Edmund, and Werner Baer. 2002. “Neo-​liberalism and Its Consequences in Brazil.” Journal of Latin American Studies 34 (4): 945–​959. Amsden, Alice Hoffenberg. 1989. Asia’s Next Giant: South Korea And Late Industrialization. New York: Oxford University Press. Arestis, Philip. 1997. Money, Pricing, Distribution and Economic Integration. New  York:  St. Martin’s Press. Baer, Werner. 2008. The Brazilian Economy: Growth and Development, 6th edition. Boulder, CO: Lynne Rienner. Baer, Werner. 2002 ed. A economia brasileira, 2nd edition. São Paulo: Nobel. Baer, Werner, and Annibal V. Villela. 1980. “The Changing Nature of Development Banking in Brazil.” Journal of Interamerican Studies and World Affairs 22 (4): 423–​440. Bast, Myriã Tatiany Neves. 2015. “Uma avaliação empírica dos efeitos dos empréstimos do BNDES aos governos municipais brasileiros.” Master’s dissertation, Faculdade de Economia, Administração e Contabilidade de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto. BNDES. 2002. BNDES: 50 anos de desenvolvimento. Rio de Janeiro: BNDES. BNDES. 2012. BNDES:  A Bank with a History and a Future. São Paulo:  Museu da Pessoa. Available at http://​goo.gl/​ru9MEv, accessed April 16, 2016. Bolle, Monica De. 2015. “Do Public Development Banks Hurt Growth? Evidence from Brazil.” Policy Brief PB15-​16. Peterson Institute for International Economics. Bonelli, Régis, and Armando C. Pinheiro. 1994. “O papel de poupança compulsória no financiamento do desenvolvimento: Desafios para o BNDES.” Revista do BNDES 1 (1): 17–​36. Cameron, Rondo, ed. 1967. Banking in the Early Stages of Industrialization:  A Study in Comparative Economic History. New York: Oxford University Press. Cameron, Rondo, ed. 1972. Banking and Economic Development:  Some Lessons of History. New York: Oxford University Press. Cavalcante, Luiz Ricardo. 2004. “Desenvolvimentismo, crise e desestatização: A atuação do BNDES entre 1952 e 2002.” Revista Desenbahia 1: 179–​201. Cavalcante, Luiz Ricardo, and Simone Uderman. 2011. “The Role Played by the BNDES in Funding Electricity Investments in Brazil.” In Energy, Bio Fuels and Development: Comparing Brazil and the United States, edited by Edmund Amann, Werner Baer, and Don Coes, 95–​ 109. New York: Routledge. Coelho, Danilo, and João Alberto De Negri. 2010. “Impacto do financiamento do BNDES sobre a produtividade das empresas: Uma aplicação do efeito quantílico de tratamento.” Encontro Nacional de Economia. Available at http://​goo.gl/​7Tsg1U, accessed April 26, 2016. Coutinho, Luciano. 2013. “Investimento, financiamento e o BNDES.” Available at http://​goo.gl/​ dt6DLB, accessed April 25, 2016. Coutinho, Luciano. 2015. “O BNDES e o Tesouro Nacional.” Available at http://​goo.gl/​9efih9, accessed April 26, 2016. Coutinho, Luciano. 2015. “BNDES: Organização e desempenho.” Presentation to the Comissão de Assuntos Econômicos e a Comissão de Serviços de Infraestrutura do Senado Federal, April 14. Development Bank of Canada (DBC). 2009. International Benchmark Study on Development Institutions. Prepared by BDC Strategy and Planning and GY Consulting. Doctor, Mahrukh. 2015. “Assessing the Changing Roles of the Brazilian Development Bank.” Bulletin of Latin American Research 34 (2): 197–​213. Ferraz, João Carlos. 2012. “Financing Development in an Uncertain World:  BNDES’s Experience.” Presentation at the conference Development Banks as Agents of Change. Doha, April.

196   Luiz Ricardo Cavalcante Ferraz, João Carlos, David Kupfer, and Felipe Silveira Marques. 2014. “Industrial Policy as an Effective Development Tool: Lessons from Brazil.” In Transforming Economies: Making Industrial Policy Work for Growth, Jobs and Development, edited by José Manuel Salazar-​ Xirinachs, Irmgard Nübler, and Richard Kozul-​Wright, 291–​305. Geneva:  International Labour Office. Gerschenkron, Alexander. 1962. Economic Backwardness in Historical Perspective: A Book of Essays. Cambridge, MA: Belknap Press. Goldsmith, Raymond William. 1969. Financial Structure and Development. New Haven, CT: Yale University Press. Hirschman, Albert O. 1968. “The Political Economy of Import-​Substituting Industrialization in Latin America.” The Quarterly Journal of Economics 82 (1): 1–​32. Jaramillo-​Vallejo, Jaime. 1994. “Comment on ‘The Role of the State in Financial Markets,’ by Stiglitz.” In World Bank Annual Conference on Development Economics, 1993, 53–​56. Washington, DC: The World Bank. Krieckhaus, Jonathan. 2002. “Reconceptualizing the Developmental State: Public Savings and Economic Growth.” World Development 30 (10): 1697–​1712. Lazzarini, Sergio G., Aldo Musacchio, Rodrigo Bandeira-​de-​Mello, and Rosilene Marcon. 2015. “What Do State-​Owned Development Banks Do? Evidence from BNDES, 2002–​09.” World Development 66: 237–​253. Levine, Ross. 1997. “Financial Development and Economic Growth:  Views and Agenda.” Journal of Economic Literature 35 (2): 688–​726. McKinnon, Ronald I. 1978. A moeda e o capital no desenvolvimento econômico. Rio de Janeiro: Interciência. Mendes, Marcos. 2014a. “Faz sentido o BNDES financiar investimentos em infraestrutura em outros países?” Published April 16. Available at http://​goo.gl/​Wk6PX5. Accessed on 5th April 2016. Mendes, Marcos. 2014b. “Quanto custam para o Tesouro os empréstimos concedidos ao BNDES?” Published September 23. Available at http://​goo.gl/​pjKcGC. Accessed on 5th April 2016. Musacchio, Aldo, and Sergio G. Lazzarini. 2015. Reinventando o capitalismo de Estado:  o Leviatã nos negócios: Brasil e outros países. São Paulo: Portfolio-​Penguin. Myrdal, Gunnar. 1960. Teoria econômica e regiões subdesenvolvidas. Belo Horizonte: Editora da UFMG (Biblioteca Universitária). OECD. 2013. OECD Economic Surveys: Brazil 2013. Available at http://​dx.doi.org/​10.1787/​eco_​ surveys-​bra-​2013-​en, accessed August 5, 2016. Ottaviano, Gianmarco, and Filipo Lage de Sousa. 2008. “O efeito do BNDES na Produtividade das Empresas.” In Políticas de incentivo à inovação tecnológica, edited by João Alberto De Negri and Luis Claudio Kubota, 361–​386. Brasília: IPEA. Ottaviano, Gianmarco, and Filipo Lage de Sousa. 2014. “Relaxing Credit Constraints in Emerging Economies:  The Impact of Public Loans on the Performance of Brazilian Manufacturers.” Development Studies Working Paper No. 369. Centro Studi Luca d’Agliano. Pinheiro, Armando Castelar, and Fabio Giambiagi. 2000. “Os antecedentes macroeconômicos e a estrutura institucional da privatização no Brasil.” In A privatização no Brasil: O caso dos serviços de utilidade pública, edited by Armando Castelar Pinheiro and Kiichiro Fukasaku, 13–​43. Rio de Janeiro: BNDES. Pinto, Rogerio Feital S. 1969. The Political Ecology of the Brazilian National Bank for Economic Development (BNDE). Washington, DC: Organization of American States.

The Brazilian Development Bank    197 Ribeiro, Eduardo Pontual, and João Alberto De Negri. 2009. “Estimating the Effect of Access to Public Credit on Productivity in Brazil.” In LACEA/​LAMES Buenos Aires 2009: Papers and Proceedings. Buenos Aires: LACEA. Shaw, Edward Stone. 1973. Financial Deepening in Economic Development. New York: Oxford University Press. Stiglitz, Joseph. 1994. “The Role of the State in Financial Markets.” In World Bank Annual Conference on Development Economics, 1993, 19–​52. Washington, DC: The World Bank. Stiglitz, Joseph, and Andrew Weiss. 1981. “Credit Rationing in Markets with Imperfect Information.” The American Economic Review 71 (3): 393–​410. Studart, Rogerio. 1995. “Saving, Financial Markets and Economic Development:  Theory and Lessons from Brazil.” In Finance, Development and Structural Change:  Post-​ Keynesian Perspectives, edited by Philip Arestis and Victoria Chick, 46–​60. Aldershot, UK: Edward Elgar. Zysman, John. 1983. Governments, Markets and Growth: Financial Systems and the Politics of Industrial Change. Ithaca, NY: Cornell University Press.

Chapter 10

The Evolu t i on of Br azil’s Bank i ng  Syst e m Gustavo S. Cortes and Renato L. Marcondes

10.1. Introduction Like most backward Latin American economies, Brazil developed a formal financial system only in the late nineteenth century.1 The credit expansion observed in this period is strongly associated with Brazilian economic growth. Along with price stability, institutional changes were fundamental in spurring credit growth. Throughout the twentieth century, we observe a particular structure of the financial system, consequently leading to particular results. Before the 1930s, there was an active financial system based on a wide variety of funding instruments, such as nonbank lending by private individuals (capitalists), bank lending, mortgages (collateralized debt obligations), pawn credit, debentures, and savings accounts, among others. Businesses could get funding at longer terms—​often outside of the banking system—​and the capital market was an important source of funding for real estate and manufacturing. This well-​functioning financial sector in Brazil faced significant limitations when institutional reforms were undertaken after the Great Depression. Several interventions were introduced by the government in the lending business, for example the usury law (Lei de Usura) and the golden clause (Cláusula Ouro). These changes remained virtually intact until the mid-​1960s and are key to understanding the atrophy of the financial system; price interventions, together with high inflation, reduced banks’ capability to provide long-​term financing for the public and private sectors. From 1964, the military regime introduced reforms that established the framework of the more recent Brazilian financial system. They include the foundation of the Central Bank of Brazil, the creation of indexation mechanisms, market segmentation, and a structure of incentives toward concentration in the banking sector. By providing a historical description of the Brazilian banking system, we aim to recover its most important institutional elements and thus to contribute to the task of rethinking and improving the current banking system.2

The Evolution of Brazil’s Banking System    199 The chapter is organized as follows. In section 10.2, we discuss credit in the Brazilian economy at the end of the colonial period (i.e., before the rise of banks). Section 10.3 discusses the surge of the first banking institutions in Brazil and their difficulties in surviving and perpetuating. Section 10.4 focuses on the period of the First Republic (1889–​1930) and its governmental interventions in the banking system. Section 10.5 details the institutional changes implemented in response to the Great Depression of the 1930s, and how they helped to recover economic growth. We also discuss how these institutional changes, combined with high inflation, hampered the functioning of banking and overall credit markets. In section 10.6, we detail the institutional reforms introduced by the military regime after 1964, which allowed economic and credit growth throughout the mid-​1960s and the 1970s. In the final two sections we describe the recent evolution of the Brazilian banking system. Section 10.7 describes the problems caused by the hyperinflation of the 1980s and early 1990s in the banking system and the eventual victory in the battle against hyperinflation after 1994. Section 10.8 focuses on the most recent decades of the banking system and analyzes its stability during the Great Recession of 2008–​2010 and the recent Brazilian fiscal crisis that began in 2014 and intensified in 2015. Section 10.9 concludes.

10.2.  Credit before Banks Currency was scarce in the first two centuries of the Brazilian colonial period (1500–​ 1822), inducing a search for alternatives.3 Beyond the use of commodity currencies (e.g., cotton cloth), the small amount of existing currency was concentrated in the hands of “a few rich and in the orphans’ vault.”4 In such a scenario, credit was, naturally, an element for mitigating currency scarcity and meeting the demand generated by investments and population growth in the colonial economy. For instance, credit operations between planters (borrowers) and merchants (lenders) established that the loan had to be paid at the moment of the harvest. The accounting registries of merchants’ books served as proof of credit and debit. Additionally, paperless loans were also contracted based exclusively on reputation and trust between lenders and borrowers (Espírito Santo 2002). Interest rates reached up to 8% by the end of the contract’s maturity, but could be even higher depending on the riskiness of the project. Even though there is a relative scarcity of information on these loan agreements, the literature documents significant heterogeneity in maturities; private lenders preferred short-​term loans repayable in one or two years, whereas institutions extended credit for longer periods (Schwartz 1985, 205). Credit had an important role in sugar production—​colonial Brazil’s main economic activity—​with foreigners representing a significant share in the supply side of this market. To deal with disagreements between borrowers and lenders, conflict resolution often took place in courts, with lawsuits to enforce claims of pledgeable assets, usually slaves and real estate. Novel laws were passed to benefit borrowers in the lending relationship. For example, after 1663—​during the stagnation period of the sugar economy—​the Portuguese Crown restricted the execution of mills, factories, and real

200    Gustavo S. Cortes and Renato L. Marcondes estate properties of sugar planters to fulfill debt contracts. Pledgeability was restricted to the income of mills, which became a major source of inefficiency in credit markets and delayed the solution of lawsuits. There were also usury limits established by the Catholic Church, capping interest rates at 6.25%. Even though the usury limit existed, it was avoided in some cases by adapting other terms of the deal (Schwartz 1985). In colonial times, the supply of credit was divided among many different agents. In addition to individual merchants, Holy Houses (Santas Casas) and religious fraternities played an important role. These institutions received donations and bequests to be used in their maintenance and to promote charity work. An important institution was Bahia’s Holy House of Mercy (Santa Casa da Misericórdia da Bahia) in which could be observed with some regularity rudimentary banking services (Flory 1978, 73; Russell-​Wood 1968). The case of Rio de Janeiro was similar to that of Bahia. Loan registries in Rio show that the Orphans’ Judge (Juízo dos Órfãos) was the main lending institution, representing one-​third of total lending extended throughout the second half of the seventeenth century (Sampaio 2000). This fact provides evidence that capital accumulation by mercantile activity was not the main determinant of the lending business, and shows the fragility of mercantile capital in that period. The amount of capital was significantly raised following the discovery of gold in Minas Gerais, eased by the increase of monetary circulation and assets’ liquidity. In this period, liabilities in post mortem inventories represented one of the most important wealth sources, along with slave ownership. Following this increase in asset availability and liquidity, nonspecialized credit institutions lost their leadership to those denominated capitalists: wealthy mercantile businessmen, negotiators, and individual lenders (Fragoso 1992). Legal restrictions played a key role in hampering the credit market’s functioning in colonial Brazil. The main interventions consisted of interest rate caps and foreclosure restrictions. Privileges given to some groups—​mill owners (senhores de engenho) and miners (mineradores)—​prohibited the foreclosure of their properties (real estate and slaves) in case of default (Brito 1923, 104).5 On the other hand, the interest rate caps were a result of religious and royal interference, especially against usury abuses in the Ultramarine domain. During the Philippine Order (Ordenações Filipinas) of the beginning of the seventeenth century, contracts considered usurious were repressed. Under the Pombaline Rule (Período Pombalino) and after the 1755 Lisbon earthquake, an interest rate ceiling of 5% per year was imposed on all loans by the 1757 Charter (Portugal 1870). Despite the developments of the seventeenth century mentioned earlier, there were still barriers to establishing specialized credit-​supplying institutions distinct from religious orders, negotiators, capitalists, and Orphans’ Judges. For this reason, the colonial period’s credit was fundamentally tied to personal relationships. Due to asymmetric information and the underlying cyclical risks of harvests or commodity prices, credit in sugar production areas maintained a regional aspect. It was mainly found in regional mercantile centers, which provided trade finance to fund the flows of the sugar production and the purchase of imported products (Faria 1998, 188).

The Evolution of Brazil’s Banking System    201 The consequences of the colonial financial system’s inefficiency are seen in contemporaneous descriptions, such as that of the judge (desembargador) João Rodrigues de Brito in the early nineteenth century. Brito lamented the “shameful judicial delays” that hampered conflict resolution and the enforcement of contracts. Delays in repaying creditors warded away valuable capital and investors that would otherwise have entered the credit market, “forcing them to keep their money for themselves, but their money is less productive in their hands than it would be in the hands of skilled farmers” (Brito 1923, 108–​109).6 Despite this perception that eliminating foreclosure restrictions would enhance the market’s efficiency, this did not happen during the colonial period (Marques 2014). Later, the government introduced changes to reduce the limitations to credit contracting terms and foreclosure. Interest rate caps fell by 1810 for maritime commerce, and later for other operations.7

10.3.  The First Banks As in Portugal, it took a long time for banking institutions to become established in Brazil (Aguiar 1960).8 Although there were previous attempts, the first bank of the Portuguese Empire was the Banco do Brasil (BB), created in 1808 by Dom João VI.9 Even with privileges granted by the Crown and limited liability to shareholders, BB had difficulty subscribing its capital and began operations around a year later. BB received demand deposits, issued money, and lent to the private sector, but its main client was the government. The Crown’s demand for financing led to the significant issuance of paper money after 1814, reaching three times the original amount of capital by 1821. At the same time, the bank’s managers appropriated private benefits through discounting letters of credit, which were not repaid. The return of Dom João VI to Portugal drastically reduced the metallic supply of the bank, threatening its solvency. Additionally, its liabilities became national debt, and consequently BB was shut down in 1829. The closure of BB a mere few years after Brazil’s independence encouraged public skepticism about the country’s banking institutions (Cardoso 2010; Franco 1984). In 1831, a private French-​inspired Caixa Economica (Savings Bank) was established in Rio de Janeiro. Various publications tried to incentivize savings more broadly among the population (Saraiva 2014). This institution was able to gather thousands of shareholders—​including slaves, who represented up to 8% of all ownership—​ demonstrating less exclusiveness in the banking business (Saraiva and Oliveira 2015). Deposited resources were invested in public securities (apólices), and interest payments were distributed to shareholders. Other savings banks were established in different provinces and cities (e.g., in Bahia in 1834). The competition with other banking institutions and the 1850s financial crises destabilized the prices of public bonds and the overall trust in Rio de Janeiro’s Caixa Economica, which shut down in 1859. Two years later, the imperial government created the Caixa Economica da Corte (Court’s Savings Bank) and the Monte do Socorro, a public institution granting loans collateralized by

202    Gustavo S. Cortes and Renato L. Marcondes valuable objects such as jewelry.10 The growing interest of the population in keeping deposits in public institutions favored the creation of savings banks in other provinces. Even though there were remuneration caps for deposits of higher values, the rationale behind the population’s willingness to keep deposits in public institutions was safety. By the end of the imperial period, various savings banks functioned in different parts of the country (Adams 2005). The Banco do Ceará was the first Brazilian private bank. It was established in 1836, but lasted for only three years due to managerial excesses (Pelaéz and Suzigan 1981, 58). In 1838, the Banco Comercial do Rio de Janeiro was founded as a result of the reunion of tradespeople and private lenders. The bank operated with deposits, discounts, and the issuance of notes that circulated as paper money. The institution was highly successful, especially in Rio de Janeiro. An issuance restriction of one-​third of the bank’s total capital was applied, and because this was strictly followed, depositors developed solid trust in the institution’s financial soundness. The bank paid dividends regularly in the 1840s, highlighting its profitability as the monopolist in the city of Rio de Janeiro. Other provinces saw the first private banks established in the following decade: Bahia in 1845, Maranhão in 1846, and Pará in 1847. The new Commercial Code and the end of the Atlantic slave trade in 1850 granted a favorable institutional framework for the establishment of new Anonymous Societies and banks. In this context, Brazilian businessman Irineu Evangelista de Souza (baron of Mauá) created in Rio de Janeiro a new Banco do Brasil in 1851. The capital gathered in this second incarnation surpassed that of the first BB, and increased competition in the banking sector. BB was authorized to issue currency, also limited to one-​third of its capital. Two years later, BB merged with the Banco Comercial do Rio de Janeiro to form yet a newer Banco do Brasil in 1853, which retained a monopoly on money issuance until 1857. This period saw heated debates between papelistas and metalistas, which represented the real-​economy and financial sectors, respectively.11 Due to a crisis in 1857, a new government staff authorized emission plurality. The power to issue paper money given to private banks persisted for three years until a new government restriction in 1860. Whenever in crisis, and during the Paraguayan War (1864–​1870), the Brazilian government held a more expansionist monetary policy (Villela 1999). Ultimately the National Treasury monopolized money supply in 1866 and remained the only issuer until the end of the imperial period. At this point, foreign banks began operating directly in Brazil—​especially in Rio de Janeiro. First the London and Brazilian Bank and then the Brazilian and Portuguese Bank (later called the English Bank of Rio de Janeiro) were created in 1862 and 1863, respectively, improving banking techniques (Guimarães 2016). They pioneered the use of paper checks in Brazil and operated exactly like domestic commercial banks (i.e. focusing on short-​term lending) (Joslin 1963). Since agriculture was the main activity of the Brazilian economy, farmers regularly complained about the scarcity of capital at a reasonable price. One law that guaranteed profits for so-​called Banks of Royal Credit made possible the creation of such institutions, intended to lend to the agricultural sector at longer maturities.

The Evolution of Brazil’s Banking System    203 By the end of the Empire, there were a few dozen banks throughout the country, mostly concentrated in Rio de Janeiro. The majority did not have affiliates or branches, and therefore most provinces and cities did not have access to banking services. The Banco do Brasil was the largest institution, followed by the aforementioned two English institutions (Goldsmith 1986, 42). The transition from slavery to free labor intensified the demand for credit, and there was great dissatisfaction among farmers since banks could not supply enough funding. This led the Empire to authorize the issuance of gold-​ convertible paper money in the final years before the Republic’s proclamation. However, despite increasing the money supply, this did not solve the problem of money shortage.

10.4.  Banks in the First Republican Period (1889–​1930) The increase in money supply gained pace with the newly established Republican government, through a government authorization for regional banks to issue without the constraint of gold reserves. This expansionary monetary policy, led by the newly appointed Finance Minister Ruy Barbosa, was an episode known as Encilhamento. Its main objective was to boost economic growth and decrease credit costs. In a period of dramatic institutional change—​the abolition of slavery and the dissemination of free labor—​lack of liquidity was a common complaint in both agricultural and industrial sectors. Monetary easing was implemented as a response to these concerns and happened mainly through novel issuing banks. This policy had pronounced effects on the real economy, boosting credit growth and the establishment of a large number of businesses and banking institutions (Levy 1980). The Banco do Brasil merged with another institution created in the boom of the Encilhamento and became the Banco da República do Brasil. However, the Encilhamento had its costs. Pronounced exchange rate devaluation, balance of payments disequilibrium, and higher inflation all resulted in pressure for the policy to change. The government tried to counteract these effects without success, only managing to do so after 1898 when a new finance minister, Joaquim Murtinho, prohibited banks from issuing money (Franco 1983, 1990; Schulz 1996). The abrupt monetary restriction produced the very first banking crisis of the Brazilian Republic, in its transition to the twentieth century. Several financial institutions defaulted, especially those domestically born in the Encilhamento period. The Banco da República faced severe distress, spurring government intervention in the bank. The intervention was unsuccessful and the bank was closed in 1901. Four years passed with no “national bank,” and only in 1905 did the government decide to create a new Banco do Brasil. This operation was a landmark in the development of Brazil’s financial system, because it created the first public-​private mixed capital bank in the country, with the majority of ownership belonging to the federal government (Triner 2000, Chapter 4).

204    Gustavo S. Cortes and Renato L. Marcondes The restructuring operation modernized the national banking system. Bank credit grew from 3.4% of gross domestic product (GDP) in 1906 to 17.6% in 1922, retaining this pace until the end of the 1920s (Triner 2000, 101). Ownership structure changed in terms of nationality as World War I  decreased foreign presence and incentivized domestic institutions. It also changed in terms of the importance of public ownership, since a large share of this credit growth occurred through public banks. Finally, the system changed geographically because the creation of commercial and savings banks owned by states helped decentralize the banking business from the capital, Rio de Janeiro (Costa Neto 2004, 35). State governments began providing financial support to mortgage banks through lending and deposits from their state-​owned savings banks, as in the case of São Paulo (Costa 1988, 80–​81). In addition to the events of World War I, the banking system was nationalized following the institutional reforms introduced in the 1920s. The Banco do Brasil became the official exchange rate market broker of the Treasury and accumulated functions of national monetary authority, such as paper money emission and rediscount operations in some years (Goldsmith 1986, Chapter 4; Topik 1987, Chapter 2). These reforms were introduced by the General Inspectorate of Banks (Inspetoria Geral dos Bancos) in the 1920s and were reinforced in the 1930s by the constitutions of 1934 and 1937. As such, foreign banks had a fundamental role in the final decades of the nineteenth century and the first decades of the twentieth, but a much more limited one after that. During the First Republic, Brazilian businesses adapted to different institutional frameworks by combining a diversified portfolio of funding sources. They used financing instruments such as the issuance of corporate bonds and equity shares. For large companies, the most common external financing source was debentures. Despite adverse economic policy and external restrictions, there was an important market for private bonds and stocks until the 1920s. Firms from various sectors issued bonds and equity in the market, protected by a pro-​creditor institutional framework that existed until at least the start of World War I. Even though the early twentieth century saw the rise of an expanding formal banking system, private capitalists and negotiators still managed many credit operations. The market for these agents was distinct, especially the mortgage credit market. Banks provided large loans under more favorable maturities and interest rates to clients mostly concentrated in public services, commerce, and manufacturing. On the other hand, capitalists and other private agents had a broader and more diversified lending portfolio, with smaller amounts per lending and less favorable contracting conditions (Cortes et al. 2014). This structure suggests considerable segmentation of the credit market in Brazil.

10.5.  The Financial System after the Great Depression (1930–​1964) Despite its impacts on exchange rate depreciation and the balance of payments, the Great Depression had relatively less dramatic effects on the Brazilian economy and its

The Evolution of Brazil’s Banking System    205 financial system. The government reacted strongly to avoid a deepening of the crisis by using two main policy instruments: the Rediscount Portfolio (Carteira de Redesconto) in 1930, and the reserves available in the Caixa de Mobilização Bancária in 1932.12 In this new economic conjuncture, the BB once again increased its share of the national banking system, mainly at the expense of foreign banks. The retraction of foreign banks’ business in Brazil occurred both because of losses in their headquarters across borders and because of the new, domestic-​friendly banking legislation. Because many credit operations defaulted, the government also passed the Usury Law (Lei da Usura) that established a ceiling of 12% on interest rates. Additionally, the Gold Clause Law (Lei da Cláusula-​Ouro) prohibited the indexation of contracts in gold or foreign currency. Finally, the Law of Economic Readjustment (Lei do Reajustamento Econômico) cut by half the pre-​1933 collateralized-​debt obligations between farmers and the federal government. The remaining debt was exchanged by specific federal government bonds known as economic readjustment obligations. The policy responses of the Brazilian government were inspired by those used in the United States and helped to mitigate the effects of the crisis and to recover economic activity as early as the 1930s. However, this set of regulations—​created in crisis times—​ persisted for over a quarter of a century further, when Brazil’s economic conditions were distinct, with higher rates of inflation and economic growth. Therefore, although these laws were fundamental for fighting the crisis, they became an obstacle for the development of the financial system in normal times. The acceleration of inflation rates, combined with usury laws and other restrictive regulations, entailed a reduction of real interest rates.13 Even though nominal returns were still positive, inflation eroded real gains significantly for most of the period. For instance, during World War II and for most of the following two decades, the inflation rate exceeded returns on public bonds and mortgage lending. These restrictions affected virtually all financial instruments—​ public bonds, debentures, insurance policies—​producing distortions in the financial system. In short, the problem of negative real interest rates was a major constraint upon the economy’s total savings and the normal functioning of credit markets. As expected, agents found alternatives to the interest rate cap and other interventions.14 The significant increase in interest rates that occurred in the 1950s led to an improvement in mechanisms to trick these restrictions. The main mechanism was the bill of exchange (letra de câmbio). Because these assets were traded on the stock exchange market, they could exhibit discounts of their nominal value. The returns usually exceeded the interest rate caps, but were still insufficient to provide positive real returns to creditors (Contador 1974). Thus, even the best financial instrument available—​which was introduced only by the end of the decade—​was not able to systematically produce real gains to investors. On top of that, it was not an easily accessible instrument since it was available only to very select groups trading in the stock exchange. Under these conditions, the banking system began to operate on a short-​term basis by collecting demand deposits to also lend short term under negative real interest rates. Demand deposits kept their share of GDP broadly unchanged throughout 1946–​1962, even though there was an expressive reduction on term deposits in that same period.

206    Gustavo S. Cortes and Renato L. Marcondes There was a significant restructuring of the banking system in order to increase the collection of demand deposits, which were important to commercial banks’ profitability. Banks raised their number of branches but shrank in the overall number of institutions, consequently increasing market concentration. The number of banks decreased from around 500 to a little over 300 between 1945 and 1964, whereas the number of branches in this period grew from around 1,500 to around 7,000 (Sochaczewski 1993, 120).15 Despite the growth in the presence of branches, demand deposits—​and consequently bank lending—​did not increase beyond GDP growth rates. Therefore, there was not stagnation of the banking system, but rather its consolidation in large, regional-​level banks (Saes 1997). Consolidation did not alleviate the difficulties faced by banks in accessing private resources that could allow long-​term lending. The foundation of the Brazilian Development Bank (Banco Nacional de Desenvolvimento, BNDE) in 1952 aimed at mitigating these restrictions by lending—​at long-​term maturities—​a mandatory share of income tax revenues.16 However, the BNDE’s lending supply was clearly insufficient for the size of the Brazilian economy. In order to keep its capital, the BNDE managed to gain exemption from usury laws.17 Finally, the necessity of escaping the interest cap led to the creation of extremely persistent distortions, such as the indexation mechanisms in the real economy and the financial sector. As a result, long-​term credit was restricted to the BNDE and savings banks in a strongly rationed manner.

10.6.  The Financial System under the Military Government (1964–​1985) The military government introduced reforms that radically changed the monetary and financial system. Some remarkable institutional changes were the creation of the Central Bank of Brazil (Banco Central do Brasil, BCB), the National Monetary Council (Conselho Monetário Nacional, CMN), and the reform of capital markets.18 Additionally, the monetary correction mechanism was introduced for Readjustable Obligations (Obrigações Reajustáveis) of the National Treasury, and later for housing contracts, private bonds, and finally for savings accounts in 1967.19 Thus, these assets were guaranteed to yield a positive real interest rate, since they were readjusted by the inflation rate. The reform resulted in the possibility of the Banco do Brasil withdrawing resources from the BCB, as long as authorized by the CMN. This was known as the Movement Account (conta movimento) and allowed the Banco do Brasil to act as monetary authority, until the extinction of this account in 1986. Finally, according to its originating law, the BCB was responsible for managing the public debt of the National Treasury, but the Federal Constitution of 1988 eliminated this possibility (Barbosa 1995; Simonsen

The Evolution of Brazil’s Banking System    207 and Campos 1979). This new institutional framework allowed an economic growth takeoff after stabilization, especially in the post-​1967 period. The financial reform was based on the principle of banking specialization and segmentation, with the BCB authorizing and regulating all financial institutions. There was a link between fundraising and the application of resources (e.g., bills of exchange for the financeiras and longer-​term deposits for investment banks).20 The Financial System for Housing (Sistema Financeiro da Habitação, SFH) was created to fund public investment in real estate by using resources from the job tenure severance fund (Fundo de Garantia do Tempo de Serviço, FGTS).21 The reform of capital markets provided fiscal and institutional incentives to disseminate securities through the stock market and its agents (brokers and dealers).22 Beyond the necessity of modernizing fiscal institutions, the new government aimed at decreasing the financial fragility of the Brazilian economy in that period. Credit growth was not concentrated in a unique financial instrument, but rather reflected a diversity of financing tools. Bills of exchange were key for funding household credit, and the SFH had a high level of resources to fund construction. These sectors underwent enormous growth in this period and pushed national GDP growth rates. Other instruments also emerged, such as real estate bills of exchange (letras imobiliárias) and certificates of deposit, among others. Two instruments specially benefited from the reforms: internal debt (dívida interna) and mortgages (hipotecas). In the period of the “economic miracle,” the Brazilian government encouraged mergers and acquisitions in the financial sector in order to increase scales of operation, reduce costs, and enhance the overall efficiency of financial institutions. Although the number of nonbanking institutions (especially brokers and dealers) increased in that period, the number of banks decreased from 328 to 106 between 1964 and 1976 (Toneto Jr. 1992, 116). However, market concentration did not reduce the share of the banking sector in the economy: financial intermediation grew from 3.1% of GDP in 1964 to 7.9% in 1976 (IBGE 2000). The issuance of new charter letters was suspended so that one bank had incentives to acquire another bank in order to increase its size.23 Many banks became financial conglomerates acting in a diverse range of financial services, such as investment banking, insurance, broker-​dealers, and so on. The restructuring of the financial sector allowed a recovery of term deposits until 1973. Up to the mid-​1960s, the vast majority of deposits were demand deposits due to the interest rate restrictions and the erosion caused by high inflation. With the introduction of monetary correction, there was an incentive to lengthen maturities. In the transition to the 1980s, inflationary acceleration reduced demand deposits by much more than term deposits. The opportunity cost of holding money significantly increased in this period, making term deposits more attractive for their greater protection against inflation. In the transition from the nineteenth century to the twentieth, the Brazilian credit market grew mainly based upon the existing instruments (i.e., public and private bonds). Later, by the mid-​twentieth century, these bonds became unattractive in a context of high inflation and government price controls. The consequence of this was a lower capacity for financing firms, households, and government. After the institutional

208    Gustavo S. Cortes and Renato L. Marcondes reforms introduced in 1964, credit was freed from its repressed conditions and grew rapidly until at least the early 1980s. Monetary stability was shown to be a necessary (albeit not sufficient) condition for the development of Brazilian financial markets throughout the twentieth century. The two events moved in tandem: periods of high monetary stability were accompanied by high credit growth, whereas times of high inflation and monetary instability saw stagnation and retraction in credit. The mechanisms of indexation aimed at protecting investments from high inflation were not enough to guarantee credit in a context of high instability, but rather generated even more distortions.

10.7.  Bank Restructuring during and after the Inflationary Period (1986–​2005) Several years of high inflation transformed the activity of securing deposits to finance public debt purchase into one of the most important credit operations in banks’ balance sheets. Banks used short-​term deposits, imperfectly protected from inflation, to finance the purchase of securities, yielding much higher interest rates than the ones paid to depositors. Inflation-​based revenues became a disproportionate part of banks’ balance sheets, especially in state-​owned banks (Beck et al. 2005, 2227). For example, inflation-​ related revenues accounted for 4% of GDP by 1992, representing 41.9% of total bank revenues, and dropped to practically zero in 1995 (Baer and Nazmi 2000, 8). Indeed, when inflation suddenly ceased, many banks found themselves in financial trouble (Beck et al. 2005; McQuerry 2001). The lack of an easy return activity gave rise to excessive risk-​taking by banks, which entailed an increase in the share of nonperforming loans: from 5% in mid-​1994 to about 15% throughout 1997. This impact was higher in state-​owned banks, because their main activity was to bridge state treasuries’ credit shortfalls and they therefore lacked the know-​how to manage credit risk properly (Baer and Nazmi 2000, 11). Such a conjuncture provided the incentives for the government to accelerate the liberalization of the repressed financial system, mainly through privatization of the banking system.24 Prior to the post-​inflation period, the Central Bank of Brazil passed in 1988 a resolution to establish the universal banking model adopted in Germany, moving away from the segmented model used up to that point (Cardim de Carvalho et al. 2009, 875).25 The resolution eliminated entry barriers such as the bank charter (carta patente) restriction on the establishment of new banks, and allowed the flourishing of banking institutions interested in inflation-​related revenues. By the mid-​1990s, Brazil’s financial sector was strongly dominated by banking institutions. For example, in 1995 banks accounted for 86% of the income generated by the financial sector, 30.6% from commercial banks and 55.8% from universal banks (Barbosa 1995; Cardim de Carvalho 1998; Minella 1996).26

The Evolution of Brazil’s Banking System    209 The defeat of inflation was a necessary condition for credit growth, although international and regional macroeconomic factors also contributed to delaying credit expansion in Brazil. For example, the average spread of non-​earmarked (non-​directed) credit operations experienced a remarkable drop throughout 1995, as a result of the price stability achieved by the Real Plan (see Figure 10.1, Panel A).27 Nevertheless, in 1998–​1999 the currency crisis of the Brazilian real had increased the average cost of lending.28 It was only in the first decade of the twenty-​first century that the average spread began a slow-​paced decay to lower levels. Even in the period of decrease, some events affected the cost of lending by Brazilian banks, such as the Argentinian crisis of 2001–​2002, and the spike in political uncertainty in 2003 due to the election of left-​wing presidential candidate Luiz Inácio Lula da Silva. On a different front of liberalization, the 1990s saw an unprecedented wave of bank privatizations in Brazil and Latin America in general, led by mergers and acquisitions undertaken by foreign-​owned banks (Domanski 2005; Goldberg 2009). However, unlike the ideological switch toward market-​friendly policies observed in other Latin American countries (e.g., Argentina, Chile, and Uruguay), the privatization process in Brazil was embarked upon mainly for pragmatic reasons.29 The fiscal situation of private and especially state-​owned banks deteriorated after many years of excessive reliance on inflation-​based earnings and political misuse (in the case of state-​owned banks). To mitigate the risks of a banking crisis, the Central Bank of Brazil passed two programs designed to restructure the national banking system: the PROER for private banks in November 1995, and the PROES for state-​owned banks in August 1996.30 The Central Bank of Brazil benefited from special legislation—​the RAET—​granting it powers to intervene extra-​judicially in problematic banks.31 The PROER liquidated extra-​judicially private banks that were too fragile, and divided into two categories some healthier but still fragile banks: the “good banks” and the “bad banks.” The “good” contingent was sold to other players in the market, while the “bad” part was liquidated extra-​judicially B. Foreign banks: % of total number of banks

A. Credit cost: Average spread 150%

40% 30%

100%

20% 50%

10% 0%

0% 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013

1995 1997 1999 2001 2003 2005 2007 2009 2011 2013

Figure 10.1.  Average spread of non-​earmarked credit operations (Panel A) and participation of foreign banks in the Brazilian banking system (Panel B). Sources: Panel A, Banco Central do Brasil’s Time Series System, series #3955. Panel B, Claessens and Van Horen (2014). Notes: In Panel A, the spread is measured in percentage points and is given by the difference between the bank’s borrowing interest rate (i.e., the policy interest rate in Brazil—​SELIC) and the bank’s lending interest rate. In Panel B, the share of foreign banks is measured as a percentage of the total number of banks operating in Brazil each year.

210    Gustavo S. Cortes and Renato L. Marcondes to redeem losses to managers and shareholders (Maia 1999, 112). The PROES offered the following options to states: (1) liquidation; (2) allow the federal government to privatize them; (3) privatize themselves; or (4) restructuring (Beck et al. 2005, 2224). Most privatizations occurred as a direct consequence of the inability of states to maintain their banks. In 1997, 22 out of 27 states negotiated their debt with the federal government, in a deal that had the controlling rights of state banks as its central element: a state could maintain control of its bank if and only if it provided at least 50% of its restructuring costs. Since for most states the restructuring costs were much higher than the political benefits of keeping the state bank, the majority decided to transfer ownership to the federal government, which later privatized them (Beck et al. 2005, 2228–​2289). The participation of public banks in total deposits of the Brazilian banking system shrank from 19.3% in mid-​1996 to only 6.54% in mid-​1998 (Baer and Nazmi 2000, 16, Figure 7). Following the Latin American wave of privatization, Brazilian banks were targeted by cross-​border merger and acquisition operations. The federal government encouraged such operations from both domestic and foreign buyers by providing tax incentives and credit facilities to accelerate bank consolidation (Baer and Nazmi 2000, 13). The period saw a substantial increase in the participation of foreign banks in the Brazilian banking system. In 1995, foreign banks accounted for about 7% of the Brazilian banking system’s assets, whereas in 2000 they represented roughly 25% of total assets (Crystal et al. 2002, 2, Figure 1). In terms of the number of foreign banks operating in Brazil, the share of foreign banks went from 22% in 1995 to around 34% in 2000, representing the period with the highest growth rate of foreign ownership participation throughout the entire series (Figure 10.1, Panel B).32 Indeed, the most important state-​owned bank, the São Paulo State Bank (Banespa) was acquired by the Spanish-​headquartered Banco Santander in 2000 in a more than US$3.6 billion deal, nearly four times the minimum established by the auction rules.33 The consolidation process naturally affected the market structure of the banking industry.34 It was believed by the government that economies of scale driven by larger banks would increase the overall efficiency of the banking system (Cardim de Carvalho et al. 2009, 877–​887). Another perception that had by that time been spread by multilateral institutions such as the International Monetary Fund (IMF) and the Bank for International Settlements (BIS) was that foreign banks would contribute toward higher productivity by imposing tougher market discipline and risk management expertise (Cardim de Carvalho et al. 2009; Domanski 2005). However, on the question of whether or not the higher share of domestic and/​or foreign private banks increased the efficiency of the banking system, the empirical evidence for Brazil is not conclusive. It favors the hypothesis of a positive correlation between private bank ownership and efficiency, but not so much that of a correlation between foreign bank ownership and efficiency.35 Another aspect of the banking consolidation to evaluate is whether the new banking system became more resilient to negative shocks. There is positive evidence in this direction, from the perspective of both a more globalized (larger share of foreign banks) and a more concentrated (smaller number of banks) banking sector. Empirical results suggest that the banking consolidation pursued

The Evolution of Brazil’s Banking System    211 by the Brazilian government at the turn of the twenty-​first century was beneficial in terms of financial stability. 36

10.8.  Banks before and after the Great Recession (2006–​2015) The post-​hyperinflation bank restructuring process described in the preceding section exhibits a decisive turning point regarding public banks’ share of total credit at the outset of the recent financial crisis. As depicted in Figure 10.2, the share of public banks in total lending shrunk up to 2008, at which point this trend reversed, up to an impressive share of 57% in 2015. To capture the magnitude of the change in government policy toward public banks’ participation, it is important to describe what led to the reversion of the banking consolidation process. As is to be expected, the key players in this phenomenon were the federal government–​owned banks (Banco do Brasil, Caixa Econômica Federal, and BNDES37), since state-​owned banks were no longer significant after the PROER and PROES programs. More important, BNDES—​whose main objective is to supply earmarked credit—​performed a leadership role in the process. As the Great Recession hit, the federal government decided to use the federal banks’ balance sheets expansion as an instrument of countercyclical economic policy (De Bolle 2015; Jäger 2012). The expansion was unprecedented; the total balance of the public banks with the Treasury was only 0.5% of GDP in 2007, whereas it increased almost threefold in 2008 (to 1.4% of GDP) and then by more than threefold in 2009 (to 4.6% of GDP). The balance reached 9.71% of GDP in 2013 and kept increasing, up to 9.9% of GDP as of 2014 Treasury data until July.

Share of total lending

100% 75% 50% 25% 0% 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 Public

Private domestic

Private foreign

Figure 10.2.  Participation in total lending by bank ownership (1988–​2015). Source: Banco Central do Brasil, Time Series System, series #21300, #21301, and #21302. Notes: Original series start from 2011. To construct the series for the entire period in which the data is available, we used BCB Metadata instructions available in the Time Series System of the BCB under each series number.

212    Gustavo S. Cortes and Renato L. Marcondes BNDES alone accounted for significant shares of these figures each year. For example, in 2014 BNDES held with the Treasury an amount equivalent to 6.8% of GDP according to BNDES data. The money transferred to BNDES was used to grant earmarked credit operations, as evidenced by both the levels and growth rates depicted in Figure 10.3. The highest growth rate in disbursements was observed, indeed, in 2008, as if to confirm the countercyclical nature of the public credit-​led expansion. However, critics point to the high fiscal cost and the inefficiency of BNDES-​subsidized loans, since they are mainly directed to large firms that could have as easily found funding in domestic or international private capital markets.38 The disproportionate magnitude of BNDES loans—​reaching one-​quarter of Brazilian GDP in 2010—​has attracted the attention of political economy scholars for both the positives and the “dark side” of national development banks. 39 With the benefit of hindsight, the public bank-​led credit boost seems to have gone too far in the case of Brazil, as the fiscal cost of such operations is rearing its head in the current Brazilian crisis that began in 2014. In terms of market concentration, the end of 2008 saw a major event, when two leading private banks merged into the largest private bank in Brazil.40 The share of total lending of the top four banks rose from 60% to 70% between mid-​2008 and mid-​2009, and this level has increased steadily to 75% by the end of 2015 (BCB 2015, Figure 4.3.2). These levels place Brazil’s banks as one of the most concentrated banking sectors in Latin America, together with Mexico’s, where the top three banks have retained nearly 70% of total lending since 2002 (Corbae and D’Erasmo 2015, Figure 1). Another aspect to analyze in the context of the recent financial crisis is whether the Brazilian banking system as a whole was also negatively affected by the Great Recession. In other words, how well did banks in Brazil endure the crisis? The emerging market economies—​including those of Latin America—​are deemed by analysts to have resisted the crisis relatively well compared with other crises in the past.41 This seems to be true

A. BNDES disbursements: (%) of GDP

B. BNDES disbursements: growth rate 80%

25% 20%

40%

15% 10%

0%

5% –40%

0% 1995

2000

2005

2010

2015

1995

2000

2005

2010

2015

Figure 10.3.  Total earmarked credit: Levels and growth rates (1995–​2015). Source: Banco Central do Brasil, Time Series System, series #7415. Notes: Panel A shows the amount of earmarked credit as a percentage of GDP. Panel B shows the annual growth rate of earmarked credit.

Liquidity ratio index

The Evolution of Brazil’s Banking System    213

4 3 2 1 0 2006

2007

2008

2009

2010 Public

2011

2012

Private National

2013

2014

2015

Private Foreign

Figure 10.4.  Financial fragility of the banking system: Liquidity ratio index (2006–​2015). Source: Banco Central do Brasil, Time Series System, series #21850, #21851, and #21852. Notes: The Liquidity Ratio Index (LRI) aims to ensure that a bank maintains an adequate level of unencumbered, high-​quality liquid assets that can be converted into cash to meet its liquidity needs for a 30-​calendar-​day time horizon under a significantly severe liquidity stress scenario. The dashed horizontal line represents the safety lower bound for banks’ liquidity (LRI = 1). The first shaded area represents the Great Recession (2008–​2009) period according to NBER business cycles dating, and the second shaded area represents the Brazilian crisis period, defined as the year in which GDP growth was negative (2015).

for Brazil and its banking system. Several measures of financial stability suggest that banks in Brazil have faced the crisis with a relatively stable level of financial soundness.42 For example, Central Bank regulatory data on the banking system’s liquidity (Liquidity Ratio Index, LRI) show that banks went through some degree of liquidity shortage during the Great Recession, but their good conditions prior to the crisis ensured that the safety lower bound (LRI = 1) was met on average by public and private foreign banks throughout the entire time span (see Figure 10.4).43 The only exception was October 2008, when the LRI of private domestically owned banks surpassed the lower limit of 1, although the safety level was recovered in the next month. Overall, the evidence shows that the Brazilian banking system is solid and performed well during the international financial crisis. However, it is possible to detect a decreasing long-​run trend on the LRI since at least 2010, especially for public banks. It remains to be seen whether the proven resilience of the Brazilian banking system will endure the domestically generated fiscal crisis that began in 2014 and intensified in 2015.

10.9. Conclusion The financial sector has enjoyed expressive participation in the Brazilian economy since the nineteenth century. Non-​banking credit shrank and increasingly gave rise to banking credit at the turn of the century. While foreign banks were an important source of funding in the pre–​World War I period, it was the domestic banking system that became more significant for funding economic activity after that. The Great Depression of

214    Gustavo S. Cortes and Renato L. Marcondes the 1930s did not cause bank defaults in Brazil, but induced a significant number of institutional reforms in the financial system throughout subsequent decades. Banks’ adaptations to these institutional changes allowed their expansion in the postwar period. The reforms introduced by the military regime in the late 1960s spurred strong credit expansion and increased concentration in the banking system. Inflationary acceleration modified the incentives structure and consequently the activity of banks; credit supply was reduced, and was substituted by the intermediation of sovereign debt bonds. This increased the participation of the financial sector in the national income. The banking crisis caused by 1994’s Plano Real (Real Plan) and its defeat of hyperinflation allowed the Central Bank of Brazil to accelerate the financial consolidation process and to correct inefficiencies such as the excessive participation of state-​owned banks and their politically guided objectives. Privatization by domestic and—​especially—​ foreign groups occurred in line with the trend observed across Latin America in the transition to the twenty-​first century, even though this process occurred with less intensity in Brazil relative to similar countries such as Argentina and Chile. In recent years, Brazil has consolidated its banking system as a solid and sophisticated performer relative to its Latin American neighbors.

Notes 1. We are grateful for helpful comments from Teresa Marques, Gail Triner, John Turner, and the editors. All errors are our responsibility. 2. Previous work of a similar nature includes that of Goldsmith (1986) and Levy (1994) for the Brazilian banking system, and more recently Turner (2014) for the British banking system. 3. Commerce in the Rio da Prata was an alternative route for the supply of Spanish currency to São Paulo and the entire colony itself (Canabrava 1984). 4. Authors’ translation from Alcântara Machado (1980, 143). 5. The indivisibility of real property was one of a wider set of constraints. It was compounded by the mandated divisibility of personal estates (Triner 2015, Chapters 2–​3). 6. Authors’ translation. 7. Respectively, the Charter (Alvará) of May 5, 1810, and the Law of October 24, 1832. 8. Nevertheless, Brazil established its first banks earlier than other major Latin American countries (e.g., Argentina and Mexico), and had built a relatively more developed banking network by around the mid-​nineteenth century (Marichal 1994). 9. Charter (Alvará) of October 12, 1808. 10. Decree 2.723 of 1861. 11. The papelistas were inspired by the English Banking School and advocated more freedom (plurality) of issuance without constraining the supply of money to fixed gold reserves or convertible currencies. The metalistas were followers of the Currency School and preferred a conversion rate fixed by metallic reserves or convertible currencies. 12. Decree 19,525/​1930 and Decree 21,499/​1932. 13. The first government of President Getúlio Vargas—​known as the Estado Novo—​created in 1945 the SUMOC (Superintendência de Moeda e Crédito), by decree 7,293/​1945, to be the regulatory agency of the Brazilian financial system. SUMOC is considered to have

The Evolution of Brazil’s Banking System    215 been the embryo of the Central Bank of Brazil and edited a variety of regulations and distortionary norms into the banking system, such as a nominal interest rate cap paid to depositors at 12%, depending on the type and maturity of the investment. 14. For instance, banks included the differential between desired and permitted interest rates in the price of other banking services. They also required positive balances of checking accounts within a certain threshold in order to grant loans. Sochaczewski (1993, 36) analyzes banks’ balance sheets to calculate banks’ effective rates (paid and charged), reaching 30%–​40% by the beginning of the 1960s. Even with negative interest rates, banks were able to maintain their profitability by conciliating a strong demand for funds by the productive sector with the pursuit of demand deposits among savers. However, this favored the establishment of an informal, parallel shadow banking market, and the profit reinvestment of firms. 15. This emphasizes the process of banking concentration through mergers and acquisitions (M&As) and the disappearance of small institutions. The Banco do Brasil increased its number of branches from 260 to 600 and the Caixa Econômica Federal from 150 to 450 between 1945 and 1964. In 1946, BB alone had more than one-​fifth (21.6%) of demand deposits (Goldsmith 1986, 270). 16. Law 1,628/​1952. 17. Law 2,973/​1956. 18. Law 4,595/​1964. 19. Law 4,357/​1964. 20. The Societies of Credit, Funding, and Investment (Sociedades de Crédito, Financiamento e Investimento) known as financeiras were created by decree 7,583/​1945. Even though they were not able to take demand deposits at first, they were allowed to receive deposits from their associates (Decree 9,603/​1946). They began their business in short-​term lending (6–​24 months) through participation funds, from which their associates were depositors. 21. Law 4,380 from August 21, 1964, established the SFH and the National Bank of Housing (Banco Nacional da Habitação, BNH). 22. Law 4,728/​1965. 23. Other measures favoring banking concentration were the requirement of minimum levels of capital and reserves, as well as interest rate controls (Carvalho et al. 1989; Tavares and Carvalheiro 1985). 24. For a precise definition of government-​led financial repression, see, e.g., Mathieson and McKinnon (1981). 25. The segmented model that prevailed until 1988 was more similar to the US banking model, as proposed by the Glass-​Steagall Act. 26. Barbosa (1995), Minella (1996), and Cardim de Carvalho (1998). Universal Banks in Brazil are called “multiple” banks. 27. Earmarked credit refers to operations directed to specific types of debtors, such as in BNDES loans. We exclude them in calculating the average spread because interest rates charged in earmarked loans are defined by the government, rather than being determined by market conditions as is the case in standard non-​earmarked credit operations. 28. Baer and Nazmi (2000) detail how the currency crisis and the consequent increase of the baseline interest rate (SELIC) to extremely high levels affected banks’ financial conditions. 29. Evidence of the relatively less intense privatization process in Brazil is that its two largest banks in terms of assets—​then (1995) and now (2015)—​are still public (Banco do Brasil

216    Gustavo S. Cortes and Renato L. Marcondes and Caixa Econômica Federal), but are owned by the federal government rather than any state government. 30. PROER (established by Resolution 2.212/​ 95) was the Programa de Estímulo à Reestruturação e ao Fortalecimento do Sistema Financeiro Nacional, or Program of Incentives to Strengthen the National Financial System. PROES (Provisional Measure 1.514/​96) stands for Programa de Incentivo à Redução do Setor Público Estadual na Atividade Bancária, or Program of Incentives to Reduce States’ Participation in Banking Activity. 31. RAET was the Regime de Administração Especial Temporária, Temporary Special Administration Regime. 32. Banks like HSBC, ABM Amro, and Santander took advantage of this period to enter the Brazilian market. 33. “Banco Santander Wins Banespa Sale, Topping Rivals with $3.6 Billion Bid,” Wall Street Journal, November 12, 2000. 34. For an analysis of the heterogeneous competition structure of the recent Brazilian banking system, see Coelho et al. (2013). 35. Micco et al. (2007) find a positive impact of private bank share on bank productivity for a panel of countries including Brazil, and Nakane and Weintraub (2005) find the same result using exclusively Brazilian data. Cardim de Carvalho (2002) and Guimarães (2002) find no productivity impact of the increase in foreign private banks in Brazil, whereas CGFS (2004, 2005)  reports positive impacts of foreign participation on bank productivity in general. Some caveats pertaining to the empirical evaluation of privatization events are discussed in Cardim de Carvalho et al. (2009, 885–​886). 36. On the positive effects of market concentration on financial stability, see Beck et al. (2006) for country panel evidence, and Chang et al. (2008) for the Brazilian case. On the positive impact of foreign bank ownership in three Latin American economies (Argentina, Chile, Colombia), see Crystal et al. (2002). Crystal et al. (2002) do not analyze Brazil, due to the low share of foreign banks in the pre-​liberalization period (i.e., there is an insufficient mixture of foreign and domestic banks), a fact that would hamper their identification. 37. The BNDE (National Economic Development Bank) eventiually became the BNDES (National Economic and Social Development Bank) to reflect a broadening of its functions in the development sphere. 38. Over the period 2001–​2015, the share of total lending going to large firms was never less than 60%, reaching more than 80% in 2008, the first year of the crisis. See De Bolle (2015). 39. The upside of a government-​owned bank or a national development bank can be rationalized in a general context by the argument of poor institutional frameworks and incomplete financial markets observed in emerging market economies (e.g., Andrianova 2012). In terms of downsides, in the specific Brazilian case, Lazzarini et al. (2015) show evidence that firms donating to winning political candidates are more likely to get BNDES funding. Carvalho (2014) also finds evidence that politicians in Brazil use bank lending to shift employment toward politically useful regions and away from less expedient regions. 40. DealBook, “Brazilian Banks Itaú and Unibanco to Merge,” New York Times, November 3, 2008. 41. Didier et al. (2012) find that emerging market economies suffered a growth collapse comparable to that seen in advanced economies, contrary to common perception. However, they show that the performance of emerging markets in comparison to those same

The Evolution of Brazil’s Banking System    217 markets in past crises has improved; they recovered faster due to the availability of more and better policy instruments. 42. A common measure inversely related to financial fragility is the Z-​Score, also known as Distance to Default. The calculation of annual country-​level Z-​Score is described in detail by Adrianova et al. (2015). A high Z-​Score means that a country’s entire banking system is distant from insolvency (i.e., the banking system is less fragile). Brazilian Z-​Scores are relatively well placed when compared to other similar countries in Latin America over the period 1998–​2012. 43. The liquidity ratio index is calculated by the Banco Central do Brasil and aims to ensure that a bank maintains an adequate level of unencumbered, high-​quality liquid assets that can be converted into cash to meet its liquidity needs for a 30-​calendar-​day time horizon under a significantly severe liquidity stress scenario, such as deposit runoff and potential losses on liquid assets and derivative positions. This ratio can be expressed by the following formula: Brazilian Liquidity Ratio for Banks = (Unencumbered Liquid Assets)/​ (Stressed Cash Flow), provided that the Stressed Cash Flow represents the liquidity level that each institution needs to keep so as to withstand funding volatility and losses under market stress (BCB 2015).

References Adams, A. A. 2005. “The Caixa Econômica:  A Social and Economic History of Popular Banking in Rio de Janeiro, 1821–​1929.” PhD dissertation, Department of History, Harvard University. Aguiar, P. 1960. “Bancos no Brasil colonial: Tentativas de organização bancária em Portugal e no Brasil até 1808.” Salvador: Progresso. Alcântara Machado, J. 1980. Vida e morte do bandeirante. São Paulo: Itatiaia/​EdUSP. Andrianova, S. 2012. “Public Banks and Financial Stability.” Economics Letters 116 (1): 86–​88. Andrianova, S., B. Baltagi, T. Beck, P. Demetriades, D. Fielding, S. Hall, S. Koch, R. Lensink, J. Rewilak, and P. Rousseau. 2015. “A New International Database on Financial Fragility.” Working Paper, University of Leicester. Baer, W., and N. Nazmi. 2000. “Privatization and Restructuring of Banks in Brazil.” The Quarterly Review of Economics and Finance 40 (1): 3–​24. Barbosa, F. H. 1995. El sistema financiero brasileño. Fundación Konrad Adenauer/​Centro Interdisciplinario de Estudios sobre el Desarrollo Latinoamericano. BCB. 2015. Relatório de estabilidade financeira. Brasilía: Banco Central do Brasil. Beck, T., J. M. Crivelli, and W. Summerhill. 2005. “State Bank Transformation in Brazil—​ Choices and Consequences.” Journal of Banking & Finance 29 (8): 2223–​2257. Beck, T., A. Demirgüç-​Kunt, and R. Levine. 2006. “Bank Concentration, Competition, and Crises: First Results.” Journal of Banking & Finance 30 (5): 1581–​1603. Brito, J. R. 1923. A economia brasileira no alvorecer do século XIX, Vol. 2. Salvador: Progresso. Canabrava, A. P. 1984. “O comércio português no Rio da Prata (1580–​1640).” São Paulo: Itatiaia/​ EdUSP. Cardim de Carvalho, F. J. 2002. “The Recent Expansion of Foreign Banks In Brazil:  First Results.” Latin American Business Review 3 (4): 93–​119. Cardim de Carvalho, F. J. 2013. “The Real Stabilisation Plan and the Banking Sector in Brazil.” Banca Nazionale del Lavoro Quarterly Review 51 (206): 291–​326.

218    Gustavo S. Cortes and Renato L. Marcondes Cardim de Carvalho, F. J., L. F. de Paula, and J. Williams. 2009. “Banking in Latin America.” In The Oxford Handbook of Banking, edited by A. N. Berger, P. Molyneux, and J. O. S. Wilson, 868–​902. Oxford: Oxford University Press. Cardoso, J. L. 2010. “A New Contribution to the History of the Bank of Brazil (1808–​1829): Chronicle of a Foretold Failure.” Revista Brasileira de História 30 (59): 167–​192. Carvalho, D. 2014. “The Real Effects of Government-​Owned Banks:  Evidence from an Emerging Market.” Journal of Finance 69 (2): 577–​609. Carvalho, L. C.  P., G. L. Garófalo, and J. M.  G. Del Mônaco. 1989. Concentração e desconcentração dos bancos comerciais no Brasil. São Paulo:IPE-​USP. CGFS. 2004. “Foreign Direct Investment in the Financial Sector of Emerging Market Economies.” Committee on the Global Financial System. Report submitted by a working group established by the CGFS, Basel. CGFS. 2005. “Foreign Direct Investment in the Financial Sector: Experiences in Asia, Central and Eastern Europe and Latin America.” Committee on the Global Financial System. Summary of Follow-​up, CGFS, Basel. Chang, E. J., S. M. Guerra, E. J. A. Lima, and B. Tabak. 2008. “The Stability-​Concentration Relationship in the Brazilian Banking System.” Journal of International Financial Markets, Institutions and Money 18 (4): 388–​397. Claessens, S., and N. Horen. 2014. “Foreign Banks:  Trends and Impact.” Journal of Money, Credit and Banking 46 (1): 295–​326. Coelho, C. A., J. M. P. De Mello, and L. Rezende. 2013. “Do Public Banks Compete with Private Banks? Evidence from Concentrated Local Markets in Brazil.” Journal of Money, Credit and Banking 45 (8): 1581–​1615. Contador, C. R. 1974. Mercado de ativos financeiros no Brasil:  Perspectiva histórica e comportamento recente. Instituto Brasileiro de Mercado de Capitais. Corbae, D., and P. D’Erasmo. 2015. “Foreign Competition and Banking Industry Dynamics: An Application to Mexico.” IMF Economic Review 63 (4): 830–​867. Cortes, G. S., R. L. Marcondes, and M. D. M. Diaz. 2014. “Mortgages for Machinery: Credit and Industrial Investment in Pre-​World War I Brazil.” Financial History Review 21 (2): 191–​212. Costa, F. N. 1988. “Banco do Estado: O caso Banespa.” PhD dissertation, Institute of Economics, State University of Campinas. Costa Neto, Y. C. 2004. Bancos oficiais no Brasil: Origens e aspectos do seu desenvolvimento. Brasilía: Banco Central do Brasil. Crystal, J., B. G. Dages, and L. S. Goldberg. 2001. “Does Foreign Ownership Contribute to Sounder Banks in Emerging Markets? The Latin American Experience.” Federal Reserve Bank of New York’s Current Issues in Economics and Finance 1 (8): 1–​6. De Bolle, M. 2015. “Do Public Development Banks Hurt Growth? Evidence from Brazil.” Peterson Institute for International Economics Policy Brief, No. PB15-​16, September. Didier, T., C. Hevia, and S. L. Schmukler. 2012. “How Resilient and Countercyclical Were Emerging Economies during the Global Financial Crisis?” Journal of International Money and Finance 31 (8): 2052–​2077. Domanski, D. 2005. “Foreign Banks in Emerging Market Economies:  Changing Players, Changing Issues.” BIS Quarterly Review, December: 69–​81. Espírito Santo, C. C. 2003. “A economia da palavra: Ações de alma nas Minas setecentistas.” Master’s thesis, Department of History, University of São Paulo. Faria, S. C. 1998. A colônia em movimento:  Fortuna e família no cotidiano colonial. Rio de Janeiro: Nova Fronteira.

The Evolution of Brazil’s Banking System    219 Flory, R. J. D. 1978. “Bahian Society in the Mid-​Colonial Period: The Sugar Planters, Tobacco Growers, Merchants, and Artisans of Salvador and the Reconcavo 1680–​1725.” PhD dissertation, University of Texas at Austin. Fragoso, J. L. R. 1992. Homens de grossa aventura: Acumulação e hierarquia na praça mercantile do Rio de Janeiro (1790–​1830), Vol. 1. Rio de Janeiro: Arquivo Nacional, Orgão do Ministério da Justiça. Franco, B. S. 1984. Os Bancos do Brasil: Sua história, defeitos da organização atual e reforma do sistema bancário. Brasilía: Universidade de Brasília. Franco, G. H. B. 1983. “Reforma monetária e instabilidade durante a transição republicana.” Master’s thesis, PUC-​Rio, BNDES. Franco, G. H. B. 2014. “A primeira década republicana.” In A ordem do progresso: Dois séculos de política econômica no Brasil, edited by M. P. Abreu, 29–​44. Rio de Janeiro: Elsevier. Goldberg, L. S. 2009. “Understanding Banking Sector Globalization.” IMF Staff Papers 56 (1): 171–​197. Goldsmith, R. W. 1986. Brasil 1850–​1984: Desenvolvimento financeiro sob um século de inflação. São Paulo: Banco Bamerindus do Brasil. Guimarães, C. G. 2016. “Foreign Direct Investment in Imperial Brazil and the Activities of British and Portuguese Banks: Colonial Banking versus Imperial Banking?” In Colonial and Imperial Banking History, edited by N. Valério, 39–​62. London: Routledge. Guimarães, P. 2002. “How Does Foreign Entry Affect the Domestic Banking Market? The Brazilian Case.” Latin American Business Review 3 (4): 121–​140. IBGE. 2000. Sistema de Contas Nacionais:  Referência 2000 (Anual). Instituto Brasileiro de Geografia e Estatística. Available at Ipeadata, www.ipeadata.gov.br, accessed August 16, 2016. Jäger, M. 2012. Brazilian Banking Sector: A View from 30,000 feet. Frankfurt: Deutsche Bank Emerging Markets Research Briefing. Joslin, D. 1963. A Century of Banking in Latin America. Oxford: Oxford University Press. Lazzarini, S. G., A. Musacchio, R. Bandeira-​de Mello, and R. Marcon. 2015. “What Do State-​ Owned Development Banks Do? Evidence from BNDES, 2002–​09.” World Development 66: 237–​253. Levy, M. B. 1980. “O Encilhamento.” In Economia brasileira: Uma visão histórica, edited by P. Neuhaus, 191–​255. Rio de Janeiro: Campus. Levy, M. B. 1994. “El sector financiero y el desarrollo bancario en Rio de Janeiro (1850–​1888).” In La formación de los bancos centrales de España y América Latina (Siglos XIX y XX), Vol. II: Suramérica y el Caribe, edited by P. Tedde and C. Marichal, 61–​83. Madrid: Banco de España. Maia, G. 1999. “Restructuring the Banking System:  The Case of Brazil.” BIS Policy Papers 6: 106–​123. Marichal, C. 1994. “Modelos y sistemas bancarios en América Latina en el Siglo XIX (1850–​ 1880).” In La formación de los bancos centrales de España y América Latina (Siglos XIX y XX), Vol. I: España y México, edited by P. Tedde and C. Marichal, 131–​157. Madrid: Banco de España. Marques, T. C. N. 2014. “Eram os senhores de engenho caloteiros? Reflexões sobre o crédito e os direitos de propriedade no mundo Luso.” História Econômica & História de Empresas 17 (1): 147–​175. Mathieson, D. J., McKinnon, R. 1981. “How to Manage a Repressed Economy.” Princeton Essays in International Finance 145: 1–​26.

220    Gustavo S. Cortes and Renato L. Marcondes McQuerry, E. 2001. “Managed Care for Brazil’s Banks.” Economic Review-​Federal Reserve Bank of Atlanta 86 (2): 27. Micco, A., U. Panizza, and M. Yanez. 2007. “Bank Ownership and Performance: Does Politics Matter?” Journal of Banking & Finance 31 (1): 219–​241. Minella, A. C. 2002. “Empresariado financeiro:  Elementos discursivos na decada de 1980.” In História de empresas e desenvolvimento econômico, edited by T. Szmrecsányi and R. Maranhão, 165–​188. São Paulo: Hucitec/​EdUSP. Nakane, M. I., and D. B. Weintraub. 2005. “Bank Privatization and Productivity: Evidence for Brazil.” Journal of Banking & Finance 29 (8): 2259–​2289. Peláez, C. M., and W. Suzigan. 1981. História monetária do Brasil. Brasilía: Universidade de Brasília. Portugal. 1870. Ordenações Filipinas:  Aditamentos. Instituto Filomático do Rio de Janeiro. Available at http://​www1.ci.uc.pt/​ihti/​proj/​filipinas/​l4inda.htm. Russell-​Wood, A. J. R. 1968. Fidalgos and Philanthropists: The Santa Casa da Misericórdia of Bahia, 1550–​1755. Berkeley: University of California Press. Saes, F. A. M. 1997. “Bancos e banqueiros: A consolidação dos grandes bancos de âmbito regional no Brasil, 1930–​1964.” Annals of the 2nd Brazilian Meeting of Economic History, ABPHE, 106–​128. Sampaio, A. C. J. 2000. “Na curva do tempo, na encruzilhada do império: Hierarquização social e estratégias de classe na produção da exclusão (Rio de Janeiro, c. 1650–​c. 1750).” PhD dissertation, Department of History, Fluminense Federal University. Saraiva, L. F. 2014. “O homem benfazejo: Bibliotheca Constitucional do Cidadão Brasileiro, Mercado Editorial, Cidadania e a Construção do Império Brasileiro, Rio de Janeiro 1831–​ 1832.” Outros Tempos 11 (18): 208–​232. Saraiva, L. F., and T. A. Oliveira. 2015. “A primeira Caixa Econômica do Rio de Janeiro: 1831–​ 1858, notas de pesquisa.” Annals of the 11th Brazilian Meeting of Economic History, ABPHE, 1–​26. Schulz, J. 2008. The Financial Crisis of Abolition. New Haven, CT: Yale University Press. Schwartz, S. B. 1985. Sugar Plantations in the Formation of Brazilian Society: Bahia, 1550–​1835. Cambridge: Cambridge University Press. Simonsen, M. H., and R. Campos. 1979. A nova economia brasileira. Rio de Janeiro: J. Olympio. Sochaczewski, A. C. 1993. Desenvolvimento econômico e financeiro do Brasil, 1952–​1968. São Paulo:Trajetória Cultural. Tavares, M. A.  R., and N. Carvalheiro. 1985. O setor bancário brasileiro:  alguns aspectos do crescimento e da concentração. São Paulo: IPE–​USP. Toneto, R., Jr. 1992. “Estado, bancos e acumulação financeira no Brasil: 1964–​1984.” Master’s thesis, Department of Economics, University of São Paulo. Topik, S. 2014. The Political Economy of the Brazilian State, 1889–​1930. Austin: University of Texas Press. Triner, G. 2000. Banking and Economic Development: Brazil, 1889–​1930. London: Palgrave. Triner, G. D. 2015. Mining and the State in Brazilian Development. London: Routledge. Turner, J. D. 2014. Banking in Crisis: The Rise and Fall of British Banking Stability, 1800 to the Present. Cambridge: Cambridge University Press. Villela, A. 1999. “The Political Economy of Money and Banking in Imperial Brazil, 1850–​1870.” PhD dissertation, Department of Economic History, London School of Economics and Political Science.

Chapter 11

Br azil’s M acroe c onomi c P olicy Insti t u t i ons , Quasi-​S tagnat i on, a nd t he Interest Rat e –​ E xchange Rat e  T ra p Luiz Carlos Bresser-​P ereira

11.1.  The Conundrum of Ongoing Quasi-​Stagnation In a 2007 book, I argued that the Brazilian economy had been deindustrializing and quasi-​stagnant since 1981, due, first, to the major foreign debt crisis and high inflation, and second, from the early 1990s on, to a macroeconomic trap of high interest rates and an overvalued currency over the long term, which discouraged investment and hampered economic growth.1 However, at the time of the book’s publication it seemed to make no sense, in light of the satisfactory growth rates between 2006 and 2010, which were driven by the large increase in the prices of exported commodities (the “China effect”). At that time it was common for distinguished Brazilian or foreign economists, whether liberal or developmental, and for representatives of the national and international financial system to declare that Brazil had “resumed growth,” was one of the BRICs (Brazil, Russia, India, and China), and was destined for greatness. However, the country was benefiting only from a boom in commodities, as the low growth rates between 2011 and 2014 and a major recession in 2015–​2016 soon confirmed. In fact, the Brazilian economy has been quasi-​stagnant since 1981. The average rate of per capita gross domestic product (GDP) growth between 1981 and 2014 was 1.2% per annum; if we exclude an exceptionally negative period (the 1980s, when the country

222   Luiz Carlos Bresser-Pereira stagnated due to a major financial crisis) and the commodity boom (2006–​2010), the rate has been even lower: 0.78% per annum. What is the explanation for deindustrialization and these low growth rates, which, for a developing country aiming to “catch up,” represent quasi-​stagnation? Why did an economy that between 1950 and 1980 grew at a per capita rate of 4.5% per year then grow so slowly from 1981?2 The reason for stagnation in the 1980s is well known: it was the foreign debt crisis, which resulted from the policy of growth with foreign indebtedness (“foreign savings”) adopted by the Geisel government (1974–​1979), and from the high and inertial inflation this crisis unleashed, insofar as the government was compelled to undertake two maxi-​devaluations (1981 and 1983) in an economy that had been formally indexed from 1964.3 The oft-​heard alternative explanation, that the exhaustion of the import-​substitution model explains the stagnation, has little to support it and arguably is used more to make a point with ideological intent. That model was exhausted as early as the start of the 1960s; it is true that import tariffs remained high after the model was abandoned, but the important point is that in 1967 Brazil began a highly successful period of growth led by the export of manufactured goods, on which I offer some numbers in the following. But why, after the 1994 Real Plan had brought inflation under control, did the Brazilian economy continue to grow so slowly? Why did the investment and savings rate continue to be so low? To answer these questions we need significant new historical facts. Four simple and decisive new facts meet this need: (1) the fall of public savings with the debt crisis, (2) the exhaustion of an unlimited supply of labor due to a fall in fertility rates, (3) the 1990 trade liberalization that dismantled the mechanism that neutralized the Dutch disease, and (4)  the extremely high rates of interest under the Real Plan. These four historical facts reduced both public and private investment and pushed the Brazilian economy into long-​term quasi-​stagnation. Before the major recession of 2015–​2016 there was already some uneasiness among the Brazilian economic and political elites, and with good reason: an appreciated exchange rate except during financial crises, when it would depreciate; the basic interest rate set by the Central Bank at a very high level; unsatisfactory profit rates in manufacturing industry; premature deindustrialization; low savings and investment rates; and low growth rates, far below what would be needed for “catchup.” What is the explanation for these elites’ inability to solve these problems? Why might they be reluctant to frame a development project starting from the aim of overcoming the macroeconomic trap of high interest rates and an overvalued currency? I wish to suggest that this is essentially for two fundamental reasons. The first is that the elites, along with the general public, have lost the idea of nation; this leads them to be unduly swayed by the recommendations and pressures of wealthy countries, without subjecting these views to critical interrogation. The second is that Brazilian society as a whole is strongly characterized by a preference for immediate consumption. More specifically, politicians, businessmen, economists, and economic journalists—​whether liberal or developmental, left or right—​refuse to lower the basic interest rate on the justification that it is required to control inflation, and refuse to depreciate the exchange rate because this will cause, in the short term, a

Brazil’s Macroeconomic Policy Institutions    223 temporary reduction in revenues and an increase in inflation. Moreover, such leaders have proved unable to increase the state’s investment capacity—​whether because those on the right see public savings and investments as unnecessary, if not dangerous, or because both the left and the right prefer to increase social expenditures that produce electoral dividends. Brazilians have been subject to several economic disappointments since the 1985 transition to democracy. The first came in the José Sarney administration (1985–​1989), with the collapse of the Cruzado Plan in 1987, a major setback that demonstrated that the opposition that had fought and won over the military regime lacked a project to promote growth and development. In its place was an unsophisticated Keynesianism, characterized by fiscal and exchange rate populism (i.e., high fiscal and current account deficits). Besides an economic crisis, the collapse of the Cruzado Plan also caused the demoralization of the politicians who had led the transition to democracy, and thus paved the way for the election of a young and unknown politician, Fernando Collor de Mello. Collor, following the global pressures of the time (the Washington Consensus), swiftly switched the economic policy regime from a developmental regime (in place in various forms since 1930, first under Getúlio Vargas and later under the military) to a liberal or neoliberal one, through the opening of trade and finance. In 1994, following the impeachment of Collor and subsequent assumption of the presidency by Itamar Franco, the Real Plan—​a heterodox stabilization plan based on the theory of inertial inflation—​succeeded in controlling inflation. However, this major success was followed by a second disappointment, during the Fernando Henrique Cardoso administration (1995–​2002). After the Brady Plan (1990) had resolved the financial crisis of the 1980s, and the Real Plan (1994) had very effectively controlled high inflation, many expected that the economy would commence rapid growth. What came instead were the second and third phases of the installation of the liberal economic policy regime; in 1995, with the privatization and denationalization of monopolist public services, and in 1999, with the adoption of a liberal-​orthodox macroeconomic system where the basic interest rate was kept very high to attract capital. At the same time, the  crawling peg system was abandoned,  while the exchange rate floated and remained highly overvalued, with state expenditures increasing substantially between 1995 and 1998. The outcome of this exchange rate and fiscal populism was a major financial crisis—​a currency crisis—​at the turn of 1999, while growth rates were mediocre.4 The third disappointment came during the government of the Partido dos Trabalhadores (PT), a social-​democratic political party that was in office from January 2003 to April 2016. By criticizing neoliberal policymaking, the PT created an opportunity for economic development, but this did not materialize. The PT—​which had at some points defined itself as “social developmental”5—​was relatively successful with its social commitment, but not with its developmental ambition. It was unable to change the liberal policy regime, or to form a developmental class coalition associating the industrial bourgeoisie with workers and the public bureaucracy, and, thus, it failed to lead the country into resumed growth (Boito Jr. 2012; Boito Jr. and Berringer 2016; Bresser-​ Pereira 2017; Bresser-​Pereira and Diniz 2016). The PT’s great achievement was to secure

224   Luiz Carlos Bresser-Pereira social inclusion, which occurred due to the substantial increase in the minimum wage and the expansion of cash transfers to the poor, allowing a significant portion of the population access to mass consumption. In the 13 years that the PT was in office, the liberal policy regime was not altered. Lula (2003–​2010) did not attempt to; Dilma Rousseff (2011–​2016) tried by sharply reducing the interest rate in July 2011, but a slight rise in inflation and a large protest from rentiers and financiers—​the major beneficiaries of the neoliberal policy framework—​led her to stop. Thus, the basic interest rate returned to 6%–​7% yearly in real terms, and the long-​term overvaluation of the exchange rate was not resolved; on the contrary, it was exacerbated. In January 2003 the Lula administration inherited a highly depreciated exchange rate from the previous government, R$5.30 per dollar (at third quarter 2016 prices)6—​a consequence of a second currency crisis in the Cardoso administration. Benefiting from such a highly devaluated real, the Lula administration allowed the national currency to appreciate greatly during his eight years in office, reaching R$2.20 per dollar at the end of 2010. Dilma Rousseff ’s successor government achieved some real devaluation, but the real remained highly overvalued. Figure 11.1 shows the last full exchange rate cycle (2002–​2014), where the real remains overvalued for seven years, between the second semester of 2007 and the first semester of 2014. As a consequence, the manufacturing industry first stopped exporting, and second, lost the domestic market to foreign competitors (i.e., deindustrialization—​see Figure 11.2), heavily indebting manufacturing companies. January 2015 saw the outbreak of a new financial crisis and a major recession—​this time not a currency crisis, but a financial crisis of manufacturing enterprises. Lula’s administration saw five years of satisfactory growth, driven by a rise in the price of commodity exports (a typical commodity boom), which, combined with 7.00 6.00 R$: US$

5.00 4.00 3.00 2.00 1.00

14 20 15 20 16 *

13

20

12

20

11

20

10

20

09

20

08

20

07

20

06

20

05

20

04

20

03

20

02

20

01

20

00

20

99

20

98

19

97

19

96

19

19

19

95

0.00

Real effective exchange rate (at 2016 3Q prices) Industrial equilibrium effective exchange rate (at 2016 3Q prices)

Figure 11.1.  Real exchange rate and the industrial equilibrium, 1996–​2016. Note: 2005 = 100. Source: Center for New Developmentalism/​EESP-​FGV.

Brazil’s Macroeconomic Policy Institutions    225 40.00% 35.00% 30.00% 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Figure 11.2.  Share of the manufacturing industry in GDP, 1947–​2015. Source: IBGE.

much-​needed distribution policies, expanded the domestic market. This involved a trade-​off for the manufacturing industry: it lost foreign markets due to the appreciation of the real, but gained a stronger domestic market. This was hailed as an achievement by the developmental defenders of the wage-​led strategy. But this kind of strategy only works when the country is closed to imports; it is of the import-​substitution model. Since the Brazilian economy is an open economy, which is supposed to be competitively integrated into global markets, this trade-​off was short-​lived. Within around three years, importers of manufactured goods organized themselves, and imported goods flooded the domestic market; as a result, the Brazilian manufacturing industry lost its domestic market, accelerating the deindustrialization process (on the deindustrialization of Brazil, see Bresser-​Pereira 2009; Nassif 2011; Oreiro and Feijó 2010). Figure 11.3 demonstrates indirectly this leakage of domestic markets to imports by comparing physical production of the manufacturing industry and retail sales. Figure 11.4 shows how exports of manufactured goods stagnated while exports of commodities continued to increase. When Dilma Rousseff assumed the presidency in January 2013, with the real exchange rate at R$2.50 per dollar (third quarter 2016 prices), she faced an impossible task. She did not have power to depreciate the real by more than 50%—​to R$3.80, the competitive equilibrium per dollar at that time. All that was achieved was a 20% depreciation in the first two years of the Rousseff administration, while the Central Bank lowered the interest rate substantially. However, manufacturing enterprises did not start investing. The interest rate required a high expected rate to make investments viable, but the overvalued national currency made local manufacturing firms non-​competitive, their expected rate of profit remaining very low, if not negative. The low rates of growth

220 200 180 160 140 120 100

Retail sales

Aug. 2016

Jan. 2016

Jun. 2015

Nov. 2014

Apr. 2014

Sep. 2013

Jul. 2012

Feb. 2013

Dec. 2011

Oct. 2010

May. 2011

Mar. 2010

Jan. 2009

Aug. 2009

Jun. 2008

Apr. 2007

Nov. 2007

Sep. 2006

Jul. 2005

Feb. 2006

Dec. 2004

May. 2004

Oct. 2003

Mar. 2003

Jan. 2002

Aug. 2002

80

Manufacturing industry

Figure 11.3.  Physical production of the manufacturing industry and retail sales, 2002–​2012. Note: January 2002 = 100. Source: IBGE—​Monthly Industrial Survey and Monthly Retail Survey. Observ: seasonal adjustment.

180 160 140 120 100 80 60

Basic products

Sep. 2016

Nov. 2015

Jan. 2015

Mar. 2014

May. 2013

Jul. 2012

Sep. 2011

Nov. 2010

Jan. 2010

Mar. 2009

May. 2008

Jul. 2007

Sep. 2006

Nov. 2005

Jan. 2005

Mar. 2004

May. 2003

Jul. 2002

Sep. 2001

Nov. 2000

Jan. 2000

Mar. 1999

May. 1998

Sep. 1996

Jan. 1995

Nov. 1995

20

Jul. 1997

40

Manufactured products

Figure  11.4.  Exports of manufactured goods against exports of basic products (quantum), 1995–​2016. Note: “Basic products” include primary goods and the output of the extractive industry. Source: Funcex.

Brazil’s Macroeconomic Policy Institutions    227 surprised the government, and the president made a last-​resort decision that amounted to a major mistake: she adopted an “industrial policy” involving a substantial reduction in taxes for a large number of manufacturing firms.7 Once again, industrial enterprises did not resume investing, because an industrial policy is no substitute for a competitive exchange rate and a reasonable interest rate. The country remained caged in the high interest/​overvalued currency trap; manufacturing business continued without net (of interest) expected profits. Besides, the governing party became involved in the major Mensalão corruption scandal. Thus, industrial entrepreneurs, who since 2003 had been called on by Lula and Dilma to form a developmental class coalition with workers, gave up and opened the way for the liberal hegemony of the rentier capitalists, including the traditional middle class and the financiers who manage the rentiers’ wealth. Despite the opposition of the economic elites, Dilma Rousseff was re-​elected at the end of 2014 with support clustered socioeconomically among the poor and geographically in the Northeast region. However, when she took office in January 2015, the economy was entering a recession, while inflation had risen to 8% a year, the primary surplus had deteriorated from 2% of GDP positive in 2013 to 0.6% negative in 2014, and the current account deficit reached 4.6% of GDP.8 The recession was triggered by falls in the international prices of the two major commodity exports (soybeans and iron ore) in the second semester of 2014, and, principally, by a Minskyan financial crisis caused by the fall in profits and the high indebtedness of the manufacturing industry after years of a highly overvalued exchange rate and high interest rates.9 Moreover, a second major scandal, this time involving involving Petrobrás, exploded at the end of 2014. Consequently, on assuming office for the second time, in January 2015, President Dilma faced an acute economic and political crisis—​a generalized loss of confidence. This was swiftly exacerbated by the failure of the liberal economists who entered the Finance Ministry at that point to appreciate the full depth of the crisis, assuming that it was merely a fiscal problem, and engaging in a major fiscal adjustment while the economy faced a deep recession. In fact, the increase in expenditures and the tax exemptions in 2013 and 2014, coupled with major drops in GDP (3.8% in 2015 and around 3.5% in 2016) and the subsequent fall in state revenues led Brazil into a large primary deficit. As a result of both the crisis and her own lack of political savvy, President Rousseff was impeached in April 2016. The radically liberal administration that replaced her failed—​ as its developmental predecessor had—​to understand the real origin of the quasi-​ stagnation that Brazil is facing from the Real Plan, namely the interest and exchange rate trap. Instead, it identified the fiscal problem as being the cause of the recession, when it was essentially its consequence.

11.2.  Four New Facts Setting aside the short-​term adjustment problem faced by the Brazilian economy, there are a number of important longer-​term questions. What are the new historical

228   Luiz Carlos Bresser-Pereira facts that keep the Brazilian economy growing so poorly—​that is, quasi-​stagnant? Why do financial advisors, whose forecasts are consolidated in the Focus Report of the Central Bank, expect GDP growth up to 2018 to reach a maximum of 2% per year? Of the four explanations that are most often proposed—​insufficient household savings, a low level of basic education, lack of strong institutions, and lack of investment in infrastructure—​only the last is useful. These problems are long-​standing; they are always being confronted and never satisfactorily resolved, but they haven’t held Brazil back from strong growth in the past. To explain this quasi-​stagnation, then, I propose four new historical facts: (1) the reduction in public savings and, therefore, the lessening of the state’s capacity to invest in infrastructure since 1980; (2) the end of the unlimited supply of labor; (3) a very high (although decreasing) interest rate level since the Real Plan; and, last and most important, (4) the large competitive disadvantage that Brazilian manufacturing businesses have faced since the trade opening in 1990, an episode that involved the dismantling of the anti–​Dutch disease mechanism. These four factors caused the fall in both public and private investment, and explain why the historical per capita growth rate from 1990 onward has been only a quarter of the rate between 1950 and 1980. This section considers these in turn.

11.2.1.  Public Savings As shown in Table 11.1, public savings reached high levels in the 1970s (an average of 3.9% of GDP), but plummeted in the 1980s and have remained negative since then; in the first decade of the 2000s they were negative by 2.8% of GDP. The fall in public savings originates from two policies pursued by the Geisel government in the second half of the 1970s: the use of the prices of state-​owned enterprises to control inflation, and the decision to grow with current account deficits which would be financed by “foreign savings.”

Table 11.1 Public Savings and Investments in Decades as Percentage of GDP (Averages from the 1970s to the 2000s) Public Savings

Investments

1970s

3.9

21.4

1980s

–​1.5

22.1

1990s

–​0.8

18.2

2000s

–​2.8

17.1

Source: Elaborated by the author from statistics provided by IBGE.

Brazil’s Macroeconomic Policy Institutions    229 On the first point, from the start of the military regime in 1964, the profits of state-​ owned enterprises had been used to finance government investment in infrastructure;10 however, 10  year later, using their prices to control inflation was a serious mistake, similar to the use of the exchange rate as an anchor to control inflation. This decision reduced the profits of state-​owned enterprises, and public savings fell. On the second point, the first OPEC oil shock in 1973 led all rich countries into recession. On taking office in the following year, President Geisel declared that Brazil would nevertheless continue to grow in accordance with his Second National Development Plan. This goal was pursued by accumulating current account deficits and making the Brazilian economy indebted in foreign currency (Bonelli and Malan 1976; Bresser-​ Pereira 1990; Coes 1995), a self-​defeating policy since it appreciates the national currency, involves substitution of foreign for domestic savings, and leads a country into recurrent currency crises. The average growth rate, which during the 1968–​ 1973 Brazilian “miracle” had been 11.3% yearly, fell to a still high rate, an average of 6.9% a year between 1974 and 1979, but at the cost of a high increase in foreign indebtedness (the foreign debt rose from US$6.4 billion in 1973 to nearly US$54 billion in 1980), which led the country into a major financial crisis and the stagnation of per capita income in the 1980s. With the second oil shock, in 1979, the United States dramatically raised interest rates, and countries indebted in foreign currency, including Brazil, found themselves in big trouble. The state was forced to bail out businesses that were highly indebted in foreign currency, which represented a second blow to the fiscal health of the country, besides the loss of revenues derived from state-​owned enterprises. As a consequence of these two mistakes, public savings became negative and the state’s capacity to invest declined, while Brazil faced for the next 10 years a severe currency crisis. From this moment on, the country has had serious difficulties financing required infrastructure projects, a difficulty yet to be resolved. Public savings recovered somewhat in the 1990s, but then in the first decade of the 2000s deteriorated further, for several reasons: first, the Brazilian government, captive to neoliberal thinking in the 1990s, privatized monopolistic state-​owned enterprises, whose profits had financed investment; second, since the 1985 transition to democracy, governments had given priority to social spending to the detriment of investment in infrastructure; and third, the engineering capacity that a developmental state must have to develop infrastructure projects was seriously damaged by the many years of low public investment. In light of the extreme economic inequality that prevailed at the close of the dictatorship, it is perhaps understandable that emphasis was placed on the social state over the developmental state, but the change went too far. The tax burden increased from 22% of GDP in 1985 to 36% by 2014, but, of this 14  percentage-​point increase, around 11 percentage points were accounted for by the social area (education, health care, social security, social assistance, and culture), and the rest in financing the high interest rates that the treasury pays to rentiers. Social spending is a fair and highly efficient way of increasing indirect incomes. In fact, this increase in social spending was the result of a momentous political agreement—​the 1977 Democratic Popular Pact—​that, besides calling for democracy, was committed to reducing social inequality. The fact is,

230   Luiz Carlos Bresser-Pereira however, that public investment lost the priority that it had had in the 1970s, and this is one reason for the subsequently lower investment and growth rates.

11.2.2.  The End of “Unlimited Labor” The second new fact that had a negative impact on investment and growth was demographic: the exhaustion of the “unlimited supply of labor” that exists in developing countries. According to the classical model of Lewis (1954), this depresses wages but keeps them sufficiently high to allow for the transfer of labor from agriculture to the manufacturing sector, which can pay low wages, while the productivity of the country increases. As enterprises benefit from low wages and plow the resulting profits into investment and technical progress, economic growth accelerates. This simple model explains some of the industrialization in developing countries, including Brazil. But fertility rates fell strongly in Brazil after the 1980s, resulting in a strong decrease in the labor supply in the early 2000s (when the country reached the “Lewis turning point”). This was the main cause of the sharp rise in formal employment that began at that moment, and is one of the reasons why wages began to increase faster than productivity in several industries.

11.2.3.  The Increase in Real Interest Rates The third new fact that explains Brazil’s long-​term quasi-​stagnation is an increase in real interest rates, which were very low if not negative in the 1970s, but became extremely high from the Real Plan on. It is true that the level of interest rates has been falling throughout the period since 1994, but it is still very high; in June 2015, it was 6% a year in real terms. What is the explanation for this? A common response states that high interest rates are required to control inflation. Certainly, when inflation is rising, an increase in the interest rate is the first thing that should be done. But monetary policy need not fluctuate around a 5% real rate of interest, as it currently does in Brazil; it can proceed around a midpoint of a 1%–​2% real rate of interest. The high interest rates in Brazil reflect the political power of rentier capitalists and financiers, who have a seigniorage of around 5%–​6% over GDP. Since the collapse of the Cruzado Plan (1987), rentier capitalists and financiers have become very powerful in Brazil, and their influence only increases insofar as large numbers of industrialists sell their businesses to multinationals and become rentiers. When, in 2011, the Central Bank substantially lowered the basic interest rate, President Dilma Rousseff gained the support of the manufacturing industry. But the political power of industrialists has long been waning in Brazil, due not only to the process of deindustrialization, but also to the process of denationalization: since the 1990s the number of manufacturing businesses sold to multinationals has consistently increased.

Brazil’s Macroeconomic Policy Institutions    231

11.2.4. The Dismantling of the Anti–​Dutch Disease Mechanism The fourth new historical fact explaining Brazil’s quasi-​stagnation is the dismantling of the mechanism for neutralizing the “Dutch disease.” This occurred in 1990, within the framework of the trade liberalization then realized. This was a major mistake. In the 1990s Brazil no longer had an “infant” manufacturing industry, and could open its economy and become more competitive. However, it still should have been important not to ignore the threat of Dutch disease—​a competitive disadvantage for the non-​ commodity tradable goods sector—​which is the major cause of the long-​term overvaluation of the Brazilian currency. The mechanism that neutralized the Dutch disease was built into Brazil’s foreign trade system. In 1990, Brazil’s average import tariff was reduced from 45% to 12% of GDP, and the subsidy to exports of manufactured goods, also 45% of GDP, was eliminated. Through this action, the government was not only opening the economy; it was also dismantling the mechanism that neutralized the Dutch disease, unknowingly creating a major competitive disadvantage for Brazilian manufacturing firms. For 60 years from the 1930s, the developmental economists who managed economic policy somehow instinctively or intuitively neutralized the Dutch disease without fully grasping the concept. Multiple exchange rate regimes, or high import tariffs combined or not with export subsidies for manufactured goods, did this job; in one stroke this neutralization, wrongly understood as protectionism, was discarded, giving rise to a major competitive disadvantage for Brazilian firms. The Dutch disease can be defined as a permanent appreciation of the exchange rate and, therefore, as a competitive disadvantage caused by the export of commodities using abundant and cheap natural resources; these commodities can be exported profitably at a significantly more appreciated exchange rate than the rate that would be necessary to make competitive both existing and potential producers of tradable goods and services that use world state-​of-​the-​art technology. The commodities that generate the Dutch dis­ease set the “current equilibrium”—​the value of foreign currency that guarantees the intertemporal equilibrium of the current account—​while the value required to render other competent tradable businesses competitive is the “industrial equilibrium.” The greater the difference between these two equilibriums, the more severe will be the Dutch disease. In oil-​exporting countries like Venezuela or Saudi Arabia, where the cost of production is very low, the disease is very serious, while in countries like Brazil or Argentina the disease is moderate but enough to cause deindustrialization and—​more than that—​ to prevent the vast majority of potential industrial projects in Brazil from being realized. The takeoff of industrialization in Brazil in the 1930s benefited from the depreciation of the national currency caused by the Great Depression and the long-​term fall in coffee prices. From the early 1950s, the Dutch disease was neutralized by a disguised tax on commodity exports, mainly coffee at that time. Originally, this export tax was embedded in multiple exchange rate regimes, involving a more appreciated rate for exporters of commodities. What can be called the “Delfim Netto model”—​after the

232   Luiz Carlos Bresser-Pereira minister of finance during the “miracle”—​is the mechanism that, from 1967 to 1990, neutralized the Dutch disease. It was embodied in the Brazilian foreign trade system of high import tariffs and substantial subsidies for exports of manufactured goods. The coffee exporters knew that this was a disguised export tax and called it “exchange rate confiscation” (confisco cambial), although ultimately they paid nothing because they recovered their tax payments through the depreciation of the currency. It was a large tax, amounting to 31% of commodity prices—​more than is required today to neutralize the Dutch disease. With this mechanism, the competent manufacturing enterprises that Brazil was building became competitive, and exports of manufactured goods soared; they accounted for 6% of total exports in 1965 and for 62% in 1991. Today they represent only 36% of Brazilian exports.

11.3. The Theory The last two “new historical facts” that explain the low investment and growth rates in Brazil since the early 1990s (the high level of interest rates and the non-​ neutralization of the Dutch disease) may be more clearly understood in light of the “new developmentalism” framework of macroeconomics developed in the literature since around 2003 (Bresser-​Pereira 2010, 2016; Bresser-​Pereira, Marconi, and Oreiro 2014), which focuses on the exchange rate and the current account, instead of on the interest rate and the budget deficit. In short, according to this view, economic development depends on investment, which depends on the expected profit rate and the interest rate; and the expected profit rate, in turn, depends on the exchange rate. The theoretical novelty in this understanding is the exchange rate, which is not considered either in Keynesian or neoclassical macroeconomics, because both assume that it is volatile but floats around the equilibrium exchange rate. New developmentalism drops this assumption and claims that in developing countries the exchange rate tends to be overvalued in the long term insofar as it exhibits tendencies to cyclical and chronic overvaluation. Thus, when businesses evaluate their investment opportunities, they take into consideration the ongoing exchange rate, which most of the time is overvalued, and conclude that the investment will not be competitive even if they utilize, or plan to utilize, world-​leading technology. The competitiveness of a country depends on the evolution of this equilibrium exchange rate, which, for its part, depends on the comparative index of unit labor costs, that is, the wage rate over the productivity of the country compared with the unit labor cost of a basket of countries. When this index rises, the equilibrium exchange rate goes up and the national currency depreciates so that its enterprises may remain competitive; when it falls, the exchange rate appreciates with no harm to local businesses. Whenever the exchange rate does comply with this expected market behavior, only diverging from it in the short term (the volatility problem of the exchange rate), it should not be a cause for concern. However, this is not true when the exchange rate remains overvalued in the

Brazil’s Macroeconomic Policy Institutions    233 long term, as we suppose it does in developing countries due to the cyclical and chronic tendencies of exchange rates observed in such countries. In this case, investment will cease, and the economy will immediately face deindustrialization and low rates of growth, if not stagnation. There are four causes for this, one structural cause—​a non-​neutralized Dutch disease—​and three habitual policy causes:  the policy of growth with current account deficits or “foreign savings,” the use of an exchange rate as an anchor to control inflation, and a central bank that conducts its monetary policy around a high level of interest rate. While the Dutch disease pulls the market exchange rate to the current equilibrium, these other three policies—​widely used in developing countries other than those East Asian countries that rank as competent developmental states—​explain the current account deficits. These policies, together with expansive and irresponsible fiscal policies, cause not only low investment and growth rates, but also lead the country into increasing indebtedness in foreign currency and into recurrent balance of payment crises. The policy of growth with current account deficits (“foreign savings”) to be financed by foreign loans or the investment of multinational companies automatically appreciates the exchange rate. Since there is a direct relationship between current account deficits and the exchange rate—​the higher the current account deficit, the more appreciated the exchange rate, and vice versa—​this policy appreciates the exchange rate, discourages domestic investment, and involves a high rate of substitution of foreign for domestic savings. Foreign investment does not add to, but rather replaces, domestic investment, except when the country is growing very fast and there is already a very high expected rate of profit.11 For the case of Brazil, see Figure 11.5.12 Meanwhile, the exchange

30

8 7

25 Foreign Debt (% GDP)

5 20

4 3

15

2 1

10

0 –1

5

Current Account Deficit (% GDP)

6

–2

Foreign Debt

2007

2004

2001

1998

1995

1992

1989

1986

1983

1980

1977

1974

1971

1968

1965

1962

1959

1956

1953

–3 1950

0

Current Account Deficit

Figure 11.5.  Current account deficits and foreign debt 1950–2007. Source: CBB (2010); Bresser-​Pereira, Gala, and Araújo (2015).

234   Luiz Carlos Bresser-Pereira rate anchor policy means maintaining a relatively fixed exchange rate while inflation continues to occur, causing inflation to fall. This is a perverse way of fighting the symptomatic evil that inflation is, since it does so at the cost of distorting the most strategic price that exists in a national economy, namely the exchange rate. As for the high level of the interest rate, that benefits only the rentiers and financiers, typically associated with foreign interests. Large fiscal deficits are an expression of fiscal populism and unsophisticated Keynesianism, not of liberal orthodoxy; instead, large current account deficits associated with a long-​term appreciation of the exchange rate are a manifestation of exchange-​rate populism, whether developmental or orthodox.13 These two forms of economic populism have always been present in Brazil since the transition to democracy. Exchange-​rate populism was present in the Fernando Henrique Cardoso administration, was even more intense under Lula, and recurred in the final two years of the Dilma administration. Fiscal populism was absent from 1999 to 2013, but returned in 2014.

11.4. Recession After seven years of overvaluation of the real between 2007 and 2014, the last exchange rate cycle ended in the second semester of 2014. Between 2015 and 2016, income per capita has fallen 9.0% and unemployment achieved an all-​time high of 12%. A sharp fall in the prices of the two main export commodities acted as a trigger for the recession. As the model of the tendency to the cyclical and chronic overvaluation of the exchange rate predicts, after a long-​term overappreciation, the cycle ends with a financial crisis and a sharp devaluation of the national money. This had happened in 1999–​2002, and was repeated in 2014, but the financial crisis was not a currency crisis, as usually happens in developing countries, nor a banking crisis like the 2008 crisis, but a financial crisis of business enterprises. It was not a currency crisis, because the country had accumulated huge and expensive reserves in the boom years; it was not a banking crisis, because banks were monopolist, were included in favorable government policies, and were well managed. It was a crisis of the manufacturing industry because for seven years it confronted an adverse exchange rate, its profit rates fell sharply, and manufacturing firms became heavily indebted, paying a very high interest rate. Thus, the causes of the recession were just an aggravation of the causes of the long-​term quasi-​stagnation. In the last long seven years in which the real was highly overvalued (2007–​2014), the average real effective exchange rate was R$2.80 per dollar, while the industrial equilibrium rose from R$3.80 to 4.00 per dollar due to the deterioration of the unit labor cost in comparison with key competitors. Considering an average R$3.92 per dollar industrial equilibrium between 2007 and 2014, the real would have required a 40% devaluation to become competitive. Such a huge overvaluation of the currency was not just the consequence of the Dutch disease, but also of the three habitual policies that appreciate the exchange rate of developing countries: the practice of a high level of the real

Brazil’s Macroeconomic Policy Institutions    235 interest rate around which the central bank conducts its monetary policy; the growth-​ cum-​foreign indebtedness (“savings”) policy; and the use of an exchange rate anchor to control inflation. The depreciation occurred only in the second semester of 2014, but the international market was pressing in that direction before that. The depreciation did not happen before because the Central Bank had bought back reserves (actually, swaps) from August 2013 to avoid it and the consequent increase of the rate of inflation. In February 2017 the real exchange rate was around R$3.10 per dollar, significantly below the industrial or competitive equilibrium, which is around R$4.00 per dollar.14 At the height of the financial crisis, in September 2015, the exchange rate peaked at R$4.40 per dollar, but soon, confirming the tendency to cyclical and chronic overvaluations of the exchange rate, it re-​appreciated to R$3.10 per dollar, owing to commodity prices’ partial recovery and the lessening of the fear of a balance of payment crisis. It would have depreciated more, had it not been for the real basic interest rate, which, despite the recession, is around 8% a year, reflecting the power of rentiers and financiers in Brazil.

11.5.  Long-​Term Solution The present major recession will end sooner or later. Conversely, the long-​term quasi-​ stagnation requires a long-​term solution, which necessitates the recovery of the fiscal equilibrium, lost from 2014, and overcoming the tendency to cyclical and chronic overvaluation of the exchange rate, which involves the neutralization of the Dutch disease and the rejection of the three habitual policies that lead to the overvaluation of the exchange rate. The severity of the Dutch disease in Brazil (the difference between the industrial and the current equilibrium) is difficult to measure, because the estimations of the industrial equilibrium are only approximations, and because a series for the current equilibrium is not available. We can say, at least, that the situation is not as severe as those in Venezuela or Saudi Arabia. Research on the value of the industrial equilibrium suggests that the average severity of the Dutch disease in Brazil is around 15%, ranging from 8% to 25% in accordance with the variation of the international prices of the commodities that the country exports.15 When the price rises, the current equilibrium goes down and the Dutch disease worsens; the reverse happens when commodity prices fall, as happened in the last quarter of 2014, when the Dutch disease almost disappeared. The solution to the problem—​the correct way of neutralizing the Dutch disease—​is to levy a variable export tax on the commodities that originate the disease, equal to the severity of the disease. As this tax increases the cost of production, exporters will require a more depreciated exchange rate, and since, given the foreign demand, it is the supply of the commodities (not of manufactured goods) that determines the exchange rate, the supply curve will shift to the left as the cost plus reasonable profit fall, and the exchange rate will duly depreciate. If the tax is equal to the severity of the disease, the current equilibrium will become equal to the industrial equilibrium, and the neutralization will

236   Luiz Carlos Bresser-Pereira be complete. In consequence, all technically competitive tradable industries (not only commodities benefiting from Ricardian rents) will be economically competitive. To explain the chronic overvaluation of the real we should consider, besides the Dutch disease, the three habitual policies that appreciate it: the growth with current account deficits policy; the use of the exchange rate as an anchor to control inflation; and the high level of the interest around which the Central Bank conducts its monetary policy. The weight of this second type of cause may be huge. When, between 2007 and 2014, the average overvaluation was around 60%, we can estimate that the Dutch disease accounted for one-​third of this difference, while the three habitual policies accounted for the remaining two-​thirds. Such long-​term overvaluation of the exchange rate since 1990 is more than enough to explain Brazilian industry’s loss of competitiveness, and deindustrialization in motion—​the deindustrialization that we see in the falling shares of the manufacturing industry in employment, GDP, and total exports, and in the increasing trade deficit of the manufacturing industry. In fact, deindustrialization was not still greater only because the “Brazilian automotive regime” initiated in 1995 imposed an import tariff of around 35% on the auto industry. Thus, in relation to this industry, which is crucial to the Brazilian economy, the government fully neutralized the Dutch disease, but only in relation to the domestic market; the competitive disadvantage remained in exports. The rationale for the adoption of the program was the importance of planning the production chain, but its good results reflected the fact that tariffs are a form of exchange rate, and its increase led to the neutralization of the Dutch disease in relation to imports. To counteract the tendency to cyclical and chronic overvaluation of the exchange rate and to make the exchange rate competitive, the government must neutralize the Dutch disease and radically reject the three habitual and populist policies. But before that, Brazilian economists and politicians would be well advised to, in the first place, understand the Dutch disease and why an export tax proportional to the severity of the disease neutralizes it. This is a serious problem because few economists in Brazil or in other countries are aware of this, in the same way that only a few economists knew of inertial inflation between 1980 and 1994—​a knowledge that ultimately was essential to the price stabilization that the 1994 Real Plan achieved. Once the knowledge problem is overcome, there is then also a political problem. It is politically difficult to neutralize the Dutch disease. First, the once-​and-​for-​all depreciation involved causes a temporary but unpopular fall in all real revenues, and a temporary increase in inflation—​something that Brazilians are loath to accept. Second, powerful commodity exporters, supported by liberal-​orthodox economists who do not believe that productive sophistication is a condition for growth, would resist the tax (even though their net cost would be zero, since what they paid in taxes they would receive back in depreciation). The first difficulty was solved by the fall in the commodity prices and the financial crisis that happened in the second semester of 2014 and caused both recession and a strong depreciation of the real. What the administration had to do was simply adopt (or avoid) the particular policies to avoid re-​appreciation of the Real. As for the second difficulty, a possible solution would be that the law establishing the

Brazil’s Macroeconomic Policy Institutions    237 tax also includes a table defining the relation between its international price and the percentage tax for each commodity. Currently, since the fall in international prices was huge, this percentage would be zero. In this way, the commodity exporters would not depend on the will of the government. If the international price of a commodity fell, the tax would be reduced, up to zero. However, note that an export tax will not assure the competitiveness of the manufacturing industry if the government continues to adopt the habitual policies that appreciate the national currency. This happened in Argentina from 2007 onward. In the 2001 financial crisis, a tax (retención) was imposed on commodity exports, which neutralized the Dutch disease and for six years allowed the economy to grow at a very high rate. However, in 2007, given the rise of inflation, the government decided to adopt the exchange-​rate anchor policy to control inflation. In consequence, despite the tax, the exchange rate has appreciated, industry has lost competitiveness, and the growth rate has fallen, at the same time that the country has failed to achieve a current account surplus. This shows that it is futile to neutralize the Dutch disease with an export tax and then adopt policies that appreciate the national currency. In the case of Brazil, after the 2014 depreciation, an export tax remained a proposal “off the agenda,” and the three habitual policies were not changed. Consequently, the real again became overvalued, as we have seen.

11.6. Conclusion This chapter has used the ideas of new developmentalism and developmental macroeconomics models to explain the quasi-​stagnation of the Brazilian economy. In short, after the mechanism that neutralized the Dutch disease was dismantled with the 1990 trade opening, the exchange rate appreciated chronically, varying the overvaluation from 8% to 25%, except in the cyclical moments of financial crisis, when it sharply depreciated. In addition to this structural cause, there are three habitual policy causes: the growth with current account deficits (“foreign savings”) policy; an exchange rate anchor policy to control inflation; and a high level of the interest rate both to attract capital and to control inflation. The non-​neutralization of the Dutch disease and the three habitual policies have reduced the productivity and competitiveness of Brazil’s manufacturing industry in monetary terms and also in technological terms, because the lack of investment hinders the modernization of machines and equipment. Second, the interest-​rate level has remained very high since the Real Plan. Third, from the late 1970s public savings became negative, which substantially decreased the investment capacity of the Brazilian state and so rendered infrastructure obsolete. Finally, in the early 2000s the country reached the “Lewis turning point” in that the unlimited supply of labor ended. Of these four policies, the first two, which raise interest rates and result in a currency that is overvalued over the long term, are the most important causes of Brazil’s low investment and growth rates. They represent a serious problem, but neither the liberals

238   Luiz Carlos Bresser-Pereira nor the developmental economists have conducted a serious debate about this macroeconomic trap. The political economy causes for this are as clear as they are burdensome. The necessary exchange rate devaluation displeases both groups. The developmental macroeconomics models relating to the Dutch disease and to the critique of growth with foreign savings remain generally unknown. Therefore, instead of discussing how to carry out devaluation, and what the economic and political obstacles to doing so are and how to overcome them, both camps reflexively argue that devaluation is either unnecessary, unfeasible, or both. Regardless of whether they are developmental or liberal, economists reject the required initial and once-​and-​for-​all devaluation on the basis that in the short term it would reduce wages (which it would) and would increase inequality (which it would not, because it would reduce not only wages but also all kinds of income). Indeed, attempting to reduce extreme inequality in Brazil through the exchange rate makes little sense. A better way to reduce it is through progressive taxation, a minimum-​income policy, low interest rates, and an expanded social state. Progressive taxation explains, for example, why Sweden has a far more equal distribution of income than the United States. The Gini index pre-​taxation is almost equal in the two countries, but post-​taxation it is very different; while taxation is progressive in Sweden, it is not in the United States. Liberal economists also reject devaluation, both because it temporarily increases inflation and reduces the real interest rate—​unacceptable to rentier capitalists—​and because it would create difficulties for companies indebted in dollars and therefore for the creditor banks. Like the developmental economists of the left, right-​wing liberals hold an instinctive horror of currency devaluation. To the left it implies inaction, and to the right it implies fiscal austerity, which will result in an “internal devaluation” as unemployment grows and wages fall while the incomes of rentiers will remain untouched. Insofar as the two sides focus only on the difficulties associated with the proposed policy, they avoid responsibility for defending the temporary reduction of incomes and the temporary increase in inflation which devaluation involves. Thus, because of this “active omission” on the part of the economic elites, society is uninformed about the real causes of the stagnation of the Brazilian economy since the early 1990s. Since the reluctance is shared across left and right, government remains paralyzed on the issue no matter which political party holds office. Economists, businesspeople, and politicians continue to lack an understanding of the fundamental and detrimental role of the long-​term overvaluation of the exchange rate and the ensuing competitive disadvantage suffered by non-​commodity tradable industries on the processes of growth and internationally “catching up.”

Notes 1. Macroeconomia da Estagnação [Macroeconomics of Stagnation] was the name of the 2007 edition; in English it was published two years later with the title Developing Brazil: Overcoming the Failure of the Washington Consensus (Boulder, CO: Lynne Rienner, 2009). 2. Between 1930 and 1980 the per capita growth rate was 2.8% a year.

Brazil’s Macroeconomic Policy Institutions    239 3. Inflation becomes “inertial” when economic agents index their prices and wages to previous inflation formally and informally, and inflation becomes independent from demand. For the first works on inertial inflation, see Simonsen (1970) and Pazos (1972); for the fully developed theory, see Bresser-​Pereira and Nakano (1987) and Lara-​Resende and Arida (1985). 4. Note that the Real Plan was successful because it was the outcome of a heterodox economic theory (the theory of inertial inflation) developed by Brazilian economists. 5. See the 2010 PhD dissertation of Senator Otávio Mercadante, at that time leader of the government in the Federal Senate. 6. All exchange rates in this chapter are expressed in real terms, in third quarter 2016 prices; and they are the “effective” exchange rate, which considers 16 currencies instead of only the dollar. 7. The administration had already failed to undertake a fiscal adjustment when, in the second semester of 2011, the Central Bank significantly lowered the interest rate. 8. The explanation for the rise in inflation is given in the previous note. A substantial reduction in the interest rate must be accompanied by a fiscal adjustment. 9. On the Minskyan financial crisis, see Renato Resende (2016). 10. In 1964 the liberal and highly competent Planning Minister Roberto Campos nationalized foreign companies to incorporate them in two major state-​owned enterprises, Telebras and Eletrobras, and immediately increased their consumer prices—​a policy that allowed the two companies to self-​finance their much-​needed investments. 11. For the theory, see Bresser-​Pereira and Gala (2007). There is a large empirical literature on “savings replacement,” which demonstrates empirically this key New Developmentalist model on the substitution of foreign for domestic savings. 12. On the case of Brazil, in addition to Figure 11.5, see Bresser-​Pereira (2009, Chapter 7); Bresser-​Pereira, Araújo, and Gala (2014). 13. Liberal orthodoxy is closely associated with exchange-​rate populism in developing countries insofar as their economists have a positive view of current account deficits, which, in most cases, finance consumption. 14. Figures on the industrial equilibrium at December 2015 prices arrived at through research with Nelson Marconi at the Center of New Developmentalism of the São Paulo School of Economics, Getúlio Vargas Foundation. Studies by Nassif, Feijó, and Araújo (2015) and Oreiro, Basílio, and Souza (2014) arrive at similar numbers. For the method adopted by myself and Nelson Marconi, see Marconi (2012). 15. For the estimate of the industrial equilibrium see Marconi (2013), Nassif, Feijó, and Araújo (2015), and Oreiro, Basilio, and Souza (2014).

References Boito, Armando, Jr. 2012. “Governos Lula: A nova burguesia nacional no poder.” [“Lula administration: the new bourgeoisie in power.”] In Política e classes sociais no Brasil dos anos 2000, edited by Armando Boito Jr. and Andreia Galvão, 69–​104. São Paulo: Alameda. Boito, Armando, Jr., and Tatiana Berringer. 2014. “Social Classes, Neodevelopmentalism, and Brazilian Foreign Policy under Presidents Lula and Dilma.” Latin American Perspectives 41 (5): 94–​109. Bresser-​Pereira, Luiz Carlos. 1990. “The Perverse Macroeconomics of Debt, Deficit and Inflation in Brazil.” Journal of Post Keynesian Economics 12 (4): 503–​518.

240   Luiz Carlos Bresser-Pereira Bresser-​Pereira, Luiz Carlos. 2009. Developing Brazil: Overcoming the Failure of the Washington Consensus. Boulder, CO: Lynne Rienner. Bresser-​Pereira, Luiz Carlos. 2010. Globalization and Competition:  Why Some Emerging Countries Succeed While Others Fall Behind. New York: Cambridge University Press. Bresser-​Pereira, Luiz Carlos. 2016. “Reflecting on New Developmentalism and Classical Developmentalism.” Review of Keynesian Economics 4 (3): 331–​352. Bresser-​Pereira, Luiz Carlos. 2017. The Political Construction of Brazil. Boulder, CO:  Lynne Rienner. Bresser-​Pereira, Luiz Carlos, Eliane Araújo, and Paulo Gala. 2014. “An Empirical Study of the Substitution of Foreign for Domestic Savings in Brazil.” Revista Economia 15: 54–​67. Bresser-​Pereira, Luiz Carlos, and Eli Diniz. 2016. “Industrial Entrepreneurs, Democracy and Political Power.” In The Political System of Brazil, edited by Dana de La Fontaine and Thomaz Sthnken, 183–​200. New York: Springer. Bresser-​ Pereira, Luiz Carlos, and Yoshiaki Nakano. 1987. “The Theory of Inertial or Autonomous Inflation.” In The Theory of Inertial Inflation, by Luiz Carlos Bresser-​Pereira and Yoshiaki Nakano, 65–​ 82. Boulder, CO:  Lynne Rienner. [Originally published in Portuguese in 1983 as Fatores aceleradores, mantenedores e sancionadores da inflação.] Bresser-​Pereira, Luiz Carlos, José Luis Oreiro, and Nelson Marconi. 2014. Developmental Macroeconomics. London: Routledge. Coes, Donald V. 1995. Macroeconomic Policies, Crises and Growth in Brazil, 1964–​1990. Washington, DC: World Bank. Lara-​Resende, André, and Persio Arida. 1985. “Inertial Inflation and Monetary Reform.” In Inflation and Indexation:  Argentina, Brazil and Israel, edited by J. Williamson, 27–​45. Washington, DC: Institute for International Economics. Marconi, Nelson. 2012. “The Industrial Equilibrium Exchange Rate in Brazil: An Estimation.” Brazilian Journal of Political Economy 32 (4): 656–​669. Mercadante, Aloísio. 2010. “As bases do novo desenvolvimentismo no Brasil.” PhD dissertation, Unicamp. Nassif, André. 2011. “Overcoming the ‘Impossible Trinity’: Towards a Mix of Macroeconomic Policy Instruments for Sustaining Economic Development in Brazil.” Brazilian Journal of Political Economy 31 (5): 912–​927. Nassif, André, Carmem Feijó, and Eliane Araújo. 2015. “Overvaluation Trend of the Brazilian Currency in the 2000s: Empirical Estimations.” Brazilian Journal of Political Economy 35 (1): 3–​27. Oreiro, José Luis, Flávio A. C. Basilio, and Gustavo J. G. Souza. 2014. “Effects of Overvaluation and Exchange Rate Volatility over Industrial Investment: Empirical Evidence and Economic Policy Proposals for Brazil.” Brazilian Journal of Political Economy 34 (3): 347–​369. Oreiro, José Luís, and Carmen A. Feijó. 2010. “Desindustrialização:  Conceituação, causas, efeitos e o caso brasileiro.” Revista de Economia Política 30 (2): 219–​232. Passos, Felipe. 1972. Chronic Inflation in Latin America. New York: Praeger. Rezende, Felipe. 2016. “Financial Fragility, Instability and the Brazilian Crisis:  A Keynes-​ Minsky-​ Godley Approach.” Minds (Multidisciplinary Institute for Development and Strategies), Discussion paper 1.  Available at http://​www.minds.org.br/​media/​papers/​wp-​ minds-​201601-​rezende57aa276042e28.pdf. Simonsen, Mário Henrique. 1970. Inflação:  Gradualismo x tratamento de choque. Rio de Janeiro: ANPEC.

Pa rt  I I I

T H E P RODU C T I V E SE C TOR S

Chapter 12

Evolu tion and Se c tora l C om petitive ne s s of the Brazi l ia n Manufacturin g I ndu st ry Paulo César Morceiro 1

12.1. Introduction After a lengthy period of near-​stagnation from 1980 to 2003, the Brazilian gross domestic product (GDP) rebounded in the new millennium, increasing more than global GDP between 2004 and 2013,2 the majority of this growth spurred by domestic demand. Although almost all sectors benefited, the manufacturing sector’s share of GDP declined, a situation referred to by various economists as “deindustrialization.” However, the Brazilian manufacturing industry has followed a variety of sectoral trajectories, resulting from patterns of household consumption and from the sector’s level of competitiveness. After presenting the evolution of production and demand in Brazilian manufacturing, the objective of the chapter is to verify whether the sector experienced demand leakage in favor of foreign producers; that is, whether the sustained growth of domestic demand in the period 2004–​2013 was captured by domestic suppliers, or by foreign suppliers in the form of imports. In addition, the chapter seeks to assess the competitiveness of Brazilian manufacturing in the twenty-​first century using a set of sectoral indicators and a measure of technological intensity. Through this, the chapter aims to diagnose the competitiveness of the manufacturing sector in the recent period. Section 12.2 discusses Brazilian manufacturing’s long-​term performance. Section 12.3 presents the main determinants of Brazilian economic growth in the first decade of the 2000s, which affected industrial adjustments. Section 12.4 considers the evolution of manufacturing in terms of technology intensity, and highlights a remarkable demand leakage to other countries through imports. Section 12.5 presents a data set revealing

244   Paulo César Morceiro competitiveness and the main sector highlights for GDP growth, productivity, employment, production, demand, trade coefficients, and trade balance. Section 12.6 discusses the 2014–​2016 crisis in Brazil and its impact on industrial production. Finally, section 12.7 presents the conclusions of the study.

12.2.  Brazilian Manufacturing: The Longer-​Term Context Brazilian GDP grew 7.4% per annum, between 1950 and 1980, almost 3 percentage points above the globally reported rate. This period saw intense industrialization in Brazil, and the manufacturing sector’s share of GDP doubled (Figure 12.1). The period was also characterized by strong state planning, through the Plano de Metas (Goals Plan, 1956–​ 1960) and the Second National Development Plan (1975–​1979), which resulted in the establishment of plants for durable, intermediate, and capital goods, sectors considered the most difficult to internalize. 24

23.02 21.67

22 20 17.87

18 16 13.57

14

16.37 13.69 13.58 11.79

12 10

17.79

12.48

13.81

11.40 11.82

10.18 10.17

1947 1949 1951 1953 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015

8

Manufacturing (% of GDP)

Manufacturing (% of total employment)

Figure  12.1.  Brazilian manufacturing industry participation in GDP (%) (basic and current prices, 1947–​2015) and in total employment (1950–​2013). Note: GDP series from 1947–​1994 and employment series from 1950–​1999 were chained to the current methodology proposed by the Brazilian System of National Accounts (SNA)—​Reference 2010, which is used by IBGE. Homogeneous data from the System of National Accounts was used to chain the GDP series. Sources: Contas consolidadas para a Nação—​Brasil, 1980–​1993 (IBGE, DECNA, October 1994) and Sistema de contas nacionais—​Brasil, 2003 (IBGE 2004). The participation level decreases were as follows: 1989–​1990: 32.39% to 29.08%; 1994–​ 1995: 26.79% to 23.91%. Homogeneous data series from the Groeningen Growth and Development Centre (1950–​1990) and from Sistema de contas nacionais—​Brasil, 2003 (IBGE 2004) were used to chain the employment series. Elaborated using data from IBGE (2016a) and the Groeningen Growth and Development Centre.

Competiveness of Brazilian Manufacturing Industry    245 By the mid-​1980s, Brazil had an integrated and diversified industry (Castro 1985). However, since then the contribution of manufacturing to GDP has decreased; in 2015, it was inferior to that prior to the Goals Plan period (Figure 12.1). Moreover, GDP growth significantly decreased due to the loss of manufacturing dynamism. Between 1980 and 2015, Brazil’s GDP evolved at an average of 1 percentage point below the global economy. Consequently, Brazil lost its importance on the world stage; in 1980, the country accounted for 2.5% of global manufacturing, decreasing to 1.8% in 2000, and to 1.5% in 2014, according to UN data. Its rank decreased from the eighth to the twelfth position as a leading manufacturing nation between 1980 and 2014. However, Brazil is still a significant player in several global value chains, such as food, products intensive in natural resources, and regional aviation. Economists define deindustrialization as a sustained decrease of the absolute level of manufacturing and/​or its contribution to total employment and/​or GDP (Singh 1987, 302). Other authors also define deindustrialization as the deterioration of the balance of payments, especially of the balance of trade in manufacturing (Morceiro 2012).3 Deindustrialization is considered early or premature (Palma 2008; Ricupero 2005) when the manufacturing sector’s share of GDP decreases long before the country’s per capita income reaches US$25k (at purchasing power parity and constant prices for 2015). Low per capita income limits the expansion of sophisticated service sectors with high income elasticity, which would have been capable of absorbing occupations lost with the decline of manufacturing. In such cases, the economy discards its main dynamic sector without another sector taking up the mantle, and falls into a trap of low growth of per capita GDP. Deindustrialization has a number of negative consequences, especially for a developing country. In terms of long-​term growth trajectories, manufacturing still has development-​boosting characteristics, such as (1)  stronger intersectoral linkages; (2) being a key source of innovation and diffusion of new technologies, since around two-​thirds of research and development (R&D) investments come from manufacturing (McKinsey 2012); (3) decreasing pressure on the balance of payments (Thirlwall 2002), since at least two-​thirds of exports are manufacturing products (McKinsey 2012); (4)  providing an important platform for other sectors by purchasing an significant portion of their production (Morceiro 2012); and (5) contributing significantly to an economy’s productivity growth, with strong static and dynamic economies of scale (Kaldor 1967; McKinsey 2012). Rodrik (2007) mentions a number of “stylized facts” of manufacturing and reports that the highest GDP growth rates were obtained during the industrialization stage. UNIDO (2015) and Szirmai and Verspagen (2015) recently reviewed these and also suggested other characteristics favorable to manufacturing.4 Thus, manufacturing is highly important—​a fact that became apparent following the 2008 crisis, when developed countries adopted reindustrialization measures. The United States focused on advanced manufacturing reconversion (National Science and Technology Council 2012), and the Eurozone adopted explicit goals to increase manufacturing’s contribution to GDP from 16% to 20% by 2020 (European Commission 2012).

246   Paulo César Morceiro However, deindustrialization has generated contrasting views in Brazil. Manufacturing has significantly lost participation in GDP and the country has built up a high trade deficit in manufactured goods. On the other hand, there was no absolute reduction in manufacturing production, and employment increased in absolute and relative terms since the late 1990s (Figure 12.1). The manufacturing sector’s contribution to total employment has seen small oscillations since 1950, particularly during industrialization and the most recent phases (Figure 12.1).5 Discussions about Brazilian deindustrialization have been disproportionately focused on the manufacturing industry’s behavior at the aggregate level and its macroeconomic effects, with very few sectoral approaches (an exception is Morceiro 2012). The account in this chapter acknowledges the deep sectoral heterogeneity of Brazilian manufacturing and the sectoral adjustments that have taken place recently. The following section presents the determinants of growth that have influenced the industrial sector during the period under consideration.

12.3.  Determinants of Economic Growth in Brazil in the Twenty-​first Century Brazilian GDP increased 2.8% per annum from 2001 to 2015, below the global average. There was continued growth of 4.0% per annum for 10 years from 2004 to 2013. During that period, GDP per capita increased 2.9% per annum, a rate only bettered during the 1970s (1968–​1980). Growth was interrupted in 2014 and, since then, the country has undergone a severe crisis. This section focuses on the 2000–​ 2013 period, and particularly 2004–​2013. The recent crisis (2014–​2016) is addressed in section 12.6. From 2000 to 2013, various factors contributed to GDP growth and to the expansion of industrial demand. From an aggregate demand perspective, household consumption was the most important growth component since 2004 (Miguez 2016, 50–​77; Sarti 2015, 524). Gross fixed capital formation (GFCF) and exports had secondary roles in distinct periods, while government consumption had a marginal contribution. Exports were the component of aggregate demand that contributed most to GDP growth between 2001 and 2003, and were the second most dynamic component in 2004 and 2005. During this period, exports expanded significantly, as a result of the 1999 currency devaluation and increasing Chinese demand for agricultural and mineral commodities. Various commodity prices boomed from 2002 due to Chinese industrialization and urbanization; as such, the terms of trade favored Brazil for almost the entire period up to 2013. This contributed to a reversal of the trade balance deficit that had existed since 1995 as soon as 2001, and to the accumulation of a high and growing trade surplus from then until the mid-​2000s. High international liquidity and extremely high

Competiveness of Brazilian Manufacturing Industry    247 Brazilian interest rates attracted foreign capital, which led to currency appreciation, but also to a massive foreign exchange reserve accumulation.6 Foreign capital attraction and trade surplus eliminated the vulnerability and external economic growth restrictions that had marked the two previous decades. Export growth during the first half of the first decade of the 2000s was the growth catalyst, sustained by household consumption and background investments. Household consumption increased during 45 consecutive quarters, from the fourth quarter of 2003 to the fourth quarter of 2014, according to data from Instituto Brasileiro de Geografia e Estatística (IBGE 2016b). Household consumption increased 61.1% (4.9% per annum) from 2004 to 2013, while GDP increased 48.2% (4% per annum), according to IBGE (2016b). This resulted in an increase of the contribution of household consumption to GDP from 59.6% to 66.0% during the same period. Job creation, real increases in the minimum wage, credit expansion, and currency appreciation all contributed to domestic demand increases as well. Significant employment generation in almost all economic sectors was the most important aspect. From 2000 to 2013, 23.6 million new jobs were created, increasing the total national job count to 102.5 million (IBGE 2016a). It is important to highlight that this was the most significant episode of job creation in Brazilian history, and contributed to a decrease in unemployment rates nationally, from 10.5% to 7.1% between 2003 and 2013 (IBGE 2016c). Informal occupations had historically been the majority in Brazil, but a remarkable occupation quality improvement was observed during this period. Formal occupations increased from 38.2% to 53.8% of the total between 2000 and 2013 (IBGE 2016a); 26.3 million formal jobs were created. Informality is harmful for workers, insofar as they are not supported by labor laws7 and commonly face a range of problems, including excessive working hours. It is also harmful for employers, who have no guarantee of employees’ commitment and regularity, and for the government, which will receive lower tax revenues. New formal employees—​who were previously marginalized—​ began to have regular access to credit channels. As such, the substantial reduction in informality benefited the economy in numerous ways. Real increases in the minimum wage were observed every year from 2000 to 2014. The minimum wage displayed a 73.7% (5.7% per annum) increase in real terms between 2004 and 2013. The highest increases occurred during the Lula administration, through a wage improvement policy, which increased public, private, and retired employees’ purchasing power. The Brazilian minimum wage is the minimum legal wage for formal and retired employees and both directly and indirectly affects all economic wage adjustments. Credit granting boomed during this period. The credit operations balance increased from 24.3% to 56.0% of GDP from December 2003 to December 2013, according to Central Bank of Brazil (BCB) data.8 Moreover, there was an improvement in credit conditions, such as terms, extended limits, extended coverage9 (Borça Jr. and Guimarães 2015), and the real interest rate decreased from 11.9% to 2.0% between 2003 and 2013, according to BCB data. Some institutional and financial risk mitigation innovations,

248   Paulo César Morceiro such as payroll loans implementation (effective since 2004)  and real estate credit innovations, also contributed to a strong credit expansion. The increase was seen in almost all credit lines. The 83.8% currency appreciation between 2003 and 2011, when the BRL/​USD exchange rate decreased from 3.08 to 1.67, improved import conditions and limited domestic price increases. This situation contributed to an expansion in price-​sensitive domestic demand.10 Baer (2014, 174–​186) and Amann and Baer (2012) emphasize the appreciated exchange rate’s use in inflation control since the implementation of the Real Plan. Tax reductions for significant products (vehicles, appliances, furniture, construction products, and capital goods), implemented during the 2008 international crisis and extended until 2013 had a positive impact on price-​sensitive domestic demand in most cases. Some measures and government transfers also impacted household consumption (Tesouro Nacional 2016). These include, for example, (1) Programa Minha Casa Minha Vida (My Home My Life Program, PMCMV), a low-​cost home construction program, which directly stimulated the construction sector and its productive chain from 2009; (2) cash transfer programs (e.g., Bolsa Família),11 which contributed to the development of poor and rural regions (mainly in the North and Northeast parts of Brazil, where the majority of this program’s resources are committed); (3) Luz para Todos (Light for Everyone), which benefited around 15 million people living in isolated rural and non-​ electrified regions, which now have access to house lighting, and are able to diversify their consumption pattern, including appliances (e.g., refrigerators), TVs, cell phones, and other items (MME 2014). Initially, exports and household consumption contributed to decreasing idle capacity in the majority of sectors and reactivating the investment cycle. The economy’s overall GFCF increased 65% in real terms between 2003 and 2011, and the investment rate increased by 3.5 percentage points, to 20.6% in 2011 (IBGE 2016a). The 2008–​2009 international crisis had a slight impact on investments, but countercyclical measures implemented by the federal government, such as PMCMV and the aforementioned tax reductions, led to a quick recovery in 2010. Despite the high performance observed after 2003, the investment rate is still low compared to the global average, and especially to China, India, and South Korea. In addition to demand, demographics played a significant role as well. Brazil is still a “young country,” differing from other nations where population growth has stagnated. From 2004 to 2013, the Brazilian population increased 11.3% (or 1.08% per annum), or from 180 to 201 million inhabitants (IBGE 2016d) (i.e., 20 million new consumers). The country had a “demographic bonus,” that is, during this period the working-​age population (between 16 and 64 years old) increased proportionally in relation to the dependent population (children and the elderly). The demographic bonus resulted from a reduction in fertility and mortality rates. The population between 16 and 64  years old increased from 63.5% to 66.7% of the total population, that is,, from 115 million to 134 million people of working age, who are capable of increasing consumption and capital formation.

Competiveness of Brazilian Manufacturing Industry    249 In summary, various factors contributed to stimulating domestic demand and the consequent economic growth. Although almost all manufacturing sectors benefited, the impact was distinct depending on the income-​elasticity of demand and sectoral competitiveness, as discussed in subsequent sections.

12.4.  Evolution of Production and Demand for Manufacturing by Technological Intensity The gap between industrial production and demand presented in Figure 12.2 shows the recent loss of Brazilian manufacturing competitiveness.12,13 The gap between these series started in 2005 and is increasing. Brazilian industrial production increased 33.9% from 2004 to 2013, and total demand increased 55.3%. The difference between these rates explains the “alligator mouth shape” at the end of the period, a gap between production and total demand, which has been filled with imports (Figures 12.2 and 12.3). Approximately 55% of demand growth leaked to other countries through imports since 2009. Some sectors—​such as parts and accessories for motor vehicles, apparel, metallurgy, electrical equipment, machines and equipment—​experienced particularly intense “demand leakage,” as presented in Appendix 12.1.

160.4

160

155.3

150 140

133.9

130 120 110 100 90

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Production

Total demand

Domestic demand

Figure 12.2.  Production, domestic demand, and total demand of the Brazilian manufacturing industry, 2000–​2013. Note: All variables are measured using basic and constant prices (2003 = 100). Source: Author’s calculations using IBGE (2016a) data.

250   Paulo César Morceiro Low and medium-low technology

140

133.5 130

123.6

120 110 100

Total demand

Medium-high and high technology

190

13

12

20

11

20

10

20

09

20

08

Production

20

07

20

06

20

05

20

04

20

03

20

02

20

01

20

20

20

00

90

189.2

170 153.1 150 130 110

Production

13

20

12

20

11

10

20

20

09

20

08

07

20

20

06

20

05

20

04

03

20

20

02

20

01

20

20

00

90

Total demand

Figure 12.3.  Production and demand of Brazilian manufacturing industry by technology intensity, 2000–​2013. Note: All variables are measured using basic and constant prices (2003 = 100). Source: Author’s calculations using IBGE (2016a) data.

Distinct evolutionary patterns can be observed when categorizing the manufacturing industry in two groups of technology intensity.14 Total manufacturing demand for high-​ and medium-​high technology (HT and MHT) goods was 2.7 times higher than for low-​and low-​medium technology goods (LT and LMT), as presented in Figure 12.3. The gap between total demand and industrial production is higher for HT-​MHT industries (Figure 12.3), and thus larger demand leakages have occurred in the more technologically dynamic manufacturing sectors. Starting from 2004, there have been three stages of demand increases. Initially, industrial production followed demand growth until 2005, with no demand leakage; at this point, production increased mainly due to idle capacity reduction in the majority of sectors. Subsequently, moderate demand leakage can be observed from 2006 to 2008,

Competiveness of Brazilian Manufacturing Industry    251 occurring with the expansion of industrial production and investment. Finally, from 2009, a great demand leakage period began, with a magnitude of 20 percentage points (Figure 12.2), or 46 percentage points for HT-​MHT (Figure 12.3), with additional deceleration of the production growth rate and escalating demand. The factors explaining these gaps, including at the sectoral level (see Appendix 12.1), are still not fully understood in the literature. Several analysts identify as the main cause currency overvaluation throughout most of the period, and the related increase in import competitiveness nationwide. Despite currency appreciation, Bastos et  al. (2015) verified that the 453 largest manufacturing companies presented good profitability and increased investments during the 2004–​2010 period. However, those investments were not sufficient to increase capital stock in most sectors and to meet all demand growth.15 Messa (2015) calculates the sectoral capital stock of manufacturing from 2002 to 2010 and verifies that, surprisingly, it decreased in most sectors. Bielschowsky, Squeff, and Vasconcelos (2015) suspect that the industrial investment cycle in the first decade of the 2000s was focused on modernization and depreciation recovery, with little contribution to expansion and productive diversification. Therefore, they argue, currency appreciation discouraged investments in expansion and diversification, but not in modernization. With this investment profile, labor saving was expected, but various sectors presented significant manpower increases (see Appendix 12.1), and it is hard to imagine that there was no increase in the productive capacity, although it was insufficient to meet fully the strong demand increase. Another hypothesis regards transnational corporations (TNCs) established in Brazil (Bastos et al. 2015; Sarti 2015). After the 2008 crisis, there was a strong decrease in demand from developed economies. Thus, TNCs with subsidiaries in various countries began redirecting their production and industrial inventories to countries with strong demand, such as Brazil. This global productive scale optimization strategy contributed to a decrease in industrial investments in Brazil during the post-​2008 crisis period, when the highest demand leakage was observed. It is important to highlight that, especially in more technology-​intensive sectors, there is a predominance of foreign TNC subsidiaries in Brazil. It is worth emphasizing that Brazilian manufacturing production did not experience any significant increase in the 1980s and 1990s, especially in HT-​MHT industries. But in the period 2004–​2013, manufacturing production grew 33.9% and the HT-​MHT industries grew 53.1% (see Figure 12.3 and Appendix 12.1). As mentioned in section 12.3, manufacturing growth was mainly boosted by significant formal job creation, real minimum wage increases, a credit boom, population growth, and the increase in social programs. These boosted domestic demand, which increased more than total demand (see Appendix 12.1).16 The manufacturing industry’s domestic demand increased more than the entire economy’s, especially in HT-​MHT sectors (see Appendix 12.1). The distinct evolutionary pattern of the two different technology intensity groups may be explained by the significant improvements in income distribution and by suppressed HT-​MHT product demand (Medeiros 2015). The economic classes defined in Brazil as A, B, and C significantly expanded from 2003 to 2013.17 The consumption profile of the

252   Paulo César Morceiro Brazilian A–​B classes is similar to that of the North American middle class, owning two cars and two dogs, but does not characterize the worldwide middle class, represented by the C class (or “new middle class”) (Neri 2014). In 2003, 98.85 million individuals (54.7% of the Brazilian population) belonged to D–​E classes; by 2013, this number had dropped to 62.08 million (30.9%) (Neri 2014, 24). C class increased from 67.89 million individuals (37.6%) to 112.56 million (56.0%) during the same period, and A and B classes also increased from 13.89 million individuals (7.7%) to 26.40 million (13.1%) (Neri 2014, 24). Therefore, 57 million individuals—​more than the entire populations of England or South Korea—​moved into the A, B, and C classes between 2003 and 2013. Formal job creation and real increases of the minimum wage resulted in an increase in the number of families in the A, B, and C classes. Higher income classes have a consumption pattern focused on more income-​elastic products, such as vehicles, computers, and higher quality services, which also demand manufacturing inputs. On the other hand, lower income classes tend to focus their consumption on essential, low income-​elasticity products, such as home, food, and clothes. In this sense, demand income-​elasticity is higher for HT-​MHT industries than for LT-​LMT goods. The increase of A, B, and C class families explains the higher demand for income-​elastic products. Additionally, credit growth also contributes to explaining the higher demand for HT-​MHT products with a higher unit value, especially for computers, electronics, and vehicles.

12.5.  Competitiveness and Industrial Productivity Brazilian manufacturing competitiveness in the twenty-​first century can be evaluated through the evolution of trade coefficients, the trade balance, and labor productivity. These variables are presented in Table 12.1 by technology intensity, and in Appendix 12.1, for the 22 manufacturing sectors. The import penetration coefficient (IPC) for the manufacturing industry doubled between 2003 and 2013, reaching 26.8% with a general sectoral increase (see Appendix 12.1). During this period, the LT-​LMT industries’ IPC increased from 5.9% to 13.6%, and the HT-​MHT industries’ IPC increased from 25.2% to 41.2%. LT-​LMT industries’ IPC is lower than that of HT-​MHT industries for two reasons: first, Brazil is competitive in natural resource industries, which compose the LT-​LMT aggregate; second, HT-​MHT industries are assembly-​intensive and have more internationally fragmented value chains. Inputs and components imports are beneficial for promoting productive effectiveness and export increases. During the analyzed period, imports gained participation in the domestic demand for finished and intermediate goods (Morceiro 2012, 111–​150; 2016). IPC significantly increased in the majority of productive sectors between 2003 and 2013, with a 100% increase in the manufacturing industry, thus certainly improving its productive efficiency. However, the export coefficient remained the same and even decreased in more technological sectors (see Appendix 12.1), that is, imports did not contribute to export competitiveness.

Table 12.1 Performance Indicators of the Brazilian Manufacturing Industry (2003–​2013) (%, 2003 Base Year) Annual Growth Rate, 2004–​2013

Total manufacturing

Export Share of Production

Import Penetration

Trade Balance as a Share of Production

GDP

Workers

Labor Productivity

2003

2013

2003

2013

2003

2013

2.8

3.3

–​0.5

15.8

14.7

13.5

26.8

2.7

–​16.5

LMT and LT

1.4

2.8

–​1.4

15.3

13.1

5.9

13.6

9.9

–​0.5

HT and MHT

4.7

5.0

–​0.2

17.0

17.2

25.2

41.2

–​10.9

–​40.8

Total economy

3.8

2.0

1.7

7.2

7.6

6.2

11.5

1.1

–​4.4

Note: All variables are measured using basic and constant prices. Source: Author’s calculations using IBGE (2016a) data.

254   Paulo César Morceiro It is evident that the significant manufacturing production increase, especially in HT-​ MHT, occurred with an increase in the import content, that is, with a lower physical/​ chemical/​biological transformation per product unit. The increase of import content was a partially reactive strategy adopted by Brazilian industry in order to protect the internal finished goods market. Pereira (2016) analyzes production multipliers and Hirschman-​Rasmussen inter-​sectoral linkage indices, and concludes that Brazilian industry is weakened, while Cassiolato and Fontaine (2015) declare that there is a tendency for productive structure emptying. Industry’s trade balance deteriorated, including for LT-​LMT sectors, which had traditionally had a positive balance. In 2003, the manufacturing industry presented a positive trade balance of around 2.7% of manufacturing production, which dropped to –​16.5% in 2013. Among HT-​MHT industries, the production deficit increased 30 percentage points, from –​10.9% to –​40.8% (Table 12.1). Almost all sectors, except for transport equipment, experienced a substantial trade balance deterioration during this period. International competitiveness loss “is not necessarily due to inefficiencies but rather due to the anomaly of an appreciated exchange rate resulting from the dictates of inflation targeting” (Amann and Baer 2012, 417), which has been hindering the manufacturing sector (Baer 2015, 13). In 2013, 15 of 22 industry sectors presented trade balance deficits. The computers, electronic, and optical products sector was particularly (negatively) notable due to its trade balance reaching –​93.6% of manufacturing production in 2013 (see Appendix 12.1). The machinery and equipment and electrical equipment sectors presented a trade balance deficit of approximately –​58.0% of production; auto parts around –​48.9%; and chemicals around –​ 43.1%; that is, the highest deficits occurred in sectors with wages two times higher than those of the average manufacturing industry. Therefore, imports resulted in the transfer of qualified jobs and technology development potential to other countries (Morceiro 2012). Between 2004 and 2013—​during the period of the fastest manufacturing production increase—​manufacturing labor productivity decreased by –​0.5% per annum, while GDP and employment increased by 2.8 and 3.3% per annum, respectively. Various recent studies, based on distinct data, verified decreases or negative evolutions of labor productivity or total factor productivity (e.g., De Negri and Cavalcante 2015; Silva, Menezes Filho, and Komatsu 2016). Fourteen of 22 sectors showed decreases in labor productivity, and the remaining sectors presented a small increase. The best scenario for any developing country with a growing population is of productivity increases with job count stability or increases; the worst scenario is the opposite. Brazil presents an intermediate situation, with adverse productivity evolution and increasing employment. Among 22 sectors, only four presented a decrease in GDP. Therefore, the low productivity growth was obtained with an elevated number of new jobs in almost every sector. As Cavalcante and De Negri (2015) state, there is consensus about the low productivity evolution in the twenty-​first century, but not about its causes. Generally, labor productivity has a positive correlation with per capita capital stock, technological development, and human capital at the sectoral level. Three hypotheses regarding what explains the negative evolution of productivity are as follows. First, there was no increase of per capita capital stock. Despite the good investment cycle—​even when considering that a substantial part has been used to increase

Competiveness of Brazilian Manufacturing Industry    255 productive capacity—​the labor force expansion has been very robust, so that the per capita capital stock has probably not increased in the majority of sectors. Second, various manufacturing sectors, mainly in more dynamic industries, increased their assembly intensity with minimal operations, at the expense of transformation per product unit. In Brazilian HT-​MHT industries, where production and employment increased significantly, high technology is not developed or produced; products are merely assembled using imported technology, with rare exceptions. The country has increasingly imported a large quantity of technological inputs and components. In 2014 some sectors, such as computers, electronic and optical products, other transportation equipment, and pharmaceutical products, imported more than half of their inputs and components (Morceiro 2016). The low and stalled technological development and innovation indicators, compared with those for supplier countries’ technologies, are evident. Moreover, between 2004 and 2013, 3.4 million new workers earning up to two minimum wages were hired. However, 1.2  million workers earning more than two minimum wages were dismissed across almost all manufacturing sectors, according to data from the Ministry of Work and Jobs’s General Register of Employment and Unemployment (CAGED, Cadastro Geral de Empregados e Desempregados). These low-​wage jobs are linked to more simple manufacturing operations and to the lower status roles that Brazil’s domestic industry has acquiesced in. In summary, assembly using imported inputs and components probably had a limited contribution to value-​added and productivity increases. Third and finally, the acute currency overvaluation during most of the period contributed to the factors described in the previous hypotheses and limited price increases for manufacturing products, thus decreasing the productive sector’s potential revenue. Inflation in manufacturing product imports was very low between 2004 and 2013: 1.24% per annum, and negative or close to zero for some HT-​MHT sectors, based on IBGE (2016a) data.18 Other structural factors have been present in Brazilian industry for a long time. According to the Brazilian Association for Machinery and Equipment (ABIMAQ, Associação Brasileira de Máquinas e Equipamentos), the typical Brazilian industrial park is relatively old, with few automatic machines. The average was 17 years old in 2015, twice that of highly industrialized countries’ industrial parks, for instance, eight years in Germany. Moreover, the capital stock per Brazilian worker is low, approximately a quarter of the American or Japanese.19 With the exception of the post-​2004 period, Brazil has had very few episodes of growth during the last 35 years that could have justified a wide-​ranging technological update. Systemic and institutional obstacles have accumulated and have impaired manufacturing productivity. These especially include outdated and inadequate infrastructures (transportation modes, in particular); poor quality of education; an unfavorable business environment; a widening technological gap between Brazil and leading countries (Amann and Baer 2012; Baer 2015); and the high and complex tax burden on manufacturing products (Morceiro 2012, 218). In addition to such severe contextual difficulties facing the Brazilian business sector, most companies suffer from bad management practices, and highly educated personnel are relatively scarce and are not necessarily trained in areas relevant to innovative activities, such as engineering and technical or natural sciences (Menezes Filho et al. 2014).

256   Paulo César Morceiro In summary, the loss of competitiveness of Brazilian industry during the twenty-​first century has been widespread, regardless of sector.

12.6.  Sectoral Performance of Manufacturing during the Recent Crisis (2014–​2016) The Brazilian economy entered recession in 2014, when GDP increased by only 0.10%. It decreased by 3.85% in 2015, and by 3.6% in 2016, and forecasts indicated growth below 1% in 2017 (according to estimates from the December 2017 Focus Report). From 2011, Brazilian economic growth lost its dynamism. Serrano and Summa (2015) argue that the restrictive nature of changes in fiscal and monetary policies, such as increases in the basic interest rate, macro-​prudential measures to control consumer credit, and public investment reduction, reduced the aggregate demand growth rate. The 2014 presidential election was disputed and generated uncertainty about the probable winner and the economic policy to be adopted. This explains the decrease in investment in 2014. The political crisis deepened in 2015 with the initiation of the process of impeachment of Dilma Rousseff, which continued until August 2016, when her mandate was eventually removed. This political crisis caused various uncertainties for economic decisions, deepening the crisis. After 10 years of high growth, household consumption increased only 1.3% in 2014 and then decreased 4.0% in 2015, and GFCF decreased 4.5% in 2014 and 14.1% in 2015 (IBGE 2016b). Furthermore, unemployment increased. After reaching a low of 6.2% in the final quarter of 2013, it had increased to 11.3% by the second quarter of 2016 (IBGE 2016c). Manufacturing physical production has been decreasing since 2014 due to the retraction of domestic demand. A –​18.2% retraction can be observed for the manufacturing industry comparing the first seven months of 2016 with those of 2013 (Table 12.2). A decrease is observed for all sectors, except for food, which has remained constant. The sharpest retraction occurred in higher income and credit-​elastic sectors such as vehicles and computers, and in fixed investment sensitive sectors such as machinery and equipment (Table 12.2). In 2016, approximately half of sectors operated with production below the 2003 level, and almost all sectors featured high idle capacity. In summary, the continuous growth of 2004–​2013 came to an end, and the subsequent crisis partially eliminated the achievements of previous cycles. Manufacturing production decreased to the 2004 level. As such, it is crucial that the current crisis be better analyzed and understood, particularly in relation to the structural characteristics that have led to an investment decrease since the most recent moment of high demand in 2012.

HT and MHT

LT and LMT

Table 12.2 Physical Industrial Production Index

Activity Sector

2013

2014

2015

January–​ July 2016

Food

112

111

109

112

Beverages

Variation between 2013 and January/​ July 2016 (%) 0.3

151

152

145

143

–​5.0

Tobacco

88

86

77

70

–​20.8

Textiles

83

77

66

62

–​24.8

Clothing apparel

91

89

78

73

–​20.3

Leather, leather products, and footwear

77

74

68

67

–​12.6

Wood and products of wood, except furniture

88

86

82

83

–​5.7

Pulp, paper, and paper products

126

124

123

125

–​0.2

Refined petroleum, coke, and biofuels

123

125

118

111

–​9.3

Rubber and plastics

118

114

103

96

–​18.3

Other non-​metallic minerals

138

134

124

113

–​17.9

Basic metals

106

99

90

83

–​22.0

Metal products, except M&E

121

109

96

87

–​27.9

Furniture

134

125

106

97

–​28.0

Other manufacturing

122

116

111

103

–​15.9

Soap, detergents, cleaning, cosmetics, perfumes and personal hygiene

137

141

136

136

–​0.6

Other chemicals

122

117

111

110

–​9.4

Pharmaceutical

157

161

141

140

–​11.0

Computers, electronic and optical products

126

122

85

69

–​45.3

Electrical equipment

144

134

119

112

–​22.7

Machinery and equipment (M&E)

142

134

115

101

–​28.7

Vehicles, trucks, and buses

183

153

113

98

–​46.3

Other transportation equipment

235

235

213

174

-​–​25.9

Total manufacturing

128

123

111

105

–​18.2

Note: 2003 = 100, with seasonal adjustments. Source: Elaborated using the Monthly Industrial Survey, IBGE Physical Production.

258   Paulo César Morceiro

12.7. Conclusion After a long period of low and irregular growth, the Brazilian economy began to grow at above the global average rate from 2004 to 2013. This was mainly the result of heightened domestic demand, which stimulated various industries. From the preceding discussion, it is possible to make a number of concluding observations. First, domestic demand for manufacturing products increased more than total demand during 10 years of continuous growth. However, a significant leakage of demand to other countries can be observed in the form of imports. The demand leakage was sectorally generalized, with higher intensity in HT-​MHT industries. Regardless, a significant increase in production and employment in various manufacturing sectors took place, especially in HT-​MHT sectors, and in a few LT-​LMT sectors. In summary, domestic demand for manufacturing products, and production, employment, and demand leakage were all higher in more technology-​intensive and income-​elastic sectors. Second, given that the increase in domestic demand centered disproportionately on manufacturing products, we might expect that the manufacturing sector’s proportional contribution to GDP would increase, instead of decreasing. However, this did not occur. Therefore, Brazilian deindustrialization is a symptom of a loss of manufacturing competitiveness, rather than of a change in the composition of demand in favor of the services sector, as has happened in developed countries. Third, from a supply perspective, manufacturing growth was mainly achieved through the massive incorporation of new formal workers. There was no increase in the stock of per capita capital, technological advances, or human capital performing in higher status roles. Instead, there was an increase in employment, given Brazil’s abundance of relatively cheap labor. Fourth, the Brazilian manufacturing industry lost competitiveness during the first 15 years of the twenty-​first century. The majority of manufacturing sectors registered negative labor productivity growth, and increasing and high trade balance deficits for more technology-​intensive sectors. Even the various traditionally surplus LT-​LMT sectors began to experience substantial trade balance deficits. The manufactured product import penetration coefficient doubled, tripled, or significantly increased for some sectors, but the export coefficient remained stable or decreased. Growing imports did not result in export competitiveness. Consequently, Brazilian industry could be considered a typical “introverted fragmentation” case generating limited exports. Fifth, despite the increase in production, a lower physical/​chemical/​biological transformation per product unit can be observed. The increase in jobs paying up to two minimum wages and dismissal of all higher salary positions, increases in imported inputs, and widening of the technological “gap” when compared with leading countries are all indicators suggesting that the Brazilian manufacturing sector has performed fewer manufacturing and more assembly operations, with low value added; that is, it has specialized in low-​order tasks in sustaining productivity growth. However, pursuing increases in import content was a strategy adopted by industry to partially protect the domestic market against finished goods imports.

Competiveness of Brazilian Manufacturing Industry    259 Sixth, in more technologically intensive sectors, with a predominance of foreign transnational corporations, the regression of manufacturing transformation per product unit reinforces the thesis that Brazil does not produce HT-​MHT products, but rather only assembles them. With rare exceptions, transnational corporations’ subsidiaries are limited to adapting and performing specific improvements and quality-​control procedures in Brazil as their highest-​status activities, and also import the majority of technological components. These corporations perform in Brazil a mere fraction of the R&D their peers perform in the origin countries, and with a lower technological effort compared to national corporations (Cassiolato and Fontaine 2015). They operate with a high input and technological components import index, impeding the domestic development of innovation (Cassiolato, Szapiro, and Lastres 2015) and, as described by Baer (2014, 187–​218), foreign multinational corporations generate both benefits and costs for their host countries. In Brazil, these corporations generate limited benefits for the productive sector and limited technological development. Nonetheless, Brazil remains an important player in global industry. Brazilian manufacturing is diversified, and complex products are manufactured, such as cars, airplanes, pharmaceuticals, chemicals, and electronics. From a technological perspective, most of this production is a statistical illusion, since it generates little technological development and transformation power due to the high import content of the technological components. The country is not specializing in high value-​adding and ability-​ intensive tasks such as design, marketing, and R&D. This is a barrier for the progress of a populous country. The recent crisis demonstrates that Brazil requires rapid advances in its structural and systemic agenda for improving the business environment and regaining the manufacturing sector’s competitiveness. Consequently, it is necessary to reduce the tax burden and its administrative complexity, to modernize and expand infrastructure, improve management practices, and maintain a competitive currency rate. These are urgent actions for the short and medium terms. It is also vital to improve the quality of education and to establish a focused industrial policy, involving counterparts.

Notes 1. The author is grateful to Milene Simone Tessarin, Vinícius Rena Pereira, and Joaquim Guilhoto for their contributions in discussions, and also to Milene, Celso Pereira Neris Jr., and Victor Prochnik for reading and commenting on the first version of this text. He also appreciates the volume editors’ comments and suggestions. The author is entirely accountable for any remaining errors and inaccuracies. 2. Brazilian and global GDP increased 4.01% and 2.93% per annum, respectively, between 2004 and 2013, according to World Bank data. 3. There are a variety of definitions of deindustrialization, accounts of its causes and consequences, and distinct approaches according to a country’s development level. For the main Brazilian and international reviews, see Morceiro (2012), Hiratuka and Sarti (2015), and Britto (2015). 4. Knowledge-​intensive services have recently been recognized in the literature for some of their manufacturing characteristics. 5. From an employment perspective, it is observed that Brazil has a distinct trajectory behavior when compared to developed countries, whose industrialization absorbed an

260   Paulo César Morceiro important part of the economy’s employment (approximately 30%). Brazil’s latecomer industrialization was capital intensive and labor saving, since the country adopted technologies that are similar to those developed by industrialized countries with low population growth. 6. Brazilian foreign exchange reserves increased from US$37.8 to US$373.1 billion from 2002 to 2012. 7. E.g., health assistance, maternity leave, unemployment insurance, and retirement by years of service. 8. The highest growth occurred in 2006–​2009, when credit (in % of GDP) increased by 15 percentage points. 9. The real estate credit term was extended from 20 to 35 years, and it was occasionally possible to purchase a vehicle using a 96-​month loan term without a down payment. Borça Jr. and Guimarães (2015) report that the average loan term more than doubled in 2003–​2013 and more than tripled for some credit lines. 10. Currency devaluations of 16.7% and 10.4% occurred in 2012 and 2013, respectively. However, these devaluations were insufficient for improving Brazilian tradable goods’ competitiveness. The Economist’s Big Mac index indicated that the Brazilian currency was still overvalued. 11. This income transfer program served 6.6 million beneficiaries in 2004 and 14.1 million in 2013 (Tesouro Nacional 2016). However, in 2013 Bolsa Família was only accountable for 0.75% of Brazilian household consumption. 12. In this section and subsequent ones, monetary variables such as total demand, domestic demand, industrial production, export, import, and gross added value are measured using basic and constant prices. IBGE (2016a) provides these variables in current and previous years’ prices since 2000. Therefore, it is possible to calculate the corresponding sectoral deflators. 13. Total demand is a proxy variable obtained from the combination of production and imports. The effective demand difference is in inventory. We assume that inventory variation has little impact on effective demand, since we are working on a longer period (2000–​2013). 14. This study adopts the industrial categorization by technological intensity from the Organisation for Economic Co-​operation and Development (OECD). However, the LT-​ LMT and HT-​HMT categories were grouped. 15. From a business perspective, the industry’s profitability was obtained from the productive and financial sectors, and from the resale of imported products. A significant factor in inducing productive investments is productive sector profitability, which is difficult to calculate. Therefore, it is impossible to confirm if productive sector profitability was sufficient to induce investment expansion. 16. The difference between total demand and domestic demand is that total demand includes exports. 17. In January 2014, E class included families with an income lower than 1.7 minimum wages, D class between 1.7 and 2.8, C class between 2.8 and 11.9, B class between 11.9 and 15.6, and A class above 15.6. A typical Brazilian family was composed of three individuals in 2013. 18. Author’s calculation based on IBGE (2016a) data. Annual sectoral imports deflator (negative, in some cases) from 2004 to 2014: other transportation equipment (–​3.0%); metal products (–​1.5%); electrical equipment (–​1.4%); computers, electronic and optical products (–​0.9%); machinery and equipment (–​0.5%), and vehicle parts and accessories (–​0.2%). 19. Unpublished technical note:  Brazilian Industrial Park Modernization Plan (MODERMAQ—​Plano de Modernização do Parque Industrial Brasileiro) (Abimaq 2014).

Appendix 12.1 Performance Indicators of the Brazilian Manufacturing Industry between 2003 and 2013 (%, 2003 base year) Share in Manufacturing Annual Growth Rate Production 2004–​2013 Labor Workers Productivity

Sectors of activity

2003

2013

GDP

LMT and LT

65.3

60.3

1.4

Food and beverages

20.6

18.4

0.8

4.3

–​3.3

0.7

0.5

–​0.6

0.2

–​0.9

Tobacco

2.8

–​1.4

Cumulative Growth 2004–​2013

Export Share of Production

Import Penetration

Trade Balance as a Share of Production

Domestic demand

Total demand

2003

2013

2003

2013

2003

2013

38.0

33.5

15.3

13.1

5.9

13.6

9.9

–​0.5

19.5

24.8

22.9

17.7

16.9

3.8

6.9

14.5

10.7

–​4.9

–​1.8

8.9

38.1

53.3

27.7

47.1

14.3

11.6

Production 23.6

Textiles

2.8

2.1

0.5

1.6

–​1.1

0.4

23.7

20.0

8.6

6.7

7.8

23.6

0.9

–​22.1

Wearing apparel

2.6

2.0

0.4

2.1

–​1.7

3.5

32.5

23.5

8.1

1.5

2.3

18.2

5.9

–​20.4

Leather, leather products, and footwear

1.9

1.2

–​0.7

0.9

–​1.6

–​13.7

0.2

–​8.2

30.7

25.7

2.5

10.1

28.9

17.4

Wood and products of wood, except furniture

1.3

0.9

–​0.5

–​0.6

0.1

–​5.3

24.6

–​4.6

35.3

15.1

2.5

2.8

33.7

12.7

Pulp, paper, and paper products

3.4

3.5

2.7

2.9

–​0.2

34.7

35.9

41.1

25.1

29.4

8.2

14.2

18.4

17.7

Printing and reproduction of recorded media

1.3

1.1

0.6

2.0

–​1.4

10.2

10.6

10.4

0.2

0.0

0.1

0.3

0.0

–​0.3

Refined petroleum, coke, and biofuels

12.4

13.5

–​1.1

6.3

–​7.0

45.3

54.0

52.0

6.1

5.0

6.1

10.3

0.0

–​5.9

Rubber and plastics

3.7

3.4

1.7

4.9

–​3.1

21.8

46.6

43.9

8.1

7.3

10.2

24.7

-​2.3

–​23.1

Other non-​metallic minerals

2.8

3.0

3.4

4.2

–​0.8

43.4

71.5

62.2

11.2

6.6

5.5

16.9

6.0

–​12.4

Basic metals

5.8

4.7

1.3

3.8

–​2.3

9.7

31.9

22.5

33.0

30.4

10.2

22.4

25.3

10.2

Metal products, except M&E

3.5

3.7

3.9

3.7

0.1

42.5

68.0

66.5

6.7

6.8

6.8

21.0

-​0.1

–​18.0 (continued)

Appendix 12.1 Continued Share in Manufacturing Annual Growth Rate Production 2004–​2013 Sectors of activity Furniture and other manufacturing * MHT and HT

Export Share of Production

Import Penetration

Trade Balance as a Share of Production

Total demand

2003

2013

2003

2013

2003

Cumulative Growth 2004–​2013

GDP

Labor Workers Productivity

2.4

3.0

1.4

1.6

26.7

46.5

40.8

7.2

3.6

4.3

14.0

2003

2013

2.6

Production

Domestic demand

2013

3.1 –​12.1

34.7

39.7

4.7

5.0

–​0.2

53.1

94.4

89.2

17.0

17.2

25.2

41.2

–​10.9 –​40.8

Chemicals

9.8

8.7

2.3

1.1

1.2

18.8

44.6

42.3

12.3

12.4

25.4

38.8

–​17.6 –​43.1

Pharmaceuticals

2.1

2.5

5.1

2.9

2.1

55.2

65.3

68.2

3.5

6.2

25.1

31.7

–​28.9 –​37.3

Computers, electronic and optical products **

4.2

5.5

6.0

4.6

1.3

74.5

148.1

128.9

14.1

3.5

36.9

50.2

–​36.2 –​93.6

Electrical equipment

2.6

2.6

3.3

5.9

–​2.4

34.4

79.9

75.3

11.3

10.8

24.7

43.4

–​17.7 –​57.5

Machinery and equipment (M&E) ***

5.9

6.7

3.5

7.0

–​3.3

52.5

114.0

100.8

17.3

13.1

26.5

45.0

–​12.6 –​57.9

Motor vehicles and bodies (coachwork)

5.4

8.1

7.1

5.1

1.9

100.6

168.9

135.3

23.9

14.1

8.5

23.0

Parts and accessories for motor vehicles

2.7

2.8

4.2

3.8

0.4

34.8

104.8

82.8

29.3

23.9

27.9

48.8

Other transport equipment

2.0

2.9

7.1

7.1

–​0.1

99.4

35.2

141.1

31.6

83.2

19.2

70.7

Total manufacturing

100.0

100.0

2.8

3.3

–​0.5

33.9

60.4

55.3

15.8

14.7

13.5

26.8

2.7 –​16.5

Total economy

—​

—​

3.8

2.0

1.7

45.5

53.6

53.6

7.2

7.6

6.2

11.5

1.1

Note: All monetary variables are measured in basic and constant prices. * Except medical-​dental instruments; ** And medical-​dental instruments; *** And maintenance, repair, and installation. Source: Author’s calculations from IBGE data (2016a).

16.8

–​11.5

2.1 –​48.8 15.3

42.6

–​4.4

Competiveness of Brazilian Manufacturing Industry    263

References Amann, E., and W. Baer. 2012. “Brazil as an Emerging Economy: A New Economic Miracle?” Brazilian Journal of Political Economy 32 (3): 412–​423. Baer, W. 2014. The Brazilian Economy:  Growth and Development, 7th edition. Boulder, CO: Lynne Rienner. Baer, W. 2015. “Institutional Obstacles to Brazil’s Economic Growth and Development.” Revista de Desenvolvimento e Políticas Públicas 1 (1): 6–​17. Bastos, C. P., R. Bielschowsky, E. C. Pinto, N. Guaita, and S. Guaita. 2015. “Rentabilidade, investimento e produtividade na indústria de transformação brasileira:  2000–​2009.” In Produtividade no Brasil: Desempenho e determinantes, Vol. 2, edited by F. De Negri and L. R. Cavalcante, 209–​254. Brasília: ABDI and IPEA. Bielschowsky, R., G. C. Squeff, and L. F. Vasconcelos. 2015. “Evolução dos investimentos nas três frentes de expansão da economia brasileira na década de 2000.” IPEA Texto para Discussão 2063. Brasília: IPEA. Borça, G., Jr., and D. Guimarães. 2015. “Impacto do ciclo expansionista de crédito à pessoa física no desempenho da economia brasileira.” Revista do BNDES 43: 119–​159. Britto, J. 2015. “Ajuste produtivo e dinâmica setorial da indústria brasileira no período 1996–​2013: Uma análise dos dados da PIA-​IBGE.” Mimeo. Rio de Janeiro, UFF http://​pdf.blucher.com. br.s3-​sa-​east-​1.amazonaws.com/​engineeringproceedings/​1enei/​031.pdf. Cassiolato, J. E., and P. Fontaine. 2015. “O papel das empresas transnacionais no sistema nacional de inovação brasileiro.” In Sustentabilidade socioambiental em um contexto de crise, Vol. 1, edited by J. E. Cassiolato, M. G. Podcameni, and M. C. C. Soares, 233–​262. Rio de Janeiro: Epapers. Cassiolato, J. E., M. Szapiro, and H. Lastres. 2015. “Dilemas e perspectivas da política de inovação.” In Indústria e desenvolvimento produtivo no Brasil, edited by N. Barbosa, N. Marconi, M. C. Pinheiro, and L. Carvalho, 377–​416. Rio de Janeiro: Elsevier. Castro, A. B. 1985. “Ajustamento x transformação: a economia brasileira de 1974 a 1984.” In A economia brasileira em marcha forçada, edited by A. B. Castro and F. E. P. Souza, 11–​98. São Paulo: Paz e Terra. Cavalcante, L. R., and F. De Negri. 2015. “Consensos e dissensos sobre a evolução da produtividade na economia brasileira.” In Produtividade no Brasil: Desempenho e determinantes, Vol. 2, edited by F. De Negri and L. R. Cavalcante, 541–​563. Brasília: ABDI and IPEA. De Negri, F., and L. R. Cavalcante. 2015. Produtividade no Brasil: Desempenho e determinantes, Vol. 2. Brasília: ABDI and IPEA. European Commission. 2012. A Stronger European Industry for Growth and Economic Recovery. Brussels: European Commission. Hiratuka, C., and F. Sarti. 2015. “Transformações na estrutura produtiva global, desindustrialização e desenvolvimento industrial no Brasil: Uma contribuição ao debate.” Texto para Discussão no. 255. Campinas: IE/​Unicamp. IBGE [Instituto Brasileiro de Geografia e Estatística]. 2016a. Sistema de Contas Nacionais—​ Brasil—​referência 2010: Série retropolada, 2000–​2013. Rio de Janeiro: IBGE. IBGE. 2016b. Contas nacionais trimestrais. Rio de Janeiro: IBGE. IBGE. 2016c. Pesquisa nacional por amostra de domicílios contínua—​PNAD Contínua. Rio de Janeiro: IBGE. IBGE. 2016d. Projeção da população do Brasil por sexo e idade: 2000–​2060. Rio de Janeiro: IBGE. Kaldor, N. 1967. Problems of Industrialization in Underdeveloped Countries. Ithaca, NY: Cornell University Press.

264   Paulo César Morceiro McKinsey. 2012. Manufacturing the Future: The Next Era of Global Growth and Innovation. McKinsey Global Institute. https://​www.mckinsey.com/​business-​functions/​operations/​ our-​insights/​the-​future-​of-​manufacturing Medeiros, C. A. 2015. Inserção externa, crescimento e padrões de consumo na economia brasileira. Brasília: IPEA. Messa, A. 2015. “Determinantes da produtividade na indústria brasileira.” In Produtividade no Brasil: Desempenho e determinantes, Vol. 2, edited by F. De Negri and L. R. Cavalcante, 23–​41. Brasília: ABDI and IPEA. Miguez, T. H. L. 2016. “Evolução da formação bruta de capital fixo na economia brasileira 2000–​2013: Uma análise multissetorial a partir das Matrizes de Absorção de Investimento (MAIs).” PhD dissertation, IE/​UFRJ, Rio de Janeiro. Menezes Filho, N. A., B. Komatsu, A. Lucchesi, and M. Ferrario. 2014. “Políticas de Inovação no Brasil.” Policy Paper no. 11. São Paulo: Centro de Políticas Públicas (CPP) do Insper. MME [Ministério de Minas e Energia]. 2014. Impactos do Programa Luz para Todos:Apresentação de 10 anos. Brasília: MME. Morceiro, P. C. 2012. Desindustrialização na economia brasileira no período 2000-​2011: Abordagens e indicadores. São Paulo: Cultura Acadêmica. Morceiro, P. C. 2016. “Esgarçamento do tecido industrial brasileiro: Uma análise setorial dos coeficientes de insumos e componentes importados.” Mimeo, São Paulo: Universidade de Sao Paulo. National Science and Technology Council. 2012. A National Strategic Plan for Advanced Manufacturing. No. 20502, February. Washington, DC: Executive Office of the President. Neri, M. 2014. “Social e renda: A classe média brasileira.” Assuntos Estratégicos 1, Nov 2014. Brasília: Secretaria de Assuntos Estratégicos da Presidência da República. Palma, J. G. 2008. “‘De-​industrialization,’ ‘Premature’ De-​Industrialization and the Dutch Disease.” In The New Palgrave Dictionary of Economics, edited by S N. Durlauf and L. E. Blume, 2nd edition, 401–​410. London: Palgrave Macmillan. Pereira, V. R. 2016. “As transformações na estrutura produtiva brasileira na era do Real: uma abordagem insumo-​produto comparada.” XXI Encontro Nacional de Economia Política. São Bernardo do Campo: ENEP/​SEP. Ricupero, R. 2005. “Desindustrialização precoce: Futuro ou presente do Brasil? Síntese das principais teses e demonstrações do relatório Trade and Development Report 2003—​ UNCTAD.” São Paulo: Mimeo. Rodrik, D. 2007. “Industrial Development: Stylized Facts and Policy Directions.” In Industrial Development for the 21st Century: Sustainable Development Perspectives. New York: United Nations. Sarti, F. 2015. “Padrão de crescimento e desenvolvimento industrial.” In Indústria e desenvolvimento produtivo no Brasil, edited by N. Barbosa, N. Marconi, M. C. Pinheiro, and L. Carvalho, 519–​543. Rio de Janeiro: Elsevier. Serrano, F., and R. Summa. 2015. Aggregate Demand and the Slowdown of Brazilian Economic Growth from 2011–​2014. Washington, DC: Center for Economic and Policy Research, CEPR. Silva, F., N. A. Menezes Filho, and B. Komatsu. 2016. “Evolução da produtividade no Brasil:  Comparações internacionais.” Policy Paper No. 15. São Paulo:  Centro de Políticas Públicas (CPP) do Insper. Singh, A. 1987. “Manufacturing and De-​industrialization.” In The New Palgrave Dictionary of Economics, Vol. 3, edited by J. Eatwell, M. Milgate, and P. Newman, 301–​ 308. London: Macmillan.

Competiveness of Brazilian Manufacturing Industry    265 Szirmai, A., and B. Verspagen. 2015. “Manufacturing and Economic Growth in Developing Countries, 1950–​2005.” Structural Change and Economic Dynamics 34: 46–​59. Tesouro Nacional. 2016. Gasto social do governo central: 2002 a 2015. Brasília: Secretaria de Política Econômica. Thirwall, A. P. 2002. The Nature of Economic Growth:  An Alternative Framework for Understanding the Performance of Nations. Cheltenham, UK: Edward Elgar. UNIDO [United Nations Industrial Development Organization]. 2015. Industrial Development Report 2016: The Role of Technology and Innovation in Inclusive and Sustainable Industrial Development. Vienna: UNIDO.

Chapter 13

The Agricu ltu ra l  Se c tor Carlos José Caetano Bacha

13.1. Introduction Since the arrival of the Portuguese at the end of the fifteenth century, agricultural production in Brazil has continually expanded, although the sector is no longer the most important in the Brazilian economy. In 1952, agriculture was responsible for 25% of Brazilian gross domestic product (GDP), a percentage that had declined to 5.2% by 2015 according to the Brazilian Institute of Geography and Statistics (IBGE). However, agriculture has stimulated other economic activities. The Brazilian Center for Advanced Studies in Applied Economics (CEPEA) determined that Brazilian agribusiness contributed 21.5% to the country’s GDP in 2015. Over the last five centuries, agricultural and agro-​processed product exports have been an extremely significant contributor to Brazil’s trade balance. While their contribution to total exports has decreased as the manufacturing and services sectors have matured, they still represented 37.7% of all Brazilian exports in 2013, which corresponded to 5.5% of total world agricultural and agro-​processed exports (FAO database). The Brazilian agricultural sector’s expansion has been advanced by the addition of new crops and livestock products, the development of innovative technologies, and the opening of new farming areas. This chapter aims to explain Brazil’s agricultural evolution by concentrating on seven factors that have stimulated and supported the sector’s expansion: (1) Brazil’s vast territory, diversified soils, and favorable weather; (2) the construction of infrastructure to transport and/​or store agricultural and agro-​processed products; (3) increased foreign-​ market orientation; (4) agricultural policies geared toward stimulating agricultural production mainly directed to foreign markets; (5) the participation of large multinational corporations in the financing of farming activities and the marketing of Brazilian agricultural products in foreign markets; (6) advances in agricultural technology, whether imported or homegrown; and (7) a transference of crops among Brazilian regions in a

The Agricultural Sector   267 manner that helped to integrate the country, spur the development of new farming areas, and successfully adapt to market demands. The changing mix of these seven conditioning factors over time has determined the course of agricultural expansion in Brazil. There is a great deal of available literature focused on the evolution of Brazilian agriculture, but the majority of these inquiries emphasize only some of the aspects noted in the preceding, and address earlier time periods. For example, Baer (2008), Furtado (1989), and Prado Jr. (1982) analyze the agricultural cycles of Brazilwood, livestock and sugarcane production during Brazil’s colonial period, concluding that the country’s vast shoreline, good soil, and favorable climatic conditions facilitated these agricultural activities. These authors also place emphasis on the role of foreign trading companies, such as the Dutch West India Company, in opening international markets to Brazilian sugar, especially Western European markets. The aforementioned authors, as well as others such as Delfim Netto (1981), Simonsen (1938), and Albuquerque and Nicol (1987), analyze the Brazilian coffee production’s evolution during the nineteenth and early twentieth centuries, when Brazil was under the imperial regime and later became a republic. These authors argue that good soil and weather conditions, in conjunction with accessible arable lands in the Vale do Paraíba in the states of Rio de Janeiro and São Paulo, permitted an enormous expansion of the coffee plantation system. They also note that agricultural policies that the Brazilian government put in place over the first three decades of the twentieth century to support coffee grower incomes were able to soften the coffee market’s instability. Matos (1974) and Saes (1981) examine transportation infrastructure during the nineteenth and early twentieth centuries, and show that it was designed to facilitate the movement of agricultural export goods (mainly coffee exports) to the harbor in the country’s then-​capital, Rio de Janeiro, and, more important, to the harbor at Santos in São Paulo. They note that it was mainly joint ventures between domestic and foreign entrepreneurs that led to the construction of a railway network linking coffee plantations to export facilities. Albuquerque and Nicol (1987), Szmrecsányi (1990), and Bacha (2012) present studies focused on the evolution of Brazilian agriculture from the 1930s to the 1990s. Their work primarily analyzes agriculture’s traditional roles in economic development and how agricultural policies and land availability have supported agriculture in the fulfillment of these roles. The authors pay little attention to the role of larger multinational companies in the stimulation of Brazilian agriculture’s advance during the twentieth century. This chapter widens the analysis of agricultural expansion in Brazil to take in a more extensive time period than the studies noted in the preceding, although focusing particularly on the period from the 1970s onward. It also concentrates on the role played in this growth by large multinational corporations, technological advance, and changes in crops and areas under cultivation. In order to help grasp the current situation of Brazilian agriculture in the broader context, the next section addresses its early development.

268   Carlos José Caetano Bacha

13.2.  Agricultural Cycles in Colonial Times The signing of the Treaty of Tordesillas in 1494 created a demarcating meridian line that had the effect of dividing in principle the as yet largely unexplored (by Europeans) South America between Spain and Portugal. Portugal’s putative share included most of what is now Brazil. At the very beginning of the sixteenth century, Portugal sent naval expeditions to explore these lands. Failing to discover valuable metals in the first decades of this exploration, Portugal decided to attempt to extend control over the territory through agricultural exploitation (Fausto 1994, 30–​43). Portugal sought to protect its new territory by establishing permanent colonies. As a first step in this process, in the 1530s the Portuguese government began to grant enormous tracts of Brazilian land (captaincies) to royal favorites, military men, and bureaucrats, giving them the task of administration and economic development. The ruler of each captaincy (called capitão-​hereditário) allowed other Portuguese to establish agriculture in sizable land grants (named sesmarias), which could only be confiscated by the Portuguese Crown. This process is the origin of the latifundia (very large landholdings) in Brazil. Slavery was also an important part of Portuguese colony-​building. Initially, slave labor came from the indigenous workforce, but by the mid-​sixteenth century, slaves were being imported from Africa. Latifundia and slavery became the trademarks of agricultural expansion in colonial Brazil. Despite the abolition of slavery on May 13, 1888, the latifundia continued to be predominant from the nineteenth century throughout the first three decades of the twentieth century (Guimarães 1968). Portugal also adopted an exploitative policy toward its Brazilian colony. The Portuguese Crown limited Brazilian manufactured imports to products from Portugal and restricted the sale of Brazilian agricultural exports to Portuguese companies linked with the largest European trading companies. Brazilwood was the first significant agricultural export to come out of Brazil, and its exportation generated income for the Portuguese Crown during the sixteenth century. This was the first case of non-​native settlers depleting Brazilian forest resources, which would be a salient feature of all future agricultural cycles (Dean 1997). With technology imported from Portuguese colonies in Africa and production allocated to foreign markets, sugarcane became the economic foundation in two of Portugal’s seventeenth-​century Brazilian colonies, Pernambuco and São Paulo (Baer 2008; Furtado 1989; Prado Jr. 1982). Sugar mill construction in Pernambuco was financed by converted Jews (New Christians) who also helped to market the sugar in Portugal (Melo 1996). In addition, the Dutch West India Company was granted partial responsibility for sugar sales in European markets (Furtado 1989). Due to a lack of suitable harbors and a complete lack of roads penetrating the interior, late sixteenth-​century Portuguese settlers occupied only small patches of the Brazilian

The Agricultural Sector   269 shoreline and even less of the country’s interior. This began to change in the early seventeenth century, when Portuguese settlers and fortune hunters in search of valuable metals began to penetrate Brazil’s vast interior. Out of necessity, the fortune hunters, known as bandeirantes for the tiny Portuguese flags they carried, started opening small agricultural plots in their areas of exploration and thereby expanded the Portuguese colonial presence. The seventeenth century reflects characteristics of Brazilian agriculture that can be found today: small-​scale, subsistence agriculture flourishing in the country’s more isolated inland areas; large-​scale, foreign-​market-​oriented agriculture dominating in fertile areas easily linked with the shore (Guimarães 1968).

13.3.  Coffee Plantations during the Imperial Era and the First Republic To escape Napoleon’s invading forces, the Portuguese court relocated to Brazil in 1808. The regent prince then opened Brazil to trade with nations other than Portugal. In an attempt to increase export earnings, he encouraged Portuguese nobility relocating in the New World to plant coffee trees on the outskirts of the city of Rio de Janeiro by distributing coffee seeds to the recently arrived aristocrats. The new Brazil-​Portugal monarch also created the Rio de Janeiro Botanical Garden, where exotic plants, including coffee trees, were tested and adapted to local weather and soil conditions (Bacha 2012; Szmercsányi 1990). The growing European market for coffee during the imperial governments of Dom Pedro I and Dom Pedro II spurred the expansion of coffee plantations into extensive stretches of available fertile land found in the nearby Paraíba River Valley provinces of southern Rio de Janeiro and northeastern São Paulo. Over the first seven decades of the nineteenth century, most of Brazil’s increasing coffee production was destined for the foreign market and was exported from Rio de Janeiro’s harbor. Production growth was facilitated by financing obtained from exporting companies and was repaid by the growers with bags of coffee, a method of agricultural financing that is still very much in practice today. According to Szmrecsányi (1990), technological improvements developed during the nineteenth century increased efficiency in coffee production’s industrialized segment, but few advances trickled down to the plantation level, which was left to increase production by converting native forests into cultivated land, with no thought given to forest conservation (Dean 1997) In 1867, construction was completed on a railway linking the Port of Santos to the city of São Paulo and west from there to the city of Jundiaí. The system, built using British investment capital and engineering expertise, connected coffee plantations in the extensive, fertile inland areas of western São Paulo to export facilities at the port of Santos (Furtado 1989; Monbeig 1984; Prado Jr. 1982). The Brazilian railroad network grew from

270   Carlos José Caetano Bacha 1,801 kilometers in 1875 to 31,967 kilometers in 1929, an annual geometric growth rate of 4.8% over the period, with most of this expansion designed to facilitate the production and transportation of coffee. Despite the railroad network’s rapid growth and its important role in making viable the exploitation of São Paulo’s inland areas, it suffered from a handicap: different routes were built using different railway gauges, which greatly impeded the flow of traffic within the system, a problem that persists to today. By the end of the nineteenth century, degradation of once-​fertile soil in the Paraíba River Valley and the 1888 abolition of slavery brought a halt to the coffee plantation system’s rapid advance in this region (Furtado 1989). Labor scarcity caused by slavery’s abolition was so severe that from the last quarter of the nineteenth century through the first three decades of the twentieth, the imperial and later republican governments acted to increase the supply of workers by stimulating European immigration (Furtado 1989; Prado Jr. 1982). The nineteenth century ended with the state of São Paulo having surpassed the state of Rio de Janeiro as Brazil’s largest coffee-​producing state. With the addition of the United States to its export market, Brazil became the world’s largest coffee exporter by the end of the nineteenth century (Delfim Netto 1981; Simonsen 1938). Agriculture in São Paulo was still concentrated in coffee production, and that production continued to grow; however, coffee’s share of total agricultural acreage was shrinking as the production of other crops forced its way onto newly cleared farmland. Not only were coffee plantations forced to compete for land with sugarcane and cotton plantations, but inside most of the larger plantations (mostly coffee plantations) employees used land to plant crops, such as rice, beans, and corn, to satisfy their own needs. The share of cultivated land planted with coffee in São Paulo dropped from 68% in the 1894–​1895 season to 42.6% in the 1918–​1919 season, as shown in Table 13.1. The

Table 13.1 Area Occupied with Various Crops in the State of São Paulo for Selected Harvesting Years (Thousand Hectares) Crop

1894–​1895

1900–​1901

1904–​1905

1910–​1911

1914–​1915

1918–1​ 919

Cotton

4.0

8.3

8.4

19.2

12.2

148.0

Sugarcane

9.4

25.9

48.7

49.7

62.2

80.3

382.1

751.1

875.0

900.1

1,023.8

1,164.1

5.2

49.2

66.4

72.5

100.3

168.9

38.7

99.5

156.8

183.7

228.7

130.2

6.2

5.0

4.8

5.3

4.4

32.1

101.6

220.2

347.0

372.9

508.1

815.2

14.5

24.2

31.0

36.3

48.4

195.6

561.7

1,183.4

1,538.1

1,639.7

1,987.9

2,734.4

Coffee Rice Beans Tobacco Corn Other crops Total

Source: Elaborated from Szmrecsányi (1990, 68).

Table 13.2 Share of Some Agricultural Goods of Total Brazilian Exports (Mean by Decade)

Cocoa

coffee

mate herb

Tobacco

Other Agricultural and Non-​agricultural Goods

3.07

0.94

45.87

1.21

3.08

15.75

5.52

1.22

56.32

1.50

3.46

10.53

4.54

7.61

1.71

60.49

1.08

2.83

11.18

2.50

14.21

1.36

65.38

1.32

2.02

7.08

1.55

2.39

25.62

2.83

53.16

2.72

2.60

9.13

2.42

1.63

16.37

3.48

52.06

3.17

2.58

18.29

2.33

2.69

2.89

3.21

67.25

2.66

2.12

16.86

0.52

11.14

0.97

3.99

56.25

2.03

1.66

23.42

1940–​1945

0.54

10.01

2.22

3.35

31.91

0.94

1.09

49.94

1946–​1964

2.25

8.34

0.28

4.62

56.21

0.76

1.52

26.03

Decade

Sugar

Cotton

Rubber

1860–​1869

12.34

17.74

1870–​1879

11.79

9.66

1880–​1889

10.56

1890–​1899

6.13

1900–​1909 1910–​1919 1920–​1929 1930–​1939

Source: Elaborated by the author from IBGE (1990)—​Estatísticas Históricas do Brasil.

272   Carlos José Caetano Bacha shift away from the coffee monoculture shown in Table 13.1 was also a reaction to the outbreak of World War I and a strategy to avoid the food shortages that took place at the gold camps in the neighboring state of Minas Gerais during the eighteenth century. During the nineteenth and early twentieth centuries, foreign-​market-​oriented agriculture appeared in the present-​day Brazilian states of Bahia (cotton), Pernambuco (sugar), Maranhão (cotton), and Pará (rubber), according to Furtado (1989), but this development was patchy. In these and other provinces (which became states during the republican era), small-​scale, subsistence agricultural production was the norm, maintaining the dichotomy between subsistence farmers and large-​scale, foreign-​market-​oriented agriculture found in Brazil throughout the nineteenth century and later (Szmrecsányi 1990). The initial effort to develop agricultural technology within Brazil came from Dom Pedro II during the second half of the nineteenth century. Brazil’s ruler created six imperial institutes to advance agricultural expertise during his reign, but only two made important contributions: the Imperial Agronomical Institute of Campinas in the state of São Paulo focused on coffee, and the Imperial Institute of Agriculture in Rio de Janeiro focused on sugarcane (Szmrecsányi 1990). Rarely did these institutes devote resources to assisting the small-​scale subsistence farmer. From the late nineteenth century through the first three decades of the twentieth century, coffee products were Brazil’s most valuable export. Although responsible for only 46% of exports in the 1860s, that percentage reached 65% in the 1890s and was never less than 50% through 1940 (see Table 13.2). Coffee plantations were concentrated in three states: São Paulo, Rio de Janeiro, and Minas Gerais. Large quantities of cocoa, sugar, cotton, tobacco, mate, and rubber were also exported during the period, but the great majority of the production of these items took place outside the coffee region, and their production growth was sporadic. Ranching was also an important activity in the country’s uncultivated, inland areas, but meat was rarely exported. The main livestock market was internal, mainly the southeastern states, and the products demanded in this booming market were mule and salted meats (Furtado 1989).

13.4.  Diversification and Inland Occupation from the 1930s into the 1960s The 1929 financial crisis toppled coffee prices from £5.02 per bag in 1928 to £0.90 per bag in 1939 (Lima et al. 1983, 57). As a consequence, farmers shifted production to other, more profitable agricultural activities with support from federal and state governments. Farmers in the state of São Paulo shifted their focus from coffee to sugar and cotton production. As the price of coffee fell during the 1930s, some of the larger coffee plantations began to fragment. Even former coffee plantation employees bought pieces of the plantations

The Agricultural Sector   273 they had worked on or became co-​owners (Albuquerque and Nicol 1987). This process led to an increase in the number of farms in Brazil and a simultaneous decrease in their average size, as shown in Table 13.3. Between 1920 and 1940, the number of farms in Brazil almost tripled, but their average size more than halved, from 270 hectares in 1920 to 104 hectares in 1940. Over the same period, 1920 through 1940, coffee’s share of total export value fell, but it continued to account for more than half of total Brazilian export earnings. As Table 13.2 shows, coffee’s share of Brazilian agricultural product export dropped 11 percentage points from its 1920s average to its 1930s average. Although a rather sizable fall in percentage, coffee still accounted for 56.25% of Brazilian exports in the 1930s. That decade also saw the rise of cotton and the export market shares of “other agricultural and non-​ agricultural goods” (see Table 13.2), with cotton replacing coffee in some São Paulo plantations. Brazilian cotton exports’ share in total Brazilian exports increased by a factor of four from the 1920s to the 1930s (IBGE 1990). This impressive growth was spurred by technological advances coming out of São Paulo’s agricultural research centers, such as the Agronomic Institute at Campinas, and the growing demand for cotton from the domestic textile industry and international buyers. During the 1940s and the 1950s, a new agricultural frontier was opening in São Paulo’s southern neighbor, the state of Paraná. Driven by a migration of São Paulo’s coffee culture to Paraná’s north and additional farm labor coming from the states of São Paulo and Minas Gerais, the trickle of new farming enterprises entering the state in the 1940s became a wave in the 1950s. In 1940, the state of Paraná contained 64,397 farms, holding a total area of 6,252,480 hectares. Ten years later, the number of farms reached 89,461 and they held 8,032,743 hectares. By 1960, the state contained 269,146 farms, holding 11,384,934 hectares. Between 1940 and 1960, the amount of land held by Paraná’s farmers almost doubled, and the number of farm holdings more than quadrupled. Growth in the number of Paraná farms was matched by growth in the total area under cultivation with temporary and permanent crops, which also more than quadrupled between 1940 and 1960. Crops in Paraná covered 764,370 hectares in 1940, 1,358,222 hectares in 1950, and 3,440,971 hectares in 1960. This increase in agricultural activity implied an increase of 593,852 hectares under cultivation of temporary and permanent crops in the state between 1940 and 1950. Over that period, the area under similar cultivation throughout Brazil increased only 259,627 hectares. Therefore, cropland expansion in Paraná more than offset agricultural contraction in the rest of Brazil. Although this disparity decreased over the decade of the 1950s, the growth in the area planted with permanent and temporary crops in Paraná was still impressive, representing 22% of the entire increase in Brazil (Bacha 2011). Brazilian coffee production rebounded strongly from 1945 to 1964, with the area under cultivation increasing at a geometric rate of 3.75% annually and productivity growing 4.74% annually. Paraná led this expansion, surpassing São Paulo as Brazil’s coffee production center by the start of the 1960s. In the two-​year period from 1946 to 1948, Paraná, then Brazil’s third-​largest coffee-​producing state, accounted for 9.6% of

Table 13.3 Selected Census Indicators for Brazilian Agriculture 8. Indicator

1920

1940

1950

1960

1970

1980

1985

1995–​1996*

2006

Number of agricultural establishments

648,153

1,904,589

2,064,642

3,337,769

4,924,019

5,159,851

5,801,809

4,859,865

5,175,489

Total farming area (thousand of hectares)

175,105

197,720

232,211

249,862

294,145

364,854

374,925

353,611

329,941

Number of employed people

6,312,323

11,343,415

10,996,834

15,633,985

17,582,089

21,163,735

23,394,919

17,930,890

16,567,544

Annual cropland (ha)

6,642,057 12,873,660

14,692,631

20,914,721

25,999,728

38,632,128

42,244,221

34,252,829

48,234,391

5,961,770

4,402,426

7,797,488

7,984,068

10,472,135

9,903,487

7,541,626

11,612,227

Number of cows

34,271,324 34,457,576

44,600,159

56,041,307

78,562,250

118,085,872

128,041,757

153,058,275

171,613,337

Number of pigs

16,168,549 16,849,570

22,970,814

25,579,851

31,523,640

32,628,723

30,481,278

27,811,244

31,189,339

Number of poultries

52,940,010 62,912,437

77,830,259

136,391,313

218,937,380

421,933,117

446,924,170

735,399,000

1,401,341,000

8,372

61,345

165,870

545,205

665,280

803,742

820,673

Permanent cropland (ha)

Number of tractors

1,706

3,380

Source: Elaborated by the author from Agricultural Censuses of Brazil. * Information about area, employed people and tractors refers to December 31, 1995. Information about number of animals refers to July 31, 1996.

The Agricultural Sector   275 the Brazilian coffee production, while São Paulo accounted for 49.5%. In the 1958 to 1960 period, coffee production in Paraná easily surpassed that of São Paulo, accounting for 41.7% of Brazil’s coffee production compared with São Paulo’s 31.6% (Bacha 1988). The 1960s saw coffee cultivation in Paraná reach its apex. In 1969, Paraná contained 45.3% of Brazil´s coffee acreage, São Paulo held 30%, followed by Minas Gerais with 11.5%, and Espírito Santo with 8.6% (IBGE data set). Rail route-​miles slightly increased from 1930 to 1964, growing from 31,967 kilometers in 1929 to 34,262 kilometers in 1964, which represents a miniscule 0.41% annual geometric rate of growth (1/​12 the growth rate from 1875 to 1929). As a track gauge standard was never instituted in Brazil, the difficulties in linking different railway systems encountered in earlier decades continued. The railways were generally unprofitable, and they were under increasing state or federal government control until the 1980s. The system was then mostly privatized in the mid-​1990s. Since the 1950s, Brazil’s emphasis has been to stimulate road transportation. According to the IBGE data set, there were 302,147 kilometers of roadways in Brazil in 1952 and 548,510 kilometers in 1964, a geometric rate of growth of 4.3% per year. Although most of the roadways were unpaved in the 1950s and through the first half of the 1960s (only 3.4% of all roads were paved in 1964; IBGE Statistical Yearbook), agriculture was able to expand in northern Paraná, the new agricultural frontier during this period, and in the southern portion of the Center-​West region. The Brazilian government initiated construction of an extensive highway system in the early 1960s to open Brazil’s Center-​West and North regions to development. This system was designed to connect these regions with the country’s Southeast and Northeast regions. For agriculture to flourish in the landlocked Center-​West region, access to the developed harbors of the Southeast and Northeast was a priority. Road construction intensified during the 1970s, and the Brazilian drive to open the vast areas in the Center-​West and North has continued into the new millennium (Becker 2001).

13.5.  Agricultural Expansion from 1964 through 1986: New Frontiers and Food Imports From 1964 to 1985, Brazil was run by military governments that funded a set of measures to stimulate agriculture. During the period, both crop and meat production increased, new agricultural frontiers were opened, and the production of specific crops shifted across the country. Both the military government and multinational corporations directed their attention to products destined for foreign markets and largely ignored small-​scale farming operations producing crops for the domestic market. This export-​ oriented focus had a foreseeable consequence: shortages in the domestic food supply that resulted in food importation in some years.

276   Carlos José Caetano Bacha

350

1000

300 250

800

200

600

150

400

100

200

50 2015

2013

2011

2009

2007

2005

2003

2001

1999

1997

1995

1993

1991

1989

1987

1985

1983

1981

1979

1977

0 1975

0

Millions of tons (excluding sugarcane)

1200

1973

Milllions of tons (including sugarcane)

Figures 13.1 and 13.2 show the evolution of Brazilian crop and meat production from 1973 through 2015. From 1977 to 1985, meat production increased by 1.6% per year. The growth in crop production was much more pronounced, increasing by 6.4% per year from 1973 to 1986; however, these gains were not equally distributed among crops. The production of agricultural products to meet export demand grew substantially, while basic foods needed to meet domestic demand had to be imported. For example, Brazil imported beans and onions every year from 1972 to 1976. In 1973, Brazilian onion imports equaled 4.3% of domestic production; in 1976 Brazilian bean importation equaled 7.4% of national production (Melo 1981). During the 1980s, Brazil also imported milk (in powdered form) and corn. Both milk and corn had been supplied by small-​ scale farming operations that were now unable to meet domestic demand.

Year Without sugarcane

With sugarcane

Figure  13.1.  Evolution of crop production in Brazil (including and excluding sugar cane), 1973 –​2015. Source: IBGE.

30,000,000 25,000,000

Tons

20,000,000 15,000,000 10,000,000

2015

2013

2011

2009

2007

2005

2003

2001

1999

1997

1995

1993

1991

1989

1987

1985

1983

1981

1979

0

1977

5,000,000

Year Cattle

Pigs

Chicken

Total

Figure 13.2.  Brazilian meat production, 1977–​2015.  Source: IBGE.

The Agricultural Sector   277 Table 13.4 Annual Geometric Rates of Area and Productivity for Selected Crops (Values in Percentages) 1964–​1986 Crop

Domestic

Rice

1.41

0.46

–​4.48

3.90

–​2.16

4.05

Beans

2.42

–​2.58

–​1.82

4.65

–​1.76

3.02

Manioc

0.83

–​1.63

–​1.35

0.07

0.21

0.85

Potato

–​1.23

3.48

0.19

2.06

–​1.46

3.79

1.63

1.64

0.03

4.11

1.51

4.09

Foreign market Cotton

–​1.05

1.24

–​11.02

4.70

2.62

5.32

Cocoa

1.28

3.53

0.93

–​4.50

0.02

0.70

Coffee

–​1.31

1.52

–​4.44

2.14

–​0.18

–​0.27

4.45

1.75

1.09

0.96

5.14

0.65

Orange

8.30

1.74

2.27

1.57

–​1.83

1.81

Soybean

17.83

2.46

1.24

3.59

5.48

1.40

6.24

1.81

–​11.60

0.29

2.75

2.37

Sugarcane

Wheat

Productivity Area

1997–​2015

Main Market

Corn

Area

1987–​1996

Productivity Area

Productivity

Source: Elaborated by the author from data collected from Brazil’s Statistical Yearbook—​several issues.

Data in Table 13.4 show percentage change in area under cultivation and percentage change in yield per hectare for domestic and export-​oriented crops over different time periods from 1964 through 2015. While both increases and decreases in productivity per hectare and area under cultivation can be found in each crop set from 1964 to 1986, some crops destined for foreign markets showed spectacular gains in both productivity and area under cultivation—​gains that far outstripped gains for products destined for the domestic market, especially in area under cultivation. For example, from 1964 to 1986, the area under cultivation and productivity of the soybean crop increased by 17.83% and 2.46% per year, respectively, while the corresponding percentages for manioc were 0.83% and –​1.63%. The data for cotton were skewed by various export bans that discouraged its production. Agricultural development expanded from Brazil’s southern states into the North and especially Center-​West regions during the 1970s and 1980s, due to an abundance of available arable land, new technology, supportive government policies and settlement programs, new roads, and the role of larger multinational companies. The soybean and corn crops were at the forefront of this spread. The expansion began in the southern state of Rio Grande do Sul, progressed northward to the states of Santa Catarina, Paraná, Mato Grosso do Sul, and eventually reached the states of Mato Grosso and Goiás, the heart of the Center-​West. The North and Center-​West contain 45.3% and 18.9% of Brazil’s territory, respectively. Soil in the Center-​West has a high aluminum content that limits plant growth.

278   Carlos José Caetano Bacha A Brazilian network that included state-​funded agricultural research institutes, public universities, and private organizations, and was led by the federally funded Brazilian Agricultural Research Corporation (EMBRAPA), developed a soil-​conversion process (Goldin and Rezende 1983) and new corn and soybean seed strains (Bacha 2011) that made profitable these crops’ expansion into the Brazilian savannah areas of the North and Center-​West (the Cerrado). Before these developments, the cultivation techniques and inputs in use were adapted to soybean and corn crops grown in cooler southern Brazilian states and were unsatisfactory in the hotter North and Center-​West. In 1970, the North and Center-​West regions contained 7.9% and 23.1% of Brazil’s total farm area, respectively; however, they were responsible for only 3.1% and 7.5% of the gross value of the country’s agricultural production, respectively, according to IBGE’s Agricultural Census. At that time, most farming areas in the North were forested, and those in the Center-​West were dedicated to ranching. In 1985, the North and Center-​ West regions were responsible for 4% and 9.8% of the gross value of Brazil’s agricultural production, respectively. Soybean producers located in southern Brazil generally operated small-​scale farms. New operations opening in the Center-​West (most of them established by migrants from the South region) have generally been large-​scale farms. As such, a dichotomy arose between Southern and Center-​West soybean farming operations, and this dichotomy has continued to grow over time. In 1980, the average size of soybean plantations was 15.4 hectares in the South region and 105.7 hectares in the Center-​West region. The 2006 Brazilian Agricultural Census, the most current available, indicates that the average size of a soybean plantation is 37.3 hectares in the South and 560.7 hectares in the Center-​West. During the second half of the 1960s and throughout the 1970s, Brazil’s federal government put into effect a set of agricultural policies to stimulate foreign-​market-​oriented agricultural production (Coelho 2001; Gasques and Conceição 2001; Goldin and Rezende 1993; Mueller 2010). With reluctant support from mid-​and large-​sized farming enterprises as well as from industrial tycoons, the federal government passed the Rural Labor and Land Statute in 1964, extending to the rural worker the same rights that had existed for urban employees since 1943. These statutes democratized the working relationship in agriculture but also pushed up labor costs, causing farmers to focus on more profitable, market-​oriented activities. In 1965, the federal government created the National System of Rural Credit (SNCR), which enforces each year the setting up of low-​interest loans to farmers for the purchase of industrial inputs, such as fertilizer, pesticides, herbicides, and agricultural equipment and machinery. The SNCR was one of the key factors in increasing agricultural productivity. From 1965 to 1985, the SNCR assisted mainly mid-​and large-​scale, market-​ oriented farmers. These farmers used subsidized rural credit to buy products from domestic industry, which was a sound reason for the industrial sector’s support for the rural credit program (Goldin and Rezende 1993; Kageyama and Silva 1983). From 1965 through 1985, the federal government also improved the effectiveness of its agricultural minimum prices programs, such as the federal government’s purchases program (AGF) and loans program for farmers themselves to store their production

The Agricultural Sector   279 (EGF). Both AGF and EGF were more effective as a support for mid-​and large-​scale market-​oriented farming operations (Goldin and Rezende 1993). In another effort to mitigate agricultural risk, the federal government inaugurated in 1974 a new agricultural insurance scheme called the Guarantee Program for Agricultural Activity (PROAGRO). This was initially linked with rural credit, and, once again, benefited mid-​and large-​scale farms during its operation until the 1990s, rather than small-​scale farming operations (Bacha 2012; Goldin and Rezende 1993). Completing the range of government agricultural policies implemented in the 1970s, the government created the Brazilian Enterprise for Agricultural Research (EMBRAPA) in 1973; and one year later, the Brazilian Enterprise for Rural Technical Assistance and Extension (EMBRATER) was created to oversee rural extension programs. During the military government’s dictatorship (1964 to early 1985), both EMBRAPA and EMBRATER focused on market-​oriented farming. Large multinational companies also had an important role in helping to expand Brazil’s agricultural frontier. As new farmers arrived in the Center-​West from the South, the large multinationals provided them with agricultural inputs (such as seed, fertilizer, pesticides, and herbicides) and received later products (after harvesting) in return. This system evolved into one that was managed using a “Green Soybean” contract (Contrato Soja Verde) with which the farming operation received an advance on future production, as well as guarantees that there would be a market for their crops. In 1994, the Green Soybean system was replaced with the Rural Product Note (CPR; Cédula de Produto Rural), which was governed and defined by Brazilian law (Law 8.929/​94). Two important crop shifts took place in Brazil during the 1970s and 1980s. First, the focus of coffee production moved from the state of Paraná to the states of Minas Gerais and Espírito Santo, as shown in Table 13.5; and second, soybean and corn production expanded considerably in the Center-​West. Table 13.5 Distribution of Coffee-​Producing Areas among Brazilian States for Selected Years (Values in Percentages) State

1969

1980

1990

1995

2000

2005

2010

2015

Paraná

44.7

26.1

14.6

5.2

6.2

4.6

3.8

1.8

São Paulo

29.7

33.1

19.3

12.2

9.2

9.5

9.4

10.3

Minas Gerais

11.3

19.0

33.5

42.8

43.6

44.7

47.5

50.5

Espírito Santo

8.5

12.5

17.3

23.6

22.8

23.4

21.9

22.3

Bahia

2.3

3.6

4.7

5.4

5.1

6.5

7.0

8.3

Other states Brazil

3.5

5.7

10.6

10.8

13.1

11.3

10.4

6.8

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

Source: Elaborated by the author from data provided by Brazil’s Statistical Yearbooks (various years) (IBGE). Note: A productive area is one able to generate production immediately or in normal climate conditions.

280   Carlos José Caetano Bacha In 1969, Minas Gerais ranked third in Brazil for area under coffee cultivation (holding 11.3% of Brazil’s coffee production area) while Espírito Santo ranked fourth. In 1990, Minas had jumped up to first position (33.5%) and Espírito Santo was on the verge of passing São Paulo into second, a position it has subsequently retained since the mid-​1990s. The importance of the state of Bahia as a Brazilian coffee-​producing state also increased over the period. In 1969, Bahia contained 2.3% of Brazil’s coffee-​ production area, a percentage that had more than doubled by 1990 (4.7%). The shares of Brazilian territory under coffee cultivation in both Paraná and São Paulo moved in the opposite direction, with their shares decreasing between 1969 and 1990: Paraná’s from 44.7% in 1969 to 14.6% in 1990, and São Paulo’s from 29.7% in 1969 to 19.3% in 1990. Former coffee areas in Paraná were allocated to soybean and corn production, while São Paulo’s previous coffee areas were allocated to sugar cane and orange plantation. The coffee culture’s movement through Brazilian territory is noteworthy. During the first seven decades of the nineteenth century, Rio de Janeiro was the largest coffee-​producing state in Brazil. From the end of the nineteenth century until the 1950s, São Paulo took the lead. São Paulo’s reign as coffee producer was usurped by the state of Paraná during the 1960s. Over the 1990s and 2000s, Minas Gerais and Espírito Santo became the country’s first-​and second-​largest coffee-​producing states, respectively. Those shifts were led by farmers looking for the most suitable areas to plant coffee trees and were supported by the federal government’s agricultural policies. As previously noted, soybean production also moved, from the South to the Center-​ West. Table 13.6 shows the evolution of soybean production among the Brazilian macro-​regions. In 1970, the South region was responsible for 92.7% of Brazil’s soybean production and the Center-​West region only 1.3%. Fifteen years later, in 1985, the Center-​ West produced 31% of Brazil’s soybeans, becoming the country’s second-​largest soybean producer among macro-​regions. As the figures in Table 13.6 indicate, this increase in the Center-​West’s share continued through the 1990s and the 2000s.

Table 13.6 Distribution of Soybean Production among Brazilian Macro-​Regions for Selected Years (Values in Percentages) Macro-​region

1970

1980

1985

1990

2000

2015

North

0.0

0.0

0.5

0.0

0.0

4.4

Northeast

0.0

0.0

0.0

1.1

6.3

8.6

Center-​West

1.3

12.6

31.0

32.6

47.6

45.1

Southeast South

6.1

9.2

10.1

8.5

8.0

6.1

92.7

78.2

58.5

57.8

38.2

35.8

Source: Elaborated by the author from Brazil’s Statistical Yearbook (several issues).

The Agricultural Sector   281

13.6.  From 1987 until 2015: Export-​L ed Growth with Less Government Assistance

100,000 95,000 90,000 85,000 80,000 75,000 70,000 65,000 60,000 55,000 50,000 45,000 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 0

1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013

US$ millions

Brazilian government funding to support agriculture declined between 1987 and 2015, but both crop and meat production and exportation continued to steadily increase, as illustrated in Figures 13.1 and 13.2. From 1987 to 2015, crop production increased by an average of 4.3% annually, and meat production grew by an average of 6.9% annually. The export value of Brazilian agricultural and agro-​processed products increased even more significantly in the 1990s and 2000s, as shown in Figure 13.3. From 1990 to 2013, the value of Brazilian agricultural and agro-​processed exports grew from US$10.2 billion to US$91.2 billion, rising by an average of 7.9% annually during the 1990s and by 15.5% annually from 2000 to 2013. Food importation also increased over the period, although not as a reaction to the food scarcity in Brazil mentioned in the previous section. In preceding decades, large quantities of basic products such as rice, corn, cotton, and powdered milk had to be brought in to supply the country; it is now much more common to find these Brazilian goods being exported. Since the 1990s, an increasing share of Brazil’s agricultural and agro-​processed production has been exported. In 1990, 4.3% of Brazil’s agro-​based production was

Year Exports

Imports

Figure  13.3. Evolution of Brazil’s agricultural and agroprocessing exports and imports, 1961–​2013.

282   Carlos José Caetano Bacha exported; this share reached 9.8% in 2013, according to FAO data. Also in 2013, some products made in Brazil had more than half of their production allocated to foreign markets, such as cotton, soybean, sugar, coffee, orange juice, and pulp. Corn was imported in some years during the 1980s and had one-​third of its production exported in 2013. These figures demonstrate the importance of foreign markets for Brazil’s agricultural and agro-​processed production. During the second half of the 1980s and throughout the 1990s and 2000s, budget restraint led Brazil’s government to materially reduce funding for agricultural support programs. According to Gasques and Bastos (2014), federal government expenditures on agriculture shrank from R$50.6 billion in 1987 to R$14.3 billion in 2003 (values in 2013 reais, R$). Government expenditure on agriculture has since increased, reaching R$27.2 billion in 2013; however, as a share of federal spending, expenditures on agriculture still show a severe reduction between 1987 and 2013, from 11.9% of total government expenditures in 1987 to 1.4% in 2013. During the first 15 years of the twenty-​first century, the Brazilian government’s agricultural assistance programs have been handled by two separate ministries: one directed to family farmers, the other to non-​family farmers. At the end of 1999, the responsibilities of the Brazilian Ministry of Agriculture, Livestock and Food Supply (MAPA) were redefined to support the county’s non-​family-​farming sector, and the Ministry of Agrarian Development (MDA) was created to support the family-​farming sector. Both MAPA and MDA have supported Brazilian agriculture using the same policies (rural credit, minimum prices, rural extension, and subsidized insurance), but with programs tailored to their respective sectors (non-​family and family farms). For example, in 2003 MAPA created the Subsidy for Rural Insurance, a new insurance program needed because the existing agricultural insurance program, PROAGRO, was increasingly dedicated to family farming. In the same year, MDA created the Food Acquisition Program (PAA), a new version of the federal government’s purchase program to guarantee minimum prices (AGF). Farmers from either sector can receive AGF support, while PAA serves only family farmers. In 2004, MAPA broadened the scope of privately supplied, government rural credit by creating CDA, WA, CDCA, LCA, and CRA.1 The first two (CDA and WA) allow farms to trade agricultural products without issuing an invoice or offering them as guarantee for rural loans. CDCA, LCA, and CRA are new bonds based on previous rural credit loans. By selling CDCA, LCA, and CRA, the private sector can make money quickly and offer a larger amount of rural loans to the farmers. Additionally, new minimum price programs were created and run by the private sector linked in cooperation with the federal government, such as the Private Option Risk Premium program (PROP) set up in 2004. Any farmer (especially non-​family farmers) can buy a price insurance from a private company and, if granted by PROP, will receive from MAPA a specific amount for any loss during this operation. For example, say a private company has a PROP premium of R$2 per bag for a specific agricultural product. It sells a price insurance policy to a farmer. If the loss in this operation is up to R$2 per bag, the private company will receive this amount from MAPA. If no loss

The Agricultural Sector   283 appears in this operation, the private insurance company will hold the premium paid by the farmer and will receive no payment from the federal government. In 2006, the government instituted the Equalizing Premium Paid to Growers program (PEPRO), which guaranteed that the producer would receive a government-​set minimum price for its product. The farmer sells the product on the market, and the federal government pays to that farmer the difference between the minimum price and the market price if the former is higher than the latter. All of these relatively new credit and minimum-​price programs have allowed large domestic and multinational agribusiness companies to commit the resources needed to expand Brazil’s agricultural frontier, especially into underutilized territory located in the country’s North and Northeast regions. Despite the implicit grants in some agricultural policies, Brazil is one of the countries that least subsidizes its agriculture. The OECD’s indicator of policy transfers to agricultural producers, measured at the farm gate and expressed as a share of gross farm receipts, known as PSE/​FGR, shows an average of –​9.19% from 1995 through 1999, rising to 4.98% from 2000 through 2004, and reaching 6.07% during 2005–​2009. It decreased to 4.31% during 2010–​2014 and reached to 2.57% in 2015. The same indicator was, in 2015, 9.44% for the United States, 18.92% for European Union countries, 43.07% for Japan, 21.34% for China, 13.79% for Colombia, and 3.34% for Chile. Subsidies are greater in MDA-​ supported programs than in MAPA-​ supported programs. During the 1990s and 2000s, MDA programs took an increasing share of the federal government’s agricultural expenditures. Family farming programs accounted for 4.33% of the federal government’s agricultural expenditures in 1990, enlarging this share to 21.51% in 2000 and reaching 46.9% in 2008, but then declining until reaching 23.26% in 2014. In May 2016, MDA was transformed into the Special Secretariat for Family Farming and Agrarian Development (SEAD). This Ministry is now in charge of the government’s programs in support of family farming. Despite the division of responsibilities between MDA/​SEAD and MAPA, both family and non-​family farming do not show productivity differences in terms of the countrywide average, although the two types of farming have different areas of specialization (Bacha and Stege 2015). According to Brazil’s 2006 Agricultural Census (the most recent), family farming accounted for 33.2% of Brazilian agriculture’s gross production value, and non-​family farming was responsible for the other 66.8%. The family-​farming sector frequently achieves market dominance in a specific product’s market. In 2006, family farmers were responsible for 87% of the country’s manioc production, 46% of its corn production, 38% of its coffee production, 58% of milk production, and 59%, 50%, and 30% of its swine, chicken, and cattle herds, respectively. The majority of family farmers’ production is market-​oriented and is sold to larger domestic and multinational agribusiness companies. While Brazilian agriculture has continued to expand in traditional areas, there has been a significant push into new territory since 1987. One notable area of recent

284   Carlos José Caetano Bacha agricultural expansion is located in the northern and northeastern Brazilian states of Maranhão, Tocantins, Piauí, and Bahia. Although not officially demarcated as a region, agricultural researchers commonly refer to it using the portmanteau MATOPIBA. Covered with vegetation common to the Cerrado, MATOPIBA encompasses 73.2  million hectares, and, if regulations related to conservation and indigenous populations are complied with, 46 million hectares of this land can be farmed (Miranda et al. 2014). MATOPIBA’s potential arable area is 70% larger than the entire state of São Paulo, Brazil’s largest agricultural state in terms of production and production value. A underlying theme throughout this chapter has been that Brazil’s agricultural frontier is constantly expanding. From colonial times to the present, farming has spread into fallow territory. With this spread has come a shift in the focus on cultivation of many important crops, especially soybean, planted forest, and coffee crops from 1987 through 2015. Soybean production has maintained its push from the South to the Center-​West region and, since the late 1990s, into MATOPIBA. According to the IBGE Statistical Yearbook, from 1987 through 2015, the South region’s share of Brazilian soybean production shrank from 54.6% to 35.8%, the Center-​West’s share of soybean production grew from 34.3% to 45.1%, and MATOPIBA’s share soared from 0.9% to 11.1%. Tree plantations have also appeared in new Brazilian locations, particularly in the country’s Northeast and Center-​West, where eucalyptus and, to a lesser extent, pine are the preferred newcomers. According to ABRAF (2013), in 1987 the Northeast and Center-​West were responsible for 5.4% and 8.4% of Brazil’s planted forests, respectively, which increased to 12.3% and 10.7% by 2012. There are two reasons for this growth: first, farmers have been acquiring cheap land to grow food crops and trees, and second, much larger agribusiness companies (such as Suzano, Fibria, and Eldorado Florestal) are looking to increase pulp production, a product most often sold to foreign markets. Throughout the 1990s and 2000s, Brazil’s coffee plantation continued its expansion in the states of Minas Gerais, Espírito Santo, and Bahia, as shown in Table 13.5. In 1990, the states of Minas Gerais, Espírito Santo, and Bahia, respectively, contained 33.5%, 17.3%, and 4.7% of Brazil’s coffee tree acreage. By 2015, those shares were 50.5%, 22.3%, and 8.3%, respectively. Infrastructure investments were in part responsible for Brazil’s agricultural expansion. Roadway construction increased the system’s extension from 1,487,000 kilometers to 1,736,000 kilometers between 1987 and 2008, displaying a geometric growth rate of 0.6% per year. Although this growth rate is much lower than the annual growth rate between 1965 and 1986 (2.9%), it is from a much larger base and often over more rugged terrain. Unfortunately, road surfaces continue to be very basic. In 1987, only 14.2% of Brazil’s roadways were paved; by 2008, that percentage had grown a mere 0.2 percentage points, to 14.4% Brazilian rural electrification initiatives led to a tripling of coverage between 1980 and 1990, from 20.6% of the rural population served to 63.2%, respectively, according to World Bank Indicators and the IBGE Statistical Yearbook. By 2012, 97% of the rural population was connected to the country’s electric grid. This increase in the percentage of rural homes with access to electricity was due to federal-​and state-​funded programs

The Agricultural Sector   285 supporting rural electrification and a decrease in the rural population, which fell from 42.2 million inhabitants in 1980 to 30.6 million in 2012; the fewer the people, the easier it is to provide them with electricity. Storage capacity for the products of Brazilian agriculture increased from 66.9 million tons in 1987 to 152.4 million tons in 2015 (Conab 2016), but capacity is still grossly insufficient. In 2014, the soybean and corn harvests alone were 167 million tons, 9.6% greater than total Brazilian crop storage capacity. Insufficient capacity explains the long line of motionless trucks waiting to unload their cargo at major harbors’ export facilities during the corn and soybean harvests: the wait to offload soybeans can often be counted in weeks. Agribusiness enterprises are among the very largest companies in Brazil, and they also play an important role in Brazil’s agricultural expansion. According to EXAME magazine’s ranking in 1995, 30 of the 100 largest enterprises in Brazil were agribusiness-​ oriented. Due to mergers among the larger agribusinesses during the 1990s and 2000s, that number had fallen to 20 by 2013. Of the 50 largest exporting companies operating in Brazil in 2005, 10 were agribusinesses. Brazilian-​owned large agribusinesses have also emerged over the last 30 years. In 2005, eight of the 10 largest agribusinesses exporting products from Brazil were owned by Brazilian entities; in 2013 eight out of the 20 largest agribusinesses operating in Brazil were owned by Brazilian entities. Larger domestic and foreign agribusiness enterprises and their subsidiaries are involved in financing and developing agricultural production, as well as delivering and trading this production into the domestic and foreign markets. They are also responsible for running many of the Brazilian government’s rural-​support programs concerning rural credit (such as CPR, CDCA and CRA) and minimum-​price programs (like PEP and PROP).

13.7.  Final Considerations Agriculture has expanded in Brazil due to the seven factors listed in the introduction and analyzed throughout the chapter. It is most probable that these factors will continue to drive Brazil’s agricultural sector for at least the next two decades. The following considerations support this hypothesis. First, Brazil still possesses a very large expanse of fallow arable land. In 2010, Brazil contained 85.3  million hectares of fallow land that could be successfully cultivated without encroaching on environmentally protected or indigenous areas. At current productivity levels, this idle area could be used to double crop production. Second, the federal government has made an effort to increase the number of public-​ private roadway concessions during the second decade of the twenty-​first century. Transferring road operation to the regulated private sector will lead to improved road surfaces and maintenance, thereby facilitating the transportation of agricultural production to exporting ports. Many of these concessions are directed toward the consolidation of agriculture in the Center-​West and MATOPIBA regions, as well as the other consolidated agricultural regions.

286   Carlos José Caetano Bacha Third, since the 1990s, the European Union has reduced agricultural sector subsidies and has increased forest conservation efforts, bringing about a decrease in farmed area. Productivity gains have allowed for a marginal increase in total EU agricultural production, but the trend toward fewer subsidies and increased conservation should improve Brazilian agriculture’s competitive position in many foreign markets currently served by EU farmers. With an increasing share of Brazil’s agricultural production sold in world markets, the country’s agricultural sector is now more vulnerable than ever to uncontrollable outside forces. World economic growth, especially that of China and the European countries, is a necessity if the Brazilian agricultural sector is to continue expanding and improving efficiencies. EMBRAPA-​led research programs have focused on agricultural and agro-​processing technologies, but most Brazilian agricultural inputs continue to be produced by foreign companies or their Brazilian subsidiaries. These overseas entities are a very strong force in the domestic inputs market and represent another uncontrollable factor that affects local farmers’ earnings and Brazil’s balance of trade. It would be valuable for Brazil’s state-​funded research institutions to devote more resources to developing technologies that put control of agricultural inputs into domestic hands.

Note 1. CDA = Certificate of Agricultural Deposit; WA = Agricultural Warrant; CDCA = Certificate of Agricultural Credit Rights; LCA = Notes of Agribusiness Credit; CRA = Certificate of Agribusiness’s Receivable Assets.

References ABRAF. 2013. Anuário estatístico ABRAF 2013 ano base 2012. Brasília: Associação Brasileira de Produtores de Florestas Plantadas. Available at http://​www.ipef.br/​estatisticas/​relatorios/​ anuario-​ABRAF13-​BR.pdf, accessed September 30, 2016. Albuquerque, M. C. C., and R. Nicol. 1987. Economia agrícola: O setor primário e a evolução da economia brasileira. São Paulo: McGraw-​Hill. Bacha, C. J. C. 1988. “Evolução recente da cafeicultura mineira: Determinantes e impactos.” PhD dissertation, College of Economics and Business Administration, University of São Paulo. Bacha, C. J.  C. 2011. “The Evolution of Brazilian Agriculture from 1987 to 2009.” In The Economics of Argentina and Brazil: A Comparative Perspective, edited by W. Baer and D. Fleischer, 97–​129. Northampton, MA: Edward Elgar. Bacha, C. J. C. 2012. Economia e política agrícola no Brasil, 2nd edition. Atlas: São Paulo. Bacha, C. J. C., and A. L. Stege. 2015. “Spatial Differences between Family and Non-​Family Farming in Brazilian Agriculture.” Paper presented at Congress of European Regional Science Organization, Lisbon, August 25–​28th. Available at http://​www-​sre.wu.ac.at/​ersa/​ersaconfs/​ ersa15/​e150825aFinal01673.pdf. Baer, W. 2008. The Brazilian Economy: Growth and Development, 6th edition. Boulder, CO: Lynne Rienner.

The Agricultural Sector   287 Becker, B. K. 2001. “Revisão das políticas de ocupação da Amazônia:  é possível identificar modelos para projetar cenários?” Parcerias Estratégicas 6 (12): 135–​159. CEPEA. Centre for Advanced Studies on Applied Economics, University of São Paulo, Brazil. Website, http://​www.cepea.esalq.usp.br/​english/​gdp/​, accessed July 31, 2016. Coelho, C. N. 2001. “70 Anos de Política Agrícola no Brasil (1931–​2001).” Revista de Política Agrícola 10 (3): 3–​58. CONAB. “Capacidade estática dos Armazéns, Companhia Nacional de Abastecimento.” http://​ conab.gov.br/​detalhe.php?a=1077&t=2, accessed September 2, 2016. Dean, W. 1997. With Broadax and Firebrand: The Destruction of the Brazilian Atlantic Forest. Berkeley: University of California Press. Delfim Netto, A. 1981. O problema do café no Brasil. São Paulo: IPE/​USP. FAO. Food and Agriculture Organization. Website, http://​www.sidra.ibge.gov.br/​bda/​cnt/​default.asp?z=t&o=15&i=P. Fausto, B. 1994. História do Brasil. São Paulo: EDUSP. Furtado, C. 1989. Formação econômica do Brasil, 23rd edition. São Paulo: Companhia Editora Nacional. Gasques, J. G., and J. C. P. R. Conceição. 2001. “Financiamento da Agricultura: Experiências e propostas.” In Transformações da agricultura e políticas públicas, edited by J. G. Gasques and J. C. P. R. Conceição, 95–​155. Brasília: IPEA. Goldin, I., and G. C. Rezende. 1993. “A agricultura brasileira na década de 80: Crescimento numa economia em crise.” Série IPEA 138, 119. Rio de Janeiro. Guimarães, A. P. 1968. Quatro séculos de latifúndio. Rio de Janeiro: Paz e Terra. IBGE. Multiple years. Brazil’s Statistical Yearbook. Brasília: Instituto Brasileiro de Geografia e Estatística. Lima, J. L., I. D. N. Costa, and F. V. Luna. 1983. Estatísticas básicas do setor agrícola no Brasil. São Paulo: Instituto de Pesquisas Econômicas. Kageyama, A. A., and J. G. Silva. 1983. “Os resultados da modernização agrícola dos anos 70.” Estudos Econômicos 13 (3): 537–​559. Matos, O. N. 1974. Café e ferrovias: A evolução ferroviária de São Paulo e o desenvolvimento da cultura cafeeira. São Paulo: Alfa-​Ômega. Melo, F. H. 1981. “Abertura ao comércio exterior e estabilidade de preços agrícolas.” Revista Brasileira de Economia 35 (2): 189–​205. Melo, J. A.  G. 1996. Gente da Nação:  Cristãos-​Novos e Judeus em Pernambuco, 1542–​1654. Recife: Massangana. Miranda, E. E., L. A. Magalhães, and C. A. Carvalho. 2014. “Proposta de Delimitação Territorial do MATOPIBA.” Campinas:  Embrapa—​Grupo de Inteligência Territorial Estratégica—​ GITE. [Technical note.] Available at https://​www.embrapa.br/​gite/​publicacoes/​NT1_​ DelimitacaoMatopiba.pdf, accessed September 29, 2016. Monbeig, P. 1984. Pioneiros e fazendeiros de São Paulo. São Paulo: Hucitec/​Polis. Mueller, C. C. 2010. “A política agrícola no Brasil: Uma visão de longo prazo.” Revista de Política Agrícola 19: 9–​23. Prado, C., Jr. 1982. História econômica do Brasil, 27th edition. São Paulo: Editora Brasiliense. Saes, F. A. M. 1981. As ferrovias em São Paulo 1870–​1940. São Paulo: Hucitec. Simonsen, R. 1938. Aspectos da história econômica do café. São Paulo:  Instituto Histórico e Geografico Brasileiro. Szmrecsányi, T. 1990. Pequena história da agricultura no Brasil. São Paulo: Editora Contexto. World Bank. 2016. “World Bank Indicators.” Available at http://​data.worldbank.org/​indicator, accessed August 25, 2016.

Chapter 14

T r aditional Ag ri c u lt u re an d L and Di st ri bu t i on in Bra z i l Charles C. Mueller

It was once widely believed in Brazil that concentrated access to land was an obstacle to the modernization of the country’s agriculture. Accordingly, taking into consideration the highly skewed pattern of land tenure prevailing in the 1950s and the substantial political and economic changes that took place after the return to democracy in 1985,1 an uninitiated observer might presume that the remarkable expansion and modernization of Brazilian agriculture over the last half a century was preceded by, or at least accompanied by, agrarian reform that, through deconcentrating access to land, created the basic conditions for modernization. A  more detained scrutiny, however, reveals that the strategies and policies adopted amidst a variety of quite different political and economic settings helped to generate substantial modernization without significantly altering the pattern of land tenure. Addressing this apparent paradox, this chapter examines the events that have taken place since the early 1950s. It begins by focusing on the progression of access to agricultural land in Brazil, and outlines the conception of the agricultural frontier employed here. Following that, the performance of Brazilian agriculture in the post–​World War II period is examined, through three substantially different periods: a phase of horizontal expansion, prevailing up to the early 1970s, in which all agriculture remained traditional and low productivity; a phase of substantial but conservative modernization, from the early 1970s to the late 1990s; and the subsequent period of consolidation of a modern and highly productive agricultural sector. We then focus on the development of both the modern agricultural sector, typically composed of large farm units, and of “traditional agriculture,” usually consisting of small farms, and discuss their contribution to the supply of food for domestic markets. Following this, we focus on the events that have sustained the concentrated pattern of land tenure, affecting both large-​scale agriculture

Traditional Agriculture and Land Distribution in Brazil    289 and small farms, and highlight the substantial changes that have affected these two groups, before offering final concluding remarks.

14.1.  The Evolution of Land Tenure in Brazil Table 14.1 presents data on land tenure in Brazil, from the agricultural censuses of 1950, 1970, 1980, and 2006.2 It reveals a similar pattern of distribution of land over time: at one extreme, there are large numbers of agricultural establishments quite small in size; and at the other, small numbers of very large establishments, encompassing substantial proportions of the overall area of agricultural establishments. Between 1950 and 2006, the number of agricultural establishments increased almost 2.4 times, from 2.06 million to 4.92 million units; the total area of agricultural land rose a little more than 1.4 times in the period, from 232.2 million to 333.7 million hectares. As for the proportions of the number and total area by area groups, these did not markedly change during this period of more than a half-​century. If we consider, for instance, all farms less than 100 hectares as small and all farms larger than 100 hectares as large, the proportion of the number of small farms was 83.4% in 1950 and 90.4% in 2006, but the overall area they occupied was only 16.6% of the total area of farms in 1950, increasing not markedly to 21.2% by 2006. As is seen in Table 14.1, the modest increases in the proportional significance of small farms essentially took place in the “less than 10 hectares” area group (very small farms) and between 1950 and 1970. After 1970, the proportional figures for this group display little variation. Looking at the large farms group, its proportion of the total number of farms was 14.6% in 1950 and 9.6% in 2006, but this group accounted for 83.4% of the total area in 1950 and 78.8% in 2006. The declines in this group’s proportional significance took place mostly due to changes within the very large segment (farms with more than 1,000 hectares), whose proportion of the total number of farms fell from 1.6% in 1950 to 1.0% in 2006, and their proportion of total area from 50.9% in 1950 to 45.0% in 2006. However, their overall area increased remarkably, from 118.1 million hectares in 1950 to 150.1 million hectares in 2006. Contrast this with the combined area, in 2006, of 7.8 million hectares for the 2.5 million farms with less than 10 hectares. In short, land tenure in Brazil has remained highly concentrated up to the present day. However, the stability of this concentration throughout a period of considerable political and economic upheaval was due to a large extent to the conformation of agricultural modernization, by demographic factors and by phenomena occurring at the agricultural frontier. To understand these changes, after an outline of the concept of agricultural frontier, we focus on the major features of the transformations that occurred from the end of World War II to the present, before examining the main impacts of these changes on land distribution.

Table 14.1 Brazil, Agricultural Land Distribution by Area Groups: Number of Establishments (1,000 Units) and Area (1,000 Hectares), 1950, 1970, 1980, and 2006 1950

1970

1980

2006

Number

Area

Number

Area

Number

Area

Number

Area

Area groups (hectares)

1,000 units

1,000 hectares

1,000 units

1,000 hectares

1,000 units

1,000 hectares

1,000 units

1,000 hectares

Less than 10 hectares

710.9

3,025.4

2,519.6

9,083.5

2,598.0

9,004.3

2,477.2

7,798.7

1,052.6

35,562.7

1,934.4

60,069.7

2,016.8

64,494.3

1,971.6

62,894.0

268.2

75,520.7

414.7

108,742.7

488.5

126,799.2

424.3

112,844.2

32.6

118,102.3

36.9

116,249.6

47.8

164,556.6

47.6

150,143.1

2,064.3

232,211.1

4,905.6

294,145.5

5,151.1

364,854.4

4,920.7

333,680.0

10 to less than 100 hectares 100 to less than 1,000 hectares 1,000 hectares or more TOTAL

Proportion of the number and of the total area of agricultural establishments by area group 1950 % number

1970 % of area

% number

1980 % of area

% number

2006 % of area

% number

% of area

Less than 10 hectares

34.4

1.3

51.4

3.1

50.4

2.5

50.3

2.3

10 to less than 100 hectares

51.0

15.3

39.4

20.4

39.2

17.7

40.1

18.9

100 to less than 1,000 hectares

13.0

32.5

8.4

37.0

9.5

34.8

8.6

33.8

1.6

50.9

0.8

39.5

0.9

45.1

1.0

45.0

1,000 hectares or more

Source: IBGE, Agricultural Census, 1950, 1970, 1980 and 2006.

Traditional Agriculture and Land Distribution in Brazil    291

14.2.  Brazilian Agriculture after the End of World War II Brazilian agriculture has undergone considerable changes over the last half-​century, which, along with the expansion of the agricultural frontier, have played important roles. In fact, in the 1950s and 1960s frontier expansion was the major determinant of agricultural growth, but starting in the early 1970s modernization accelerated substantially, and a significant portion of the expansion of the frontier went along with it.

14.2.1.  The Concept of the Agricultural Frontier In line with Sawyer (1984) and Mueller (1984), the agricultural frontier of a region at a given moment is regarded as that portion of its habitually unoccupied space that achieves the potential for agricultural exploitation. This usually occurs due to factors such as investments in transportation; the penetration of markets for agricultural products, services, labor, and inputs; accessibility of finance; speculative schemes; and, importantly, institutional changes affecting the potential space. It particularly makes sense to focus on the agricultural frontier in the context of countries or regions that have had considerable “unoccupied” portions of land, such as Brazil. Land to be incorporated by economic agents must be available; access to it may be either legal (through purchase or through grant by an agrarian policy), semi-​legal (through squatting), or illegal (obtained by violent means). Non-​agricultural activities—​such as logging—​tend to affect the availability of land for agricultural activities. The implementation of policies for frontier areas, and of technical change policies, may also alter the potential space of the agricultural frontier. These factors establish conditions for the expansion, into the potential space of the frontier, of activity fronts. An activity front is a concrete undertaking that unfolds in the space of the frontier. It could be the expansion of subsistence agriculture, the formation of cattle ranches or of large agricultural ventures, the spread of small-​scale official or private agricultural settlements, or land speculation, among other things (Mueller 1984). Agricultural activity fronts evolve in many ways. They may start precariously but progress into established endeavors; they can also fade away. They may expand sustainably or may tend to overlook future consequences. Moreover, different activity fronts may evolve simultaneously in the space of the frontier, increasing the incidence of dispute and violence. With the passage of time, the potential space occupied by an agricultural frontier will undergo changes as a result of the progression of the previously mentioned factors. This was so in the case of Brazil. In the 1950s, the location of the country’s agricultural frontier involved a relatively small portion of territory in its South and Southeast. Over time, however, the amount of substantial unused or underused in that

292   Charles C. Mueller area declined and the agricultural frontier moved toward the center of Brazil. In fact, large areas of the savannas in Central Brazil (the Cerrado), which, until the early 1980s, were considered unfit for modern agriculture, became portions of the frontier and are now significant segments of the country’s modern agricultural sector. Nevertheless, Brazil still has substantial areas undergoing incorporation into its agricultural frontier (Mueller 1984, 2012).

14.2.2. Three Phases in Post–​World War II Brazilian Agriculture Since the end of World War II, Brazilian agriculture has passed through three distinctive phases: a phase of horizontal expansion, up to the early 1970s; a phase of conservative modernization, from the early 1970s to the mid-​1990s; and a phase of the consolidation of modernization, from the mid-​1990s onward. In the phase of horizontal expansion, the growth of agricultural production was essentially due to the expansion of the agricultural frontier. Outside limited islands of modernization, Brazil’s agriculture was mostly traditional and low productivity. An analysis of the performance of Brazilian agriculture in this period must consider the urban bias of the import-​substitution industrialization strategy (ISI) adopted after World War II (Baer 2008). At that time, the sector was identified as backward, deserving attention only because of the key roles it played in ISI. In fact, the development strategy was able to achieve the transfer of income from agriculture to the urban-​ industrial sector, chiefly through the manipulation of relative prices against agriculture (Bacha 1975; Oliveira 1981). The foreign exchange—​then primarily generated by agricultural exports—​was maintained in a state of consistent overvaluation, and the prices of agricultural products for the domestic market were artificially compressed, in contrast to prices of internally produced industrial goods, which were the subject of protectionist measures. Nevertheless, the performance of agriculture was satisfactory in this period. Production increased enough to ensure that, by and large, sectoral terms of trade did not negatively affect the urban-​industrial sector, despite the rapid pace of import-​ substitution activities and of the growth of urban demand for food (Mueller 2011). In fact, in the booming 1950s, regardless of a consistently overvalued foreign exchange, agriculture contributed most of the country’s export earnings, essential for input and equipment imports. It is important to stress, however, that until the early 1970s Brazilian agriculture remained “traditional,” regardless of the size category of farms. Yields were very low and remained so throughout the period. Policies to modernize agriculture were almost nonexistent (an exception being efforts to advance the production of coffee, cotton, and sugarcane by organizations of the state government of São Paulo, the effects of which were limited mostly to that state; Pastore et al .1976). The main policy favoring the expansion of production in this period was road building (Nicholls 1970). New and better

Traditional Agriculture and Land Distribution in Brazil    293 roads enlarged the potential space of the agricultural frontier, enabling production to grow even with the same traditional methods. By the end of this first phase, however, the availability of unoccupied fertile land in the agricultural frontier had diminished significantly. There was unused land in the Cerrado (central Brazil’s savannas) and in the Amazon, but technologies for their productive exploitation had yet to be developed (Cunha et al. 1994). At that time, official programs affecting the frontier in these biomes had primarily geopolitical objectives (Mueller 2012). Oddly, most of the agrarian studies then conducted by social scientists tended to regard the country’s agriculture as an obstacle to progress, due to the high disparity of land distribution. According to the prevalent structuralist view they adopted, the large landowners—​the latifundistas—​were concerned with political power and land speculation, not with productivity and efficiency. As for the small farmers, they were considered too weak and oppressed to make a difference. Accordingly, the main structuralist policy prescription was that of agrarian reform, expropriating land from large landholdings and transferring it to small farmers and agricultural workers. The latter group was regarded as more responsive to the requirements of urban-​industrial development. Indeed, this outlook was at that time also commonly held by significant portions of urban public opinion.3 Because of the priority given to ISI, there were scant efforts to introduce policies for agricultural modernization in this first phase, and there were no effective measures to tackle the inequality of land distribution. In the second phase, of conservative modernization (from the early 1970s to the mid-​ 1990s), Brazilian agriculture experienced considerable expansion and modernization, but this took place without a prior land reform. By the end of the 1960s, the potential space of the frontier of Brazil’s South and Southeast regions had been largely exhausted. Fearing problems resulting from an inadequate agricultural performance, the 1964–​ 1985 military government adopted a series of measures and policies, outlined in the following, which resulted initially in a more intensive use of land in already settled areas, but subsequently reached the borders of the Cerrado savannas accessible to markets. There were also policies aimed at the occupation of land in more remote Cerrado areas and in parts of the Amazon, but, again, with predominantly geopolitical rather than agricultural focus (Mueller 2012). The development strategy then adopted was essentially a high-​powered ISI model (Baer 2008), in which the roles attributed to agriculture—​the provision of adequate domestic supplies of food and fibers and the generation of foreign exchange—​were considered vital. The foundations of the modernization of this second phase, whose impacts are still felt today, can be detailed as follows: Development of an effective research system in tropical agriculture: The first steps toward this were taken in the late 1960s, but efforts were intensified in the 1970s and afterward. The constitution of this system involved the assembly of an ample physical infrastructure (research facilities), the hiring and training of the personnel needed to advance the process, and the institution of a scheme to coordinate, manage, and

294   Charles C. Mueller expand the system. To this end, the federal government established a public enterprise, the Empresa Brasileira de Pesquisa Agropecuária (EMBRAPA) (Martha Jr., Contini, and Alves 2012; see also Chapter 15 in this volume). In view of the considerable geographical size of Brazil, and the diversity of the country’s habitats and social design, EMBRAPA was instituted as a decentralized research network, composed of units spread throughout the country, together with special thematic centers. Furthermore, EMBRAPA strove to enlist the collaboration of other organizations involved in agricultural research (state research units, universities, private organizations). Results began to be felt already by the 1970s, but initially new technologies involved improvements in production processes. With time, however, more complex developments arose, such as the creation of plant varieties adapted to the conditions of specific regions. Modernization, which accelerated through the 1980s and beyond, owes a lot to the implementation of this strategy of technical change. Availability of an active class of entrepreneurial farmers:  If the typical agricultural producers of this period were the folkloric absentee landlords, the impacts of the research effort would no doubt have been modest. However, in portions of the South and Southeast regions there was an important reserve of entrepreneurial farmers willing to innovate. They were vital for the modernization that accelerated through the second phase and beyond. In the settled areas of the South and Southeast existed extensively used land, together with portions of the agricultural frontier that were adequately served by transport infrastructure, with a high potential for modernization. These lands were mobilized by those entrepreneurial farmers who implemented new technologies, prompted by incentives policies adopted from the late 1960s (see later discussion). They have been fundamental for the modernized expansion of agriculture in areas of the more recent frontier. Expansion of a dynamic agribusiness sector:  Throughout the first phase, Brazil’s agricultural sector was essentially a producer and exporter of a few unprocessed commodities. This changed significantly through the second phase, thanks to the growth of a dynamic agribusiness sector (Montoya and Guilhoto 2000). An agribusiness segment comprises a set of economic activities operating in tandem with agricultural or livestock production. It involves three main aspects: enterprises and activities providing inputs and services to farms; agricultural activities proper; and businesses purchasing, transporting, processing, transforming, selling, and exporting the products generated by the agribusiness complex. By the mid-​1970s, the more advanced agricultural areas of the state of São Paulo and of the south of Brazil had incipient agribusiness complexes linked to a few agricultural or livestock segments (Mueller 1992). From this period onward, several new agribusiness complexes were formed and expanded, spurred by market conditions, by incentives provided by import substitution policies, and by the spread of new technologies in increasingly diversified agricultural segments of broader geographical areas, reaching new agricultural and livestock products. Many of the major agribusiness complexes feature the significant participation of foreign multinationals, but a few large Brazilian-​led agribusinesses have already emerged.

Traditional Agriculture and Land Distribution in Brazil    295 Inducement policies: Assuming that the adoption of modern agricultural technologies would necessitate powerful inducements, the agricultural strategy of the second phase made substantial financial resources increasingly available to agricultural producers willing to follow this path. The main policies to this effect were as follows: The establishment of the National System of Rural Credit (NSRC): In the late 1960s the NSRC began providing abundant financing, on very generous terms, to modernizing farmers. Among other things, it financed the purchase of modern inputs (equipment, fertilizers, pesticides and insecticides, selected seeds), much of which initially was imported but gradually became domestically produced. Agricultural credit offered by the system was highly subsided in the sense that its interest rates were maintained far lower than the growing rates of inflation, and the principal was usually not corrected for inflation. The amount of NSRC rural credit expanded considerably in the 1970s, reaching US$16 billion in 1974 and remaining above US$20 billion in all years of the 1975–​1982 period.4 It should be noted that until the mid-​1980s, the ultimate source of financial resources for the credit policy was the Treasury, which did not hesitate to tap the Central Bank to create money for this. The credit bonanza was maintained up to the mid-​1980s. Sharp cutbacks in subsidized agricultural credit have occurred since, and provisions for the correction of amounts due for inflation became the norm. In this period, Brazil was frequently at the brink of hyperinflation and international insolvency. Moreover, in 1986 the almost automatic link between the Treasury and the Central Bank was curtailed and the use of federal funds was restricted. The Constitution of 1988 reinforced this tendency. Improvement and expansion of the minimum price policy:  In the 1970s, an already existing minimum price apparatus was reformed and the role of the policy increased. However, until the mid-​1980s the credit policy prevailed in the modernization strategy. The aforementioned changes in this policy and the official determination to continue to extend financial inducements to agriculture led to an escalation of the use of the minimum price policy (Goldin and Rezende 1993; Rezende 2003, Chapter  1). However, throughout the second half of the 1980s and the early 1990s, problems increasingly arose from the administration of the policy. Sharply growing public expenditures with minimum prices, on the one hand, and the substantial accumulation of inventories of products together with ensuing logistical problems, on the other, led to the imposition of restrictions to the policy. By the late 1980s, the policy had become an instrument of regional development (Rezende 2003). Setting up nationally unified—​and usually remunerative—​minimum prices of commodities such as soybeans, the expansion of agriculture in frontier areas of the savannas of central Brazil (the Cerrado) was stimulated. The tropicalization of crops such as soybeans and cotton achieved by the EMBRAPA system (see Chapter 15 of this publication) contributed to successful cultivation of these crops in the Cerrado, previously considered unsuited (Cunha et al. 1984; Rezende 2003). The main difficulty for the expansion in cultivation in this area was the high transportation costs resulting from a deficient transportation infrastructure. To overcome this, official minimum prices

296   Charles C. Mueller assured the frontier producers nearly the same compensation as those of producers located near markets or export channels. Since in most of the new Cerrado areas minimum prices substantially exceeded market prices after the deduction of transportation costs, producers there were induced to sell their output to the minimum price organization. Over the second half of the 1980s, considerable portions of the output of new areas (mainly soybeans) became publicly owned and the government sustained the growing costs of transportation and storage of products, which ended up being disposed of with substantial losses. Dumping portions of surpluses on markets was also employed as a measure to help contain inflation. This brings us to the third and final phase, that of the consolidation of modernization after the mid-​1990s. As stated, a feature of the 1980s and early 1990s was a chaotic agricultural policy setting, which engendered turbulence for the sector. As extensively documented by Dias and Amaral (2000) and Rezende (2003), the frequent policy changes were caused chiefly by macroeconomic constraints and shifting priorities. The efficacy of the agricultural credit system in inducing output growth had weakened, and the system became regarded as wasteful and distorting (Sayad 1984), and as an obstacle to the implementation of monetary policy (da Mata 1982); hence the cuts in official agricultural credit and the rise in real interest rates on farm loans. Similarly, it became increasingly evident that the minimum price policy of the late 1980s could not be maintained. Consequently, various reforms were carried out in a trial-​and-​error fashion. It became evident that the public sector could not continue to finance, as it previously had, an increasingly complex and diversified modern agriculture. This was magnified by changes imposed by the Constitution of 1988.5 Thus, there was a gradual but considerable change in direction of the agricultural strategy. An important feature of the recent evolution of Brazil’s modern agriculture was the liberalizing trend of the 1990s (Mueller and Mueller 2016). In this decade, the country’s productive sectors—​including agriculture—​were increasingly exposed to international competition. Tariffs were reduced, export prohibitions and import quotas ceased to be employed, and the foreign trade bureaucracy was streamlined. There were several main changes. The direct governmental funding of commercial agriculture was contained, and interest rates were kept positive, although often below market rates. Official financing was channeled to small farmers and to land reform settlement projects; for most of commercial agriculture, other sources of finance, usually private, emerged. As for the minimum price policy, the government ceased to use it to transfer resources to producers, and reduced significantly the purchase of commodity surpluses. Modern, more agile instruments were created, avoiding the untenable practices of the past. The policy changes of the 1990s evolved with ups and downs, bringing some turmoil for the sector. Starting in 1994, for instance, there was a period of officially induced strong appreciation of the real—​a measure implemented by the administration of the Real Plan (Baer 2001, 199–​220)—​adversely affecting agricultural exports and stimulating agricultural imports, in a period of slack international commodity prices. In contrast, agriculture received an important boost starting in 1999, when the foreign exchange rate was allowed to float freely, producing a sharp depreciation of the real.

Traditional Agriculture and Land Distribution in Brazil    297 Together with increasing trends in world commodity prices, this led, in the first decade of the new millennium, to the consolidation of the expansion and diversification of agricultural production and exports. Examining the sector`s performance from the early 1990s onward, we observe that, instead of being contained by the “hands-​off ” policy changes sketched in the preceding, modern agriculture—​ prompted by a favorable institutional setting (Mueller and Mueller 2016)—​became driven largely by market conditions. Between 1991 and 1998 (i.e., the period of foreign exchange artificial appreciation), the annual rate of growth of real agricultural gross domestic product (GDP) averaged a modest 2.4%, in line with the rate of growth of GDP for the economy (2.8% annually). But from 1999 to 2004, after the foreign exchange was allowed to depreciate, the annual agricultural growth rate averaged an impressive 5.4%. Moreover, growth remained substantial in almost every year of the period. In contrast, aggregate real growth averaged only 1.8% in this period (Mueller and Mueller 2016). The performance of modern agriculture had important repercussions on Brazil’s international trade. By 2011, the country had become a world leading exporter of soybeans, sugar, meat from poultry and cattle, coffee, and orange juice, and a significant exporter of soy meal, oil from soybeans, poultry, pork, corn, and cotton. This represents quite a change from the performance of agriculture based on coffee and a handful of other commodities of the 1950s.

14.3.  The Outcomes of Modernization From the 1970s onward, there were noticeable increases in the production of the modern segment of agriculture.6 The harvest of grains and oilseeds7—​a proxy of Brazilian agricultural performance—​embarked on a trend of steady growth, which became much steeper after the late 1990s. In the 22 years between 1977 and 1999, production grew 121.1%, from 42.2 million to 93.4 million tons; but then grew 110.6% in only 13 years between 2000 and 2013, reaching 196.6 million tons. The 1977–​1999 expansion of production took place with a modest addition to the total land under cultivation, from 37.1 million to 38.6 million hectares. In the 2000–​2013 period, however, there was a marked area increase (45.4%). In the first period, production increased mostly by means of a more intensive use of land in settled areas of the Southeast and South regions, whereas in the 2000–​2013 period the incorporation of land in the frontier—​notably in the Cerrado—​played an important role (see later discussion). Most of the substantial increase in output was due to gains in yield made possible by technological change. It is worth noting that modernization reached not only the grains and oilseeds segment, but also crops such as sugarcane, coffee, and—​with notable outcomes—​the beef, poultry, pork, eggs, and milk segments. Also notable is that, in addition to technological change, the recent evolution of the modern sector occurred during years of very favorable external markets.

298   Charles C. Mueller It should be stressed that the performance of modern agriculture was attained in spite of a significant decline in official backing. According to an Organisation for Economic Co-​operation and Development (OECD) survey of policy support (OECD 2006), between 2002 and 2004 Brazil provided support to agricultural producers amounting to only 3% of its gross value of agricultural output, contrasting with the levels of support granted by the European Union (34% of the gross value of agricultural output), the United States (17%), the average of OECD countries (30%), and Japan (58%). Of the major agricultural countries, only New Zeeland had a lower level of support (2%). A similar state of affairs held more recently (OECD 2011). Further aspects of recent developments can be found in Chapter 13 of this publication. There we see that modern agriculture, functioning mainly within agribusiness complexes, transformed Brazil into an international powerhouse of agricultural production and exports. However, this occurred without major changes in the pattern of land tenure (see Table 14.1).

14.3.1.  Impacts of Modernization on Employment The combination of growth based on modern technologies and the recent incorporation of land in the frontier did not produce favorable impacts on unskilled rural employment. Modern technology heavily involved mechanization, which brings much lower labor requirements. In some places, and with some crops, the agricultural expansion made possible by modernization actually displaced large amounts of unskilled labor and, at the same time, generated shortages of skilled labor. One consequence of this has been the formation of large pockets of unemployment in urban centers and, at least in some areas (such as the Cerrado plateaus), a low human presence in vast rural areas. The replacement of crops such as coffee in the state of Paraná in the 1960s, caused by a succession of frosts together with low world prices, generated unemployment and was another factor contributing to rural migratory flows (see later discussion). But mechanization has certainly been a major element in this. Rural labor absorption and unemployment in Brazil are important matters to mention as part of the overall context covered here; they are also highly complicated issues deserving of a deeper analysis in their own right, but that is beyond the scope of this chapter.

14.3.2.  Small Agricultural Producers Table 14.1 reveals the persistence and indeed growth over time of very large numbers of small farms (units smaller than 100 hectares). In 1950 there were some 1.8 million of these; in 2006 they totaled more than 4.4 million units.8 However, the increase was not simply a case of a smooth growth in number; there was also considerable elimination of small farms through consolidation into larger units, both by legal purchase and

Traditional Agriculture and Land Distribution in Brazil    299 by violent means (see note 17). This took place both in already settled regions and in areas of fairly recent frontier expansion, notably in the Amazon region, as discussed later. Countering this, however, has been a simultaneous widespread creation of small units. Two factors have affected these patterns more recently: the impact of agrarian policies as they evolved from 1950 onward, and the expansion of the agricultural frontier. The following takes these in turn. In the post–​World War II period until the mid-​1960s, almost nothing was done in terms of agrarian policy.9 Land reform was an prominent issue politically in the 1950s and 1960s, but legislation made it cumbersome to execute initiatives in this area. At the outset of the 1964–​1985 military regime, important changes in the legislation governing land reform were introduced, along with institutional changes. Nevertheless, this did not translate into effective outcomes.10 As seen earlier, modernization occurred without major changes in land distribution. After the return of democracy in 1985, land reform again became an important policy objective but, in spite of the pressure of public opinion and of landless movements, little was accomplished until the early 1990s. Between 1986 and 1994, for instance, an annual average of only 14,637 families were settled by land reform programs (INCRA 2016), but this changed considerably afterward. Continued public opinion pressure, declining prices of agricultural land resulting from the implementation of stabilization (the Real Plan),11 and the actions of landless movements led to a substantial increase in families settled by official land reform programs. Between 1995 and 2006 (the year of the most recent census) the total number of families settled in land reform projects was 922,123, an average of 76,844 families annually, much higher than the 1986–​1994 average. As argued by Mueller and Mueller (2010), however, these numbers may give a false impression. In the 1995–​2006 period the land reform policy was characterized by the purchase of often distant land, with poor soils, and by the haphazard settlement of people unprepared for agricultural undertakings. Settlers were often poor and uneducated, and technical and managerial support for them was scarce. Many of the settlers pressured for land not because they aimed at becoming emancipated and productive farmers, but because they stood to receive land and credit from the government; official credit was made available with effectively no strings attached (Mueller and Mueller 2010). The results of these settlement programs in terms of increases in production have not been impressive, although there are exceptions (Alves and da Silva e Souza 2015). It is worth outlining the origin and nature of the landless movements pressuring for settlements. A significant factor in this was the change that took place in the 1960s and beyond in Brazil’s migration patterns. In the 1950s, there were substantial flows of migrants from densely populated areas, chiefly in the poverty-​stricken Northeast region, to urban-​industrial centers booming as a result of the ISI strategy (Baer 2008). However, in the 1960s industrialization began to lose pace, resulting in sharp declines in urban labor requirements. The demographic pressure continued, but now it engendered increasing migratory waves from areas of demographic surplus to rural areas with

300   Charles C. Mueller available land to occupy. Migrants focused initially on areas already settled in the Southeast and South regions, but later they concentrated increasingly on areas of subsistence fronts, mainly in the so called Legal Amazon region in the North of the country. This form of land occupation was initially spontaneous. In the 1980s, however, it led to the creation of organizations to pressure for land to settle, the most prominent being the Movement of Landless Workers (MST in the Portuguese acronym) founded in 1985. Typically, MST and others would search for (what they argued were) “unproductive” farms and carry out their invasion by encouraging followers to camp on the land and nearby roads, pressuring the government to expropriate the invaded area to establish a settlement. Such actions, which often involved a degree of violence, ended up gaining the attention of the media and inducing the government to expropriate the land to establish settlement projects. MST aimed initially at unproductive lands in previously settled areas in the South, the Southeast, and in parts of the Northeast regions, but, once such land became scarce in those areas, it also turned its attention to public land in the agricultural frontier. The settlement of the landless in the recent frontier began taking place in the 1970s, although not only as a result of organized initiatives of MST and similar organizations. In fact, as shown in the following, there were significant waves of spontaneous land occupation by small settlers in remote public areas, notably in the Amazon. In sum, a complex set of actions have taken place through recent decades, helping to maintain a high number of small farms, a substantial portion of which have remained traditional. The following section examines the role of the segment of small farms of all types in the production of food.

14.3.3. The Current Role of Traditional Agriculture in Supplying Domestic Markets General opinion in Brazil has harbored for some time the notion that traditional family farms are responsible for feeding the country; recently this was highlighted by a “70% supposition,” which emerged from a study based on a special tabulation of the 2006 agricultural census (IBGE 2009), a partnership of IBGE, responsible for the census, and the Ministry of Agrarian Development (MDA)12. As shown by Hoffmann (2014), the government claimed that this study had established that family farms produced 70% of the total food consumed in the country.13 The IBGE-​MDA study, however, does not make such a claim; it merely stresses the participation of family farms in the supply of several individual food items. After the study was released, official statements made the “70% supposition” gain momentum. However, as Hoffmann (2014) argues, not only does the “70% supposition” not have concrete support in the IBGE-​MDA study, it would actually be extremely difficult to construct any such kind of estimate, owing to the complexity involved in the aggregation of a highly heterogeneous set of food items. The same author also raises issues with the very definition of family farms used; the IBGE-​MDA study relies on a legal definition

Traditional Agriculture and Land Distribution in Brazil    301 (Law n. 11326, July 24, 2006), which Hoffman considers inconsistent, especially in the context of recent changes in Brazilian agriculture.14 Hoffmann’s conclusion is that “the claim that family farms produce 70% of the food consumed in Brazil is baseless, and worse, it is meaningless. The recognition of the importance of family agriculture does not require fictitious data” (2014, 420; author’s translation). Nevertheless, the view that small traditional farms are the main suppliers of staples for the domestic markets and that large commercial farms are interested only in external markets still carries some weight in Brazil. This view may have been true six decades ago, but more recently substantial changes have taken place. Consider three important food items: beans, poultry, and rice. In the past, beans were supplied coarsely, unprocessed, and often precariously, but today the typical urban dweller purchases this food item cleaned, packaged, and ready to cook. Poultry is commonly sold—​whole or in parts—​packaged, ready to cook; in fact, Brazil is a leading world exporter of poultry, and its production involves cutting-​edge technology. The production of rice today comes mainly from high-​productivity irrigated farms. Of course, the production of these food items is not uncommon in small, traditional farms, usually for subsistence consumption, but the main source for growing urban consumption is farms, of various sizes, using modern technology. To circumvent the difficulty of defining family farms and of physical aggregation of dissimilar products, a research group headed by EMBRAPA’s Eliseu Alves examined aspects such as the concentration of production in small and medium farms of various types in 2006 (the year of the most recent agricultural census) (see, e.g., Alves, Souza, and Rocha 2013; Alves and Contini 2014; and Alves and da Silva e Souza 2015). Working with census microdata, they determined the value of production15 of different sets of farms. The following groups were defined: very small farms (farms that in 2006 generated less than two minimum wages monthly);16 small farms (units that in that year generated 2 to 10 monthly minimum wages); medium-​sized farms (farms with production between 10 and 200 minimum wages monthly); and large farms (farms with production of more than 200 minimum wages a month). Taken together, the very small and the small groups—​farms with values of production of 10 minimum wages or less—​encompassed 3.9 million units, or 88.6% of the total number of farms undertaking some production in 2006. Their combined value of production was only 13.4% of the total of all farms of the sample. The minute value of production of the very small group stands out; in 2006 it amounted to only 3.3% of the total and generated an average 0.52 minimum wages a month. To survive, the families of this group usually complemented their income with employment outside the farm and/​ or with supplements from government programs. A large proportion of the very small farms (57.2%) were in the Northeast, Brazil’s poorest region. Regarding the medium-​sized and the large production classes, in 2006 their combined number of farms amounted to only 11.4% of the total of the studied group, but their share in the total value of agricultural production was 86.7%. The share of the farms in the large production class alone was only 0.52% of the total farm units, but in 2006 they generated 51.2% of the aggregate value of production.

302   Charles C. Mueller According to Alves and da Silva e Souza (2015), in 2006 there was considerable concentration of production within the small and very small groups, with a modest number of units faring quite well, and a large set with low production. To highlight the differences, they studied these groups in two different regions: the South (the states of Paraná, Santa Catarina, and Rio Grande do Sul), and the dryland Semiarido region of the Northeast. In the South, this group included 142,896 farms in 2006, with an average area of 30.9 hectares per farm, and an average production of 10.7 minimum wages per hectare. But 1,634 farms there, with an average area of 26.1 hectares per farm, generated an average of 321 minimum wages per hectare, an amount more than 10 times the overall average of the small and very small group. And in the Semiarido region, in 2006 there were 33,826 farms in these two groups, with an average area of 21.9 hectares, which produced an average of 15.0 minimum wages per hectare. However, a group of only 708 farms, with an average area of 19 hectares per farm, generated an average of 557.7 minimum wages per hectare. The concentration of production within the group of small to medium-​sized farms was also very significant in the two regions. One factor explaining much of the wide difference between farms with lower productivity and the more prosperous segments of each group was irrigation. Even though the Semiarido is characterized by extreme rural poverty, it is also the location of a highly successful irrigation venture, the Petrolina-​ Juazeiro project. In the South, meanwhile, irrigation is fundamental in the modern production of rice. It is interesting to observe that, except for the very small farms, land size was not the dominant factor distinguishing the traditional farms from the more prosperous ones. The average farm size is not too dissimilar between the two sets in each of the two regions. As emphasized by Alves and Contini (2014), irrigation played an important role, but so did differences in natural conditions, education levels, market imperfections (of products, credit, and inputs), infrastructure (transportation, effective cooperatives), and government support programs, among other factors.

14.3.4. The Agricultural Frontier in the Second Phase and Beyond In relation to large-​scale agricultural units, through the second broad phase of Brazilian agriculture, described earlier as a phase of “conservative modernization” from the early 1970s to the mid-​1990s, frontier expansion ceased to be the main motor of the expansion of agriculture. Nevertheless, it continued to play an important role. One important factor was the growing accessibility of land in the huge sparsely settled savannas (the Cerrado) of Central Brazil and of portions of the Northeast and the Amazon. This was favored by regional development policies of the 1970s and 1980s designed to promote the settlement of “empty” areas of the county’s hinterland (Mueller 2012).

Traditional Agriculture and Land Distribution in Brazil    303 The more recent frontier attracted significant flows of entrepreneurial farmers. By the end of the 1980s, the Cerrado had already become an important area of expansion of Brazil’s modern agriculture. Today, it is a major producer of soybeans, corn (maize), and cotton. From the mid-​1970s, regional development policies made possible the acquisition of land in the Cerrado at low prices, and provided farmers there with very generous incentives (subsidized credit, and, as discussed earlier, remunerative minimum prices). It is important also to emphasize the role of research from the EMBRAPA system in enabling high-​yield cultivation of the acid Cerrado lands, which were low in natural fertility. Moreover, these technologies were especially conducive to mechanization. These events led to the establishment, especially in the plateau areas of the Cerrado, of several large to very large farms, which are a significant part of today`s modern agriculture in Brazil. At the end of the 1990s and in the first decade of the new millennium, this took place more intensively in the huge Center-​West state of Mato Grosso. More recently the Cerrado frontier has evolved into the savanna region to the east-​northeast of central Brazil that has come to be known as “MATOPIBA” (combining portions of the states of Maranhão, Tocantins, Piauí, and Bahia). These events help to explain the substantial increase in large agricultural establishments depicted in the data of Table 14.1. Of course, new technologies were also important for the advance of modern agriculture in the Southeast and South of Brazil, settled much earlier, but there a concentrated pattern of land distribution already existed and has not changed significantly in recent decades. Large-​scale farm units also sprung up in parts of the recent frontier, not uncommonly associated with unorthodox, illegal, and sometimes quite violent methods of land acquisition (see note 17). In relation to small farm units, policies promoting the expansion of the frontier in the 1970s and 1980s also help to explain the very large number of small farms in 2006, depicted in Table 14.1. Before the late 1960s, the occupation of land by small settlers in subsistence fronts of the frontier was largely spontaneous; there were a few official colonization projects, but their impact was negligible. This was far from an orderly process. Typically it involved public lands in remote areas, away from the shelter of authorities at any level. Violence was widespread, land grabbing and forceful eviction were common, and so was the loss of life and property. In fact, this was not a new development; it had occurred, for instance, in the 1940s and 1950s, when the western and northwestern portions of today’s settled and developed state of Paraná in the South of Brazil was being accessed and occupied (Serra 2010). Similar incidents have been observed as a result of movements of subsistence and speculative fronts in parts of the Amazon and of the Center-​West regions.17 As a rule, such events involve public lands (of the state government in the case of Paraná, and, recently, of the federal government18). Faced with the unruly state of affairs in parts of Amazônia, the military government tried to impose order. In the early 1970s, it embarked on an ambitious strategy, the goal of which was the orderly occupation and settling of semi-​empty remote Amazon areas (Mueller 1980). This involved the construction of an extended road network in the

304   Charles C. Mueller region, the introduction of a scheme of generous fiscal incentives for large-​scale ventures (especially cattle ranches), and the establishment of a series of model public colonization schemes, conceived to help redirect to “empty” areas in the Amazon part of the demographic surpluses of the Northeast. They were supposed to become model colonization projects, endowed with the wherewithal to become organized, viable settlements. The approach aimed at promoting the orderly development of portions of the northern frontier, while at the same time alleviating demographic pressures in the Northeast, but it failed to produce the expected results. In the late 1960s and throughout most of the 1970s, growing portions of the rural migratory currents of the period—​not only, or even mainly from the Northeast—​took over land in considerable areas reached by the roads under construction in the Amazon, especially in the zones of the planned colonization schemes. The government tried to contain this, but without success. Reliable numbers on the process are not available, but it is possible to say that the populations of the state of Rondônia in the northwest of Brazil, of the south and southeast of the northern state of Pará, and of the west of the state of Maranhão all received significant boosts from the rural migration of the 1970s and 1980s. A substantial part of these migrants ended up forming small farm units. Although now waning, these frontier population movements help to explain the increases in the number of small farms registered in recent agricultural censuses.

14.4.  Concluding Remarks This chapter has focused on processes that have resulted—​particularly since the early 1970s—​in complex and variegated situations regarding the “modern” and “traditional” segments of Brazilian agriculture, and on their impacts on the country’s skewed pattern of land tenure. Access to land has remained highly concentrated throughout more than 60 years. Nothing indicates that it is changing. However, data on land distribution hide substantial structural changes that have altered the agricultural sector, particularly after the early 1970s. These changes have affected both extremes of land distribution: the large farm units, agents of modernization, which were favored by technical change and had ample access to new land in the agricultural frontier; and the small farm segment, which mostly remained at the margins of modernization. The large numbers of the latter have been maintained by factors such as demographic and migratory pattern changes, agrarian policies, and movements in the agricultural frontier. The groups of small, and particularly very small, farms that remain the locus of Brazil’s “traditional agriculture” are large in number, but their role in production has been limited, and poverty is prevalent. Changing this state of affairs would require improvements in education, infrastructure, public support, and rural development. More specifically, it calls for the complex task of creating and extending modern technologies adapted to this segment.

Traditional Agriculture and Land Distribution in Brazil    305

Notes 1. Such changes include hectic federal administrations; the approval of the 1988 Constitution; economic stagnation and the path to virtual hyperinflation up to the mid-​1990s; the economic reform of 1994 and control of inflation; the election of a social democratic administration in 1995; and the rise to power in 2003 of a left-​wing political party. 2. At the time of writing, Brazil’s most recent agricultural census was in 2006. For brevity, data from the agricultural censuses of 1940, 1960, 1980, 1985 and 1996 are omitted here; they did not differ markedly from the pattern portrayed in the table. An agricultural census was due to be held in 2016, but was postponed. 3. The structuralist argument was demolished by Antonio Barros de Castro, an analyst once ideologically aligned with it (see Castro 1969). Castro identified the main weaknesses of the structuralist account by showing how it clashed with the actual performance of agriculture in the 1950s and early 1960s. However, it took considerably longer for the truth to be accepted. 4. Values expressed in current US dollars (no correction for the US inflation). Values obtained by Goldin and Rezende (1993). 5. The 1988 Constitution drastically reduced the capacity of the federal government to transfer resources to agriculture in the way that it had done previously. 6. For further details, see Chapter 13 of this publication. 7. Data on grain and oilseed production from CONAB, 2016. Crops included canola, rye, barley, beans, sunflower seeds, mamona, corn (maize), soybeans, sorghum, wheat, and triticale. 8. As show by Hoffmann and Ney (2010), the census data overestimate somewhat the number of small farms since they include recreational (nonproducing) units. Anyhow the number of small units remains very large. 9. Details can be found in Mueller (2009) and in Mueller and Mueller (2010). 10. In fact, at the height of the military period it was considered subversive to advocate land reform. 11. One of the chief assets sought by private investors to help offset the very high and increasing inflation of the 1980s and 1990s (before the Real Plan), was land (agricultural or urban). With stabilization came an increase in the land for sale. Some of the unoccupied or under-​occupied agricultural establishments faced the threat of invasion promoted by landless organizations, so many landowners sought to sell their land, leading to reductions in its real prices. 12. In 2016 the Ministry of Agrarian Development was merged with the Ministry of Social Development and Fighting Hunger to form the new Ministry of Social and Agrarian Development. 13. For instance, Portal Brasil (July 27, 2011), http://​brasil.gob.br, cited in Hoffmann (2014). 14. A similar argument is found in the thorough studies of Navarro and Pedroso (2010, 2011). 15. The value of production calculated by the research group included the amounts produced for sale in 2006 and the amounts consumed or used in the farm (e.g., manioc used to produce manioc flour). Units that failed to declare production to the census were excluded. 16. In 2006 the official minimum wage was R$300 per month. 17. Recently an investigation by a team of reporters from the major newspaper O Estado de São Paulo in the frontier states of Pará, Acre, Amazonas, Rondônia, and Tocantins, and in areas of the Center-​West states, uncovered 485 hotspots in which, between 1996 and

306   Charles C. Mueller mid-​2016, at least 1,309 assassinations related to land disputes took place (O Estado de São Paulo 2016; the findings were printed daily between July 10 and 17, 2016). This investigative coverage shows vividly that there are two portions of Brazil: one in which legal institutions operate fairly effectively, and a remaining extended area that is essentially outside the rule of law and where violence prevails. 18. In the 1960s all available public lands were transferred to the federal government.

References Alves, Eliseu, Geraldo da Silva Souza, and Daniela de Paula Rocha. 2013. “Desigualdade nos campos na ótica do Censo Agropecuário 2006.” Revista de Política Agrícola 22 (2): 67–​75. Alves, Eliseu, and Elisio Contini. 2014. “Tecnologia: Prosperidade e pobreza nos campos.” In Censo entra e, campo: O IBGE e a história dos recenseamentos agropecuário, edited by Nelson de Castro Senra, 207–​229. Rio De Janeiro: CDDI IBGE. Alves, Eliseu, and Geraldo da Silva e Souza. 2015. “Pequenos estabelecimentos também enriquecem? Pedras e tropeços.” Revista de Política Agrícola (July–​September 2015): 7–​21. Bacha, Edmar L. 1975. “O café na economia brasileira.” In Os mitos de uma década, edited by Edmar L. Bacha. Rio de Janeiro: Paz e Terra. Baer, Werner. 2001. The Brazilian Economy: Growth and Development, 5th edition. Westport, CT: Praeger. Baer, Werner. 2008. The Brazilian Economy: Growth and Development, 6th edition. Boulder, CO: Rienner. Castro, Antônio Barros de. 1979. 7 Ensaios Sobre a Economia Brasileira, Vol. I. Rio de Janeiro: Forense. CONAB. 2016. Produção e área de cerais: Séries históricas. Brasília: Ministério da Agricultura, Pecuária e Abastecimento. Cunha, Aércio, Charles Mueller, Eliseu Alves, and Euripides da Silva. 1994. Uma avaliação da sustentabilidade da agricultura no Cerrado, Vol. 1. Brasília: IPEA. Dias, Guilherme Leite da Silva, and Cicely Moitinho Amaral. 2000. “Mudanças estruturais na agricultura brasileira, 1980–​98.” In Brasil: Uma década em transição, edited by Renato Baumann, 223–​254. Rio de Janeiro: CEPAL/​CAMPUS. Goldin, Ian, and Gervásio Castro de Rezende. 1993. A agricultura brasileira na década de 1980: Crescimento numa economia em crise. Brasília: IPEA. Hoffmann, Rodolfo. 2014. Agricultura familiar produz 70% dos alimentos consumidos no Brasil? Mato Grosso do Sul: Sistema FAMASUL, July. www.senarms.org.br, accessed November 21, 2016. Hoffmann, Rodolfo, and Marlon Gomes Ney. 2010. Estrutura agrária e propriedade agrícola no Brasil, grandes regiões e unidades da federação. Brasília: Ministério do Desenvolvimento Agrário. IBGE. 2009. “Censo Agropecuário 2006: Agricultura familiar, primeiros resultados.” Available at www.ibge.gov.br/​estatística/​economia/​agropecuária/​censoagro/​agri_​familiar_​2006_​2/​, accessed November 16, 2016. IBGE. 2015. “Historical Data on Agricultural Censuses.” Available at www.sidra.ibge.gov.br, accessed February 16, 2016. INCRA. 2016. “Website” www.painel.incra.gov.br, accessed on February 3, 2016. da Mata, Milton. 1982. “Credito rural: Caracterização do sistema e estimativas dos subsídios implícitos.” Revista Brasileira de Economia 36 (3): 215–​245.

Traditional Agriculture and Land Distribution in Brazil    307 Martha, Geraldo, Jr., Elisio Contini, and Eliseu Alves. 2012. “Embrapa:  Its Origins and Changes.” in The Regional Impacts of National Policies: The Case of Brazil, edited by Werner Baer, Chapter 12, 204–​226. Northampton, MA: Edward Elgar. Montoya, M.A., and J. M. Guilhoto. 2000. “O agronegócio brasileiro entre 1950 e 1995: Dimensão econômica, mudança estrutural e tendências.” In O agronegócio brasileiro no final do século XX, edited by M. A. Montoya and L. Parré, 3–​32. Passo Fundo (Rio Grande do Sul): Editora UFP. Mueller, Bernardo. 1994. “The Economic History, Political Economy, and Frontier Settlement of Land in Brazil.”PhD Thesis, Department of Economics, University of Illinois at Urbana-​Champaign, 196. Mueller, Bernardo, and Charles Mueller. 2010. “The Political Economy of the Brazilian Model of Agricultural Development:  Institutions versus Sectoral Policy.” Quarterly Review of Economics and Finance 62: 12–​20. http://​dx.doi.org/​10.1016/​j.qref.2016.07.012. Mueller, Charles C. 1980. “Recent Frontier Expansion in Brazil:  The Case of Rondônia.” In Land, People and Planning in Contemporary Amazonia, edited by Françoise Barbira-​ Scazzocchio, 141–​ 153. Centre of Latin American Studies Occasional Publication No. 3. Cambridge: Cambridge University Press. Mueller, Charles C. 1984. “El estado y la expansión de la frontera agrícola en la Amazonia.” In Expansión de la frontera agropecuaria y medio ambiente en América Latina, edited by C. C. Mueller, 37–​78. Comisión Económica para América Latina (CEPAL) y Programa de las Naciones Unidas para El Medio Ambiente (PNUMA). Madrid: Naciones Unidas y CIFCA. Mueller, Charles C. 2009. “Agricultural, Agrarian and Environmental Policy Formation under Lula: The Role of Policy Networks.” In Brazil under Lula: Economy, Politics and Society under the Worker-​President, edited by Joseph Love and Werner Baer, 135–​150. New York: Palgrave Macmillan. Mueller, Charles C. 2011. “Inflation and Income Transfers during the Golden Phase of Import Substitution Industrialization of the 1950s: The Contribution of an Expanding Agricultural Frontier.” In The Brazilian Economy in Historical Perspective, edited by Sonia Ranincheski, Camilo Negri and Charles Mueller, 41–​52. Brasília: Verbena Editora. Mueller, Charles C. 2012. “Regional Development and Agricultural Expansion in Brazil’s Legal Amazon:  The Case of the Mato Grosso Frontier.” In The Regional Impact of National Policies:  The Case of Brazil, edited by Werner Baer, 184–​293. Northampton, MA: Edward Elgar. Mueller, Charles C., and Bernardo Mueller. 2010. “The Evolution of Agriculture and Land Reform in Brazil, 1960–​2006: Essay in Honor of Werner Baer.” In Economic Development in Latin America, edited by H. S. Esfahani, G. Faccini, and G. J.  D. Hewings, 133–​162. Basingstoke, UK: Palgrave Macmillan. Navarro, Zander, and Maria Thereza Pedroso. 2010. “A agricultura familiar no Brasil: Entre a política e as transformações da vida econômica.” In A agricultura brasileira: Desempenho, desafios e perspectivas, edited by J. Garcia Gasques, José Eustáquio Ribeiro Vieira Filho, and Zander Navarro, 185–​209. Brasília: IPEA. Navarro, Zander, and Maria Thereza Pedroso. 2011. Agricultura familiar: é preciso mudar para avançar. Brasília: EMBRAPA Informação Tecnológica (Textos para Discussão, n. 42). Nicholls, William. 1970. “The Brazilian Agricultural Economy:  Recent Performance and Policy.” In Brazil in the Sixties, edited by Riordan Roett. Nashville, TN:  Vanderbilt University Press.

308   Charles C. Mueller OECD. 2006. OECD Review of Agricultural Policies:  Brazil 2005. OECD Publishing. http://​ dx.doi.org/​10.1787/​9789264012554-​4-​en, accessed June 28, 2016. OECD. 2011. OECD Agricultural Policy Monitoring and Evaluation, 2011. OECD Countries and Emerging Economies.   http://www.oecd.org/tad/agricultural-policies/agriculturalpolicy monitoringandevaluation2011oecdcountriesandemergingeconomies.htm#access, accessed June 28, 2016. Oliveira, João do Carmo. 1981. “An Analysis of Transfers from Agricultural Sector and Brazilian Development, 1950–​1974.” PhD dissertation, University of Cambridge. Pastore, José, Guilherme L. Silva Dias, and Manoel C. Castro. 1976. “Condicionantes da produtividade da pesquisa agrícola no Brasil.” Estudos Econômicos 6 (3): 147–​181. Rezende, Gervásio de Castro. 2003. Estado, macroeconomia e agricultura no Brasil. Porto Alegre: Editora UFRGS/​IPEA. Sayad, João. 1984. Crédito rural no Brasil: Avaliação das críticas e das propostas de reforma. São Paulo: FIPE and Livraria Pioneira Editora. Sawyer, Donald. 1984. “Ocupación y desocupación de la frontera agrícola en el Brasil:  Un ensayo de interpretación estructural y espacial.” In Expansión de lafrontera agropecuaria y medio ambiente en América Latina, edited by C. C. Mueller, 79–​104. Comisión Económica para América Latina (CEPAL) y Programa de las Naciones Unidas para El Medio Ambiente (PNUMA). Madrid: Naciones Unidas y CIFCA. Serra, Elpídio. 2010. “Conflitos rurais no Paraná:  Como foi que tudo começou.” Boletim Geográfico 28 (1): 75–​89.

Chapter 15

B r azil’s Agri c u lt u ra l Moderniz at i on and Emb ra pa Geraldo B. Martha Jr. and Eliseu Alves

15.1. Introduction For most of the twentieth century, Brazilian agriculture was considered an inefficient, backward sector. Food crises repeatedly occurred up to the 1970s, and until the following decade Brazil depended heavily on imports to ensure food security for its people. Macroeconomic pressures arising from food inflation were a persistent problem. It was against all odds, then, that Brazilian agriculture was able to reinvent itself and in a period of only one generation—​from the 1970s to the 2000s—​become the sole agricultural power in the tropics. As of today, among individual countries Brazil comes second only to the United States in the world’s agricultural trade. Brazil’s ability to leave behind traditional agriculture, based on land area expansion and on limited use of modern inputs, was of course decisive in the production expansion of agriculture in the country following the 1970s, and especially following the 1990s. Furthermore, in recent decades a considerable range of institutional innovations and policies have played a decisive role in modernizing the agricultural sector. In this chapter, however, we focus on a particularly key decision that enabled that achievement, namely the country’s commitment, constancy of purpose, and perseverance during the past 40–​50 years toward an agriculture based on science. During this period, Brazil eventually became self-​sufficient in food production. As food production increased over time at higher rates than food demand, real food prices for consumers significantly decreased. Given Brazil’s central role in the world’s agriculture, such a trajectory additionally contributed to global food security. The fact that, in the aggregate, agricultural production grew predominantly through yield increases rather than area expansion also decisively contributed to the generation of impressive

310    Geraldo B. Martha Jr. and Eliseu Alves land-​saving effects that allowed millions of hectares to be spared from cultivation over the past 60 years. As argued influentially in the work of Douglass North (e.g., 2006, viii), history matters not only because one can learn from the past, but because present and future choices are shaped by past decisions and the continuity of society’s institutions. Thus, on the one hand, the story of the development and consolidation of a science-​based agriculture in Brazil has generated intense interest from other developing countries and international organizations. In particular, they wish to learn from the Brazilian experience of developing an efficient, competitive, and sustainable agriculture in the tropical belt in such a short period of time. However, on the other hand, a past full of achievements is no guarantee for a successful future. In spite of the considerable progress observed in Brazilian agriculture, it is necessary to move even further along the path of sustainability, and to solve localized problems relating to agricultural production and to environmental and social claims. In considering future perspectives for Brazilian agriculture, the multifaceted role that agriculture will play in the coming decades is clear, as is its increasing reliance on knowledge, technologies, and innovation (Embrapa 2014). A central challenge in the near future for a science-​based agriculture is to clearly and objectively identify the sequence of relevant problems that need to be solved by research to increase society’s welfare (Martha Jr. et al. 2012a). Keeping these perspectives in mind, the chapter discusses three topics in depth. The following section (15.2) surveys the origins of the modernization of agriculture in Brazil. We explore in detail the role of industrialization and key policies as drivers for the modernization of the agricultural sector, especially from the 1970s on. In section 15.3 we focus on the science-​generating dimension of Brazil’s agriculture. We consider the relevance of Brazil’s National Agricultural Research System, and then concentrate more specifically on Embrapa (Brazilian Agricultural Research Corporation). We review and expand upon key points previously presented in other studies (Alves 2010; Lopes and Martha Jr. 2016; Martha Jr. et al. 2012a; Pereira et al. 2012). In section 15.4 we present selected achievements of Brazilian agriculture and discuss its efficiency in terms of land productivity and total factor productivity. In the final section (15.5) we conclude by reviewing concisely several future challenges facing Brazilian agriculture, from both demand-​and supply-​side perspectives. We bring to the discussion an insistence on the importance of expanding science generation in Brazil to face these challenges, and of aiming to alleviate market imperfections to ensure that modern technology can be effectively adopted in a more inclusive way in farms across the country.

15.2.  The Origins of Agricultural Modernization in Brazil The importance of the agricultural sector for the Brazilian economy dates back to the colonial period of the sixteenth to nineteenth centuries; until the first decades of the

Brazil’s Agricultural Modernization and Embrapa    311 twentieth century, the Brazilian economy depended heavily on agriculture (Baer 2008; Furtado 2005), which at that time mostly relied on the factors of land and labor to increase production. Coffee and some other agricultural commodities (rubber, cocoa, cotton) destined for foreign markets accounted for over 55% of exports until the 1960s (Thorp 1998). Such an externally oriented approach, focused on a few agricultural products, ultimately translated into short-​lived periods of boom and bust (Baer 2008). As a consequence, Brazil repeatedly faced volatile economic growth and considerable external vulnerability in this period (Gremaud et al. 2004). For much of Brazil’s history, the agricultural sector was identified as backward, because it was trapped into a low-​productivity approach, heavily dependent on area expansion to increase food production.1 However, the availability of more fertile sites for agriculture in the South and Southeast regions was already becoming low in the 1960s. Meanwhile, the expansion to the Central-​West region at that time found the acidic, low-​ fertility soils of the Brazilian savannah (Cerrado) to be a major barrier to increasing food production. Such a scenario is easily illustrated by the evolution of the per capita production of major grains and oilseeds (corn, rice, beans, wheat, and soybean) in Brazil since the 1950s.2 The figures changed little from 1950 to 1970 (1950: 214 kg per capita; 1960: 201 kg per capita; 1970: 247 kg per capita).3 More consistent increases in food availability were evident only in the estimates for 1980 (336 kg per capita) and 1990 (368 kg per capita). It is worth noting that Brazil received food donations from abroad until the 1960s, and up until the 1980s the country was still one of the world’s largest food importers.4

15.2.1. An Agricultural Modernization Process Driven by the Import-​Substitution Industrialization Strategy The shift toward the modernization of Brazilian agriculture had its origins in the import-​substitution industrialization strategy adopted from the 1950s–​1960s to the early 1980s (Alves et al. 2008; Pereira et al. 2012). Until the mid-​1980s, the industrial sector was granted a series of advantages that strongly discriminated against agriculture. The ambitious import-​substitution industrialization policy was based on exchange controls, on a multiple exchange-​rate system to support capital goods imports, and on subsidized interest rates for loans for the capital goods industry (Alves and Pastore 1978; Baer 2008; Gremaud et al. 2004). This economic policy further promoted consumer goods imports and investments in energy and transport infrastructure (Gremaud et al. 2004; Baer 2008). The investments in federal and state highway systems were essential to support agricultural production expansion, in terms of an increased crop area (from the 1950s to 1970s–​1980s) and, later, in terms of higher productivity levels (transport of modern inputs). And, of course, transport infrastructure played a pivotal role in moving the increased agricultural

312    Geraldo B. Martha Jr. and Eliseu Alves production to markets in the cities. Other government priorities for investments included urban infrastructure, housing and health, and salary protection. As far as possible, food prices were kept artificially low to avoid pressure on urban salaries in this phase (Alves and Pastore 1978; Baer 2008; Gremaud et al. 2004). Politically, the distorting industrialization policy shifted power from rural areas to cities, transforming Brazil into a progressively urban society (Dias and Amaral 2000), with an accelerated rural-​urban migration process starting in the 1950s. According to the Brazilian Institute of Geography and Statistics (IBGE), the urban population percentage in Brazil in 1960, 1970, 1980, 1991, 2000, and 2010 was 45.1%, 56.0%, 67.7%, 75.5%, 81.2%, and 84.4%, respectively. In the span of only 30 years the rural population shrank from 55% of the total population in 1960 to 23% in 1991 (IBGE 2011). Migration from rural to urban areas was very intense from the 1960s to the 1980s, but then lost impetus after the 1990s. This was in part because the rural-​urbanization cycle was almost complete in the South, Southeast, and Midwest regions, but also because low economic growth rates in Brazil during the 1980s and 1990s weakened the attractiveness of cities (Alves et al. 2008).

15.2.2. Key Policies in the Modernization of Brazilian Agriculture Until at least the 1950s, most agricultural policy in Brazil was aimed at expanding the agricultural frontier. However, agricultural policy during the import-​substitution industrialization policy period was subordinated to the overriding goal of industrialization (Baer 2008). The persistent food supply crisis throughout the 1960s and the early 1970s led to the formulation of several hypotheses aiming to explain the lack of a more substantial increase in productivity, and intending to guide more efficient policy approaches. Four main lines of reasoning can be identified: (1) the need for a profound agrarian reform (Cline 1970); (2) a recognition that the agricultural problem reflected a combination of domestic and export price policies, labor-​saving industrialization, and a lack of agricultural research (Schuh 1974); (3) the idea of insufficient demand (i.e., that the inelastic nature of own-​price and income elasticities for agricultural products implied that small increases in supply determined a more significant reduction in prices, which in turn prevented the diffusion of the modernization process) (Paiva 1975; Pastore et al. 1976; Pastore and Barros 1976; Ryff 1976); and (4) Hayami and Ruttan’s induced innovation theory of 1971 (Hayami and Ruttan 1985), suggesting that the prolonged surplus of land and labor in Brazilian agriculture hampered the widespread modernization of the sector (Pastore et al. 1976; Schuh 1974). Three policies turned out to play a central role in the agricultural modernization process: (1) rural credit, mainly for capital goods and the purchasing of modern inputs; (2) rural extension; and (3) support of agricultural research.

Brazil’s Agricultural Modernization and Embrapa    313

15.2.2.1. Rural Credit In 1965 the Brazilian government established the National Rural Credit Program, which provided loans to finance modern agricultural inputs, such as improved seeds and fertilizers, and machinery and equipment, such as tractors. Interest rates were subsidized, particularly from the late 1960s to 1985 (Coelho 2001). Rural credit averaged R$.153–​154 billion per year from 1969 to 1985; R$76.40 billion per year from 1986 to 2000; and R$113.60 billion per year from 2001 to 2015.5 The role of rural credit in the modernization of Brazilian agriculture in the 1970s and the 1980s, which helped to foster modern production inputs, machinery, and equipment adoption, was remarkable. For example, apparent fertilizer consumption (nitrogen, phosphorus, potassium) in Brazilian agriculture averaged only 146,091 metric tons per year from 1950 to 1964. This more than doubled in 1965–​1969, reaching an average of 343,668 metric tons per year. The truly great increase in fertilizer consumption, however, came in the 1969–​1985 period, in which an average of 1,669,559 metric tons of fertilizer per year was registered.6 The stock of tractors in Brazilian farms also experienced profound changes in the 1970s and 1980s. The number of tractors was 165,870 in 1970, but grew at an astonishing rate of 9.70% per year up to 1985, reaching 665,280 units. Such an accelerated mechanization process in the 1970–​1985 period—​from an estimated 241 hectares of cropland per tractor, to 92 hectares of cropland per tractor—​was essential to the development of agriculture in the Cerrado.7

15.2.2.2. Rural Extension Until the early 1970s, Brazilian policymakers emphasized rural extension and neglected efforts in agricultural research, their rationale being that a vast array of technologies was already available for adoption. Fostering public and private extension services, for example, by connecting them to the emerging rural credit policy, would theoretically solve the agricultural problem. By the early 1970s, empirical evidence proved that this hypothesis was false. Indeed, the example previously discussed concerning the stagnant per capita production of major grains and oilseeds from 1950 to 1970 reinforces this argument. The belief that strengthening human capital was key to better utilizing available resources and to increasing the impact of the investments made in capital goods and modern inputs was, of course, in the right direction. The flaw emerged in not recognizing that agricultural problems—​and the demands posed to the sector—​were dynamic rather than static. A successful strategy would, of course, embrace a robust research system to continuously generate knowledge and technology to be transferred to extension services that, in turn, would be better positioned to support a sustained innovation process at the farm level. This condition reflected the fact that the stock of knowledge and technology positively responds to the accumulated experience of farmers and to their ability to import technologies and adapt them to specific conditions. However, this process

314    Geraldo B. Martha Jr. and Eliseu Alves has limited chances of success when very different environments (or ones experiencing fast-​changing conditions)—​such as the Cerrado as against Brazil’s Southeast and South regions—​are considered. In such cases, problems can only be solved through the continuous generation of new knowledge and technologies (i.e., through science-​based approaches) (Alves 1987).

15.2.2.3. Agricultural Research In 1812, King Dom João VI of Portugal recommended the creation of an applied agricultural course in the state of Bahia. In 1814, an agricultural course was established in the state of Rio de Janeiro, and by the end of the nineteenth century and the early twentieth century, several initiatives to promote the teaching and diffusion of applied knowledge in agriculture were established. The success of such initiatives was variable, meaning that some of them perpetuated and strengthened over the decades, such as the Escola Superior de Agricultura Luiz de Queiroz, from 1901, and the Escola Superior de Agricultura de Lavras, from 1908, while others were short-​lived. A greater effort to structure and develop agricultural colleges in Brazil took place in the first half of the twentieth century. In 1887, the Instituto Agronômico de Campinas was created. Its influential results for agriculture helped to shape and eventually to consolidate the research initiative at the Ministry of Agriculture by 1938. After several revisions, the research-​driven approach in Brazilian agriculture experienced a very important milestone in 1972, when the Brazilian Ministry of Agriculture commissioned a team of experts with the bold assignment of suggesting how the government could structure a vibrant and efficient agricultural research system in the country. The resulting report—​later known as the “black book” because of the color of its cover (Embrapa 2006)—​was the cornerstone for establishing the Brazilian Agricultural Research Corporation (Embrapa) in 1973,8 and for strengthening the web of agricultural research in the country. Embrapa was designed to serve as the research arm of the Brazilian Ministry of Agriculture, with a nationwide mandate. At that time, several state governments also established agricultural research organizations. Embrapa was assigned the additional mission of coordinating the Brazilian Agricultural Research System, which included state agricultural research organizations, universities, and agricultural colleges,9 and Embrapa itself. In the following decades, Embrapa became one of the largest agricultural research networks in the tropical world (Lopes 2012; Pastore and Alves 1976). In summary, the development of modern agriculture in Brazil was initially prompted by the import substitution industrialization policy from the late 1960s to the mid-​1980s. The accelerated growth in population, urbanization, and per capita income at that time posed a clear and strong demand for the agricultural sector. The expansion of agricultural output enabled larger export volumes, as well as more diverse exports,10 which in turn provided the means to finance imports of technology and capital goods for the emerging national industry. The increased opportunity cost of labor for farmers and the sustained migration from rural areas to cities additionally led to a favorable environment for agricultural growth and modernization (Alves and Pastore 1978). A science-​based

Brazil’s Agricultural Modernization and Embrapa    315 approach, based on the continuous generation of new knowledge and technologies, played a crucial role in transforming Brazilian agriculture from mid-​1970s on.

15.3.  The Emergence of an Era of Science-​Based Agriculture in Brazil 15.3.1.  The Brazilian National Agricultural Research System In the early 1970s, macroeconomic pressure coexisted with a strong appreciation among policymakers of the need to reform Brazil’s public agricultural research. Brazil improved its research structure and capacity substantially by developing a two-​tier system of federal and state-​based agencies, called the National Agricultural Research System (SNPA) (Lopes 2012). Over the decades, the SNPA has been responsible for designing, implementing, developing, and promoting a wide array of knowledge and technology to contribute to innovation in agricultural value chains. As discussed by Beintema et  al. (2001), the implementation of the SNPA concept led to the strengthening of Brazil’s agricultural research and development (R&D) capacity through improved infrastructure, enhancement of human resources, implementation of appropriate management mechanisms, and application of support policies on a national scale. The financial support of the Brazilian government to the SNPA was, of course, of fundamental importance. The accumulated level of investments in public agricultural R&D from 1981 to 2013 totaled US$61.6 billion (an average of approximately US$1.9 billion per year).11 In the 1980s, the average intensity of agricultural R&D investments was 1.21% of the agricultural gross domestic product (GDP), but in the past 25 years this figure has increased by around 50%, and averaged 1.91% of the agricultural GDP in the 1990s, and 1.75% of the agricultural GDP from 2001 to 2013.12 Embrapa is the largest component of the SNPA system. As of 2013, Embrapa represented 42% of the SNPA’s research capacity, followed by the state research organizations (29%), agricultural colleges (26%), and nonprofit organizations (3%). The full-​ time research equivalent in 2013 (FTE: 5,869.4) was composed of 72.5% of researchers with PhD qualifications, 21.5% with MSc, and 6.0% with BSc. Almost 60% of those researchers were concentrated in the 41–​60-​year-​old cohort (Flaherty et al. 2016).

15.3.2. Embrapa From the beginning, Embrapa was conceived as having the clear and important goal of structuring a vibrant and efficient agricultural research system in the country (Embrapa 2006).13 In retrospect, it is now evident that there was no single “silver bullet” explaining Embrapa’s success.

316    Geraldo B. Martha Jr. and Eliseu Alves On the one hand, Embrapa was able to adjust itself to the evolving macro-​economic and political environment. It was created at the right time, when conditions were favorable for its success. Perhaps if Embrapa had begun a few years earlier or later, it would not have been such a success. Over the decades, Embrapa was able to wisely read the signals of change and to build the necessary institutional flexibility to adjust and cope with a changing surrounding environment. On the other hand, most of Embrapa’s strength comes from its excellent human resources. The ability of its teams to anticipate the “real-​world” problems of Brazilian agriculture, and to respond promptly with the generation of knowledge and technologies relevant to a given context, coupled with a focus on science-​based solutions and long-​ term commitment, provided Embrapa with the means to thrive and grow stronger over its history.

15.3.2.1. The Embrapa Model: A Public Corporation Learning from approximately 40 years of experience in public research at the Ministry of Agriculture (e.g., the Departamento Nacional de Pesquisa Agropecuária) led to the organization of Embrapa, a federal agency under the Ministry of Agriculture, as a public corporation. This strategy was intended to release it from bureaucratic rules used in public administration, and thus to give it the flexibility to administer resources and personnel, to plan, to assess performance, to implement the budget, and to disseminate results, all in a transparent manner. By choosing private-​sector rules for its employees (e.g., CLT; Consolidação das Leis do Trabalho), Embrapa gained flexibility in personnel administration, construction of careers—​especially that of researchers—​and in designing and implementing a personnel evaluation policy. Furthermore, as a public corporation, the relationship with the outside world and with the private sector would be easier.

15.3.2.2. The Embrapa Model: A Focused Research Approach Embrapa’s success, at the national level, came to depend on its ability to understand a variety of problems and to present solutions accordingly—​a considerable task considering Brazil’s continental size and diversity. The understanding was that Embrapa needed its own research network so that it could be directly responsible for its results, allowing it to become known and evaluated on its own merits. Once it was large, diverse, and decentralized, Embrapa would have the capability to represent the federal government in an area as important as agriculture and to receive priority both in the allocation of resources and with regard to institutional development (Alves 2010). This model further allowed Embrapa to seek cooperation with universities, research institutes, private-​ sector companies, and overseas partners as equals.14 Embrapa’s inception was founded on two major pillars: (1) a focused research model, and (2) a strong effort to develop and then continuously strengthen the capacity of its human resources. Embrapa chose a decentralized territorial model. At the national level, the model requires strong interaction with decision-​makers, at the level of the presidency of the

Brazil’s Agricultural Modernization and Embrapa    317 Republic, Congress, and ministries. Regionally, Embrapa research units (“labs”) were distributed throughout the national territory and followed three major structures: (1) products (rice and beans, beef cattle, etc.); (2) resources (Cerrado savannah, semi-​arid, etc.); and (3) themes (environment, instrumentation, etc.). By focusing on products, regions, and areas of fundamental importance for the development of Brazil, researchers gained a better sense of their responsibilities, minimizing ambiguities regarding goals and necessary actions. This model has also facilitated interaction with farmers and the public,15 who could objectively and efficiently get specific information and results for their needs. More important, perhaps, are the strong ties of solidarity and the collective spirit in which employees are committed to being a winning and respectful team (Alves 2010).

15.3.2.3. The Embrapa Model: Human Resources A major determinant of the observed success of Embrapa has been the development and strengthening of its human resources.16 In a first phase, Embrapa invested heavily in increasing the number of researchers, at an impressive rate of 5.8% per year, from 872 in 1974 to 2,146 researchers in 1990. In a subsequent phase, from 1990 to 2015, the rate of growth in the number of researchers was flat (e.g., 0.5% per year), resulting in 2,447 scientists. Similarly impressive has been the concurrent change in the profile of researchers that took place during this process. In 1974, researchers whose highest qualification was BSc, MSc, and PhD degrees represented 83%, 15.3%, and 1.7%, respectively. In 1990, 53.6% of researchers’ highest degree was an MSc degree, and 26.1% had a PhD degree. In 2015, only 0.7% of Embrapa’s researchers had only a BSc degree, while MSc was the highest degree held by 13.5%; the share of researchers with a PhD degree was 85.8%. This major change in the composition of researchers’ qualifications clearly reflects the science-​based approach pursued by Embrapa since its foundation. These scientists, working together with a competent staff of assistants and analysts17 structured to support research, have played a pivotal role in leading the development of knowledge and technologies to back a powerful tropical agriculture based on science. Across geographic regions, there is an approximately equal distribution among Embrapa’s assistants, analysts, and researchers. However, the share of researchers with a PhD degree at Northern research units lags behind (Martha Jr. et al. 2012a). In summary, Embrapa has traditionally invested heavily in capacity building and strengthening, and has kept a strong postgraduate program throughout its existence. Another specific role of this “human capital policy” is to stimulate creativity with an environment that encourages coexistence and interaction among peers and different stakeholders. Embrapa needs to be prepared to receive, interpret, and internalize the signals coming from a complex society, as well as from the international market, since the needs for interaction will increasingly cross national borders. In particular, human capital is increasingly in demand in an economy that is becoming both technologically and organizationally more complex (Sowell 2015), as is the case with agriculture and its value chains.

318    Geraldo B. Martha Jr. and Eliseu Alves

15.3.2.4. The Embrapa Model: International Cooperation Is Key Empirical evidence has shown that the interaction between people with different heritages in terms of geographical, cultural, social, and political factors is essential in fostering the development of knowledge and technology that ultimately supports sustained innovation flows (Sowell 2015). As such, the interaction between public and private agricultural research organizations should not be solely within a country’s borders. International partnerships are necessary to more rapidly advance the generation of basic and applied knowledge and technologies in support to the innovation process. Embrapa understood the lesson of the importance of international cooperation from its beginning in the early 1970s. This openness to international cooperation enabled Embrapa (Alves 2010) (a) to create a positive image abroad, thus facilitating relationships with donors, universities, and research organizations in other countries. This, in turn, was positively perceived by the federal government, which responded with increased support; (b) to have an international dimension in terms of the quality of research and in measuring scientists’ performance; (c) to help Brazil, as an instrument of foreign policy; (d) to understand that in a globalized world, science is also globalized and that it is crucial for its very existence to improve the mechanisms of interaction with other countries, universities, funding bodies, broader types of organizations, and, of course, other scientists. Embrapa sent abroad several hundred professionals to be trained, mostly in the United States,18 and to a lesser extent in the United Kingdom, Canada, Spain, the Netherlands, Germany, and Australia. Their performance helped Embrapa to form important bridges with the academic world abroad. Moreover, projects financed by international agencies were important in better equipping Embrapa and in helping it to finance training programs abroad. Because these activities were well implemented and conducted, they helped to solidify the image of Embrapa as a serious and responsible corporation. International scientific cooperation brings several benefits to partner countries: (1) it opens up new possibilities for the development and application of knowledge and innovation about agricultural and related value chains; (2) costs of research can be reduced when synergies are identified, built, and consolidated; and (3) it reduces the time required for the appropriation of knowledge and/​or cutting-​edge technologies by the productive sector, which ultimately contributes to maintaining its competitiveness and sustainability. In the mid-​1990s, Embrapa made another strategic decision: to further foster international scientific cooperation by establishing the Embrapa Labex Program, Embrapa’s Virtual Laboratory Abroad. Labex’s mission is to promote and develop international scientific cooperation opportunities in strategic areas and themes, generally at the frontier of knowledge, and to monitor science, innovative technologies, and innovation in agriculture, anticipating potential risks and opportunities. With Labex, there are opportunities to exchange knowledge and to engage senior scientists, allowing them to collaborate on mutually beneficial research in priority areas. Among other things, such

Brazil’s Agricultural Modernization and Embrapa    319 as expanded networking and funding possibilities, this strategy has reduced time and cost in the development of research. Experiences through the Embrapa Labex Program came from the United States, Europe, and Asia.19 The success of Brazilian agriculture motivates countries in the tropical belt to seek information and support for technology transfer from Embrapa. Besides the traditional instruments of support, Embrapa has decided to have researchers in several tropical countries, aiming to transfer knowledge and technology in tropical agriculture and to look for opportunities in licensing Embrapa’s technology. The goal is both to benefit the development of sustainable and competitive agriculture in recipient countries and to help them find sound solutions to improve food security for their people (Alves 2010). Both the Labex model—​ cooperating in research programs with developed countries—​and the structures for transferring technologies to developing countries are flexible models that can be expanded with new scientists or through occasional transfers of scientists among countries, according to the interests of Embrapa.

15.3.2.5. The Embrapa Model: The Continued Support of the Brazilian Government During its first decade or so of existence, Embrapa, in spite of its bold and modern design, was essentially a promise; the risk of a lack of actual achievements was always present. Embrapa’s strategy considered the importance of a research portfolio capable of providing short-​term outputs while more long-​term research, with more significant expected outcomes, was underway. Embrapa gave special attention to the dissemination of existing results, including through an interaction with the emerging media targeting issues in farming. The support coming from farmers and their associations, as well as from the media, helped Embrapa to cultivate a favorable image in the eyes of both the public and government. Without federal government support, Embrapa would not have been possible. During 1974–​1985, for example, huge investments were made in infrastructure, operational costs, and the training of personnel: around R$17.2 billion in 2016 values. Such investments were based on the promise that Embrapa was central for the modernization of Brazilian agriculture. However, crucially, from the early 1980s onward, as research efforts began to bear fruit, Embrapa was no longer merely a promise, and its success explains the government’s continued investments. The support of the Brazilian government has continued through Embrapa’s history. In 2016 values, it averaged R$0.98 billion in the 1970s, R$1.90 billion in the 1980s, R$2.21 billion in the 1990s, and R$2.46 billion from the first decade of the 2000s on. Embrapa’s spending, in its earlier years, focused on the Central-​West region. This stemmed from the need to incorporate the Cerrado into the productive process. While the share in spending for research units in the North, South, and Southeast regions were more or less in balance in the last four decades, the money directed to Northeast units as a share of total spending increased in the past decade (Martha Jr. et al. 2012a).

320    Geraldo B. Martha Jr. and Eliseu Alves Such a relative increase in the share of spending in the Northeast is consistent with the fact that the region accounts for 47.4% of the farms and 47.1% of the rural population. The annual income per farm in the region (R$11,578.44) represents only 41% of Brazil’s average (R$27,789.50), clearly indicating the need for greater assistance (Alves and Rocha 2010). The increasing importance of the environmental agenda, in light of the fact that land-​use changes, especially in the Amazon, are the main factor responsible for greenhouse gas emissions in Brazil, suggests that the spending share directed to research units in the North region will eventually increase as well.

15.3.2.6. The Embrapa Model: Science-​Policy Links Politicians represent Brazilian society, and as such it is obviously important that they play an active role in Embrapa’s operation with respect to directions, priorities for research, and institutional development. Furthermore, in a world that reveals itself as increasingly more complex and technology driven, the interaction with politicians is important as a means to bring scientific solutions to the political decision-​making process. Embrapa has historically had a close working relationship with politicians, but guided by a commitment to the country’s interests rather than to particular political forces; a degree of independence from politics has been a contributor to Embrapa’s success, resulting in a widely recognized impression of transparency. Hiring researchers and employees through an open public selection process and, likewise, opening positions at the head of research units based on open calls are instruments that have promoted a positive coexistence with politicians (Alves 2010).

15.3.2.7. Embrapa’s 2014–​2034 Vision In recent decades Embrapa has implemented several managerial innovations aiming to benefit Brazilian society through a sustained flow of high-​quality research, knowl­ edge, and technologies. In 2012, Embrapa implemented a model combining a portfolio (top-​down) strategy with a project arrangement (bottom-​up) strategy, to prioritize research themes of greatest relevance and strategic importance to Brazil. By the end of 2012, Embrapa had established its in-​house think-​tank, Agropensa (Embrapa’s Strategic Intelligence System). Agropensa led the process of exploring the “Vision 2014–​2034: The Future of Technological Development of Brazilian Agriculture” (Embrapa 2014). This document captured major drivers, challenges, and opportunities in agricultural value chains in Brazil and was a fundamental reference in Embrapa’s Sixth Master Plan, presented to the public toward the end of 2014 (Embrapa 2015). As shown in Figure 15.1, Embrapa’s Strategic Map begins, at the bottom, with inputs of strategic information and key references (“Bases for Action”). These, in turn, aim to support the agency’s decision-​making process regarding its “macro-​processes” such as human resources, international cooperation, and infrastructure for information and communication technology. For example, the document “Vision 2014–​2034” (Embrapa 2014)  identified that Brazilian comparative advantages may be strengthened along the entire agricultural

Mission Advances in sustainability

New sciences: biotechnology, nanotechnology and geotechnology

RDI challenges: macro themes and cross-cutting themes

Natural resources and climate change

Positioning at the frontier of knowledge

Food security, nutrition and health Agro-industrial Technology, biomass and green chemistry

Automation, precision agriculture and ICTS

Markets, policies and rural-urban communications

Production systems

Management innovations in production chains

RDI management

Bases for action

Productive insertion and poverty reduction

Animal and plant Health safety and agricultural defense

Family farming, organic and agroecological production

Institutional management

Contributions to public policies

Communication between rural and urban areas

Efficiency of RDI management Organizational management

Networks and partnerships

Administration finance and infrastructure

Embrapa’s production process

Impact axes

Vision

Strategic and competitive insertion in bioeconomy

Information Development and and communications managment techonology of people International Communications presence

Strategic information and public policies

Figure 15.1.  Embrapa’s strategic map, as presented at the VI Master Plan. Source: Embrapa (2015).

322    Geraldo B. Martha Jr. and Eliseu Alves value chain. The ample variety in the supply of biomass in the country offers real opportunities for the development of value chains based on high value-​added materials and substances targeted for food, feed, flavors, and nonfood uses. Chemical-​bio-​ catalytic processes lead to the development and use of microbial catalysts that directly convert raw materials into a range of products and chemical intermediates that can be subsequently converted into new products with a high value-​added potential (Embrapa 2014). Such a bioeconomy strategy may eventually boost the growth of associated capital goods industries, engineering services, and biomass suppliers in food, feed, chemistry, and pharmaceutical value chains (among others), creating opportunities for expanding higher-​value-​added exports. The goal of the “Institutional Management Level” is to support overall “Research, Development and Innovation (RD&I) Management.” Taking the bioeconomy example again, key questions to be answered at the “Institutional Management Level” would include—​although of course not be limited to—​the following: What is the demand for human resources and international scientific cooperation? What is the financing necessary to support infrastructure, research, and technology transfer activities associated with such a bioeconomy strategy? The RD&I production process is translated into “macro-​themes”20 that follow a value-​ chain approach, and cross-​cutting themes, namely (a) family farming, organic and agro-​ ecological production; (b) management in agricultural value chains; and (c) improved communication with society. With such a strategy, Embrapa aims to be fully connected with “real-​world problems.” The RD&I production process is thus designed to deliver to society outputs and outcomes that contribute mainly within five major “Impact Axes”: (1) advances in sustainability; (2) strategic and competitive insertion into the bioeconomy; (3) contributions to public policies; (4) science-​based agriculture to promote farmers’ inclusion and competitiveness in the economy, as well as poverty reduction; and (5) positioning at the frontier of knowledge.

15.3.2.8. Embrapa’s Payoff to Society There is little doubt that the payoffs to agricultural R&D have been high over the past 60 years (Alston et al. 1998; Avila et al. 2010; Pardey et al. 2006). The practice of conducting and presenting results from impact studies at Embrapa began in the early 1980s. The main motivation behind carrying out these studies was to demonstrate returns to the sizable investments made in Embrapa and its research units by the Brazilian government. Such impact studies centered on estimating economic surpluses generated by Embrapa’s technologies. Internal rates of return (IRR) were estimated, and averaged 25%–​30%, ranging from 20% to 74% in several studies (Avila et al. 2002).21 Pardey et  al. (2006) presented a detailed study evaluating the impact of soybean, dry beans, and rice varietal improvement at Embrapa as compared to non-​Embrapa investments. In the aggregate, varietal improvement in these crops from 1981 to 2003 yielded benefits of US$14.8 billion (1999 prices). Attributing all of the benefits to Embrapa, the benefit-​cost ratio would be 27 for upland rice, 15 for dry beans, and 149 for soybeans. Under alternative distribution rules, under which Embrapa was given partial

Brazil’s Agricultural Modernization and Embrapa    323 credit for the varieties developed jointly with other partners, the ratios would drop to 5, 3, and 31, respectively. At the regional level, Embrapa has assessed its impact from the 1970s to 1980s. While for Embrapa as a whole the registered internal rate of returns ranged from 34% to 41%, the internal rates of returns were comparatively smaller for the North, with 24% (Kitamura et al. 1989), and for the Northeast, with 25% (Santos et al. 1989), and were higher for the Midwest and the South, both with 43% (Lanzer et al. 1989; Teixeira et al. 1990). In the 1990s, regional impact could be indirectly estimated through the research impacts in grain and oilseed varieties, because of the regional distribution of these crops in the country. For example, Evenson and Avila (1995) found that the internal rates of returns for soybean, corn, rice, and wheat were 40%, 58%, 37%, and 40%, respectively. These crops are mainly concentrated in the Central-​South region. There is, however, imprecise evidence regarding the impacts of agricultural research over a complex array of technologies. In part this reflects the difficulty of attributing adequate weights to benefits and costs among different agents involved in the process. Gasques et al. (2009), looking at the more aggregated level, estimated that a 1% increase in Embrapa’s research expenditure increased agricultural total factor productivity (TFP) in Brazil by 0.2%. In short, Embrapa is a case of successful institutional innovation in which its model allowed it to develop its own “personality,” leading Embrapa to be considered a unique example in the field of public research on the national and international scenes. From the outset, Embrapa has always pursued a results-​oriented approach, based on solid science, which has facilitated its relationship with the government and has enabled it to gain public support. Embrapa’s strengths and achievements are explained by a varied set of factors:  a public corporation model of organization; a scale of operation at national level; spatial decentralization; specialized research units; a strong focus on capacity-​building and strengthening of human resources; and a vision of an agriculture based on science and technology. Omit one of these critical factors, and most probably Brazilian agriculture would not be the success story that we know today (Alves 2010). Over the past four-​plus decades, there have of course been a few challenges and crises for Embrapa. Generally, these arose from situations where some sort of interference represented a violation of Embrapa’s principles. However, such crises were generally overcome fairly rapidly and thus ultimately did not compromise Embrapa’s work (Alves 2010). Given that Embrapa is an organization with the ambition of persisting for a long time into the future, serving both Brazil and the wider world, it is relevant to briefly address a few key challenges and shortcomings observed in past decades. These may provide insights for overcoming future situations. One key issue is the need to constantly ensure that all teams at Embrapa understand and follow the Embrapa model, mission, and values. Embrapa’s success thus far has been to a great extent a consequence of its constancy of purpose and its long-​term commitment. If such ingredients became somewhat weakened, Embrapa would encounter more difficulty in facing future challenges.

324    Geraldo B. Martha Jr. and Eliseu Alves Still, from an internal perspective, one must consider that Embrapa’s strength comes from its excellent human resources. As such, keeping a high level of human capital capabilities, with the most up-​to-​date skills, is a necessary condition for facing future challenges and opportunities. Since Embrapa must be prepared for the unknown in a fast-​changing future environment, it will need to further strengthen investment in human capital, probably to a level that is more than would be strictly necessary to accomplish immediate needs. In this context, strong training in basic sciences will be the key to providing the foundations to increase the efficiency of problem-​solving (applied) research. This also indicates why it remains important that relationships and cooperation with universities and private research in Brazil and abroad be strengthened. From an external viewpoint, two other issues merit highlighting. Embrapa must evolve with surrounding policies and must have a close relationship with politicians, always in the nation’s interests. However, historical evidence has clearly demonstrated the importance of maintaining Embrapa’s independence from politics. To that end, transparency is key, for example through measures such as open public selection processes at levels from analysts and researchers to executive-​board positions, and by implementing strategies such as committees and councils to effectively capture politicians’, stake-​ holders’ and broader society’s needs and demands. A final point is the ongoing importance of Embrapa’s connection to real-​world problems, challenges, and opportunities, in order to continue its trajectory as a scientific organization driven by a results-​oriented approach. The dynamics of Brazilian agriculture over the past four decades has determined a range of groups of farmers that need to be understood in order to be efficiently reached by Embrapa’s research. To highlight the two most different groups: one is less numerous but very well integrated into markets and value chains, and is responsible for the majority of achievements in Brazilian agriculture; at the other extreme is a group that is much more numerous but generally more isolated and poorly connected to markets and value chains. The latter group urgently needs solutions tailored to their problems. If that vast group of farmers were able to adequately adopt modern technologies—​those already generated and those becoming available in the near future—​a new breakthrough from a supply-​side perspective will be observed in Brazilian agriculture. Facing these challenges—​or, perhaps, taking up these opportunities—​implies alleviating market imperfections so that the knowledge and technology generated by the research system might be more fully used by a far greater number of farmers.

15.4.  Science-​Based Agriculture Pays 15.4.1.  The Conquest of the Brazilian Cerrado In the early 1970s, Brazilian policymakers realized that the strategy of increasing food supply through the expansion of the total cultivated area and the adoption of practices of

Brazil’s Agricultural Modernization and Embrapa    325 limited technological content should be thoroughly revised.22 The subsistence-​farming option was rejected, and a huge effort to transform traditional tropical agriculture toward one based on science and the pursuit of productivity gains, rather than area expansion, was initiated (Alves et al. 2008; Martha Jr. et al. 2012a). The applied agricultural science developed by Embrapa, states’ agricultural research organizations, universities, and other public and private partners (in Brazil and abroad) lifted the constraints imposed by the climatic conditions of the tropics and the poor acidic soils of the Cerrado. New crop varieties adapted to low latitudes, the release of novel forage species, significant improvements in animal genetics and in the management and use of natural resources (soil, water), and the increased adoption of modern inputs (in addition to seeds) all provided the means for better and more fitted agricultural production systems. The intensification of agricultural mechanization, particularly for grain production, was also an important feature of the development of Brazilian agriculture (Martha Jr. et al. 2012a; Pereira et al. 2012). Changing the production environment to take full advantage of the Cerrado’s potential was thus perceived as a strategy to be explored in order that the enterprise would be competitive with traditional production regions in the country. The virtuous cycle of expanded and improved tropical agriculture research strengthened the innovation flow as new science-​based technologies fueled the extension service and ultimately reached farmers in the following decades. The innovation flow generated by this approach was backed by agricultural and more general public policies, and as a whole has made it possible for Brazilian agriculture to be transformed and to present high-​impact outcomes (Lopes and Martha Jr. 2016).

15.4.2.  The Sustainability of Brazilian Agriculture The various dimensions of sustainability—​technical, economic, social, environmental—​ have strong interdependent linkages and, ideally, should be met simultaneously. Focusing on a solitary dimension, such as the economic or environmental aspects, will not reflect the multiple dimensions of sustainability. Instead of such a limited view, agricultural production systems should design strategies that yield win-​win situations, that is, simultaneous gains in all sustainability dimensions. When this ideal condition is not an option, small loss/​big gain scenarios should be targeted. In Table 15.1 we summarize the evolution in productivity for major crops and beef cattle over the past six decades. From 1950 to 1970, yields were stagnant for corn, wheat, rice, and beef. Yields improved for sugarcane, and especially for soybean, but actually decreased for beans. Significant improvements were observed thereafter, especially from the mid-​1990s on. In the past 60 years, the productivity increase in Brazilian agriculture has ranged from two to nearly five times the levels registered in 1950. An outstanding benefit to the environmental dimension of sustainability arising from land productivity gains in Brazilian agriculture is the so called land-​saving effects, that is, the area of land left uncultivated due to technological progress increasing agricultural

326    Geraldo B. Martha Jr. and Eliseu Alves Table 15.1 Evolution of Productivity in Selected Agricultural Activities in Brazilian Agriculture (1950 = 100) 1950

1975

1985

1996

2006

2015

Corn

100

106

118

195

288

440

Soybean

100

206

236

311

347

404

Wheat

100

96

215

241

246

315

Rice

100

104

135

211

305

447

Beans

100

78

72

97

137

204

Sugarcane

100

160

225

231

256

275

Beef cattle*

100

107

123

227

431

489*

* Estimate. Source: Data from IBGE and Conab, calculations and elaboration in Martha Jr. (2016).

output per unit of area instead. In the past 15 years, the sizable productivity gains in pastoral systems allowed a significant acreage of pasture area to be freed up and thus to accommodate the expansion of crops, mainly soybean and sugarcane, effectively minimizing direct and indirect pressures on native ecosystems. Indeed, in the 1950–​2016 period, productivity gains resulting from increased technology adoption in crops (rice, beans, maize, soybean, wheat, sugarcane) and beef production supported a land-​saving effect of over 680 million hectares (Martha Jr. et al. 2012b; Martha Jr. 2016). This virtuous growth path in Brazilian agriculture explains how the country can be one of the world’s top agricultural producers while maintaining more than 60% of its territory untouched.23 The more widespread adoption of technologies such as integrated crop-​livestock systems, no-​till planting, recovery of degraded areas, planting of commercial forests, biological nitrogen fixation, and animal waste treatment will help to conserve natural resources (soil, water, forests, and biodiversity) and deal with global warming adaptation and mitigation issues in agriculture. However, one must consider that the levels of incentives for Brazilian agriculture are low,24 suggesting that farmers will strongly respond to market signals and will adopt technologies based on individual perceptions of benefits and costs. From another point of view, such a low level of incentives for Brazilian agriculture reinforces further its very favorable input-​to-​output ratio for society. Furthermore, agricultural value chains have been the cornerstone for positive results in Brazilian balance of trade for the past 20 years. From 1997 to 2015, agricultural value chains had an accumulated surplus in the balance of trade (export–​imports) of US$1,519.18 billion, compared to an accumulated value of US$555.68 billion for the aggregate of the Brazilian economy.25

Brazil’s Agricultural Modernization and Embrapa    327

Food basket real price index (1975 = 100)

110.0 100.0 90.0 80.0 70.0 Sept./16 = 61.0 60.0 50.0 Aug/06 = 42.2

Jan. 1975 Mar. 1976 May. 1977 Jul. 1978 Sep. 1979 Nov. 1980 Jan. 1982 Mar. 1983 May. 1984 Jul. 1985 Sep. 1986 Nov. 1987 Jan. 1989 Mar. 1990 May. 1991 Jul. 1992 Sep. 1993 Nov. 1994 Jan. 1996 Mar. 1997 May. 1998 Jul. 1999 Sep. 2000 Nov. 2001 Jan. 2003 Mar. 2004 May. 2005 Jul. 2006 Sep. 2007 Nov. 2008 Jan. 2010 Mar. 2011 May. 2012 Jul. 2013 Sep. 2014 Nov. 2015

40.0

Figure 15.2.  Evolution of relative real prices for a representative food basket, from January 1975 to September 2016 (January 1975 = 100). Note: Prices for the state of São Paulo. Source: Data from Dieese (2016; available at https://​www.dieese.org.br/​cesta/​). The inflation deflator is the IGP-​DI (September 2016). Authors’ calculations and elaboration.

It is worth mentioning that over the past four decades, real food prices for Brazilian consumers decreased by roughly half, along with an associated lower price volatility (Figure 15.2). This huge drop in food prices reflected the fact that supply increased at a much faster rate than demand. Gains due to lower food prices benefit the entire population, but poor families in rural areas and in cities received the greatest share of these benefits, because the greatest share of the poor’s income is spent on food. Thus, reducing the price of food in effect works as an income transfer to the poor without the need for direct reallocation of income within society (Pereira et al. 2012). In addition to providing food security for the Brazilian population, such a condition allowed inflationary pressures to be alleviated. Due to the sizable associated income effect, especially for the poor, the decrease in real food prices additionally helped boost other sectors in the economy (Martha Jr. et al. 2012a; Pereira et al. 2012).

15.4.3.  Total Factor Productivity in Brazilian Agriculture Agricultural production systems are quite diverse. However, at a very aggregated level, it is probably reasonable to identify two broad types of system in terms of factors of production usage: (1) subsistence farming, which almost entirely does not use modern inputs (usage of rudimentary tools and methods for cultivation, and occasionally

328    Geraldo B. Martha Jr. and Eliseu Alves improved seeds); (2) commercial farming, which operates using modern inputs such as fertilizers, agrochemicals, improved seeds, machinery and equipment, and whose intensity of use on a per area basis varies largely. Weather conditions in terms of temperature, radiation, and rainfall change according to agro-​climatic zones and obviously affect both systems. However, due to improvements in the production environment, commercial farming systems can be expected to cope better with such weather variations. It turns out that as the amount of variable inputs and capital increases, farmers are progressively linked to markets, and land productivity, in spite of being very important, loses its capacity to almost solely account for net revenue. Put differently, land productivity response is a necessary but not sufficient condition for maximizing net revenue. Given that product prices and associated relative prices and terms of trade show ups and downs over time, “rational targets” for land productivity will vary conversely. Only under very particular conditions are maximum land productivities associated with maximum net revenues. Hence, for more intensified agricultural systems, total factor productivity (TFP) becomes a better measurement of the system’s efficiency. Put simply, TFP is a productivity measure involving all the partial measures of productivity—​land, labor, and capital—​and defined as a ratio of inputs and outputs. TFP is, thus, the portion of the product not explained by the inputs. The US Department of Agriculture–​ Economic Research Service (USDA-​ ERS) 26 has provided TFP estimates for the world since 1961. Based on this database and its assumptions and methods, TFP growth in Brazilian agriculture averaged 2.31% from 1961 to 2013. We did some further calculations and found that only 6.3% of the 158 of the countries surveyed had higher estimates. We also found that for the 1961–​2013 period, 58.23% of these countries had a TFP growth lower than 1% per year, meaning that they have been unable to completely cover their costs. It is important to recognize, however, that progress is being made. From 1961 to 1970, 70.25% of countries had a TFP growth of less than 1% per year, but this distribution improved in the later period from 2001 to 2013, where 38.73% of the countries had a TFP lower than 1%. Similar results and trends were found by Gasques et al. (2010).

15.5.  Concluding Remarks Future challenges on both the demand and the supply side remain substantial for the agricultural sector. On the demand side, the projected world population growth, from 7.35 billion in 2015 to 9.73 billion by 2050 (UNPD 2015), would add an additional demand of 0.80% per year to the agricultural sector. If the variation in per capita income is additionally considered,27 demand would increase from 1.21% to 3.24% per year for products with income-​elasticities ranging from 0.2 to 1.2, respectively. Nevertheless, huge differences among countries will pose distinct challenges from a regional perspective (Table 15.2).

Brazil’s Agricultural Modernization and Embrapa    329 Table 15.2 Projected Demand Growth for Agricultural Products with Income Elasticities from 0.2 to 1.2 in Selected Countries Income Elasticity Population

Per Capita Income

0.2

0.4

0.6

1.2

2.03%

1.21%

1.62%

2.02%

3.24%

Variation, % per year World

0.80%

Brazil

0.39%

2.11%

0.81%

1.24%

1.66%

2.93%

Russia

–0.31%

2.22%

0.13%

0.57%

1.02%

2.35%

India

0.75%

4.39%

1.63%

2.51%

3.39%

6.02%

China

–0.06%

3.65%

0.67%

1.40%

2.13%

4.32%

South Africa

0.53%

3.26%

1.18%

1.83%

2.48%

4.44%

United States

0.54%

1.59%

0.86%

1.18%

1.49%

2.45%

Source: Data for population from UNPD (2015) and for GDP from OECD long-​term projections (OECD, 2016). Authors’ calculations and elaboration.

On the supply side, production needs to increase at least at the same pace as demand, otherwise agricultural (food, feed, fiber, biofuels, and feedstock) prices will increase and the poor will suffer the greatest impact. There are two ways to increase output in agriculture: yield gains and area increase. They need not be mutually exclusive; in fact, what is generally seen in practice is a combination of both strategies. Remarkable scientific advances are taking place in various fields of knowledge; genomics, nanotechnology, automation and robotics, information and communication technology. Taken together, these and other scientific advances, when properly appropriated by the private sector, will engender innovations that will boost the development of novel agricultural production systems with more potential to add value and to ensure increased productivity, safer and higher quality food, and other improvements to agricultural products and environmental services (Lopes and Martha Jr. 2016). The choice of a technology will vary according to the priority problem to be solved. Hayami and Ruttan (1985) argue that agricultural technologies can broadly focus on land-​and/​or labor-​saving technologies. In the former group are biological and chemical technologies, while the latter includes mechanical technologies. So-​called product-​ saving technologies, linked to reduced losses along the food chain, are additionally perceived as crucial to the future agricultural sector’s outcomes for society. In this context, some key technologies that will be eventually supported include the following (Pereira et al. 2012): new varieties and cultivars (adapted to non-​native ecosystems, with a higher yield and quality in a given environmental set of conditions, resistance and/​or tolerance to biotic and abiotic stresses, incorporation of biotechnology and nanotechnology tools);28 new inputs (machinery and equipment, fertilizers

330    Geraldo B. Martha Jr. and Eliseu Alves and agrochemicals); and new agricultural practices and innovative production systems (providing greater efficiency in water and nutrient use, and accommodation of multiple crop cycles in a year). Of course, the research system and the extension service must receive adequate financial support in order to sustain continuous gains in agricultural yields in farms. However, from a research viewpoint, a central future challenge, given the ample array of stakeholder pressures and funding possibilities, is clearly and objectively identifying the sequence of relevant problems to be solved by research in order to increase welfare in society (Martha Jr. et al. 2012a). From a farmers’ perspective, the economic pressures dictating the intensification strategy will vary according to region and over time, but they are decisive in influencing the systems’ choices and productivity goals. The success of any given strategy will be strongly influenced by the relative prices and terms of trade in a given region. In the short term, sharp variations in the factor’s relative prices may hamper the adoption of capital-​intensive technologies. From society’s viewpoint—​greatly strengthened after the 1990s—​yield increase is the preferable strategy to expand agricultural production. In this case, land-​saving (biochemical) technologies, associated with increased agricultural output per unit area should be targeted. More broadly, technology should focus on the development and/​or adaptation of resource-​saving technologies (for example, land, water, and nutrients) that protect the environment and use resources more efficiently. It is desirable that these novel technologies contribute to mitigating carbon emissions, and provide the necessary resilience to agricultural systems, under a green growth strategy (Lopes and Martha Jr. 2016; Pereira et al. 2012). Finally, for at least two decades, technology has been the main factor responsible for explaining income in Brazilian agriculture (Alves al. 2013). At the aggregate level, the achievements of Brazilian agriculture are in some ways proof that technology was successfully diffused. However, the concentration of agricultural production in a few farms—​27,300, concentrating 51.2% of the total gross income in agriculture—​highlights an enormous challenge for the research system to more accurately translate generated knowledge and technologies in a clear and usable format that can be assimilated by the extension service and adopted by the vast majority of farmers (Alves et al. 2012). Thus, a breakthrough in Brazilian agriculture would come from a greater dissemination—​and an effective implementation—​of modern technologies by a significant number of farmers in the country. Alves et al. (2012) estimate that 54% of the 4.4 million farms that reported income in the 2006 Agricultural Census were able to remunerate all inputs, but only 500,000 establishments had income greater than 10 monthly minimum wages. There remain, therefore, about 2 million establishments with monthly income of up to 10 minimum wages that may find feasible solutions to their livelihood in the agricultural sector. One of the greatest barriers to the effective adoption of modern technology in an inclusive way is market imperfection, which alters the relative prices for farmers and thus the return to investment in technologies. Market imperfection refers to market

Brazil’s Agricultural Modernization and Embrapa    331 power concentration (monopoly, oligopoly, monopsony, and oligopsony) and to non-​ technological asymmetries (such as the availability of infrastructure and education) that restrict a more widespread assimilation of modern technologies. Examples would include higher collateral requirements for loans, and higher prices paid for inputs as against lower prices received by small farmers, who still do not have resources to pay for private technical assistance and thus rely greatly on public extension services. Therefore, from a political perspective, the reduction of market imperfections is a necessary condition for expanding agricultural production in a more inclusive way and for increasing the effectiveness of policies targeting the adoption of technology by farmers. As a final thought, there is no future without science-​based agriculture. In the coming decades, the world might expect from Brazilian agriculture innovations that will increase our current ability to understand and respond to present and future risks and challenges in diverse areas of knowledge in tropical and subtropical environments (Embrapa 2014). In order to make such views a reality it is imperative to expand investments in human resources training. Capital restrictions embodied by the new technology are an outstanding deficiency, but they can be solved by credit policies, while the access to more complex machinery and equipment can be solved by amending the leasing legislation (Alves 2008). Human capital remains the most severe restriction on the potential productive capacity of the agricultural sector, in that this form of restriction requires considerable time to overcome.

Notes 1. Increases in agricultural production result from increased area and/​or increased productivity (e.g., yield). Generally, a combination of both of these two factors explains production levels over time. The higher the yields, the lower the demand for land for a given level of production. 2. As Connor et al. (2011) state, one way of roughly estimating food production needs might take the “Standard Nutritional Unit” as a starting point. A daily energy requirement of around 23 MJ per capita would translate into demand at the farm level of 8.4 GJ per capita, or an approximate annual demand of 500 kg of grain per person. This is sufficiently large—​ more than twice the necessary human intake of digestible energy—​to accommodate diversity in diets, since portions of arable land can be given over to the production of fruits, vegetables, and animal protein rather than to grain. For a more in-​depth analysis, we recognize that it would be necessary to further consider the balance of food exports and imports, but for the present discussion these estimates may suffice, given the sizable trends identified. 3. We estimated these values based on data from the Brazilian Institute of Geography and Statistics (IBGE), for both agricultural production (agricultural census) and population (population census). 4. In the 1990s, and especially in the first decade of the 2000s, the food availability situation in Brazil changed dramatically, and by and large for the better. The amount of grain produced

332    Geraldo B. Martha Jr. and Eliseu Alves per capita jumped to 461 kg in 2000, and again to 730 kg per capita in 2010. As of 2016 this value is estimated at 867 kg of major grains and oilseeds produced per capita. 5. For these calculations we used the database from Brazil’s Central Bank (www.bcb.gov.br/​ credito), and adjusted them to constant R$ 2016 values (the deflator used was the IGP-​ DI provided by FGV; available at http://​www14.fgv.br/​novo_​fgvdados/​default.aspx). Note that the more substantial increase in rural credit in the first decade of the 2000s and in the 2010s only became evident after 2007, particularly in the past five years. 6. Data from IBGE’s historical statistics; available at http://​seculoxx.ibge.gov.br/​economicas/​ tabelas-​setoriais/​agropecuaria. We further estimated that the average fertilizer consumption in croplands, assuming crops accounted for 90% of the total fertilizer usage, increased from 22 kg/​ha/​year in 1970, to 33 kg/​ha/​year in 1985. This change, while modest in absolute per hectare amounts, was probably a very important initial step for, during the following decades, transforming traditional agriculture in the Cerrado into a more modern and productive form. 7. Data from IBGE; available at http://​www.ibge.gov.br/​home/​estatistica/​economia/​ agropecuaria/​censoagro/​2006_​segunda_​apuracao/​default_​tab_​xls.shtm. In this example, we considered croplands accounting for 85% of the tractors, the remaining 15% distributed to other land-​use alternatives such as pasture and forestry. 8. More precisely, Embrapa was created by federal law 5851, from December 1972, and effectively installed on April 26, 1973. When Embrapa was created, it incorporated the former research structure of the Brazilian Ministry of Agriculture, the Departamento Nacional de Pesquisa Agropecuária. 9. Today, federal and state universities conduct research at more than 100 agricultural sciences colleges and schools. However, only a few private universities carry out comprehensive agricultural research, and the nonprofit sector still plays a modest role in research (Lopes 2012). 10. For example, by the end of the 1990s commodities were still important for Brazilian exports, but the two main products at that time—​soybean and iron ore—​represented 10% of total exports (Thorp 1998). 11. Calculations made using ASTI-​IFPRI’s database on agricultural research; available at www.asti.cgiar.org/​data. These values are in constant 2011 PPP dollars. 12. Pardey et al. (2016) have recently analyzed the geographical distribution of agricultural R&D worldwide. Middle-​income countries are increasingly accounting for a greater share of agricultural R&D, growing for instance from 28.8% of the world’s total in 1980 to 43.4% in 2011. Over the three-​decade period from 1980 to 2011, high-​income countries lagged behind middle-​income countries (spending, respectively, US$19 billion and 22.2 billion in 2009 PPP dollars) in terms of the additional amount of investments made in agricultural research. 13. This section benefited greatly from previous papers by Alves (2010) and Martha Jr. et al. (2012a). See also Embrapa (2006), Lopes (2012), Pereira et  al. (2012), and Lopes and Martha (2016). 14. States’ responsibilities for agricultural research and the role of science generation at Agricultural Colleges were further strengthened in the 1980s with the consolidation of the National Agricultural Research System, under Embrapa’s leadership. 15. This research model constitutes an interesting way of identifying research priorities, a typical case of induced innovation. The theory of induced innovation (Hayami and Ruttan 1985) emphasizes the interaction of farmers with researchers. This interaction indicates the

Brazil’s Agricultural Modernization and Embrapa    333 priorities for research within public research institutions. For private research institutions, the market acts directly, otherwise the technology developed would not find buyers. In public research, the market influence is indirect. It creates, among farmers, demand for a certain type of technology, for example, land-​saving technologies, and in response to that demand, farmers indicate their needs to researchers. Researchers, in turn, respond with the adaptation and/​or generation of technologies that increase land productivity. 16. This discussion is based on data and reports provided by Embrapa’s Human Resources Department. 17. At Embrapa, assistants have a high school education, and analysts generally occupy Bachelor’s degree posts. A  small percentage also hold Master’s degrees, and a handful hold PhDs. 18. Several factors contributed to the United States being the major destiny for young Embrapa scientists. The good relationship between the two countries, and the support of the US government and of American organizations (such as the Rockefeller and Ford foundations) played a fundamental role. Furthermore, the United States had been leading the expansion in the frontier of technology, innovation, and labor productivity in several areas of knowledge since 1870 (Gordon 2016), and the country’s achievements in science-​based agriculture over the previous 100 years were remarkable (Cochrane 1990). Additionally, the scale of agricultural enterprises in the two countries were somewhat comparable, and knowledge and technologies transferred to Brazil since the 1950s, especially and 1960s, were presenting a very positive response when properly managed. 19. Labex-​USA was the first instance to be established. The US Department of Agriculture’s Agricultural Research Service (USDA-​ARS) has hosted Labex-​USA in the United States since its inception in 1998. During almost two decades of intense collaboration, 27 researchers from Embrapa have served as Labex-​USA researchers, in loco interacting with peers in institutions of excellence in the United States in several locations across the country, for periods varying from two to four years. The active agreement between Embrapa and USDA-​ARS (December 2015–​December 2020) has identified the following areas for research collaboration: (1) natural resources and climate change; (2) new sciences: biotechnology, nanotechnology, geotechnologies; (3) automation, precision agriculture, and information and communication technologies; (4) animal and plant health and safety in the value chain; (5) production systems; (6) agroindustry and biomass technology and green chemistry; (7) food safety and nutrition for health; (8) agricultural competitiveness and science-​policy links. 20. The macro-​themes are (1) natural resources and climate change; (2) new sciences: biotechnology, nanotechnology, geotechnologies; (3) automation, precision agriculture, and information and communication technologies; (4) animal and plant health and safety in the value chain; (5) production systems; (6) agroindustry and biomass technology and green chemistry; (7) food safety, nutrition and health; (8) agricultural competitiveness—​ markets, policies, rural development (and science-​policy links). 21. Hurley et  al. (2014) scrutinized 2,242 investment evaluations, reported in 372 separate studies on returns to agricultural research from 1958 to 2011. They recalibrated previously reported IRR and found they were more modest than previous studies have indicated (a median of 9.8% versus 39% per year). Nevertheless, the authors stress that these recalibrated IRR are still substantial enough to question the current scaling back of public agricultural R&D spending in many countries.

334    Geraldo B. Martha Jr. and Eliseu Alves 22. This perception prevailed in spite of the fact that more than two-​thirds of the national territory at that time remained untouched and could have been—​and in fact, according to the military government of that time, should have been—​occupied. 23. In addition, one of the authors estimated that in the 1970s the rearing and finishing phases—​i.e., from weaning to slaughter—​implied an average methane emission level of 140 kg per head. Nowadays, the corresponding value of methane emission, from weaning to slaughter, would be around of 85 kg per head. Roughly speaking, the Brazilian cattle industry emits today about 50%–​60% of the methane, per unit of liveweight gain, that it emitted 40 years ago. In the cow-​calf phase the effect of a higher animal performance in reducing methane emission intensity follows the same positive trend. Thus, taking the land-​ saving effect discussed earlier as an analogy, and considering the substantial decrease in the intensity of emissions provided by animal performance improvements, over the past decades Brazilian agriculture similarly experienced a sizable “methane-​saving effect” (G. B. Martha Jr., work in progress, unpublished). 2 4. The incentives for Brazilian agriculture have been low compared to other major players. Considering the metric provided by the Organisation for Economic Co-​ operation and Development (OECD), namely the producer support estimate (PSE), Brazilian farmers received incentives averaging 1.6% of total farm receipts during 1995-​2014. The corresponding values for farmers in the United States and Europe, in the same period, were 13.5% and 28.3% of total farm receipts, respectively. Data available at https://​w ww.oecd.org/​tad/​agricultural-​policies/​producerandconsumersuppor testimatesdatabase.htm. 25. Calculated using FGV’s IGP-​DI for September 2016 as a deflator. Data available at http://​ www.agricultura.gov.br/​internacional/​indicadores-​e-​estatisticas/​balanca-​comercial. 26. This work is led by USDA-​ERS researchers Keith Fuglie and Nicholas Rada. Data available at www.ers.usda.gov/​data-​products/​international-​agriculture-​productivity. 27. The world’s per capita income growth in the 2015–​2050 period, estimated from OECD’s GDP long-​term projection (OECD 2015), was estimated at 2.03%. 28. The new technology of genome editing called CRISPR-​Cas9 promises to revolutionize the science of genetic modification, without the need for transgenics, or the transfer of genes from an organism to another. CRISPR stands for clustered regularly-​interspaced short palindromic repeats, and represents segments of bacterial DNA that, when paired with a specific guide protein, such as CAS-​9 (e.g., CRISPR-​associated protein 9), can be used to make target cuts in an organism genome (Collins et al. 2016). With this technique (and, more recently, CRISPR-​cpf1), it will soon be possible to edit genomes just like one edits a text, by removing or modifying parts of the DNA of the plant itself to modulate desirable traits.

References Alston, J. M., G. W. Norton, and P. G. Pardey. 1998. Science under Scarcity:  Principles and Practice for Agricultural Research Evaluation and Priority Setting. Wallingford: CABI. Alves, E. 1987. “Mobilizing Political Support for the Brazilian Agricultural Research System.” In Policy for Agricultural Research, edited by V. Ruttan, and C. E. Pray, 363–​376. Boulder, CO: Westview Press.

Brazil’s Agricultural Modernization and Embrapa    335 Alves, E. 2008. “Alguns desafios que a Embrapa enfrentará.” Fertbio seminar, Londrina, Paraná, September 17. Alves, E. 2010. “Embrapa: A Successful Case of Institutional Innovation.” Revista de Política Agrícola 19: 64–​72. Alves, E., E. Contini, and J. G. Gasques. 2008. “Evolução da produção e produtividade da agricultura brasileira.” In Agricultura tropical:  Quarto décadas de inovações tecnológicas, institucionais e políticas, edited by A. C. S. Albuquerque and A. G. Silva, 67–​99. Brasília: Embrapa. Alves, E., and A. C. Pastore. 1978. “Import Substitution and Implicit Taxation of Agriculture in Brazil.” American Journal of Agricultural Economics 60: 865–​871. Alves, E., and D. P. Rocha. 2010. “Ganhar tempo é possível?” In Agricultura brasileira: Desempenho, desafios e perspectivas, edited by J. G. Gasques, J. E. R. Vieira Filho, and Z. Navarro, 275–​290. Brasília: IPEA. Alves, E., G. S. Souza, and D. P. Rocha. 2012. “Lucratividade da agricultura.” Revista de Política Agrícola 21 (2): 45–​63. Alves, E. R. A., G. S. Souza, D. P. Rocha, and R. Marra. 2013. “Fatos marcantes da agricultura brasileira.” In Contribuições da Embrapa para o desenvolvimento da agricultura no Brasil, edited by E. R A Alves, G. S. Souza, and E. G. Gomes, 13–​45. Brasília: Embrapa. Ávila, A. F. D., L. Romano, and F. Garagorry. 2010. “Agricultural Productivity in Latin America and the Caribbean and Sources of Growth.” In Handbook of Agricultural Economics, edited by P. L. Pingalli and R. E. Evenson, 3713–​3768. Burlington: Academic Press. Ávila, A. F. D., and G. S. Souza. 2002. “The Importance of Impact Assessment Studies for the Brazilian Agricultural Research System.” Paper presented at the International Conference on Impacts of Agricultural Research and Development: Why Has Impact Assessment Research Not Made More of a Difference? San José, Costa Rica, February 4–​7. Baer, W. 2008. The Brazilian Economy, 6th edition. Boulder, CO: Lynne Rienner. Beintema, N. M., A. F. D. Ávila, and P. G. Pardey. 2001. “Agricultural R&D in Brazil: Policy, Investments and Institutional Profile.” Washington, DC: IFPRI, Embrapa and Fontagro, Ago. Cline, W. R. 1970. Economic Consequences of Land Reform in Brazil. Amsterdam:  North Holland. Cochrane, W. W. 1991. The Development of American Agriculture, 2nd edition. Minneapolis: University of Minnesota. Coelho, C. N. 2001. “70 anos de política agrícola no Brasil (1931–​2001).” Revista de Política Agrícola 10: 3–​58. Collins, J. P., E. Heitman, N. L. Achee, V. Chandler, J. A. Delborne, B. S. Gaut, S. Higgs, G. E. Kaebnick, A. Kingiri, W. Landis, L. Riddiford, J. Tait, L. A. Taneyhill, J. Travis, P. E. Turner, D. E. Winickoff, K. Sawyer, A. Thévenon, R. Miller, F. Sharples, and A. Kolesnikova. 2016. “Gene Drives on the Horizon: Advancing Science, Navigating Uncertainty, and Aligning Research with Public Values.” Washington, DC: The National Academy of Sciences, Report in Brief, June 2016. Dias, G. L., and C. M. Amaral. 2000. “Mudanças estruturais na agricultura brasileira, 1980–​1998.” In Brasil: Uma década em transição, edited by R. Baumann, 223–​254. Rio de Janeiro: Cepal/​Campus. Embrapa. 2006. Sugestões para formulação de um Sistema Nacional de Pesquisa Agropecuária. Brasília: Embrapa Informação Tecnológica. Embrapa. 2014. Visão 2014–​2034:  O futuro do desenvolvimento tecnológico da agricultura brasileira. Brasília: Embrapa.

336    Geraldo B. Martha Jr. and Eliseu Alves Evenson, R. E., and A. F. D. Ávila. 1995. “Productivity Change in the Brazilian Grain Sector and Agricultural Research’s Role.” Revista de Economia e Sociologia Rural 34: 93–​109. Flaherty, K., R. C. N. Guiducci, D. P. Torres, G. L. Vedovoto, A. F. Ávila, and S. Perez. 2016. Brazil: Agricultural R&D Factsheet. Available at www.asti.cgiar.org/​brazil. Furtado, C. 2005. Formação econômica do Brasil. São Paulo: Companhia Editora Nacional. Gasques, J. G., E. T. Bastos, and M. R. P. Bacchi. 2009. “Produtividade e fontes de crescimento da agricultura.” Seminário IPEA, Brasília, June 8. Gasques, J. G.; E. T. Bastos, M. R. P. Bacchi, and C. Valdes. 2010. “Produtividade total dos fatores e transformações da agricultura brasileira:  Análise dos dados dos Censos Agropecuários.” In Agricultura brasileira: Desempenho, desafios e perspectivas, edited by J. G. Gasques, J. E. R. Vieira Filho, and Z. Navarro, 19–​44. Brasília: IPEA. Gordon, R. J. 2016. The Rise and Fall of American Growth: The U.S. Standard of Living since the Civil War. Princeton, NJ: Princeton University Press. Gremaud, A., M. A.  S. Vasconcellos, and R. Toneto, Jr. 2004. Economia brasileira contemporânea, 5th edition. São Paulo: Editora Atlas. Hayami, Y., and V. W. Ruttan. 1985. Agricultural Development: An International Perspective, 2nd edition. Baltimore, MD: John Hopkins University Press. IBGE. 2010. “Indicadores sociais municipais:  uma análise dos resultados do universo do censo demográfico 2010.” Available at http://​www.ibge.gov.br/​home/​estatistica/​populacao/​ censo2010/​indicadores_​sociais_​municipais/​indicadores_​sociais_​municipais.pdf, accessed June 15, 2011. Kitamura, P. C., A. Souza, A. Conto, F. M. Rodrigues, J. Oliveira, J. C. Rezende, N. Vilela, P. Tinoco, P. M. Alves, R. Braga, and R. A. Carvalho. 1989. “Avaliação regional dos impactos sociais e econômicos da pesquisa da Embrapa: Região Amazônica.” Brasília: Embrapa-​DPU. (Embrapa-​SEP, Documentos 38). Lanzer, E. A., I. Ambrosi, D. Dossa, L. M. Freire, A. Girotto, V. Hoeflich, P. Reis, V. F. Osório, V. H. F. Porto, S. X. Souza, and A. M. Trindade. 1989. “Avaliação regional dos impactos sociais e econômicos da pesquisa da Embrapa: Região sul.” Brasília: Embrapa-​DPU, 1989. (Embrapa-​ SEP, Documentos 45.) Lopes, M. A. 2012. “The Brazilian Agricultural Research for Development (ARD) System.” In Improving Agricultural Knowledge and Innovation Systems: OECD Conference Proceedings, 323–​338. Paris: OECD. Lopes, M. A., and G. B. Martha Jr. 2016. “Embrapa: Development of Brazilian Agriculture.” In The Arab World and Latin America: Economic and Political Relations in the 21st Century, edited by F. Saddy, 305–​329. London: I. B. Tauris. Martha, G. B., Jr., E. Alves, and E. Contini. 2012a. “Embrapa: Its Origins and Change.” In The Regional Impact of National Policies: The Case of Brazil, edited by W. Baer. Northampton, 204–​226. UK: Edward Elgar. Martha, G. B., Jr., E. Alves, and E. Contini. 2012b. “Land-​Saving Approaches and Beef Production Growth in Brazil.” Agricultural Systems 110: 173–​177. Martha, G. B., Jr. 2016. “Distribuição espacial dos custos de oportunidade da agropecuária.” Beltsville: Labex-​USA—​Relatório do Programa de Cientista-​Visitante da Embrapa junto ao Departamento de Gestão de Pessoas. North, D. 1990. Institutions, Institutional Change and Economic Performance. Cambridge: Cambridge University Press, 1990.

Brazil’s Agricultural Modernization and Embrapa    337 North, D. 2006. “History Matters.” In Focus: Understanding Economics in United States History, edited by Mark C. Schug, Jean Caldwell, and Tawni Hunt Ferrarini, vii–​ix. New  York: National Council on Economic Education. OECD. National Accounts at a Glance. Paris: OECD. Paiva, R. M. 1975. “Os baixos níveis de renda e salários na agricultura brasileira.” In Tecnologia e desenvolvimento Agrícola edited by C. R. Contador, 195–​231. IPEA: Rio de Janeiro. Pardey, P. G., J. M,. Alston, C. Chan-​Kang, E. C. Magalhães, and S. A. Vosti. 2006. “International and Institutional R&D Spillovers: Attribution of Benefits among Sources for Brazil’s New Crop Varieties.” American Journal of Agricultural Economics 88: 104–​123. Pardey, P. G., C. Chan-​Kang, S. P. Dehmer, and J. M. Beddow. 2016. “Agricultural R&D Is on the Move.” Nature 537: 301–​303. Pastore, A. C.; E. R. A. Alves, and J. A. B. Rizzieri. 1976. “A inovação induzida e os limites à modernização na agricultura brasileira.” Revista de Economia Rural 14 (3): 257–​258. Pastore, A. C., and J. R. M. Barros. 1976. “Absorção de mão-​de-​obra e os efeitos distributivos do progresso tecnológico.” Revista Brasileira de Economia 30: 263–​293. Pastore, J., and E. Alves. 1976. “Reforming the Brazilian Agricultural Research System.” In Resource Allocation and Productivity in National and International Agricultural Research, edited by T. M. Arndt, D. G. Dalrymple, and V. W. Ruttan, 394–​403. Minneapolis: University of Minnesota Press. Pereira, P. A. A., Martha, G. B., Jr.; C. A. M. Santana, and E. Alves. 2012. “The Development of Brazilian Agriculture: Future Technological Challenges and Opportunities.” Agriculture & Food Security (1): 1–​12. Ryff, T. B. 1976. “A difusão da inovação tecnológica na agricultura:  mecanismos de auto-​ controle versus modernização induzida.” Revista Brasileira de Economia 30: 295–​327. Santos, R. F., G. Calegar, V. Silva, M. A. Barros, J. O. Lima, J. Motta, and J. S. Neto. 1989. Avaliação socio-​econômica das pesquisas da Embrapa na região nordeste. Brasília: Embrapa-​ DPU. (Embrapa-​SEP. Documentos, 37). Schuh, G. E. 1974. “A modernização da agricultura brasileira.” In Alternativas de desenvolvimento para grupos de baixa renda na agricultura brasileira, vol.2, 120–​184. São Paulo: IPE/​USP. Sowell, T. 2015. Wealth, Poverty and Politics:  An International Perspective. New  York: Basic Books. Teixeira, S. M., G. C. Gomes, F. P. Costa, E. P. Santana, A. M. Machado, N. A. Santos, J. M. Kruker, L. Coradin, and R. C. Vieira. 1990. Avaliação socioeconômica das pesquisas da Embrapa na Região Centro-​oeste. Brasília; Embrapa-​DPL. (Embrapa-​DPL, Documentos, 09). Thorp, R. 1998. Progress, Poverty and Exclusion: An Economic History of Latin America in the 20th Century. Washington, DC: Inter-​American Development Bank. UNDP. 2015. World Population Prospects—​The 2015 Revision: Highlights and tables. New York: UNDP.

Chapter 16

M anufac t u ri ng , Services, a nd t h e Produ ctiv i t y G a p Jorge Arbache

16.1. Introduction Productivity is widely recognized as one of the main engines of sustained economic growth and social development. In times of demographic transformation and globalization, when economic competition among countries is becoming harsher, productivity gains become even more significant as a growth determinant. How does Brazil compare to other countries in relation to productivity? Figure 16.1 presents the evolution of labor productivity in Brazil and in selected emerging economies from 1951 to 2015. Two important pieces of evidence emerge:  first, productivity in Brazil was higher until the late 1970s. Second, from then on, productivity lost pace and stagnated, and Brazil started to fall behind other countries. Table 16.1 presents the growth rate of labor productivity. A close look suggests that Brazil’s disappointing productivity can be explained mostly by poor performance in the period 1980–​2015; while productivity grew at an annual rate of 3.32% between 1951 and 1979, this fell to a mere 0.46% in the subsequent period, a fraction of the rate in other economies. One of the main causes of the productivity slowdown in the second period was the sharp decline of total factor productivity (TFP) (Ferreira, Pessoa, and Veloso 2013). Relative TFP has remained disappointing. Calculations of TFP from 2000 to 2014 show that it grew at –​0.47% per year in Brazil, while in Korea, it grew 1.69%; in India, 1.41%; in China, 0.46%; and in Malaysia, 1.14%.1 Brazil’s labor productivity lags well behind that of advanced economies as well. The Brazil to United States labor productivity ratio was 19% in 1951, but then increased slowly, reaching 31% in 1980. However, since then the ratio has undergone a long decline and in 2015 was back to 19%.2 Moreover, given the importance of productivity to

Manufacturing, Services, and the Productivity Gap    339 60,000 50,000 40,000 30,000 20,000

0

1951 1953 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015

10,000

Turkey

China

South Korea

India Thailand

Malaysia Brazil

Figure 16.1.  Labor productivity per person employed in 1990 US$. Source: The Conference Board.

Table 16.1 Labor Productivity Growth per Person Employed (Percentage), Selected Periods 1951–​2015

1951–​1979

1980–​2015

Brazil

1.74

3.32

0.46

China

6.18

3.95

7.86

India

2.98

1.45

4.22

Malaysia

2.69

2.63

2.74

Thailand

3.81

4.35

3.39

Turkey

3.55

4.88

2.47

Source: The Conference Board.

economic growth, one consequence of poor productivity performance is that Brazil’s income per capita is not catching up with that of advanced economies. On the contrary, after a decline in the first period, the income gap between the United States and Brazil began to increase; purchasing power parity (PPP)–​corrected data show that in 1980 the US gross domestic product (GDP) per capita was 2.7 times greater than Brazil’s, while by 2015 it was 3.6 times greater. Explaining such a productivity story is, of course, a major challenge because productivity is a multifaceted phenomenon. The literature on productivity performance in Brazil is rich and encompasses several perspectives. These include the following: resource misallocation at firm and sectoral levels; barriers to entry, idiosyncratic distortions, and protectionism, and the impacts of these on competition and the cost

340   Jorge Arbache of capital goods; slow within-​firm productivity growth; and labor mobility and labor market distortions (Arbache, Dickerson, and Green 2004; Busso, Madrigal, and Pagés 2013; De Negri and Cavalcante 2015; Dix-​Carneiro 2014; Ferreira, Pessoa, and Veloso 2013; Muendler 2004; Restuccia 2003; Schor 2004; among others). In this chapter, we examine a little-​explored and yet important aspect of poor productivity in Brazil:  the performance of the service sector. The characteristics of the service sector in Brazil are well known. The sector is composed mainly of small, underfunded, poorly equipped, and poorly managed firms, competition is weak in several segments, and services quality and availability are limited, while prices are generally high. Surveys such as Doing Business, the Enterprise Survey compiled by the World Bank, and the Global Competitiveness Report of the World Economic Forum (WEF) point to several unfavorable conditions for doing business in Brazil, and among these are services of various types and functions such as logistics, financial services, courts, and energy services. We establish that the meager long-​term performance of the services sector is a key factor in explaining poor aggregate productivity and manufacturing competitiveness in Brazil since the 1980s. The reasons for that are twofold. First, Brazil is experiencing a profound structural transformation in favor of the services sector and mainly at the expense of the manufacturing sector, to the extent that the service sector’s shares in output and in employment have become unusually high by emerging market standards. Second, manufacturing firms in Brazil are increasingly outsourcing all types of services to the point where the share of services in total costs has become comparable to the norm in advanced economies. The focus on the impacts of services performance on manufacturing competitiveness is justified by the empirical evidence showing that, contrary to the service sector at large, it holds the best opportunities for emerging economies to promote productivity catch-​ up with developed economies (Arbache 2012, 2016a; EC 2014; McMillan, Rodrik, and Verduzco-​Gallo 2014; OECD 2014; Rodrik 2013, 2015; WEF 2012). The following section (16.2) examines the services sector as a long-​term source of low productivity in Brazil. Section 16.3 investigates the impacts of services on manufacturing competitiveness. Section 16.4 discusses why it is critical for Brazil to improve service sector productivity. Section 16.5 provides some conclusions.

16.2.  The Service Sector as a Source of Low Productivity in Brazil The Brazilian economy has reached a stage where economic activity, investment, and job creation occur first and foremost in services. Figure 16.2 shows that the service sector remained relatively stable at around 50% of GDP until the early 1980s, and that manufacturing’s share of output peaked in the mid-​1980s, at 32%. Brazil’s economic

Manufacturing, Services, and the Productivity Gap    341 80 70 60 50 40 30 20 10

1951 1953 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015

0

Services

Manufacturing

Agriculture

Others

Figure 16.2.  Sectoral participation in GDP (percentage). Source: IBGE.

80

60,000

70

50,000

60

40,000

50 40

30,000

30

20,000

20

10,000

10

0

VNM *VEM IDN MYS CHN ***IND ***THA **PER COL KOR RUS CHL MEX TUR ZAF DEU ***CRI AUS AUT ***JPN SWE BRA ITA DNK BEL ***USA FRA GBR

0

Services (% GDP) - left axis

Per capita income(constant 2011 $) - right axis

*2012 **2013 ***2014

Figure 16.3.  Income per capita and share of services in GDP (2013). Source: Prepared by the author on the basis of data from the World Bank, WDI. Note: VNM: Vietnam; VEM: Venezuela; IDN: Indonesia; MYS: Malaysia; CHN: China; IND: India; THA: Thailand; PER: Peru; COL: Colombia; KOR: Korea; RUS: Russia; CHL: China; MEX: Mexico; TUR: Turkey; ZAF: South Africa; DEU: Germany; CRI: Costa Rica; AUS: Australia; AUT: Austria; JPN: Japan; SWE: Sweden; BRA: Brazil; ITA: Italy; DNK: Denmark; BEL: Belgium; USA: USA; FRA: France; GBR: Great Britain

structure then underwent profound changes as the services sector’s share grew rapidly, mainly at the expense of manufacturing, to reach 71% in 2015. Figure 16.3 shows that the contribution of services to output stands out as an exception among countries of a similar level of development. Brazil’s service share is similar to or higher than those of Austria and Germany, and much higher than in other emerging

342   Jorge Arbache economies such as China. In Korea, whose per capita income is many times higher than that of Brazil, services account for 59% of output. The Brazilian case seems to be an anomaly linked to the decline of the industrial sector. While East Asian economies were able to take advantage of the integration of the world economy and build dynamic manufacturing industries, Brazilian manufacturers were caught unprepared for international competition after decades of protection and import-​substitution industrialization that had focused on the internal market. The service sector is also by far the country’s main source of jobs, accounting for no less than 74.3% of total formal jobs.3 Services play a prominent role in job creation; in recent years, 8.3 of every 10 new formal jobs has originated in the services sector.4 According to the general literature on the turnover and productivity of firms, the productivity of new firms entering the market is immediately lower than that of incumbents, but survivors start learning and benefiting from economies of scale quickly thereafter and then eventually catch up; the productivity of firms exiting the market is lower than that of incumbents. This “purging” effect is expected to increase average productivity (Roberts and Tybout 1997). However, empirical evidence for the period 2001–​2009 in Brazil shows not only that the productivity of entrant firms in the services sector is lower than that of incumbents, but also that firms are not improving productivity over time; the productivity of exiting firms is similar to that of incumbents (Luna et al. 2015). This combination helps to explain the low and stagnant average productivity in the service sector. One possible explanation for such evidence is associated with the buoyant environment of the early years of the 2000s, based on expansionary credit, consumption, and government spending, and a rising middle class, which would have created a more forgiving environment for firms to survive, even with modest performance. The breakdown of services output in its main segments between 1951 and 2015 reveals many medium-​and long-​term trends.5 First, the share of public administration, education, and health in total services grew significantly from 12.3% in 1951 to 23.9% in 2015. Second, while the contribution of retail remains high, at 17.2%, the sector has lost in relative importance over time given that it accounted for 30.8% of total output in 1951. Third, other services, including business services, gained in participation, from 20.9% to 24.1%. Fourth, information services displayed significant growth from 1970, when they first appeared in the statistics, to 2015, moving from 1.2% to 4.4%. This is likely to be related to technological innovation and change. Fifth, the share of financial intermediation in total output grew from 6.5% to 10.4% during the entire period. However, the share of this sector has varied substantially over time; in 1993, for example, it reached a staggering 40.1%, which likely reflects the effects of the inflation tax accruing to the sector during the high-​inflation era. Finally, the share of transport, storage, and postal services in output over the same period decreased slightly from 7.0% to 6.1%. Brazilian households spend a higher share of their total expenditure on services than do their counterparts in other emerging countries. In 2009, the share of household budgets allocated to services was 62%, well above China (55%) and India (50%). Brazilian households differ markedly from other emerging countries in their consumption of financial services and business services, which account for 13% of household

Manufacturing, Services, and the Productivity Gap    343 40

200 180 160 140 120 100 80 60 40 20 0

35 30 25 20 15 10 5 1951 1953 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011

0

Agriculture

Manufacturing

Services

Mining

Figure 16.4.  Evolution of labor productivity (in constant BRL 1,000). Note: Mining = right axis; all others = left axis. Source: Groningen Growth and Development Center.

expenditure in Brazil, but only 5% in China and 6% in India; this is associated with the very high costs of financial services. Figure 16.4 shows sectoral labor productivity. After reaching a peak around 1980, services productivity has been decreasing ever since, and has continued to lag behind, while manufacturing productivity recovered in the mid-​1990s. Looking at productivity in an international comparative perspective also reveals a deficit in Brazilian services. Labor productivity in the aggregate services sector in Brazil is only about 19% of that of the United States, whereas Brazil reaches 32% of US productivity in manufacturing, and 77% in extractive industries.6 There are four factors, associated with formal firms’ characteristics, that contribute to low efficiency and poor adoption of new technologies and best practices. First, services firms are relatively small. The average value added per worker per month is around BRL 5,600, much lower than in manufacturing and extractive industries. Services firms employ 5.2 workers on average. Second, net wages (excluding payroll taxes) rose faster than value added per worker, and were in the order of 21% in 2011,7 significantly higher than in manufacturing. Potential explanations for this evolution are the minimum wage, which has grown substantially more than inflation and is especially influential in wage formation in this sector, and the deceleration of the growth of the working-​age population combined with the stagnation of labor force participation, which constrained the pool of workers seeking employment. Third, job turnover in the service sector is about four times higher than in the rest of the economy. The high turnover discourages investment in human capital both by employees and by employers, thus contributing to the persistence of low productivity (see Gonzaga and Pinto 2014).

344   Jorge Arbache Fourth, services firms are generally underfunded, have limited access to credit and to technology, are poorly managed, and have limited engagement with exports and other dynamic activities (Arbache 2015).8 Firm and worker characteristics differ considerably across service segments, however. In 2012, the value added per employee in air transportation was 79 times that in the food and catering industry. Average wages and firm sizes (number of employees) also vary substantially, too; in air transportation they were BRL 5,428 and 243, respectively, and in food and catering they were BRL 972 and 7.05. Inter-​industry wage differentials capture both disparities between services and other activities, and disparities within the services sector. After controlling for variables such as schooling, age, region, size of firm, and gender, among others covariates, it was found that services activities are overrepresented among those paying wages below average and underrepresented among those paying above the average. This finding can be explained by factors such as technology, market structure, and unmeasured human capital; while earnings of workers in hospitality and food and catering are well below the market average, earnings of workers in air transportation and information are well above. The notion of a unified service sector may therefore be misleading where high-​and low-​tech firms coexist in the same market, as well as high-​and low-​skilled workers. The business environment contributes to the poor performance of the service sector. The tax burden on services is heavier than in other sectors. In 2013, taxes collected on nonfinancial private services amounted to 24% of their total revenue, the largest part consisting of taxes on income and property (CNS 2014). While the average tax burden on the production and consumption of goods and services is 19.4%, it is much higher in the services segments most critical to other sectors of the economy; the tax toll exceeds 23% in transport and business services, 27% in information technology (IT) services, and over 30% in utilities. The service sector is generally highly regulated in Brazil. According to the OECD Services Trade Restrictiveness Index (STRI), which helps identify which policy measures may act as impediments to trade and investment in services at the service segment level, Brazil has a higher than global average score in all sectors except distribution and legal services, a fact explained both by general regulations affecting all sectors and by sector-​specific rules.9 Limitations on the temporary movement of people also affect service providers in all sectors. As an example, Brazil imposes a labor market test for all categories of services suppliers, according to which foreign workers can only be hired if no potential Brazilian candidate has the required skills. Furthermore, the managers of a joint-​stock company must be resident in Brazil in all sectors, and two-​thirds of the employees of any given firm must be Brazilian nationals. Limitations on cross-​border trade in services include the tax treatment of imported services. CIDE (Contribution of Intervention in the Economic Domain) is levied at a rate of 10% on imports of specific services, including royalty payments, technical services, compensation for technology transfers, and technical or administrative assistance. The tax targets technology-​related services, which are particularly relevant to support

Manufacturing, Services, and the Productivity Gap    345 upgrading and differentiation in manufacturing sectors. The procurement law also opens the possibility of discriminating against foreign service suppliers, as it allows for margins of preference of up to 25% of the price for specific goods and services defined by decree as produced in Brazil. At the sector level, several obstacles prevent competition. This is the case for the telecom, commercial banking, logistics, and transport sectors. As an illustration, foreign vessels are restricted from engaging in cabotage and have to be chartered by Brazilian shipping companies to operate in Brazilian jurisdictional waters. Cargo destined for government or state-​owned enterprises, or financed partially or totally with public funds, must be transported by Brazilian flag vessels. One consequence of such a regulatory framework is that competition is restricted and many relevant service markets are highly concentrated. As an example, the five largest banks account for more than two-​thirds of banking sector assets and 80% of the deposit base. These regulations also discourage entrance and limit the access of foreign-​owned companies to Brazilian markets, especially of low-​and medium-​sized firms. Global rankings of infrastructure from the Global Competitiveness Report 2015 show that infrastructure, in general, and transport, in particular, is poor in Brazil. This suggests that those services have made little if any contribution to reducing production and marketing costs. Infrastructure quality is of particular concern; Brazil is ranked 114th out of the 148 countries surveyed. High transportation costs have particularly significant impacts on industries that are more dependent on logistics, such as metallurgy, food, wood, pulp and paper, and other natural resource–​intensive industries. Brazil is also poorly ranked in other public services. Power for industrial purposes is particularly costly, even when compared to other emerging and developed countries. This has implications for industrial competitiveness in general and for energy-​intensive sectors in particular, such as cellulose, oil refining, chemicals, and basic metals. To put this in perspective, in 2011 the price of power for industrial purposes in China was about half of Brazil’s, while in the United States it was about 30% cheaper. Gas for industrial purposes, which is a basic input for many industries, is also costly compared to international standards, including large gas importers such as China and India.10 With the development of shale gas fields in the United States, China, Argentina, and other countries, it is likely that the relative price of gas in Brazil will further increase over the coming years, with additional negative implications for industrial competitiveness. The average cost of Internet access is also much higher than in many other emerging economies.11 The average cost of a 20-​foot sea container service is also very high by international standards, at over US$2,200 (almost three times higher than in China), which is a burden on the shoulders of the exporting firms. When combined with long average transit time and clearance of goods in ports, the relative costs of export services become even higher.12 One major consequence of such market failures is that services are not competitively priced compared to the rest of the economy. A  comparison of the services inflation against overall consumer price index (CPI) inflation (IPCA) in the period 2000–​2015

346   Jorge Arbache shows that both the IPCA-​services components inflation and the Central Bank services inflation rates were substantially higher than the headline index, in the order of 60 percentage points at the end of the 15-​year period. Among the explanations of the services inflation are lack of competition pressure in the services sector, poor investments in infrastructure, low productivity, changes in relative prices in favor of services due to increasing imports of manufacturing goods and exchange rate overvaluation observed in the first decade of the 2000s and the early 2010s, and acceleration of the real minimum wage. The increasing demand for services, especially for final consumption associated with the expansion of the middle class and the aging population, the shortage of skilled labor, and high and distorting taxes also seem to play a role.

16.3.  Services and Manufacturing Competitiveness The Brazilian manufacturing sector has been contracting almost monotonically since the early 1980s—​today, manufacturing contributes to only 11% of output, an unusual figure by large emerging economies’ standards. Long-​term explanations for this abound and include excessive protectionism that prevented competition, little innovation, modest engagement with export activities, and huge macroeconomic volatility, which discouraged investment. The poor regulatory framework and infrastructure, the heavy tax burden, and the shortage of human capital compound the problem.13 As a sector with long value chains, and that relies more heavily on outsourcing, manufacturing tends to be more exposed to the services performance. Indeed, services inputs play a strong role in manufacturing production processes in Brazil, as evidenced by a high share of services in manufacturing output compared to other emerging economies. Overall, 38% of the value added of manufacturing industries originated directly or indirectly in services in 2011 according to the TiVA database.14 The contribution of services to manufacturing is lower in China (31%), Mexico (32%), and Chile (22%). Figure 16.5 shows the ratio of productivity in agriculture, mining, and services to the productivity in manufacturing. The ratio of services to manufacturing fell over most of the period, suggesting that services have indeed constrained manufacturing competitiveness the most. There is significant variation across industries in their reliance on intermediate services. Estimates of the aggregate share of direct services inputs in gross operating revenue by industry, based on the Brazilian Annual Industry Survey (PIA-​IBGE), show that services command a sizable share—​in 2012, it was about 24%. While services in coke, petroleum, and derived products and biofuels accounted for 39% of the total in 2012, in motor vehicles they accounted for only 13%.15 Aside from idiosyncratic and institutional reasons, such variations across sectors are often associated with a variety of factors at the industry level, including relative prices,

Manufacturing, Services, and the Productivity Gap    347 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0

7.0 6.0 5.0 4.0 3.0 2.0 1.0 1951 1953 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011

0.0

Agriculture

Services

Mining

Figure 16.5.  Productivity ratio in relation to manufacturing. Note: Mining = right axis; all others = left axis. Source: Groningen Growth and Development Centre.

openness, technology, and type of services inputs most consumed. Where prices are constrained by factors such as competition, innovation, and technology, all else being constant, the share of services in gross operating revenue tends to go up. Where intermediate consumption of more sophisticated services, such as royalties and technical assistance, or of nontradable services increase, all else being constant, the share of services also tends to rise (Arbache 2014a). Breaking the shares of services inputs in gross operating revenue in 2012 into services segments shows that financial expenses, industrial services provided by third parties (such as installing industrial equipment), non-​industrial services provided by third parties (including IT and professional services), and transportation account for more than 70% of total service expenses. A substantive share of costs is devoted to royalties and technical assistance in oil and gas extraction and refining, but much less so in other sectors. Leasing of machinery, equipment, and vehicles also represents a significant share of costs for the oil and gas extraction and coke and petroleum industries, while it is almost negligible elsewhere. Among the sectors with large weight in manufacturing expenses, the financial sector deserves particular attention considering the still-​high burden of financing costs in firms’ balance sheets. Brazilian banks charge high interest spreads between lending rates and the remuneration of deposits, and these spreads have been on a rising trend more recently. Part of this phenomenon can be attributed to inflationary pressures and volatility in macroeconomic conditions, but the regulatory framework and the lack of competitive pressure on major banks are likely to play a substantive role as well. Besides, the allocation of credit is distorted by the size of the National Development Bank’s (BNDES) and other state-​owned banks’ lending, though their roles as major lending banks have been decreasing since 2013.

348   Jorge Arbache As the effect of the rise of services on the performance of manufacturing depends on a wide range of factors, different industries might see different results, depending on their characteristics and type of services inputs. To better understand the role of services, it is useful to categorize them in smaller, more homogeneous groups. Arbache (2014) proposes a division of services based on their main purpose as inputs. For this end, services are classified into two different yet complementary families. The first category, labeled cost services, refers to functions that affect production costs, including logistics and transportation, general infrastructure services, storage, repair and maintenance services, outsourced services in general, basic IT services, credit and financial services, travel, accommodation, food and catering, and distribution. The second category, labeled value services, refers to functions that add value by differentiating and customizing products and, therefore, raising their market price. The group is composed of services that usually require higher levels of human capital and other capabilities. It includes, inter alia, research and development (R&D), design, engineering and architecture projects, consulting services, software, specialized technical services, high-​end IT services, branding, advertising, and marketing. Better and/​or cheaper cost services can help a firm achieve greater production efficiency, but they do not usually contribute to product customization or value creation in competitive markets. In principle, the longer the supply chain of an industrial segment, the greater the importance of cost services for its competitiveness. One may also expect that the more homogenous a product is, the more important low costs will be for competitiveness. Cost services are thus particularly relevant for commoditized goods such as corn and soybean crops, iron ore, oil, economy cars, basic clothes, and textiles. Conversely, the more sophisticated and differentiated a product is, the greater the importance of value services. Competitiveness in customized products is therefore more reliant on value services. To investigate further whether and how the service sector contributes to enhancing manufacturing performance, we examine empirically the relationship between services and productivity in that sector.16 Correlation coefficients between labor productivity and the share of services in intermediate consumption in manufacturing suggest that higher shares of services inputs are positively and significantly associated with higher productivity levels. When the calculation is made using only the share of value services, the coefficients turn higher, suggesting a stronger association of this category of services with productivity. There is also evidence that this relationship has become stronger over time. While the correlation suggests that services and productivity are linked, little is known about the mechanisms behind this relationship, and potential explanations abound. Changes in relative prices of services and goods, rising per capita income, changes in preferences of consumers, new technologies of production, and the popularization of information and communications technology (ICT) may have influenced that relationship over time. We therefore adopt an agnostic approach and run lowess plots, that is, locally weighted regressions that use a function that attaches less weight to points further away from the mean.17 Looking at services inputs as an aggregate category, in Figure 16.6, Panel A suggests that slopes vary substantially according to service consumption and that the relationship

Panel A 14

14

12

12

11

11

10

10 .2

.3

.4

.5

.6

13 Inprod

13 Inprod

Inprod

Lowess smoother

14

13

.1

12 11 10

.1

.2

.3

.4

0

.05

bandwidth = .8

.1

.15

.2

.25

Value_serv

Cost_serv

Serv bandwidth = .8

Panel C Lowess smoother

Panel B

Lowess smoother

bandwidth = .8

Figure 16.6.  Relationship between share of services in intermediate consumption and productivity. Source: Elaboration by the author.

350   Jorge Arbache between services consumption and productivity appears to have a non-​monotonic shape of a cubic form. Panels B and C distinguish between cost and value services, respectively, to show the relationship between industry-​level labor productivity and each of these two categories of services inputs. While the input shares of value services are highly concentrated around a low mean, the share of cost services has a substantively higher mean and is more widely spread. The plots also reveal that while a large portion of data points in the cost service chart is in an almost flat slope section, a large portion of data points in the case of value services is in a section with a positive slope. It thus appears that value services are those most strongly associated with higher productivity. In order to check whether services are indeed associated with productivity and to explore that relationship further, we estimated regressions with three powers of the service variables using the following specifications: (1) ln productivityit = α + β1 Xit + β2 servicesit + εit (2) ln productivityit = α + β1 Xit + β2 cost _ servicesit + β3value _ servicesit + εit Each industry is represented by the subscript i and years are represented by t. Xit represents a vector of specific characteristics that may affect labor productivity. The control variables at the industry level are workers’ average years of schooling, tenure on the job, industry technology intensity (according to the OECD classification), import coefficient, size of firms, and share of imported inputs in intermediate consumption. These variables are drawn from Brazil’s Ministry of Labor, the Brazilian Institute of Geography and Statistics (IBGE), and the Brazilian National Confederation of Industry. A compilation of qualitative results suggests that services are positively associated with productivity after controlling for other determinants of labor efficiency. These results are robust to different model specifications, time spans, and sets of control variables. When disentangling the contributions of value and cost services, it emerges that higher shares of value services are positively associated with productivity, while cost services appear to have at best a weak influence and possibly a negative one. Holding constant the control variables as well as any other time-​invariant industry characteristics, increasing the intermediate consumption of value services would contribute to improving productivity as soon as a modest level of consumption of those services is reached (about 5% of total input consumption) and for the whole range of values observed in the data above that threshold. The very low levels of value services consumption are likely to correspond to commoditized goods where in the absence of differentiation of the final product, firms are least able to add value through innovation, customization, or branding—​making the services that serve this purpose least relevant. The contribution of value services to labor productivity reaches its maximum, holding other factors constant, where the share of value

Manufacturing, Services, and the Productivity Gap    351 services is slightly below 20%, corresponding to a high reliance on sophistication and differentiation. Considering the importance of cost services to manufacturing and the increasing share of some of these services in intermediate consumption, their lack of significance in the regression results may perhaps be surprising. However, a likely explanation is that the consumption of such services is a prerequisite in many manufacturing sectors and thus is relatively inelastic to prices. In that case, an inflated share of cost services inputs indicates high service prices—​which may reduce firms’ margins and therefore be detrimental to productivity measured with a value added criterion—​rather than the consumption of a large quantity of services. Overall, in the realm of cost services, the priority is to bring down the price of such services, including by upgrading the infrastructure to reduce delays and waste, and by fostering higher competition across providers of core services. More generally, more widely available and better quality cost services contribute to leveling the playing field but do not necessarily turn into higher relative productivity. In the days of global value chains and regional production networks, efficient transportation and finance are necessary conditions for a firm to compete, but they may not suffice to bring industrial firms’ performance to the next level. Innovation, technology, design, branding, marketing, and other differentiating services are those most likely to create value and to change firms’ prospects of success. Policy reforms seeking to enhance manufacturing performance will have to address the importance of both groups of services. On the one hand, financial services, transport, or IT services are the most critical to improve the cost competitiveness of manufacturing industries. On the other hand, royalties, technical assistance, and technology-​related services have the highest potential to sustain future manufacturing growth and upgrading. The preceding results suggest that policies that help firms deploy more and better value services have the potential to improve productivity and contribute to competitiveness. These policies may include measures such as encouraging R&D investments, increasing product innovation, modernizing patent legislation and the patent office, improving the competitive environment in ICT infrastructure and services, improving service-​specific advanced skills, opening markets, and getting firms access to state-​of-​the-​art professional services.

16.4.  Why It Is Crucial for Brazil to Improve Service Sector Productivity? If Brazil wants to grow more rapidly, and in a more sustained manner while improving well-​being, it will have to improve the productivity and competitiveness of its services

352   Jorge Arbache sector. As seen in the preceding, there are at least four reasons to support this argument. First, services are by far the largest economic activity, and they are likely to keep growing. Second, services already play a key role in the consumption basket, but their share is likely to increase even further. Third, services are a key input for manufacturing, and their contributions will probably keep rising. Fourth, services already are influential, but will become even more so in income and job creation in the twenty-​first century. Therefore, improving services sector competitiveness should be one of Brazil’s main economic policy objectives, and initiatives aimed at bridging the productivity and innovation gap should feature prominently in the policy agenda. This section argues that services reforms are an essential stepping-​stone toward enhanced productivity and stronger economic performance over the coming decades. Output of advanced economies is increasingly characterized by a symbiotic and synergistic relationship between manufacturing and services to create value. Empirical evidence shows that manufacturing value added increases when combined with services to form a third product that is neither in itself an industrial good nor a conventional service (Arbache 2012, 2016a). These are goods with a high content of services, such as smartphones, and products sold in packages, such as mainframe computers or aircraft jet engines. The changing relationship between manufacturing and services is one of the main elements of the so-​called fourth industrial revolution. The decline of the manufacturing sector’s share within GDP, as has occurred in advanced countries, does not necessarily mean that industry becomes irrelevant. In fact, increasing industrial density features a more sophisticated and influential stage of manufacturing, which is marked by the changing nature of goods and the way they are produced.18 The manufacturing sector acts as a catalyst for R&D and other advanced services (Savona 2016), as well as creating wealth and good jobs in a more complex way. The relationship between manufacturing and services in a country like Brazil tends to differ from that found in advanced economies as a result of structural constraints and market failures, including the following. First, there is limited availability of services that cut costs and add value to products. Second, the limited access to technology, credit, and markets keeps productivity low. Third, production and managerial technologies originally developed for industrialized countries have been adopted, but the supply of quality services required for them to operate properly is not often found there. Fourth, Baumol’s cost disease is usually more acute there than in developed countries due to the lower availability of human capital and low labor productivity in the service sector.19 If the development and modernization of value services are indeed linked to industrial development, and if the competitiveness of manufactured goods and services benefits from the synergistic and symbiotic relationship between them, then it is reasonable to assume that Brazil will face great challenges in upgrading to global value chains and performing competitively with higher value goods. The large and widening gap in productivity between developed and developing countries, combined with the lack of conditions required for progress in industrial development, suggests that the challenge for latecomer countries like Brazil wishing to engage in more value added goods will possibly increase.

Manufacturing, Services, and the Productivity Gap    353 Because of the growing importance of technology in determining income distribution and the segregation of countries involved in services related to cost and value added, one may expect the income gap between those countries to increase over the coming years. Therefore, Brazil needs strategies to strengthen the provision of services, increase industrial density, and upgrade to global value chains perhaps as much or even more than it needs conventional productive factors such as infrastructure, labor, tax breaks, or subsidies. The rise of the digital economy, another facet of the services revolution, is creating opportunities for brand new business models. While the global trade in goods and financial flows appear to have peaked in terms of share of GDP, data flows are growing almost exponentially. According to McKinsey (2016), between 2005 and 2014 the global data flow grew 45 times. Expansion of connectivity infrastructure, network effect, platform effect, falling computing and sensor costs, open software architectures, and deregulation of digital markets are accelerating the adoption and use of digital technologies and enabling the emergence of a whole new generation of investments and business models. Only in 2014, data flows added $2.2 trillion to global GDP directly and another $2.8 trillion indirectly. Plurilateral trade, investment, and intellectual property rights agreements are another important aspect of the growth-​services related prospects because they will likely redefine the parameters that guide investment location. Agreements such as the proposed Transpacific Partnership (TPP) and Trade in Services Agreement (TISA) will indeed change substantially the rules of the trade and investment game—​deregulation and market access, nondiscrimination to foreign companies, regulatory convergence, opening of services markets, e-​commerce, cloud computing, data storage markets, and government procurement are among the major changes. All this will change the business environment and influence the competitiveness of countries, reducing the relevance of conventional factors associated with geography and cost. In fact, trade agreements are already favoring the formation of production arrangements at the regional level, as we see happening in Asia, North America, and in the European Union. All these factors are already obliging large global companies to reassess the merits of engaging in long and complex intercontinental networks of production and in outsourcing from developing countries, while other factors are gaining importance in their decisions (Arbache 2014b, 2016b). This new dynamic of capitalism brings additional challenges for Brazil and the need to devise strategies to escape the income trap. The increasing pressure from developed countries to liberalize services will likely inflate the previously mentioned asymmetries, which, in turn, will likely have implications for Brazil’s development prospects.

16.5. Conclusions With a huge participation, by emerging market standards, in employment and output, the services sector virtually dictates the contours of the Brazilian economy and the labor

354   Jorge Arbache market alike, and whatever happens in this sector is likely to spill over into the rest of the economy. This chapter has examined the contribution and impacts of the services productivity gap in Brazil in the preceding decades. It shows that the rapid increase in the scale of the services sector, combined with its modest performance, can be identified as one of the sources of slow productivity growth, which in turn helps explain the increasing income gap with other economies. Services are also an important factor underlying other sectors’ competitiveness, and manufacturing in particular. As long as more labor is reallocated to services, aggregate productivity will likely remain stagnant or increase slowly. To counter this structural effect, Brazil will have to improve productivity within the service sector, which will help to increase aggregate productivity directly and indirectly through spillover effects. Therefore, a major challenge of policymakers is addressing the structural productivity deficit in the services sector. For that to happen, Brazil will have to undertake policies that create enabling conditions for the development of cost services, but also digital and value services, and for attracting global investments. Investments in first-​class business services, infrastructure, digital technologies, including digital competences and skills, and R&D, along with a bold regulatory and pro-​competition framework, legal and macro policy predictability, and foreign partnerships are among the areas that will encourage the development of those businesses and partnerships.

Notes 1. Data from The Conference Board, available at: https://​www.conference-​board.org/​. 2. Data from The Conference Board. 3. Data from CAGED, Ministry of Labor. 4. The share of workers engaged in informal activities in the services sector is even higher, due to the relatively low entry costs in activities such as street vending, cleaning, and similar jobs. 5. Data from IBGE. 6. Data from Timmer (2012). De Vries et al. (2012) examine structural transformation and its implications for productivity growth in the BRIC countries. They find that China, India, and Russia reallocated labor across sectors, and this contributed to aggregate productivity growth, whereas Brazil did not. For the period from 1980 to 1995, average annual productivity growth in Brazil was –​0.9%, mainly reflecting negative productivity growth rates in services. After 1995, productivity growth became positive in all sectors. However, within the services sector, labor moved to subsectors such as retail trade and the renting of machinery and equipment and other business activities, which have below average productivity levels. 7. In several sectors such as car repair, tourism and travel agencies, personal services, lodging, and food and catering services, the net wages/​value added per worker ratio was above 40%. For details, see Arbache (2015). 8. Data from the Annual Services Survey, from IBGE. For details, see Arbache (2015). Silva, Menezes Filho, and Komatsu (2016) use national accounts and document that wages in services grew more than productivity between 2000 and 2011.

Manufacturing, Services, and the Productivity Gap    355 9. For details, see Arbache, Rouzet, and Spinelli (2016). 10. Source: www.firjan.com.br. 11. Source: www.firjan.com.br. 12. For details, see the Doing Business Report (2016). 13. See, for exemple, Arbache (2012) and Bonelli and Pessoa (2010). 14. Trade in Value Added, by OECD/​WTO. 15. The PIA data set provides detailed information by sector on indicators such as intermediate consumption, revenues, and number of workers, which allow us to examine the effect of services inputs on labor productivity. 16. For details, see Arbache and Moreira (2015). The data used in this analysis come from PIA-​ IBGE, 1996–​2012. Labor productivity is calculated as value added per worker per industry at the two-​digit level. Service is calculated as the share of services in intermediate consumption per industry. 17. Lowess (locally weighted scatterplot smoothing) is a non-​parametric regression method that combines multiple regression models in a k-​nearest-​neighbor-​based meta-​model. 18. Industrial density refers to the manufacturing value added per capita in a country. For details, see Arbache (2012). 19. The Baumol cost disease refers to the phenomenon of the increase of wages of sectors that have not increased productivity as a reaction to the increase of wages of sectors that have increased productivity. The phenomenon occurs due to the competition for workers (Baumol 1967).

References Arbache, J. 2012. “Is Brazilian Manufacturing Losing Its Drive?” Department of Economics, University of Brasilia [online], http://​papers.ssrn.com/​sol3/​papers.cfm?abstract_​ id=2150684. Arbache, J. 2014a. Serviços e competitividade industrial no Brasil. Brasília:  Confederação Nacional da Indústria. Arbache, J. 2014b. “Convergência ou divergência de renda? Desafios do desenvolvimento no século XXI.” Conferências de Lisboa. http://​economiadeservicos.com/​wp-​content/​uploads/​ 2015/​04/​Arbache-​2014-​Confere%CC%82ncia-​Lisboa.pdf. Arbache, J. 2015. “Produtividade no setor de serviços.” In Produtividade no Brasil: Desempenho e Determinantes, edited by F. De Negri and L.R. Cavalcante, Vol. II, 277–​300. Brasília: Instituto de Pesquisa Econômica e Aplicada. Arbache, J. 2016a. “The Contribution of Services to Manufacturing Competitiveness in Brazil.” In Innovation and Internationalization in Latin America Services, edited by A. Hualde, R. Hernandez, N. Mulder and P. Sauvé. Santiago: ECLAC. Arbache, J. 2016b. “Digital Economy: Hopes for Brazil Overstretched.” The Brics Post, June 7, http://​thebricspost.com/​digital-​economy-​hopes-​for-​brazil-​overstretched/​#.V50BqrgrKUk. Arbache, J., A. Dickerson, and F. Green. 2004. “Trade Liberalization and Wages in Developing Countries.” Economic Journal 114 (493): 73–​96. Arbache, J., and R. Moreira. 2015. “How Can Services Improve Productivity? The Case of Brazil.” Paper presented at the 2015 REDLAS Conference, Montevideo. Arbache, J., D. Rouzet, and F. Spinelli. 2016. “The Role of Services for Economic Performance in Brazil.” OECD Trade Policy Papers, No. 193. Paris: OECD Publishing.

356   Jorge Arbache Baumol, W. J. 1967. “Macroeconomics of Unbalanced Growth: The Anatomy of Urban Crisis.” The American Economic Review 57 (3): 415–​426. Bonelli, R., and S. Pessoa. 2010. “Desindustrialização no Brasil:  Um resumo da evidência.” Texto para Discussão No. 7, IBRE–​FGV. Busso, M., L. Madrigal, and C. Pagés. 2013. “Productivity and Resource Misallocation in Latin America.” The BE Journal of Macroeconomics 13(1): 903–​932. CNS. 2014. “Principais propostas da CNS de políticas para economia e para o setor de serviços privados não financeiros.” Estudos Especiais CNS, Confederação Nacional de Serviços, Brasília. De Negri, F., and L. R. Cavalcante, eds. 2015. Produtividade no Brasil:  Desempenho e determinantes, Vol. I. Brasília: Instituto de Pesquisa Econômica e Aplicada. De Vries, G. J., A. A. Erumban, M. P. Timmer, I. Voskoboynikov, and H. X. Wu. 2012. “Deconstructing the BRICs:  Structural Transformation and Aggregate Productivity Growth.” Journal of Comparative Economics 40 (2): 211–​227. Dix-​Carneiro, R. 2014. “Trade Liberalization and Labor Market Dynamics.” Econometrica 82 (3): 825–​885. European Commission. 2014. “For a European Industrial Renaissance.” COM 14 Final, http://​ eur-​lex.europa.eu/​legal-​content/​EN/​TXT/​PDF/​?uri=CELEX:52014DC0014&from=EN. Ferreira, P. C., S. A. Pessoa, and F. A. Veloso. 2013. “On the Evolution of Total Factor Productivity in Latin America.” Economic Inquiry 51(1): 16–​30. Gonzaga, G., and R. C. Pinto. 2014. “Rotatividade do trabalho e incentivos da legislação trabalhista.” In Panorama do mercado de trabalho no Brasil, edited by R. Bonelli and F. Veloso, 181–​199. Rio de Janeiro: IBRE, Editora FGV. Luna, I., C. Hiratuka, P. S. Fracalanza, and M. R. S. Luz. 2015. “Decomposição da evolução da produtividade na indústria e nos serviços no período recente a partir de uma ótica evolucionária.” In Produtividade no Brasil:  Desempenho e determinantes, edited by F. De Negri and L. R. Cavalcante, Vol. 2, 495–​540. Brasília: Instituto de Pesquisa Econômica e Aplicada. McKinsey. 2016. “Digital Globalization:  The New Era of Global Flows.” Washington D.C.: McKinsey Global Institute. McMillan, M. S., D. Rodrik, and I. Verduzco-​Gallo. 2014. “Globalization, Structural Change and Productivity Growth, with an Update on Africa.” World Development 63: 11–​32. Muendler, M. A. 2004. “Trade, Technology and Productivity:  A Study of Brazilian Manufacturers 1986-​1998.” CESifo Working Paper No. 1148. OECD. 2014. Perspectives on Global Development 2014: Boosting Productivity to Meet the Middle Income Challenge. Paris: OECD Publishing. Restuccia, D. 2003. “The Latin American Development Problem: An Interpretation.” Economia 13 (2): 69–​100. Roberts, M. J., and J. R. Tybout. 1997. “Producer Turnover and Productivity Growth in Developing Countries.” The World Bank Research Observer 12 (1): 1–​18. Rodrik, D. 2013. “Unconditional Convergence in Manufacturing.” The Quarterly Journal of Economics 128 (1): 165–​204. Rodrik, D. 2015. “Premature Deindustrialization.” Economics Working Paper 107, Institute for Advanced Studies, School of Social Sciences. Savona, M. 2016. “Global Structural Change and Value Chains in Services:  A Reappraisal.” In Innovation and Internationalization in Latin America Services, edited by A. Hualde, R. Hernandez, N. Mulder, and P. Sauvé, 39–​63. Santiago: ECLAC.

Manufacturing, Services, and the Productivity Gap    357 Schor, A. 2004. “Heterogeneous Productivity Response to Tariff Reduction: Evidence from Brazilian Manufacturing Firms.” Journal of Development Economics 75 (2): 373–​396. Silva, F., N. Menezes Filho, and B. Komatsu. 2016. “Evolução da Produtividade no Brasil: Comparações Internacionais.” Insper Policy Paper No. 15, January. Timmer, M. P. 2012. “The World Input-​Output Database (WIOD):  Contents, Sources and Methods.” WIOD Working Paper No. 10. World Economic Forum. 2012. The Future of Manufacturing:  Opportunities to Drive Economic Growth. https://​www.nist.gov/​sites/​default/​files/​documents/​2017/​05/​09/​The-​ Future-​Manufacturing_​4_​20_​12.pdf

Chapter 17

Energy in  Bra z i l Past and Future José Goldemberg

17.1. Introduction Electricity and petroleum derivatives are essential ingredients for development. Brazil is well endowed with hydroelectric resources and began using these resources for electricity production shortly after the necessary technologies became available in the United States and Europe. By contrast, Brazil is not well endowed with petroleum reserves in its territory, and was dependent on imports until oil was found offshore in the Atlantic continental shelf. The introduction of electricity generation in Brazil was promoted in the nineteenth century by Emperor Dom Pedro II, an enlightened head of state who was well informed about world technological developments. In 1879 he granted Thomas Alva Edison, the inventor of the light bulb and founder of General Electric Company, a license to introduce lighting equipment in Brazil. Public illumination was installed in some quarters in Rio de Janeiro in 1891, only four years after Edison’s introduction of this technology in the United States. The first hydroelectric generation station was installed in the city of Diamantina, in the state of Minas Gerais, in 1883, only 17 years after electricity generators were invented in 1866 by Werner von Siemens in Germany. In rapid succession, many other hydroelectric stations were built in the Southeast region. In 1899, a Canadian company (São Paulo Railway, Light and Power Company) offered services in São Paulo and Rio, later becoming the LIGHT group, the main supplier of electricity in Brazil. In 1927, AMFORP (American and Foreign Power Company) began operation in São Paulo (Leite 2014). Brazil was an agrarian society until 1930, when policies were adopted to promote the industrialization of the country. Cheap and abundant energy was expected to induce industrial growth. These perspectives were based on Brazil’s abundant hydroelectric potential. In 1939 and 1940, the federal government established regulatory bodies for

Energy in Brazil: Past and Future    359 concessions and tariffs in the energy sector. After World War II, the great expansion of the cities of São Paulo and Rio de Janeiro increased demand for better electricity services (Leite 2014). Problems related to unrealistic tariffs set by the government led, in 1961, to the expropriation of foreign companies and the creation of Eletrobras, a state-​owned company at the federal level. The objective of this company was to take charge of the electricity system of the country. Similarly, some states established their own companies, such as CEMIG in Minas Gerais and CESP in São Paulo. Small privately owned companies were absorbed by these large state enterprises. This process ended in huge government participation in electricity generation. For almost three decades, Eletrobras and other state-​owned enterprises were very successful in extending the electricity grid to the whole country, but eventually exhibited signs of exhaustion, due to economic problems such as high inflation and reduced government capacity to invest in infrastructure. In 1996, Eletrobras was included in the federal plans for privatization, in line with the general policy of reduction of the role of the state in the economic activity of the country. Most companies in the business of distributing electricity were privatized, but generation and long-​distance transmission remained in the hands of the government. By the end of the twentieth century, electricity generation was almost exclusively hydroelectric. Thermal electricity generation using coal contributed only marginally to electricity production (Leite 2014). In 2004, under a new government elected in 2002, Eletrobras was excluded from the privatization plans and a new model for the system was adopted, reversing the privatization process and reinforcing the role of the state in the economy. The new model was not able to avoid another crisis in 2012–​2014, caused by an extended drought. Regarding petroleum production, the first wells were discovered in 1933 in the state of Bahia. All efforts to increase production had little success and as consumption increased, the country became almost completely dependent on imports. In 1954, Brazilian production accounted for only 1.7% of total consumption. This led to creation of a state-​ owned company, Petrobrás, in 1953, after a long nationalistic campaign, charged with the mission of finding petroleum. The first significant oil discoveries of Petrobrás occurred only in 1968, in the shallow continental waters of the state of Sergipe, in the Northeast, and production began to increase. In 2003, oil production reached 2 million barrels per day, and the country became practically self-​sufficient in oil and gas. The introduction of the automobile industry in 1956 led to a substantial increase of the consumption of gasoline and diesel oil. In the mid-​1970s, due to the dramatic increase in oil prices, a serious balance of payment problem was created, and the petroleum import bill threatened the balance payments of the country. In 1973, approximately one-​half the exports of the country were used to pay for imported petroleum derivatives. Further progress was achieved in 2007, with the discovery of large oil deposits at the depth of approximately five kilometers below the ocean floor (PRE-​SAL deposits), and the country became potentially an oil exporter. The enthusiasm raised by the PRE-​SAL discoveries led to changes in the regulatory system of oil exploration, which had disastrous consequences for Petrobrás.

360   José Goldemberg We will present in this chapter a description of the energy and electricity systems in Brazil, the shortcomings of this system, and the rationale for energy planning. We will discuss also the causes of the serious shortages of electricity supply in 2001 and 2012–​ 2014, and the problems faced by Petrobrás after 2008, their causes, and the policies required to solve them.

17.2.  The Brazilian Energy Matrix The main characteristic of Brazil’s energy matrix is the high contribution of renewable sources of energy and consequently a lower contribution of fossil fuels (coal, oil, and gas). As displayed in Figure 17.1, renewable energy sources represented 39.4% of the domestic energy supply in 2014, in sharp contrast to the contribution of these sources of energy in most industrialized countries, where it is usually lower than 10%. Hydroelectricity contributed 11.5% to the total energy consumption, sugarcane products 15.4%, other biomass products (firewood and charcoal) 8.1%, and other renewables 4.1%. Firewood and charcoal 8.1%

Petroleum and oil products 39.4%

Sugarcane products 15.7%

Hydroelectricity 11.5%

Natural gas 13.5%

Coal and coke 5.7%

Nuclear 1.3%

Other renewables 4.1% Other nonrenewables 1.0%

Non-renewables (60.6%) Renewables (39.4%)

Figure 17.1.  Brazil’s domestic energy supply (2014), shares of 305.6 (x106 TOE). Source: EPE (2015).

Energy in Brazil: Past and Future    361 100% 90% 80% 70% 60% 50% 40% 30% 20% 10%

1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

0%

Other

Oil Products

Ethanol

Elecricity

Bagasse

Firewood

Figure 17.2.  The evolution of the domestic energy supply in Brazil (1970–​2014). Source: EPE (2015).

As displayed in Figure 17.2, the composition of the energy matrix has not changed significantly since 1990. Before 1990, the contribution of firewood and charcoal were significant but declined significantly in more recent years. As for electricity, renewable energy sources represented 74.6% of the domestic supply in 2014. Generation is predominantly hydroelectric (65.2%), as shown in Figure 17.3; biomass contributed 7.4%; wind power 2.0%; with a total of 74.6% of renewables. The remaining 25.4% came from natural gas (13.0%), nuclear (2.5%), coal (2.9%), and oil derivatives (14.8%). Electricity generation both in government-​owned plants and self-​ producer power plans reached 590.5 TWh (terawatt hours) in 2014. The government-​ owned plants remained the main contributors, with 84.1% of total generation. The production of electricity from wind power reached 12,210 GWh (gigawatt hours). Considering oil, domestic production reached an average of 2.25  million barrels per day in 2014, of which 93% are offshore. The average daily production of petroleum was 87.4  million cubic meters per day, and the volume of imported natural gas was 52.9 million cubic meters per day, on average. The amount of biodiesel produced in the country reached 3,419,000 cubic meters in 2014.

362   José Goldemberg Hydraulic 65%

Other fossil fuel 2%

Biomass 7% Wind 2% Gas 13%

Nuclear 3%

Coal 3%

Oil derivatives 5%

Renewables 74.6% Non-renewables 35.4%

Figure 17.3.  Domestic electricity supply (2014), shares of 624.3 TWh. Source: EPE (2015).

One of the most interesting aspects of the energy matrix in Brazil is the contribution of ethanol produced from sugarcane and used to replace gasoline. In 2014, sugarcane production reached 631.8  million tons and ethanol production, 28,526,000 cubic meters. The origin of the ethanol program was the oil crisis of 1973, which seriously affected the balance of payments of the country. At that time, the cost of oil imports in hard currency represented approximately half of all exports (roughly 4 billion dollars at historical value, equivalent to US$12 billion in 2005). The increase in petroleum prices therefore exerted considerable strain on the Brazilian economy at that time. To remedy the situation, the Brazilian government embarked on an ambitious program to produce large quantities of ethanol from sugarcane (PROALCOHOL) as a substitute for gasoline (Moreira and Goldemberg 1999). Brazilian conditions are favorable for the production of ethanol. Sugarcane has been an important crop since the eighteenth century and Brazil was the world’s third-​largest sugar producer (5 million tons of raw sugar equivalent) in 1975. During the 1970s oil crisis, sugar was experiencing a long period of low prices in the international market,

Energy in Brazil: Past and Future    363 so the decision to divert some of the sugarcane to ethanol production was reasonable, considering also that the technology needed had been available for decades (Coelho et al. 2012). PROALCOHOL was launched by the government to include two variants: compulsory use of 10% anhydrous ethanol as an additive to gasoline, which did not require modifying engines, and voluntary use of 100% hydrated ethanol (95% ethanol + 5% water) in modified Otto cycle motors. The cost of production, however, was high (approximately three times the price of gasoline), and a number of government actions were needed to make its production attractive. In addition to low-​interest loans for the construction of refineries, guaranteed purchase by the state-​ owned petroleum company (Petrobrás) was introduced. Furthermore, a fixed price of ethanol was adopted at a level such that 44 liters of ethanol corresponded in value to 60 kilograms of sugar. To sugar producers, such conditions made it attractive to produce ethanol. The volume produced increased, and the ethanol was mixed with gasoline up to a share of 20%, which required only minor changes in Otto cycle motors built to operate with gasoline. This phase of the program extended from 1975 to 1980. From 1980 to 1985, another phase was implemented, based on the use of hydrated ethanol (95.5% pure ethanol and 4.5% water) in motors designed to operate with this fuel (Moreira and Goldemberg 1999). The multinational automobile industries based in Brazil introduced all the necessary engine and vehicle modifications for ethanol use. Gasohol vehicles running with up to 10% ethanol (in volume basis) require almost no changes, but more modifications are required for a larger share of ethanol in the fuel blend. With minor adaptations developed by the car manufacturers, all gasoline vehicles can run with blends ranging between 20% and 26% of ethanol. One of those changes was the change in the compression ratio, namely 12:1 for ethanol fuel from sugarcane, while regular gasoline requires a compression ratio of 8:1. A higher compression ratio signifies higher efficiency, which partly compensated for the lower energy content of ethanol. Energy equivalence between the two fuels takes into account the final energy service provided. In this way, 199 liters of pure (anhydrous) ethanol can replace one barrel of gasoline (159 liters). The necessary changes in the engines to run ethanol meant a drastic change in automobile manufacturing, but under government pressure, domestic automobile manufacturers adjusted to that. After 20  years, these manufacturers developed a flexible fuel technology, commercially available and predominant in car sales today at no additional cost (Coelho and Guardabassi 2014). Sugar producers welcomed these changes, which allowed them to divert more sugarcane to ethanol production, and to face oscillations in sugar prices on the international market in a better way. Another source of enthusiastic support came from nationalistic voices in government, who saw ethanol as an instrument of national independence. One of the consequences was that driving Brazilian cars with engines designed to run ethanol in neighboring countries (even in some Brazilian states) was a problem, because they did not have filling stations selling hydrated ethanol. The production of cars with pure ethanol engines began in August 1979, with the participation

364   José Goldemberg of manufacturers such as General Motors (Opel), Ford, Volkswagen, and Fiat. During the early 1980s, the success of ethanol vehicles was such that they accounted for 85% of all new car sales. In this period, two types of automotive vehicles were in use in the country: some running on gasoline, using a blend of up to 20% anhydrous ethanol and 80% gasoline, and others running entirely on hydrated ethanol. In 1985, the scenario changed dramatically, as petroleum prices fell and sugar prices recovered on the international market. Subsidies were reduced, and ethanol production could not keep up with demand. Therefore, by 1990, sales of cars running on pure ethanol dropped to 11.4% of the total. The production of ethanol leveled off, but the total amount being used remained more or less the same, because the decrease of hydrous ethanol cars was compensated by the increase in gasoline blending to 27%. Since March 2003, the Brazilian automotive industry has introduced the so-​called flex fuel vehicle, which means a vehicle running with any blend, from the Brazilian specification of gasohol (E18 to E27, using anhydrous ethanol) up to pure hydrous ethanol (E-​100, using hydrous ethanol). From 2003 onward, ethanol consumption started to rise again. As production of ethanol increased, costs declined, following a typical “learning curve” (Figure 17.4). Since 2003, ethanol became fully competitive with gasoline without any subsidies. It is important to emphasize that this achievement is the result of over 30 years of experience with technological gains, economies of scale, and political effort from Learning curve 1976 100

1981

Fuel price (US$/Gl)

1980

1986

1991

1996

2001

1982 1991 1992 1994 1987 1990

1979

2011

2016

2009 2007

1984 1986

1977

2006

2001

1995

2005

2010

1997 2000 2003 1999

1 0

50

100

150

200

250

300

350

400

450

Millions of Cubic Meters Produced Anhydrous ethanol

Gasoline

Figure 17.4.  Ethanol learning curve. Source: Goldemberg et al. (2004), updated by the author.

Energy in Brazil: Past and Future    365 policymakers and other stakeholders. Although since 2003 ethanol became fully competitive with gasoline without any subsidies, it is important to stress that in the period 1975–​2003, subsidies of approximately US$30 million were injected into the program (Goldemberg et al. 2004). Another interesting feature of the use of ethanol from sugarcane to replace gasoline is that the energy balance of its production is highly positive. The energy balance can be defined as the ratio of total fossil fuel energy required for the biofuel production process to the energy contained in the biofuel produced. Energy balance of biofuel produced from a feedstock might be an indicator of the sustainability when taking into account greenhouse gas emissions avoided in a life-​cycle basis. Figures from a study conducted by Macedo (2005) on sugarcane ethanol energy balance show significant differences between feedstock, where the energy balance of the ethanol produced from sugarcane is 8.9, from sugar beet, 2.0, and from corn, 1.3. The sugarcane ethanol presents lower emissions compared to fossil fuels due to the carbon cycle. In respect to land use, feedstocks with higher yields per hectare require less land. In other words, ethanol from sugarcane can be considered to a large extent a renewable source of energy. Hence, ethanol from sugarcane has a much higher energy balance due to both high agricultural and industrial productivities. The good performance of the ethanol produced from sugarcane is due to the use of sugarcane bagasse that provides the energy needs for the process, eliminating other external sources of energy. Bagasse is the most important byproduct from sugarcane and is burned in boilers producing the thermal, mechanical, and electrical energy required in the industrial process. In the case of corn ethanol, for instance, fossil fuels are demanded to produce heat and power for the process. As for other minor sources, coal and coke represented 5.7% of the domestic energy supply in 2014, totaling 17.6 million tons of oil equivalent. Half of this energy was used for power production and half for metallurgy (EPE 2015). Brazil has two nuclear plants, Angra I, with 657 MW, and Angra II, with 1,350 MW. The first is a “turn-​key” project from Westinghouse; the second resulted from a Brazil-​Germany agreement signed in 1975, under which eight large nuclear reactors were to be installed in Brazil before 1990. There is a third nuclear plant, part of the agreement with Germany, which has been under construction since 1980.

17.3.  Problems with Brazil’s Energy Matrix We discuss here in succession the problems in the electricity and petroleum sectors. Brazil has a hydroelectric generating potential of 251.5 million kilowatts of electricity, of which 77.8 are already being used. The installed capacity has grown by about 4 and 5 million kilowatts a year. From the technical point of view, it is possible to double the installed capacity, which would allow the maintenance of the expansion of hydropower

366   José Goldemberg production at a satisfactory level for the next 20 years. The expansion should occur in the North region, mainly in the Amazon region, since most of such resources in the South and Southeast regions of the country are already being used.In 2001, there were serious electricity shortages attributed to a prolonged drought and to poor energy planning. The origin of this problem is that since 1990, hydropower plants have been built with small reservoirs, that is, mainly on “run-​of-​the-​river” projects to minimize the areas flooded, as shown in Figure 17.5. The darker line shows the evolution of the installed power in GW (left scale) and the lighter line the evolution of the volume of the reservoir (right scale). Until 1985 the two lines grew at approximately the same rate, that is, increases in the installed electricity potential were followed closely by increases in reservoir capacity, which was a guarantee of supply in periods of drought. In the past the reservoirs were large enough to maintain generation capacity, even in dry periods as long as two years. The construction of hydropower reservoirs affects the local population that has to be displaced. There are also environmental problems deriving from deforestation and the sheer size of the engineering work required. On the other hand, every kilowatt installed in a hydropower plant generates approximately enough power to meet the needs of two families, who usually live thousands of kilometers away from the generation plant.

110.000

600 570 540 510 90.000 480 80.000 450 420 70.000 390 360 60.000 330 300 50.000 270 240 40.000 210 180 30.000 150 120 20.000 90 60 10.000 30 0 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025

Volume (x1,000 hm3)

Gigawatts

100.000

Year Installed Power (Gw) Volume (1,000 hm3)

Figure 17.5.  Evolution of hydroelectric generation and dam reservoirs. Source: EPE (2014).

Energy in Brazil: Past and Future    367 As an example, in the Belo Monte project, the construction of a large reservoir of 5 million hectares would allow the production of 10 million kilowatts (kW) of power, which could benefit about 20  million inhabitants. Facing strong opposition from environmentalists, the size of the reservoir was reduced to 0.5 million hectares, which is approximately equal to the area deforested every year in the Amazon by the expansion of the agricultural frontier. Such reduction converted Belo Monte basically into a “run-​ of-​the-​river” project, which will generate full capacity only in the wet season, reducing its average generating capacity to approximately 5  million kilowatts. Some 50,000 people were affected. There is a clear conflict here and a need to make tough decisions. Successive governments have been unable to solve such problems. The 2001 electricity shortages exposed the weakness of the energy supply system and contributed to a serious political crisis. In order to prevent the repetition of such problems, the new government, inaugurated in 2002, contemplated the idea of reversing the privatization model of 1996 and adopted a new model. This “new model” for the electric sector, adopted in 2004, defined new rules for the commercialization of electrical energy and created a new agency (Empresa de Pesquisa Energética—​EPE) to conduct technical, economic, and socioenvironmental planning of the activities in electrical energy, oil and gas, and renewable energy sources. The new agency was put in charge of the bidding process for new generating units and transmissions lines. It has adopted the lowest tariff as the criteria for selection of the winners in the bids for new projects. Long-​ term contracts for the sale of energy and environmental licensing as a precondition to enter in the bidding process were also introduced. The major problems of the new model were the auctions for building new power generation plants. The rule adopted in the auctions was to allow all forms of power (hydropower, thermal, wind, biomass) to compete in equal conditions with a maximum price (for kWh) established by the government. The justification for that procedure was to benefit the consumer, since competition could lead to a reduction in prices, ensuring “tariff modicity” (Goldemberg and Lucon 2007). The procedure used by EPE was technically questionable because it is not possible to produce power from different sources without considering the technological, locational, and operational differences. In practice, the result was that all alternatives to hydropower plants were practically excluded from the competition and run-​of-​the-​river projects made the system more vulnerable. As a consequence, in the 2012–​2014 prolonged period of drought, a large number of thermal generating stations, fueled mainly by fossil fuel (coal, natural gas, fuel oil, and others), were activated, as shown in Figure 17.6. The contribution of thermoelectric generation, which historically has been small (bellow 5GW average in an electric system with 60 GW average), increased to 12 GW average, to approximately 20%. Such an option led to a significant increase in electricity costs, which contributed to the economic problems of the country. The problems were aggravated by the government’s decision to reduce taxes in the acquisition of domestic equipment using electricity (refrigerators, air conditioners, etc.), which increased consumption. The rationale for such an action was the effort of the government to stimulate economic activity. As a measure to help control inflation, in

368   José Goldemberg 14,000

12,000

100TWh

10,000

75

8,000

50

6,000

25

4,000

2,000

Jan. 2010 Mar. 2010 May 2010 Jul. 2010 Sep. 2010 Nov. 2010 Jan. 2011 Mar. 2011 May. 2011 Jul. 2011 Sep. 2011 Nov. 2011 Jan. 2012 Mar. 2012 May. 2012 Jul. 2012 Sep. 2012 Nov. 2012 Jan. 2013 Mar. 2013 May. 2013 Jul. 2013 Sep. 2013 Nov. 2013 Jan. 2014 Mar. 2014 May. 2014 Jul. 2014 Sep. 2014 Nov. 2014 Jan. 2015 Mar. 2015 May. 2015

0

Figure 17.6.  Thermoelectric generation in Brazil (2010–​2015). Source: MME (2015).

2012 the government passed a law changing the procedure for granting concessions for the generation, transmission, and distribution of electricity. The main novelty was the decision to renew concessions (usually granted for 20–​30 years) only once. Regarding generation, 20 concession contracts (mainly in large hydroelectric plants) were due to expire in 2015–​2017. The government offered to renew these concessions immediately if the concessionaires agreed to lower their tariffs by 20%. These measures completely disorganized the system, since the state-​owned electricity companies of São Paulo, Minas Gerais, and Paraná refused to anticipate the end of their concessions. Regarding the problems of petroleum, the great success achieved by Petrobrás in reaching self-​sufficiency in production in the shallow water of the Campos basin did not last because of the progressive exhaustion of the wells in this region. CENPES, the research and development laboratories of Petrobrás, successfully led the effort to identify large deposits of oil at great depth in the ocean (PRE-​SAL deposits), raising hopes of turning the country into a large producer of oil. The concession scheme introduced in the 1990s was changed in 2009, increasing the role of the federal government in this area. The main changes in the regulatory system included changing the system of distribution of royalties to producing states and municipalities, which led to serious political disputes among states and other areas of government, even before the production in

Energy in Brazil: Past and Future    369 the PRE-​SAL began. It also included a mandatory 30% share of Petrobrás in all projects, which would lead the exploratory work and assume the main costs and risks of the projects, although prudence recommended that Petrobrás should try to reduce its costs and share the risks with other oil corporations with experience in this area. Petrobrás remained practically alone in the PRE-​SAL exploration, incurring heavy debts, which caused its shares to lose about 80% of their value. Further aggravating the problems of Petrobrás, in 2008 the federal government froze the prices of diesel and gasoline, and this situation remained unchanged for six years, despite wide variations in the price of petroleum in the international market, with the cost of a barrel surpassing US$150. The justification for such a policy was the need to curb inflation, which led Petrobrás to sell oil byproducts at lower prices than those imported. The result was a growing indebtedness and a sharp fall in the value of its shares. Moreover, Petrobrás has bought large volumes of liquefied natural gas in the “spot” market at high prices to operate thermo-​power plants to supplement the generation of the hydropower plants. One item of collateral damage of the price policy adopted for gasoline was the asphyxiation of sugarcane ethanol production, since the product selling price of ethanol is indexed to that of gasoline.

17.4.  Planning for the Expansion of Energy Production in Brazil Energy planning in Brazil has always been based on the “growth hypothesis,” that is, that the country’s GDP growth leads to an increase in energy production (and consumption). This is the argument used by governments to justify heavy investments in energy supply as the only way to raise GDP per capita and the standard of living of the population. Energy consumption per capita in Brazil in 2014 was 1.52 TOE (ton of oil equivalent), close to the world average, but significantly lower than in industrialized countries, where it is around 4 TOE per capita (IEA 2016). Such a vision of the problem usually originated in figures such as those shown in Figure 17.7, in which total primary energy supply (TPES) per capita is plotted against GDP (PPP) per capita. There are a number of reasons for the large scattering of points in such graphs. Climate clearly is an important factor: winters are very mild in Brazil in contrast with developed countries, where usually a third of the total energy consumption is due to residential and commercial heating. The structure of production, the proportion of the energy services sector in the economy, the modernization of the industrial park, the importance of agriculture, and other factors are also significant. Despite that, the data in Figure 17.7 suggests a roughly linear correlation between GDP and energy consumption and is frequently used by policy planners. Table 17.1 presents data from some countries with similar per capita TPES. Their energy consumption per capita varies from a minimum of 2.840 ktoe (thousand tonnes of

7.0 6.0

kgoe/capita

370   José Goldemberg

5.0 OCDE

4.0 3.0 World 2.0 Brazil

1.0 0.0 0.0

10.0

20.0

30.0

40.0

50.0

GDP/Capita (USD 1.000) 60.0 70.0

Figure 17.7.  TPES per capita versus GDP per capita. Source: Author’s elaboration based on World Bank data, 2015.

Table 17.1 Income per capita and oil consumption by country Country

(US$) GDP/​Capita

ktoe/​Capita

Iran

17,303

2.960

Belarus

17,661

2.882

Malaysia

26,891

3.020

Israel

35,431

2.971

United Kingdom

41,325

2.978

Denmark

46,635

3.107

Ireland

54,654

2.840

Source: World Bank Data, 2015.

oil equivalent) per capita (Ireland) to 3.107 ktoe per capita (Denmark), a change of approximately 10%. In contrast, GDP per capita varies between US$17,303 to US$54,654, a factor of approximately 3 (World Bank Data 2015). Another way to look at the problem is to compare countries with similar levels of per capita GDP and observe how energy consumption varies among them. Table 17.2 shows numbers for six countries with per capita income around US$13,000: South Africa, China, Serbia, Peru, Colombia, and Mongolia. Their energy consumption, however, varies between 2.656 ktoe per capita (South Africa) and 488 ktoe per capita (Mongolia).

Energy in Brazil: Past and Future    371 Table 17.2 Income per capita and oil consumption for countries with GDP per capita around $13,000 GDP (US$)

ktoe/​Capita

1

South Africa

13,165

2.656

2

China

14,239

2.216

3

Serbia

13,482

2.078

4

Peru

12,402

0.708

5

Colombia

13,801

0.669

6

Mongolia

12,189

0.488

Source: World Bank Data, 2015.

Thus, although it seems intuitive that growth in energy supply leads to an increase in GDP, the causality between these two variables has been investigated extensively, with conflicting results. Chen et al. (2012), Bruns et al. (2014), Pau and Fu (2014), and Menegaki and Tugan (2016) have applied the Granger causality test for that purpose. This test has been used extensively to better understand the relationship between energy and GDP growth.1 Chen et al. (2012) conducted a study of 174 sets of data from 39 studies and concluded that one is more likely to find that GDP growth causes energy growth in developing countries, and that energy growth causes GDP growth in industrialized countries. GDP growth seems to be the driver in large countries. Bruns et  al. (2014) analyzed a large sample of 574 pairs of causality tests from 72 studies selected from more than 500 papers in the literature. They find that the causality relations found are either fragile or ambiguous. The exceptions are the cases in which energy prices are controlled, where they found a genuine effect in which energy growth is the cause of GDP growth. This seems to be the case of Brazil, where energy producers are mainly state-​owned and tariffs are set by the government. There are cases in which one finds “bidirectional causality,” running from energy consumption to economic growth, and then the reverse, economic growth to energy growth over the years (Pau and Fu 2014). Interestingly, in many developing countries a “neutrality hypothesis” can be identified, that is, no causality exists between energy consumption and economic growth, leading to the conclusion that growth (in energy and GDP) is driven by other factors (Menegaki and Tugan 2016). Table 17.3 shows the main characteristics of the domestic energy consumption in Brazil in 1970 and 2014:  population, energy consumption, energy consumption per capita, electricity consumption, and electricity consumption per capita, as well as the growth rate of these indicators. The values projected for 2024 by EPE (Empresa de Pesquisa Energética) are included (MME 2015). The energy growth rate in the period 1970–​2014 was 3.2% per year, corresponding to a growth of GDP of 3.2% per year in the same period. Electricity, however, grew at 5.0% per year in the period. This indicates an increasing electrification rate in the Brazilian economy.

372   José Goldemberg Table 17.3 Domestic Energy Consumption (1970–​2014)

1970

2014

Growth Rate 1970–​2014, %/​year

95.7

205.3

1.7

Energy (x 10 TOE)

59.0

249.7

3.3

Energy capita (TOE per capita))

0.62

1.22

1.6

Electricity (GWh)

37.2

525.3

6.2

Electricity per capita (kWh/​capita)

389

2559

4.4

Population (x 106) 6

Source: MME, 2015.

Table 17.4 The 2030 Energy Plan Scenarios Scenarios

GDP Growth/​Year (2005–​2030)

High growth

A1

5.1%

Intermediate growth

B1

4.1%

Intermediate growth

B2

3.2%

Low growth

C

2.2%

Source: MME, 2007.

The Ministry of Mines and Energy regularly publishes an analysis of the evolution of energy and electricity consumption using the technique of scenarios. For each scenario, assumptions are made for the evolution of the different sectors of the economy (agriculture, industry, and services). For each sector, assumptions are made regarding the use of energy sources (gasoline or ethanol for automobiles, hydroelectric or thermal electricity generators, and so on). They estimate the total amount of energy and electricity needed to achieve the GDP envisaged in each scenario. In the National Energy Plan 2030 (MME 2007), four scenarios were constructed, with GDP growth ranging from 2.2% per year (low growth) to 5.1% per year (high growth), as shown in Table 17.4. MME (2007) proposed four scenarios on the evolution of energy (and electricity) consumption up to the year 2030 (Figure 17.8). Data from 1970 through 2000 are the observed levels of consumption. Total consumption of energy varies from 293.9 to 475.4 million TOE. Depending on the performance of the economy, energy consumption could be 1.6 times the lower level. The same goes for electricity, but in this case the maximum is only 1.46 times the minimum.

Energy in Brazil: Past and Future    373 Domestic energy consumption (1970–2030) 500

475.4

SCENARIO A SCENARIO B1

400

404.7 351.4

Million tep

SCENARIO B2

293.9

300 SCENARIO C

200

100 59.0 0 1970

1980

1990

2000 Year

2010

2020

2030

Domestic energy consumption 1250

1,243.8

SCENARIO A SCENARIO B1

1000

1,045.6 941.2

SCENARIO B2

847.0

Twh

750

500

SCENARIO C

250 37.2 0 1970

1980

1990

2000 Year

2010

2020

2030

Figure 17.8.  Domestic energy consumption (1970–​2030). Source: MME 2007

17.5.  The Need for New Policies Frequent alterations in the regulatory framework create an atmosphere of legal insecurity in the energy sector. This has led to a reduction in investments, mainly by foreign companies, which otherwise could invest in the expansion of power production

374   José Goldemberg in the country. In mid-​2016, immediately following the change in government, a number of policy changes in the energy sector were introduced, to correct for some of the problems pointed out earlier in this chapter. As part of these changes, Petrobrás was authorized to associate with foreign companies in oil exploration in the PRE-​SAL area, concessions were granted to foreign companies, and a large program of investment in activities such as gas distribution was devised. Eletrobras has begun a significant program of privatization. Electricity tariffs were realigned to realistic levels and the process of recovery of the sector began, particularly the use of auctions for new investments. The correct procedure would be to conduct regional auctions and to establish different maximum prices for each source (wind, biomass, gas, etc.). These maximum prices should gradually fall in the auctions conducted every year, stimulating competition and making room for renewable energies. New, cleaner sources could be subsidized in their early maturing stages. New actions are needed in other areas, though. A  largely ignored element in the evolution of the Brazilian energy matrix is the importance of energy efficiency, which is one of the major strategies adopted around the world. For example, the United States from time to time establishes the average number of kilometers per liter of fuel that automobiles have to achieve. It was set at 10.6 kilometers per liter in 1975, should reach 16.6 kilometers per liter in 2016 and 23 kilometers per liter in 2025. If the fleet does not increase much, the consumption of oil products should decrease. Another strategy is the production of biofuels (ethanol from sugarcane in Brazil, and corn in the United States). Biofuels currently replace 3% of the oil consumed in the world, but this percentage could easily reach 10%. There is here a great opportunity for Brazil to export its technology for producing sugarcane and ethanol, which has already reached a high productivity level. Until 2008, ethanol production increased by about 8% a year for several years and replaced about 50% of the gasoline that would be used if ethanol production did not exist. The federal government, however, set the selling prices for diesel and gasoline in 2008, and kept them frozen for over five years. One of the positive consequences of the expansion of ethanol production is the use of sugarcane bagasse for power generation. About 15,000 MW of power could be produced in the harvesting periods (from April to October), which corresponds to the dry period of the year, thus complementing power generation from hydroelectricity as indicated. The expansion of the electricity system has to be oriented firmly to increase the participation of renewable energy sources (mainly hydroelectric), although it seems clear that some thermoelectric participation will be necessary. The electricity-​ generating potential is expected to grow from 132.9 GW in 2014 to 206.4 GW (MME 2015). Table 17.5 shows a breakdown of the contribution of different sources to such an increase. If such an evolution materializes, Brazil will be able to preserve by 2024 the very significant fraction of renewables in its energy matrix that it displayed in 2014.

Energy in Brazil: Past and Future    375 Table 17.5 Additional Electricity Potential (2014–​2024) (GW) Source

GW

Hydroelectric

27.2

Thermoelectric (nonrenewable)

11.4

Wind

18.9

Biomass

6.8

Small hydropower plants

3.2

Solar

6.0

Source: MME, 2015.

17.6. Conclusions Production of energy in Brazil, particularly electricity, includes a high percentage of renewable sources. Electricity is generated mostly by hydroelectric plants, cars are fueled with ethanol produced from sugarcane, biofuels have good prospects for success, and wind generation of electricity is picking up slowly. Although dependent on foreign production until recently, the country has managed to become almost self-​sufficient in oil production. However, most of the production of energy is in the hands of government institutions, imposing a degree of instability on the system. This chapter has reviewed the effects of some important policies introduced at different points in time. The new policies adopted by the Brazilian government in 2016 are realistic and are steering the energy system in the right direction. The economic recession of 2014–​2017 with negative GDP growth reduced the pressures on the energy system and opened the way for a reorganization of the sector. One of the consequences of such policies is that Brazil will be able to remain a modest emitter of greenhouse gases and will thus have honored its commitments to the Climate Convention adopted at the Paris Conference in 2015.

Note 1. The Granger causality test is a statistical analysis for determining whether one time series or panel data is useful in forecasting another when the evolution of the variable X causes an evolution in variable Y, the patterns of X are approximately repeated in Y. The Granger causality test is based on two principles: (1) the cause happens prior to the effect, and (2) the cause has unique information on the future value of the effect (Granger 1969).

376   José Goldemberg

References Bruns, S. B., C. Gross, and D. I. Stern,. 2014. “Is There Really Granger Causality between Energy Use and Output?” Energy Journal 35 (4): 101–​134. Chen, P., S. Chen, and C. Chen. 2012. “Energy Consumption and Economic Growth:  New Evidence.” Energy Policy 44: 245–​255. Coelho, S. T., R. Gorren, P. Guardabassi, R. Grisoli, and J. Goldemberg. 2012. “Bioethanol from Sugar: The Brazilian Experience.” In Encyclopedia of Sustainability Science and Technology, edited by Robert A. Meyers, 86–​109. London: Springer. Coelho, S. T., and P. Guardabassi. 2014. “Brazil:  Ethanol.” In Sustainable Development of Biofuels in Latin America and the Caribbean, edited by B. D. Solomon and R. Bailis, 71–​101. New York: Springer Science + Business Media. EPE (Empresa de Pesquisa Energética). 2014. “Demanda de energia 2050: Séries estudos da demanda de energia.” Nota Técnica DEA 13/​14. EPE (Empresa de Pesquisa Energética). 2015. Brazilian Energy Balance 2015 Year 2014. Rio de Janeiro: EPE. Goldemberg, J., and O. Lucon. 2007. “Energia e meio ambiente.” Estudos Avançados 21 (59): 07–​20. Goldemberg, J., S. T. Coelho, O. S. Lucon, and P. M. Nastari. 2004. “Ethanol Learning Curve: The Brazilian Experience.” Biomass and Bioenergy 26 (3): 301–​304. Granger, C. W.  J. 1969. “Investigating Causal Relations by Econometric Models and Cron-​ Spectral Methods.” Econometrics 37 (3): 424–​438. IEA. 2016. World Energy Balances 2016. Paris: International Energy Agency. Leite, A. D. 2014. A energia do Brasil, 3rd edition. Rio de Janeiro: Lexicor. Macedo, I. C. 2005. Sugarcane’s Energy: Twelve Studies on Brazilian Sugar Cane Agribusiness and Its Sustainability. São Paulo: Berlendis & Vertecchia; UNICA (União da Agroindústria Canavieira do Estado de São Paulo). Menegaki, A. N., and C. J. Tugan. 2016. “The Sensitivity of Growth, Conservation, Feedback and Neutrality Hypothesis to Sustainability Accounting.” Energy for Sustainable Development 34: 77–​87. MME. 2007. Plano nacional de energia 2030. Ministério de Minas e Energia; colaboração Empresa de Pesquisa Energética. Brasília: MME; EPE. MME. 2015. Plano decenal de expansão de energia 2024. Ministério de Minas e Energia. Empresa de Pesquisa Energética. Brasília: MME/​EPE. Moreira, J. R., and J. Goldemberg. 1999. “The Alcohol Program.” Energy Policy 27(4): 229–​245. ONS (Operador Nacional do sistema Elétrico). 2013. RE 3/​0066/​2013—​Plano da Operação Energética 2013/​2017—​PEN 2013, Vol. I. Brasilia: Condições de Atendimento. Pao, H. Y.  Li, and H. Fu. 2014. “Causality Relationship between Energy Consumption and Economic Growth in Brazil.” Smart Grid and Renewable Energy 5: 198–​205. World Bank Data Energy Use (Kg of Oil Equivalent Per Capita). Available at http://​data. worldbank.org/​indicator/​EG.USE.PCAP.KG.OE. GDP per capita, PPP. http://​data.worldbank. org/​indicator/​NY.GDP.PCAP.PP.CD?end=2015&start=1990, accessed August 12, 2016.

Chapter 18

I nfrastru c t u re Edmund Amann, Werner Baer, Thomas Trebat, and Juan Villa Lora

As was made clear in the introductory chapter, Brazil’s growth performance over the past quarter of a century has, by comparison with the 1945–​1980 period, proved disappointing.1 Set against the contemporary growth record of its emerging market counterparts in East Asia, Brazil’s growth appears even less impressive. The roots of Brazil’s “growth problem” run deep; they can be traced to a variety of structural constraints and, from time to time, unfavorable international circumstances. The structural constraints—​many of which are explored in this volume—​include low investment in education, the high incidence of corruption, overreliance on commodities exports, and an inefficient tax system. The focus of attention in this chapter, however, centers on one of the most high-​ profile and widely recognized growth bottlenecks: infrastructure. By common consent (see Giambiagi et  al. 2016), enhanced spending in critical areas such as highway networks, power transmission, ports, airports, and sanitation is considered an indispensable component of any strategy designed to boost Brazil’s long-​run growth potential. Yet, for a number of reasons to be explored in this chapter, the necessary level of investments has not been forthcoming. Whereas infrastructure investment in Brazil averaged 5.2% of gross domestic product GDP in the early 1980s, it fell to an average of just 2.25% of GDP over the next 20 years, reaching 2.5% in 2013 (Garcia-​Escribano et al. 2015, 11). This was despite the introduction of an ambitious infrastructure plan, the Growth Acceleration Plan (PAC), in 2007. As a result of this, there has been a failure of infrastructural provision to keep pace with demand. At the same time, the productive sector faces serious physical and logistical constraints in its quest to boost and diversify exports and to respond swiftly to emerging market opportunities, be they domestic or global. In other words, shortcomings surrounding infrastructural provision have real impacts on supply-​side responsiveness and, by extension, the growth potential of the economy.

378    Edmund Amann, Werner Baer, Thomas Trebat, and Juan Villa Lora That underinvestment in infrastructure in Brazil has become a widely recognized problem seems at first curious. As indicated in the introductory chapter, and those focusing on import substitution and structuralism (Chapters 4 and 5), the entire basis of Brazilian developmentalism centers on the idea of effecting structural change through strategic, capital-​intensive investments. This process involves close articulation between the state and the private sector. It proved a model that survived, in modified form, even during the more free market–​inclined administrations of the 1990s. With the accession to power of PT (Partido dos Trabalhadores  –​Workers Party) administrations between 2002 and 2016, Brazilian developmentalism gained a new lease of life with the state playing a yet more accentuated role. During this period the authorities launched a two-​phase Growth Acceleration Plan (or PAC in Portuguese) that explicitly targeted strategic investment in infrastructure as a means of alleviating ingrained growth constraints. Thus, for example, surges of investment were planned in oil and gas, shipbuilding, railways, and highways. Yet despite the efforts of the PAC, expansion in infrastructural provision failed to keep pace with demand, leaving a serious growth constraint in place. Following the impeachment of President Rousseff in August 2016, the administration of President Temer, although far more inclined toward free-​market solutions than his predecessor, nevertheless reasserted the state’s key role as a prime mover in the facilitation of infrastructure projects. With the launch of the Projeto Crescer (Project Growth) in late 2016, the Temer administration is trying to address the shortcomings of the PAC. In doing so, the administration is battling a sharp contraction in fixed capital investment across the productive sector and the unfolding of a corruption scandal that has engulfed some of Brazil’s largest civil engineering companies. What all this represents, if nothing else, is an acknowledgment across the entire political spectrum—​and even in the extremely polarized political climate of today—​that infrastructure “matters” and that the state has a critical role to play. Yet, despite this rare point of tangency between left and right, as we have argued, infrastructure investment remains at suboptimal levels. Against this background, the purpose of this chapter is to analyze the development of infrastructure provision over the long term in an attempt to understand the factors constraining investment. This will be accomplished as follows. First, section 18.1 provides relevant historical background, offering an overview of the development of Brazilian infrastructure over the long term. This section discusses the key epochs of infrastructure development, ranging from the private sector–​led “model” of the nineteenth and early twentieth centuries to the more developmentalist postwar period. Regarding the latter, the waxing and waning of direct state infrastructure provision is highlighted, with a focus on the impacts of privatization and the rise of public-​private partnership modes of provision. Section 18.2 then examines the contemporary state of Brazilian infrastructure on a sector-​by-​sector basis, analyzing the challenges which exist in key areas such as electricity, highways, ports and airports. Following this, section 18.3 provides some general theoretical background, tracing the strong linkages

Infrastructure   379 that exist between growth and investment in infrastructure. Following this, section 18.4 moves on to consider contemporary policy responses to infrastructural challenges and the reasons why outcomes have so often fallen short of targets. Finally, section 18.5 draws together the main conclusions and looks ahead to consider potential future developments.

18.1.  Brazil’s Infrastructural Experience in Brief Historical Perspective 18.1.1.  Initial Developments The allocation of a growing role for the private sector has characterized the authorities’ more recent attempts to accelerate infrastructural investment across a range of sectors. This process began with a wave of privatizations in the public utilities sectors in the 1990s (Baer 2014) and has continued with the granting of concession contracts and the establishment of public-​private partnership (PPP) arrangements. In longer-​term historical perspective, these developments embody a surprising degree of continuity. During the nineteenth century, following independence, the Brazilian state adopted policies of relative openness to trade and foreign direct investment (Abreu 1989). Faced with thin domestic capital markets and a narrow tax base, policymakers were obliged to draw heavily on inward foreign direct investment (FDI) in order to meet Brazil’s rapidly expanding infrastructural requirements. These had become very pressing in the light of an upswing in the commodities cycle and the resultant “coffee boom,” which triggered significant expansion, immigration, and, ultimately, urbanization and industrialization in Brazil’s South and Southeast (Fajnzylber 1998) As a result, the early Brazilian railroad, urban transportation (tramways), and electricity supply infrastructure was developed and instituted by foreign investors working under concession contracts with state and municipal governments. Not only was the involvement of foreign enterprise essential given the shortage of domestic capital; the emergence of new electricity-​based technologies also required technical and innovative capacity that Brazil clearly lacked. Thus, for example, a Canadian enterprise came to build and operate (under concession) the electricity network in the city of Rio de Janeiro. As in the case of neighboring Argentina, early railroads were financed by foreign investors, among them the British, who received concession contracts from the state governments. The earnings from these investments were high enough to attract such capital, as most railroads were guaranteed a rate of return by the states (Summerhill 1998).

380    Edmund Amann, Werner Baer, Thomas Trebat, and Juan Villa Lora The early installation of tramways and electricity in some of the major cities was financed by both local and foreign groups under concession contracts with municipalities. Regulation of such contracts occurred within the context of local monopolies, which were regulated by subnational governments.2 For a while the development of these infrastructure projects seemed to have been satisfactory from both the investors’ and recipients’ points of view. The regulation of these local monopolies was relatively lax, ensuring a favorable rate of return to the private (mainly foreign) investors.3 While local monopolies prevailed,4 the infrastructure market could be said to have been fragmented, given the great number of enterprises involved in operating these concessions. However, as time wore on, increasing consolidation of the infrastructure providers resulted in a more concentrated picture.5 Until the 1930s, the involvement of the federal government with infrastructure projects was almost nonexistent. However, things soon began to change. In the case of railroads, there was a decline in the earnings of the sector and an initial disinclination of the state to subsidize them. As individual railroads hit crisis point, they were progressively taken over by the federal government.6 In the case of electrical energy, Brazil’s federal government intervened very little until 1930. In 1904 it had issued a decree (no. 5407) establishing rules for concessionary contracts for firms using hydroelectric sources. In 1931 the government introduced a new system of rules that restricted the frequency with which tariffs could be readjusted (Tendler 1968, 48–​49). In 1934 a new decree—​the Codigo das Aguas (the Water Code)—​set out new rules for the regulation of the water supply and hydro power sectors (Ferreira 2009, 49). The Code would prove to be highly influential in governing the future evolution of the regulatory structure governing infrastructure, not only in the electricity but also in other public utilities. The critical departure from earlier arrangements was that concessions were now granted mainly by the federal government, and then for a minimum period of 30 years. The Codigo laid down more exacting conditions than previous concession contracts, not only in terms of tariff levels, but also in regard to quality of service. As such, this new regulatory framework became a template on which future concession arrangements (and consequent investment in infrastructure) would be based (Baer and McDonald 1998). As Brazil moved into the 1940s and 1950s, and as its program of import-​substitution industrialization gathered pace, the state took on an increasing role as a direct provider of infrastructure, whether in the transportation, power generation and distribution, or water and sanitation sectors. In the case of railroads, many private companies began to run up operational losses that led them to the brink of bankruptcy.7 Rather than see lines close, a process of nationalization began with the creation of the Federal Railroad Corporation (RFSSA) in 1957. This subsequently incorporated a vast range of formerly privately owned lines (Baer 2008, 213). It was also the case that the federal government became more directly involved in electricity, especially in generation. Regarding distribution, many privately held entities—​for example, Rio de Janeiro’s Canadian-​owned Light—​were

Infrastructure   381 transferred to state-​government ownership. The progressive transfer of infrastructure to public ownership not only reflected the need to inject investment resources where the private sector had been unable or unwilling to do so,8 but also an ideological shift that overtook Brazilian economic affairs from the Vargas years onward. The centerpiece of this, of course, was the creation of the state-​owned oil company, Petrobrás, in 1953.

18.1.2. The Import Substitution Period: The State as a Provider of Infrastructure From the early post–​World War II period to the late 1970s, the import-​substitution model of economic development was dominant in Brazil (Baer 2014). During that time, the country relied to a large extent on foreign direct investment to develop new industries.9 Infrastructure, however, was in the hands of the state (at various levels of government) (Ferreira 2009). The ability of the state to furnish resources to meet Brazil’s fast-​expanding infrastructure needs considerably exceeded that which existed in the nineteenth century. By the 1950s and 1960s, the state’s capacity to raise taxation had greatly increased and it was also the case that, thanks to the post-​1944 Bretton Woods arrangements, Brazil was able to draw on special infrastructural lines of credit supplied by the International Bank for Reconstruction and Development (later the World Bank). Such sources of finance were vital in building up Brazil’s network of hydroelectric generating stations, for example. Although much was accomplished in terms of building power-​ generating facilities and some highways, other infrastructure, such as railroads, was largely neglected. As viable enterprises, government infrastructure enterprises, however, operated in such a way as to create substantial distortions in the economy. For instance, the state tried to use its enterprises, including infrastructure enterprises, as instruments of macroeconomic policy. In the midst of many periods of high inflation, the tariffs charged by such enterprises were not allowed to accompany rising costs. This created substantial losses for individual state-​owned infrastructure enterprises and resulted in a need for subsidization, which contributed to further bouts of inflation.10 During the “lost decade” of the 1980s, when Brazil underwent a serious debt crisis, investments in infrastructure were neglected. With the arrival of a new era of a neoliberal development model by the early 1990s,11 Brazil faced the necessity of substantially modernizing its infrastructure sector. However, its state enterprises had no means to finance such investments, given a crisis-​induced decline in tax revenues and much more restricted access to international credit, whether from official sources or private capital markets. By the second half of the 1990s and into the first decade of the twenty-​first century, Brazil’s policymakers were forced to confront the fact that the only way to deal with the need for infrastructure investment was to revert to the old model of appealing to the private sector through concession contracts.12

382    Edmund Amann, Werner Baer, Thomas Trebat, and Juan Villa Lora

18.2.  The Contemporary Infrastructure Challenges Facing Brazil For a country of its size and level of development, Brazil, in international comparative terms, performs rather poorly in terms of the quality and extent of its infrastructure. The most commonly deployed measure of infrastructure quality, provided by the World Economic Forum (Global Competitiveness Report 2014–​2015) ranks Brazil at only 120th place globally. Breaking down infrastructure by category, the quality of Brazil’s roads was ranked at 122nd place, its railroads at 95th, its ports at 122nd, its air transport at 113rd, and its electricity supply at 89th place. To gain more insight into the basis for these rankings, it is worth considering briefly each of these sectors in turn. This is accomplished in the following list: (1) Highways:  The import-​ substitution model pursued by Brazil rendered the country highly reliant on road transport and gave rise to the world’s fourth-​ largest road network. While the quality of highways in the South and Southeast—​ especially in the state of São Paulo—​is generally good, with a large motorway network, coverage in the rest of the country is patchy. This is especially true in the North, the Northeast and the Center-​West. Looking at Brazil overall, of 1.75 million kilometers of highways, only 18% are paved.13 In remote regions, key routes can become impassable during the rainy season: this is true of sections of the famous Trans Amazonian Highway, for example. All of this would matter less if the country were less reliant on road transport. However, relative to other equivalently sized countries, Brazil is especially dependent on the truck: something like 60% of all freight by weight is moved by road in Brazil, compared with 30% in the United States and 10% in China (Garcia-​Escribano et al. 2015, 10). All of this contributes to high transportation costs, the spending on which represents 15.4% of Brazil’s GDP, compared with a more typical 8%–​10% in advanced counties.14 (2) Railways:  The rail network in Brazil developed rapidly in the nineteenth and early twentieth centuries on the back of a surge of foreign direct investment. Nonetheless, Brazil’s contemporary rail network is, in mileage terms, five times smaller than its highway network.15 The country possesses just 3.4 kilometers of rail per 1,000 square kilometers compared with 14.7 kilometers in the United States.16 Another curiosity of the Brazilian rail network is that it is, outside the major urban areas, almost exclusively the preserve of freight rather than passenger transport. In this sense, it is a mirror image of modern European rail systems. Even the major cities in the Southeast—​Rio de Janeiro, Belo Horizonte, São Paulo, and Curitiba—​no longer enjoy passenger rail connections with one another. As a result, intercity public passenger transportation in Brazil is dominated by long-​distance buses and air. Plans to institute a high-​speed rail network in the

Infrastructure   383 Southeast have been put on hold by the current fiscal adjustment and look unlikely to see fruition in the medium term. (3) Ports:  The problems of Brazil’s port system have been quite extensively documented. According to Micco and Perez (2002, 159), “tariffs were three to six times higher than the international average, with long waiting times for ships.” According to Garcia-​Escribano et al. (2015, 10), “only one of Brazil’s ports—​the port of Santos (São Paulo)—​was in the top 100 list of best ports in the world in 2013, occupying the 41st position.” Despite extensive reforms and privatization in recent years, Brazilian ports remain associated with underinvestment, restrictive labor practices, and inadequate maintenance. This often results—​and Santos is typical here—​of enormous queues of trucks, lining the approach roads for kilometers waiting to unload. Given that Brazilian ports handle 95% of the country’s trade by volume and 85% by value, it is obvious that this is a pressing policy issue in terms of ramping up export performance. (4) Water and sanitation: According to Mourougane and Pisu (2011, 25), “water and sanitation is the sector where investment is probably most needed” in Brazil. In 2008, according to the World Bank Development Indicators, 80% of Brazil’s population had access to “improved sanitation facilities” compared with 96% in Chile (a regional leader), 83% in South America overall, and 97.5% for the Organisation for Economic Co-​operation and Development (OECD) countries as a whole. More strikingly, only 47% of Brazil’s population is provided with sewage collection, of which only 20% is treated.17 There are also strong interregional variations in access to water and sanitation services, as Table 18.1 reveals. A major enduring characteristic of the water and sanitation sector in Brazil is its decentralization. Until 1968, water and sanitation services were the sole responsibility

Table 18.1 Coverage Indicators for the Water and Sanitation Sector Treatment of Sewage Collected (%)

Coverage (%) Region

Water Total

North

57.5

Northeast

68.1

Southeast

91.3

South

84.9

Sewage Collection Total

Sewage Collection Urban

Total

71.8

81.0

100.0

22.4

87.1

19.6

26.1

32.0

96.6

71.8

76.9

40.8

96.0

34.3

39.9

33.4

Water Urban

Center-​West

86.2

95.3

46.0

50.5

43.1

BRAZIL

81.1

92.5

46.2

53.5

37.9

Source: Authors’ elaboration based on SNIS (2010).

384    Edmund Amann, Werner Baer, Thomas Trebat, and Juan Villa Lora of municipalities. However, as urbanization accelerated, the federal government became more proactive in driving up infrastructure provision. In 1968, with the establishment of the National Water Supply and Sanitation System (PLANASA), the government attempted to bring strategic direction and enhanced investment to bear on a sector badly in need of upgrade and reform. Other initiatives and regulatory structures followed, with the current framework being introduced in 2007. Rather than adopting the centralizing approach embodied in PLANASA, the current framework allows for a more decentralized approach, placing the municipalities at the heart of service provision, but granting state governments a role in larger urban areas and the federal government an overall coordinating function. Despite these reforms, which allow for municipalities to outsource water and sanitation activities to the private sector, the sector remains both highly public-​sector dominated and decentralized.   Despite the instability of the policy environment surrounding the sector, it is important to note that over the long term there have been notable improvements. Whereas in 1970 a mere 12.4% of Brazil’s population had access to piped water and 6.4% to sewerage services, by 1991 these percentages rose to 50.2% and 19.2%, respectively.18 As can be seen in Table 18.1, by 2010 water coverage was widely—​ though far from universally—​available, particularly in urban areas. Sanitation coverage, however, was rather thinner. Aside from the regulatory complexities affecting investment in this sector, it is important to note that it has been especially exposed to financing challenges stemming from changes in the fiscal policy arena. Following the implementation of the Law of Fiscal Responsibility in the 1990s (see Afonso et al. 2014), the ability of subnational governments—​in this case, municipalities—​to borrow to finance capital spending has been severely curtailed. Given the failure of the private sector to compensate fully via accelerated investment, it is little surprise that the sector has struggled to meet its challenges surrounding improving the quality and scope of service. (5) Airports: The physical scale of Brazil, the absence of long-​distance rail services, and the poor quality of highway infrastructure outside the South and Southeast mean that Brazil is highly reliant on air transportation. Here, as elsewhere, the infrastructure was associated with a legacy of underinvestment and poor connectivity, placing Brazil at a disadvantage in terms of international trade, investment, and tourism. Brazil’s two most significant international airports, São Paulo’s Guarulhos and Rio de Janeiro’s Galeão, date, respectively, from the 1980s and 1970s, and their capacity (at least in terms of terminal rather than runway) was severely constrained (da Silva Campos 2011). However, new terminals were added to correspond with the Rio Olympics in 2016 and the FIFA World Cup in 2014, and the improvements in capacity and quality of service, especially at Guarulhos, have been dramatic. Still, none of Rio or São Paulo’s airports is currently served by rail or metro links,19 highly unusual for cities of such size and international standing. At the same time, limited capacity at the military-​run national air traffic control system results in frequent delays and costs, which

Infrastructure   385 are passed on to passengers. In an attempt to overcome some of these issues, operating concessions were granted to private-​sector consortia across a range of key airports (including Guarulhos and Galeão) in 2012 and 2013. A rail link connecting Guarulhos is also under construction. Further airport concessions, including one affecting the airport of Porto Alegre, were announced by the Temer administration in September 2016.

18.3.  The Growth-​Infrastructure Link: Brief Reflections on the Literature The nature of the relationship between investment in infrastructure and economic growth is of critical significance in the public policy arena. Across the world and in very different political contexts, it is commonly asserted that a surge in infrastructural provision can play a key role in boosting output and employment. Thus, for example, the prioritization of infrastructural renewal occupies a key place in US president Donald Trump’s policy platform, just as it had for his predecessor. In the case of Japan under Prime Ministers Miyazawa, Hosokawa, Hata, and Murayama in the 1990s, a US$254 billion program of infrastructure spending formed the centerpiece of an effort to engineer a recovery after the post-​1989 crash (Uit 2008). Turning to Brazil, as already highlighted, the political priority attributed to infrastructure is little different. Yet, theoretically and empirically, what is the basis to suggest that infrastructure has unique growth-​promoting powers? Tracing the causal linkages here presents methodological challenges given the bidirectional nature of the relationship: just as the initial investment in infrastructure clearly represents an addition to GDP, so does the resulting accumulation of capital facilitate additional growth. Thus, researchers have to develop effective identification strategies to cope with the attendant issues of simultaneity and endogeneity. Despite the methodological complexities, several high-​quality studies have emerged in recent years. A survey of 64 of these by Straub (2008) indicated that in three-​quarters of the studies where a physical measure of infrastructure was deployed, a reasonably strong causal link between investment and output was determined. However, where alternative—​including monetary—​measures of infrastructure were employed, the relevant studies revealed a much weaker relationship. In the case of Brazil, there have been relatively few studies tracing the infrastructure-​ growth linkage. One of the most comprehensive (Ferreira 2007), established “significant impacts” on output resulting from variation in infrastructure stocks. Using satellite-​ derived luminosity data as a proxy for infrastructure stock, Amann et al. (2014) show that, at a subregional level, a firm infrastructure-​growth relationship exists. The same study found that spending on communications infrastructure yielded the highest output gains. Viewed as a whole, the empirical evidence suggests the real importance

386    Edmund Amann, Werner Baer, Thomas Trebat, and Juan Villa Lora of infrastructure as a growth-​promoting policy variable, though cautions us that its potency may be somewhat overstated in the political debate.

18.4.  Contemporary Policy Responses: The Role of the Private Sector and the Growth Acceleration Program The previous section indicated the important role that infrastructural investment can play in raising growth potential and performance. The salience of infrastructure in this connection has been recognized by consecutive administrations in Brazil. As discussed in section 18.1, however, by the 1990s had come the realization that, faced with binding fiscal constraints, the Brazilian state’s role as the primary direct provider of vital infrastructure could not be sustained. Instead, the private sector would need to play an increasingly significant role. Under the administration of President Fernando Henrique Cardoso (1995–​2002), Brazil embarked on what at the time became the world’s largest privatization program (see Chapter 31). Central to this, as the chapter indicates, was the sale of critical public utility state-​owned enterprises (SOEs) operating infrastructure in telecommunications, electricity generation and transmission, ports, and railways. As a result of this process, well over US$50 billion of foreign investment flowed into these sectors during the 1990s and the first decade of the 2000s. In tandem with the privatization process, a program of concession contracts was launched.20 This initially centered on highways. Under the concession model adopted, in return for specified levels of investment and operational input, private entities would take control (but not ownership) of infrastructure for a specified period. However, as we have seen, despite the shift to a more private sector–​focused model, infrastructural provision failed to keep pace with the needs of an expanding economy. With this in mind in 2007, the administration of President Lula launched what was arguably the most ambitious infrastructure initiative since the Second National Development Plan of 1974. The new program, the PAC (Programa de Crescimento Acelerado—​Growth Acceleration Program), ran in two phases (PAC 1 and PAC 2).21 The total sums contemplated by the program were impressive: R$503.9 billion for PAC 1 and R$958.9 billion (around 2.7% of 2010 GDP per year) for PAC 2. PAC-​designated infrastructural investments were facilitated through six key subprograms:  My House, My Life (housing); Water and Light for All (water, sanitation, and electricity); Bringing Citizenship to the Community (public safety and social inclusion); Better Cities (urban infrastructure); Transportation (railways, highways, and airports); and Energy (renewables, oil and gas). Regarding these sectors, there was a notable shift of priorities between PAC 1 and PAC 2. According to data supplied by the PAC

Infrastructure   387 secretariat and Morgan Stanley, during PAC 1 (2007–​2010) transportation and logistics, energy, and social and urban projects, respectively, accounted for 14.9%, 45.7%, and 39.5% of planned investments. By PAC 2 (2010–​2014), however, the proportions had changed radically, with energy projects accounting for no less than 92.4% of the total. The shares for transportation and social and urban projects had fallen to 7.2% and 0.4%, respectively. The huge resources being shifted into energy, as subsequent events have proven, provided ample opportunities for corrupt diversion of funds to political parties. This, of course, resulted in the Lava Jato (or Car Wash) crisis (see Chapter 35). Beyond the large scale of the projects involved, was there anything qualitatively novel about PAC? Were new modalities of infrastructural provision introduced? The answer here is a partial “yes.” While the concession contract model from the 1990s was maintained, for some projects a more direct financial role for the public sector was facilitated with the introduction of a public-​private partnership (PPP) framework. Under the PPP model—​which embraces Build-​Own-​Operate, Build-​Operate-​Transfer, and Build-​Own-​Operate-​Transfer formats—​under the terms of 2004 enabling legislation, the public sector is allowed to support financially private-​sector investment (Mourougane and Pisu 2011, 14). This would be typically channeled through the BNDES (National Development Bank). Thus, in a certain sense, the PAC can be seen as a latter-​ day manifestation of a long-​established feature of Brazilian developmentalism: the close articulation between the state and private capital in an attempt to overcome structural obstacles to economic progress. Viewed in outline terms, the PAC was quite successful in meeting its investment targets: 82% of PAC 1’s targeted investments were completed, while 82.3% of PAC 2’s planned investments had been concluded by 2013.22 Yet problems remained. In particular, the pivot to energy in PAC 2 left serious infrastructure gaps in the urban transportation and sanitation sectors unaddressed. In the energy sector itself, the global fall in oil prices, combined with the fallout from the Lava Jato scandal, has resulted, since 2013, in the freezing and cancellation of investment projects in the oil and gas industry. This has helped to push the state of Rio de Janeiro (where much of the offshore oil sector is located) into steep recession. It has also brought the Rio de Janeiro state government to the brink of insolvency. Despite all the achievements of the PAC program, the reality is that it did not push the needle of national infrastructure spending very far. As Garcia-​ Escribano et al. (2015) point out, by 2013 infrastructure spending—​at 2.5% of GDP—​ was still less than half that realized in the early 1980s. As an indicator of the lingering difficulties, popular concerns around infrastructure23 triggered violent riots in Rio de Janeiro and São Paulo in 2013. Recognizing that PAC has not provided a lasting solution to Brazil’s infrastructure problem, the administration of President Michel Temer has, since mid-​2016, announced a series of fresh initiatives under the banner of Projeto Crescer “Project Growth.” These will be briefly examined in the next section. For the remainder of this, we consider some of the recurrent obstacles to accelerated infrastructural investment in Brazil that the current and subsequent administrations will need to tackle.

388    Edmund Amann, Werner Baer, Thomas Trebat, and Juan Villa Lora Perhaps the most obvious constraint bearing on infrastructural investment centers on the availability of finance. With a savings rate of just 14% of GDP (World Bank 2017) Brazil remains very constrained in the domestic resources that can be mobilized to finance investment. Added to this, the high intermediation costs associated with the Brazilian private financial system and the elevated base rates deployed to combat inflation mean that financing major infrastructural projects through domestic private-​ sector financial markets remains the exception rather than the rule. For this reason, for all intents and purposes, infrastructure projects in Brazil are funded through three channels: foreign direct investment; direct public sector capital spending; and official credit channeled through Brazil’s development bank, the BNDES.24 Compared with other major economies, China and the United States, for example, the channels for funding infrastructure are therefore more restricted. At the same time, the relative role of the public sector is more extensive. The latter feature renders infrastructural spending in Brazil especially vulnerable to the impacts of periodic fiscal adjustment.25 This is a major reason why infrastructure spending as a proportion of GDP has fallen by half since the start of the 1980s. However, even where financial resources have been made available—​as they increasingly were under PAC 1 and PAC 2—​ attempts to ramp up infrastructure spending rapidly may be frustrated. This is due to a second critical characteristic of the Brazilian economic and political environment: regulatory complexity and uncertainty. Mourougane and Pisu (2011, 17) argue that the latter features play a critical role in delays attaching to Brazilian infrastructure projects. The situation facing would-​be investors is complicated by the presence, in many instances, of overlapping regulators from the federal and subnational governments. Difficulties are also created by a curious feature of the Brazilian legal system: the nonbinding nature of decisions made by higher courts on lower courts. The lack of case law precedence in Brazil—​a key feature of a common law system in the United Kingdom and the United States—​means that it can be very difficult for investors to gauge the way their cases may be determined in the event that disputes emerge with regulatory bodies. A special area of concern surrounds environmental regulation, where overlapping regulators, uncertainty, and attendant delays can prove very costly. According to one study (World Bank 2008) 15%–​20% of the total outlays involved with hydroelectric projects are linked with environmental licensing costs. A third key set of factors constraining infrastructural development in Brazil surrounds corruption. As has become clear, following the eruption of the Lava Jato scandal since 2014, the wave of infrastructure spending unleashed under the PAC created significant opportunities for graft. Many of these were deftly exploited by unscrupulous individuals who took advantage of public-​sector procurement projects—​ especially in oil and gas—​to channel kickbacks from contractors to political parties. As a result of a widening series of inquiries by the federal police and aggressive public prosecutors, several senior politicians have been formally accused. These now include the current President, Michel Temer. At the same time, senior executives from leading infrastructure investors and contractors—​most significantly, Petrobrás and Odebrecht

Infrastructure   389 (Brazil’s largest construction company)—​have been arrested, tried, and jailed. A direct consequence of these startling developments has been a sharp reduction in investment in the oil and gas sector. However, there may have been more insidious effects over the longer term: according to a study26 published by the São Paulo industrial association, FIESP, the amount of money lost to corruption during PAC 1 could have built 124% more kilometers of roads and 525% more kilometers of railways. One fortunate long-​term consequence of the current wave of anti-​corruption actions by prosecutors is likely to be that the burden of corruption will lie much less heavily on future infrastructure projects. With greater transparency and reduced costs, it is conceivable that more outside investors—​ including those with no established political connections—​will in the future be drawn to participating in Brazilian infrastructure projects.

18.4.1. The Infrastructure Challenge: Responses by the Temer Administration Partly as a consequence of the burgeoning corruption scandals surrounding the Lava Jato scandal, President Dilma Rousseff was formally impeached in August 2016. The new centrist administration of President Michel Temer has moved swiftly to address the issues surrounding underinvestment in infrastructure that had affected his predecessors. In September 2016, a new private sector–​focused initiative known as Projeto Crescer (Project Grow) was launched. At its core are 34 infrastructure projects, which are to be awarded to the private sector under the concession model (BrazilGovNews, September 13, 2016). These range from airport and highway projects to oilfield auctions to the sale of new mining rights. A number of hydroelectric projects are also envisaged through the sale of assets to the private sector. In one sense, there is very little distinction between this initiative and the PAC. However, there are some important developments around the technicalities of the concession contracts. Among them, under Projeto Crescer, are the enforcement of legal certainty and the setting of rates of return and fees according to market forces and technical studies, rather than political fiat. Furthermore, a presidential decree signed in November 2016 allows for greater flexibility in concession contracts, permitting, for example, the authorities to renew concessions early in return for new investments (Reuters Business News, November 24, 2016). Investors are also to be allowed to apply for voluntary termination of concessions in the event of unforeseen technical or environmental challenges. As previously, the BNDES development bank will provide loans to support the new projects (an initial R$30 billion). Whether these measures collectively unleash a new wave of infrastructural investment remains to be seen. However, what is noticeable is that the emphasis, in this latest in a long line of policy initiatives, has shifted toward an attempt to deal with regulatory obstacles. As this chapter indicates, there is a good case to be made that it is the latter which represent the most severe barrier to enhanced infrastructural investment.

390    Edmund Amann, Werner Baer, Thomas Trebat, and Juan Villa Lora

18.5. Conclusions Hirschman (1958), in a classic treatise on economic development, argued that infrastructural investments by their very nature tend to be lumpy and resource intensive. Certainly, they represent a significant opportunity cost. Therefore, he claimed, unless there is an acute shortage of capacity, investments tend to be postponed until a crisis situation is reached. At that point, political justification for large amounts of investments will be acceptable to society. It may be the case now that, given the depth of its economic crisis, Brazil has reached such a point. The question is whether the current modalities of infrastructural provision will prove able to help propel Brazil out of this “Hirschman trap.” Despite increasing private-​sector participation, issues surrounding financing, the regulatory frameworks applied to performance, ill-​considered environmental rules, corruption, and the political climate have all conspired to retard progress. For this reason, increases in infrastructural provision have not been sufficient to meet rapidly rising demand and expectations of the population at large. Confronting this reality, the new administration of President Temer is belatedly focusing on the long-​standing regulatory and legal issues that will need to be overcome if Brazil’s renewed quest for accelerated infrastructure provision is to be a success.

Notes 1. This chapter is partly based on an earlier working paper (Amann et al. 2014), which was the result of research funded by the UK Department for International Development 2. Baer and McDonald (1998). 3. Ibid. 4. According to Baer and McDonald (1998, 506), “Although Light dominated the electric power supply in Sao Paulo and Rio de Janeiro other cities were supplied by locally owned firms and by 1920 343 electric power firms were operating in Brazil on a concession basis with municipalities. Their owners were usually local owners or landowners who owned local concessions.” 5. Baer and McDonald (1998). 6. Baer and Sirohi (2013). 7. Ibid. 8. Very often because the tariffs set in compliance with toughened regulatory frameworks would not provide the resources or incentives for such investments. 9. For details, see Baer (2014). 10. Baer (2014), Chapters 6 and 11. 11. Amann and Baer (2002). 12. Ferreira (2009). 13. World Bank (2012). 14. Ibid. 15. Ibid., 78.

Infrastructure   391 16. Ibid., 79. 17. Ibid. 18. Arretche (2004) states that “although PLANASA proved useful in expanding water and sanitation services, the expansion of water services was favored because it was less costly to develop and water charges created returns sooner than sanitation. Much of PLANASA’s funds went to the wealthier regions of Brazil, especially the better neighborhoods of wealthier areas.” 19. Although a rail link to São Paulo Guarulhos Airport is now under construction. 20. For an in-​depth discussion of concession contracts and their financing modalities, see Pompermayer and Silva Filho (2016), pp. 1–49. 21. PAC 1 ran from 2007 to 2010 under President Lula, and PAC 2 ran from 2010 to 2014 under his successor, President Rousseff. 22. Investorideas.com, February 20, 2014. http://​investorideas.com/​news/​2014/​international/​ 02204.asp 23. In this case surrounding the quality and price of urban public transportation. 24. For more details concerning the role of the BNDES, see Torres and Zeidan (2016). 25. For an in-​depth discussion of the impacts of fiscal adjustment on investment and discretionary spending, see Amann and Baer (2006). 26. For details, see Funmi Oji and Allison Everhardt, “Brazilian Infrastructure and Corruption,” America’s Business Intelligence, Washington DC, October 11, 2013.

References Afonso, José Roberto, and Elaine Araújo. 2014. “Institutions for Macroeconomic Stability.” IRIBA Working Papers. Amann, Edmund, and Werner Baer. 2002. “Neoliberalism and Its Consequences in Brazil.” Journal of Latin American Studies 34, Part 4 (November): 945–​959. Amann, Edmund, Werner Baer, T. Trebat and J. Villalora,. 2014. “Infrastructure in Brazil’s Development Process.” IRIBA Working Papers. Amann, Edmund, and Armando Barrientos. 2016. “Introduction:  Brazil’s Development Model.” Quarterly Review of Economics and Finance 62 (November): 7–​11. Arretche, M. T. 2004. Water Supply and Sanitation. Associação Nacional de Transporte Público (ANTP) 2012. Sistema de informações da mobilidade urbana: Relatório 2011. Brasília: ANTP. Baer, Werner. 2014. The Brazilian Economy: Growth and Development, 7th edition. Boulder, CO: Lynne Rienner. Baer, Werner, and Curt McDonald. 1998. “A Return to the Past? Brazil’s Privatization of Publid Utilities: The Case of the Electric Power Sector.” The Quarterly Review of Economics and Finance 38 (3): 503–​523. Baer, Werner, and Rahul A. Sirohi. 2015. “Transportation Infrastructure and Economic Development:  A Comparative Analysis of Brazil and India.” Global and Local Economic Review 19 (2): 37–​60. Biderman, Ciro, et  al. 2009. “Mobility, Integration, and Accessibility:  Transport Policies in the São Paulo Metropolitan Region.” Chapter 6 in London School of Economics, Cities and Social Equity: Inequality, Territory, and Urban Form, 6-​2–​6-​37. London: LSE.

392    Edmund Amann, Werner Baer, Thomas Trebat, and Juan Villa Lora Block, P. J. 2008. “An Assessment of Investments in Agricultural and Transportation Infrastructure, Energy, and Hydroclimatic Forecasting to Mitigate the Effects of Hydrologic Variability in Ethiopia.” CGIAR Challenge Program on Water and Food, Working Paper. BNDES. 2012. “Transporte Público: O Papel do BNDES no Apoio da Solução dos Principais Gargalhos de Mobilidade.” In BNDES 60 anos: Perspectivas setoriais, Vol. 2, 313–​348. Rio de Janeiro: BNDES. Cusack, T. R. 1997. “Partisan Politics and Public Finance:  Changes in Public Spending in the Industrialized Democracies, 1955–​ 1989.” Public Choice 91:  375–​ 395. doi:  10.1023/​ A:1004995814758. Campos Neto, Carlos Alvares da Silva, Jean Marlo Pepino de Paula, and Frederico Hartmann de Souza. 2010. Rodovias brasileiras: Políticas públicas, investimentos, concessões e tarifas de pedágio. Rio de Janeiro: IPEA, Texto Para Discussão 1668. Djiofack-​Zebaze, C., and A. Keck. 2009. “Telecommunications Services in Africa: The Impact of WTO Commitments and Unilateral Reform on Sector Performance and Economic Growth.” World Development 37: 919–​940. Federal Government of Brazil. 2013. PAC2: A gente faz um Brasil de oportunidades, 6o Balanço, 2011–​2014, Ano II. Brasilía: Federal Government of Brasil. Ferreira, P. C. 2007. Growth and Fiscal Effects of Infrastructure Investment in Brazil. Rio de Janeiro: FGV Working Paper Ferreira, Tiago Toledo. 2009. Arranjos institucionais e investimento em infra-​Estrutura no Brasil. São Paulo:  USP, Dissertação apresentada ao Departamento de Economia para obtenção de mestre em Economia. Freeman, P., and K. Warner. 2001. “Vulnerability of Infrastructure to Climate Variability: How Does This Affect Infrastructure Lending Policies?” World Bank, Disaster Management Facility, ProVention Consortium. Hirschman, A. 1958. The Strategy of Economic Development. New Haven, CT:  Yale University Press. IPEA. 2010. Infraestrutura econômica no Brasil: Disgnósticos e perspectivas para 2015. Livro 6, Vol. 1. Brasília: IPEA. IPEA. 2012a. Comunicado 128: A nova lei de diretrizes da política nacional de mobilidade urbana. Brasília: IPEA. IPEA. 2012b. Sistema de indicadores de percepção social: Mobilidade urbana. Brasília: IPEA. IPEA. 2012c. Comunicado indicadores de mobilidade urbana da pesquisa nacional por amostra de domicílios 2012. Brasília: IPEA. IPEA. 2012d. Transportes e metrópoles: Um manifesto pela integração. Brasília: IPEA. IPEA. 2013a. Nota técnica no.  2:  Tarificação e financiamento do transporte publico urbano. Brasília: IPEA. IPEA. 2013b. Territorio metropolitano, politicas municipais. Brasília: IPEA. IPEA. 2013c. Nota técnica no.  4:  Transporte integrado social:  Uma proposta para o pacto da mobilidade urbana. Brasília: IPEA. IPEA. 2013d. Comunicado no.  161:  Indicadores da mobilidade urbana da PNAD 2012. Brasília: IPEA. IPEA. 2013e. Texto para discussão no. 34: Transportes e mobilidade urbana. Brasília: IPEA. Koetse, M. J., and P. Rietveld. 2009. “The Impact of Climate Change and Weather on Transport: An Overview of Empirical Findings.” Transportation Research Part D: Transport and Environment 14: 205–​221. doi: 10.1016/​j.trd.2008.12.004.

Infrastructure   393 Lazana, Antonio, and Luiz Martins Lopes. 2011. “Desafios da infraestrutura e expansão does investimentos: 2011/​2014.” Informações FIPE (Setembro): 28–​37. Lima, Iêda Maria de Oliveira. 2012. “As batalhas para a politica nacional de mobilidade urbana.” Revista dos Transportes Públicos—​ANTP 1: 115–​123. Micco, Alejandro, and Natalia Perez. 2013. “Determinants of Maritime Transport Costs.” Inter-​American Development Bank, Policy Research Working Paper 2781, pp. 1–​37. http://​ documents.worldbank.org/​curated/​en/​451431468766755364/​pdf/​multi0page.pdf Milesi-​ Ferretti, G. M., R. Perotti, and M. Rostagno. 2002. “Electoral Systems and Public Spending.” The Quarterly Journal of Economics 117:  609–​ 657. doi:  10.1162/​ 003355302753650346. Morgan Stanley. 2010. Brazil Infrastructure: Paving the Way. (May 5). Mourougane, A., and M. Piso 2011. “Promoting Infrastructure Development in Brazil.” OECD Economics Department Working Papers No. 898, 1–​33. Neri, Marcelo. 2012. A nova classe media:  O lado brilhante da base da pirámide. Rio de Janeiro: FGV. Pompermayer, F., and E. B. da Silva Filho. 2016. “Concessões no setor de infrasestrutura: propostas para um novo modelo de financiamento e compartilhamento de riscos.” IPEA Texto Para Discussão 2177 Fevereiro, 1–49. Rio de Janeiro. Ros, Jaime, ed. 2012. The Oxford Handbook of Latin American Economics. Oxford:  Oxford University Press. Senna, Luis dos Santos. 2013. O transporte público de passageiros e seu financiamento. Unpublished presentation. Available at http://​www.al.rs.gov.br/​FileRepository/​repdcp_​ m505/​ComEspMobilidade/​Transp_​publ_​%20L.Afonso_​Senna.PDF, accessed January 2014. Serafim, Maria Clara Silva. 2009. “Análise das políticas para infraestrutura de transporte no Brasil a partir da década de 90.” Universidade de São Paulo, Escola Superior de Agricultura “Luiz de Queiroz,” Piracicaba. Spilki, Marcelo. 2012. “Public-​Private Partnerships and the Role of Internal Control in the State of Rio Grande do Sul in Brazil.” The Institute of Brazilian Business and Public Management Issues, The Minerva Program, George Washington University, Spring. Straub, S. 2008. “Infrastructure and Growth in Developing Countries: Recent Advances and Research Challenges.” ESE Discussion Paper No. 179. Edinburgh School of Economics, University of Edinburgh. Summerhill, William R. 1998. “Railroads in Imperial Brazil, 1854–​1889.” In Latin America and the World Economy since 1800, edited by John H. Coatsworth and Alan M. Taylor, 383–​405. Cambridge, MA: Harvard University Press. Tendler, Judith. 1968. Electric Power in Brazil. Cambridge, MA: Harvard University Press. World Bank. 2012. How to Decrease Freight Logistics Costs in Brazil. Washington, DC: World Bank

Chapter 19

Trade P ol i c y from t he 1930s to t h e Pre se nt Simão Davi Silber

19.1. Introduction: Trade Policy Background The objective of this chapter is to analyze Brazilian trade policies, over the course of nearly a century, and their effects on trade performance, competitiveness, and growth. The period covered, from the 1930s to the present, is rich with episodes that changed the profile of the Brazilian economy, and where a more integrated world economy opened new trade and investment opportunities. It has also been a turbulent century, with recurrent crises ranging from the Great Depression to the 2008 financial crisis. The chapter analyzes Brazilian trade policy changes against this backdrop of large swings in the world economy, as well as numerous critical domestic factors influencing trade policy design. In the second section (19.2) of the chapter we discuss Brazilian trade policy regimes, divided into four periods, and the section ends with a discussion of the trade liberalization policy of the 1990s, a radical departure from the previously inward-​looking import policies. The results were important in terms of slashing tariffs and nontariff controls that fostered increased competition, modernization of the industrial sector, and significant productivity growth (Kume, Piani, and Souza 1988). Historically, Brazilian trade and investment policies have been characterized as highly protectionist to foster inward industrialization growth. A large array of tax and credit subsidies for industrial investments and high tariffs and nontariff barriers (NTBs) protected the domestic market from import competition, leaving for export-​promotion policies to mitigate the anti-​export bias of trade policies (Cardoso and Fishlow 1990). Even the liberalization attempts of the late 1980s and early 1990s were partial and not enough to change the domestic market focus of the Brazilian development model. Brazil is almost an autarchy in terms of trade flows.

Trade Policy from the 1930s to the Present    395 The third section (19.3) is devoted to analyzing trade performance and comparing with world trends. The composition of world exports has evolved gradually in the direction of high-​technology manufactured goods, but the Brazilian trade structure has moved in the opposite direction during the last decade. At an aggregate level, Brazil lost ground in world markets for most product categories, except commodities. In particular, in agriculture trade and iron ore, the country became a “large player” in several markets, and the increased exports contributed to the large trade surplus during the last decade. In the fourth section (19.4), we analyze the links between the current account, capital flows, and the real exchange rate. In the Brazilian case, we found empirical evidence that high foreign liabilities triggered large exchange rate devaluation to correct balance of payments disequilibrium. The final section (19.5) summarizes some main conclusions and discusses future challenges in trade policy.

19.2.  Brazilian Trade Policy Regimes The events that shaped the world economy in the twentieth and twenty-​first centuries can be roughly divided into two periods. In the first, ranging from 1914 to 1945, there was a strong setback in trade and international capital flows due to two world wars and the Great Depression. This phase can be considered the end of the first great wave of globalization begun a century earlier during the Pax Britannica. In the second period (1945–​ 2015) came a gradual opening of the world economy to international flows that lasted for 70 years and—​again—​the world economy became highly interdependent. This section covers Brazilian trade policy changes, having as a background these large shifts in the world economy, and domestic factors influencing trade policy design.

19.2.1. The Great Depression and World War II: 1930–​1945 At the outbreak of crisis in 1929, the coffee sector was the most important of the Brazilian economy. It was responsible for more than 60% of export revenue and had an almost monopoly position in the world market. Since 1830, the rapid expansion of coffee production had been an important factor behind the expansion of the domestic market and for an incipient industrial growth in sectors isolated from foreign competition by transport costs. Prior to 1930, the most important industrial sectors were textiles, apparel, and food processing. Coffee exports were the engine of growth of the Brazilian economy, representing 9% of gross domestic product (GDP) at the end of the 1920s. The international demand and supply for coffee are price inelastic, and since 1906 government intervention in the coffee sector had been successful in exploiting the Brazilian monopoly position in the international market, sterilizing excess supply

396   Simão Davi Silber through government purchase and stocks. It is well known in trade theory that a country with monopoly power in international markets can turn relative prices in its favor by restricting supply in the world market and maximizing short-​run profits. This “rent-​seeking” policy for coffee growers was not accompanied by production controls, and there was a strong incentive to increase production and sell it to the government. The country started to reap large crops when external demand shrunk sharply after 1929 and foreign loans dried up for the following 20 years. Nominal coffee prices fell by 63% between 1928 and 1931 and export revenue by 49% in the same period. Imports fell even more sharply by 70%, due to the domestic recession (GDP fell 5.3% in 1930 and 1931), high import duties, and rationing of scarce foreign exchange to essential imports. Since the imperial period, Brazil had adopted a system of specific tariffs that were very high at the outbreak of the 1929 crisis. According to Mata and Love (2008), at the outset of the 1930s the average tariff was above 30% of the import value. High tariffs were an essential tool to repress the demand for imports and maintain an overvalued exchange rate, favorable to coffee interests. It is worth mentioning that tariffs were the main item in the government revenues, representing half of total receipts. With a large production surplus, the federal government started an important support program for the coffee sector, through massive purchases and destruction of excess supply. This program lasted up to 1937, and it is estimated that the equivalent of three annual crops were burned during the period (Abreu, Bevilaqua, and Pinho 1996b). This is typically a case of expansionary fiscal policy that was followed, after 1933, by credit expansion by Banco do Brasil to finance coffee purchases by the government. Even so, terms of trade worsened dramatically during the 1930s: between 1929 and 1940 Brazilian terms of trade fell by 52%. Starting in 1941, there was a gradual terms of trade recovery that lasted up to 1954, due to reduced coffee crops and the increase of demand for commodities during World War II and the Korean War. This was the golden age for terms of trade gains. A complete picture of terms of trade performance in the long run (1920–​2015) is shown in Figure 19.1. It is worth mentioning that there is no statistically significant tendency of worsening of terms of trade in the long run as postulated by the ECLAC (Economic Commission for Latin America)-​originating “dependency theory” of center and periphery economic growth asymmetries. Given the collapse of foreign exchange markets, Brazil quickly abandoned the gold exchange regime in the mid-​1930s, followed by a sharp devaluation of the Brazilian currency. The estimated real exchange rate devaluation between 1929 and 1931 was 83%.1 After this initial devaluation, the real exchange rate saw minor fluctuations in the 1930s. It is now well documented in the Brazilian empirical economic history literature that exchange rate devaluations have a downward pressure in international price of commodities, when the country has market power in the world market (Abreu and Bevilaqua 1996a). This was the Brazilian case up to the mid-​1960s, when coffee was still the main source of foreign exchange and the country a major player in world markets. The government used extensively overvalued exchange rates to maximize export revenue, and import controls depended increasingly on tariffs and non-​tariff barriers (NTBs). Additional

Trade Policy from the 1930s to the Present    397 180.0 Coffee shortage

160.0

Korean war

Coffee 140.0 “valorization” 120.0

Oil shocks Great depression

China

100.0 80.0 60.0

20.0

1920 1922 1924 1926 1928 1930 1932 1934 1936 1938 1940 1942 1944 1946 1948 1950 1952 1954 1956 1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

40.0

Figure 19.1.  Brazilian terms of trade, 1920–​2015. Note: (2006 = 100). Source: IPEA DATA.

factors to explain exchange rate controls were the burden of government debt service in foreign currencies and the exchange rate as a nominal anchor to control inflation (Abreu and Bevilaqua 1996b). High tariffs, real exchange rate devaluation, and expansionary fiscal and monetary policies were essential to changing the level and composition of demand in favor of domestic industrial production. Average industrial growth in the period 1930–​1945 was 6.1% per year, and GDP growth averaged 3.9%. This period can be considered the first stage of import-​substitution industrialization (ISI). But Brazil remained a rural country, with 69% of the population in the countryside in 1940. It had become clear to the government that to tie the economy to commodities exports would leave the country vulnerable to balance of payment crises, foreign debt service defaults (like that of 1937), and radical downturns in economic growth. There was room for additional stages of inward-​looking ISI based on shifting underemployed supply of cheap labor from agriculture to the industrial sector, typical of the Lewis development model.

19.2.2. The Heyday of Import Substitution as Explicit Policy: 1946–​1963 Industrial growth and trade surplus during World War II was a consequence of new export possibilities and unavailable manufactured goods for imports. Brazil’s exports

398   Simão Davi Silber became more diversified, including products such as beef, cotton, rice, and textiles and clothing. Manufactured exports to Latin America and Africa were substitutes for Northern Hemisphere exports, and foreign reserves increased. Industrial production increased 5.7% per year between 1939 and 1945. Two years of import liberalization (1946 and 1947) were enough to dry up the convertible reserves, and import licensing controls were imposed once more. Import licensing was chosen, since the specific tariff of 1900 was completely eroded by inflation. In 1946—​within the Bretton Woods regime—​Brazil adopted a fixed exchange rate of Cr$/​US$18.50,2 equal to the 1939 level, even though domestic inflation was 153% during the war years. Inflation increased due to government expenditure during the war, lack of imports, and increased exports of staples. It was the chosen anchor to control high inflation, which reached 15% in 1945. Overvalued exchange rates and import controls favored domestic industrial production through access to intermediate and capital goods imports at lower exchange rates. Additionally, the overvalued exchange was an important device to reduce coffee supply, and there was a strong recovery of coffee prices and terms of trade that lasted up to the Korean War. In 1953, SUMOC3 enacted a foreign exchange reform, through Normative Instruction 70, establishing a more flexible regime with multiple exchange rates for exports and imports, even though the official exchange rate remained fixed at Cr$/​US$18.50. Auctions of foreign exchange were the instruments to ration imports instead of the outmoded specific tariffs of 1900. There were five categories of exchange rates for imports and US$ were allotted according to essentiality:  oil, wheat, and newsprint could be imported at the fixed “official rate” of Cr$18.50, and luxury goods had the highest exchange rates. There were preferential exchange rates for intermediate and capital goods (Abreu and Bevilaqua 1996b). The 1957 Tariff reform eventually introduced escalating ad-​valorem duties that could reach 150% and became an important instrument by which to restrain foreign competition to domestic industrial production, while the “Law of the Similar” gained a new status to prohibit imports with domestic industrial production. The rationale for the ISI growth strategy during this period can be summarized as follows: an overvalued exchange rate was decisive in keeping coffee prices higher, and thus maximizing foreign exchange revenue. An overvalued exchange was also the nominal anchor to control inflation. The rationing of imports through small amounts of foreign exchange in specific auctions of finished consumer goods isolated the industrial sector from international competition, and the preferential access to import intermediate and capital goods created a very favorable environment for investments and high profits in the industrial sector. Foreign direct investment (FDI) inflows increased in the second half of the 1950s, attracted by oligopoly profits and subsidies inaugurated in 1955 by SUMOCs Normative Instruction 113, which allowed the direct import of capital goods without foreign exchange outlays (Cardoso and Fishlow 1990). In the second half of the 1950s, the auto industry was the main target for industrial promotion in Brazil. A  vast array of protective devices, ranging from prohibition of imports of “similar to nationals,” high import tariffs, requirement of 95% of domestic

Trade Policy from the 1930s to the Present    399 content, credit, and imports of capital goods subsidies, created a domestic oligopoly isolated from international competition. Strikingly, 60 years later the same sector is still the most protected sector of the Brazilian economy; while the 1950s saw the invoking of the “infant industry argument” in favor of protecting the auto industry, we can perhaps say that today this should be updated to the “senile industry argument” to protect the same sector. Trade policy was inward looking, with a strong anti-​export bias due to the overvalued exchange rate and strong incentives to divert production into domestic absorption. During the first decade after the war, terms of trade increased, relaxing the constraint of foreign private and official financing shortage. In the second half of the 1950s, balance of payment deficits were financed mainly by FDI. Industrial growth was at its peak between 1946 and 1960, when industrial GDP grew at an average annual rate of 9.3%. Even with this impressive industrial growth, exports of commodities still represented 85.4% of the total exports in 1964 and manufactured goods only 6.2%, as shown in Figure 19.2. Coffee exports represented 50% of the total. At the beginning of the 1960s, the burden of distortions of ISI was high, and Brazil experienced a period of reduced economic growth, increased inflation, and balance of payment crises. An overvalued exchange rate led to stagnant exports and trade deficits; FDI diminished after a wave of investments in the protected domestic market, and government deficits were increasingly financed by inflation tax. In the 1961–​1963 period, annual industrial growth was down to 2.6% and inflation up to 50%. The 1964 military coup was a consequence of these economically and politically turbulent years.

90.0% 85.4% 80.0%

82.9% 79.3% 74.8% 66.4%

70.0%

57.6%

60.0% 50.0%

31.7%

40.0%

10.0%

51.5% 51.0% 47.2%

55.5% 54.2%

60.8%

55.3%

59.0%

39.0% 33.3%

29.8% 22.5% 15.2%

30.0% 20.0%

60.5%

54.9% 54.4% 48.7% 45.6% 46.8% 36.9%

27.8%

27.6%

27.3% 22.8%

39.4% 38.1%

29.3%

10.7% 6.2% 8.7%

19

64 19 66 19 68 19 70 19 72 19 74 19 76 19 78 19 80 19 82 19 84 19 86 19 88 19 90 19 92 19 94 19 96 19 98 20 00 20 02 20 04 20 06 20 08 20 10 20 12 20 14

0.0%

Basic

Manufactured

Figure 19.2.  Composition of Brazilian exports*. Note: * Exports of semi-​manufactured goods excluded. Source: Ministério do Desenvolvimento Indústria e Comércio.

400   Simão Davi Silber

19.2.3. Export Promotion, Oil Shocks, and Foreign Debt Crisis: 1964–​1987 At the outset, it is worth mentioning export performance prior to 1964 to highlight the challenge facing policymakers in the early 1960s. In 1946, export revenue reached US$1 billion; it went up to US$1.4 billion in 1963—​a dismal annual increase, in nominal terms, of 1.2%. In real terms, exports fell on average 1.3% per year. In this period, exports fell from 12% to 6% of GDP. After the Korean War, terms of trade worsened by 35.3%, and the balance of payment deficit increased after 1960. The big government challenge was the design of new trade policy instruments to boost exports of manufactured goods. The chosen trade policy target was manufactured export promotion, through fiscal and credit subsidies and a new “crawling peg” regime to avoid the frequent overvaluation of the Brazilian currency. The instruments to foster manufactured exports were exemptions of manufactured exports from the domestic tax burden and subsidies through tax rebates on domestic sales based on export performance and a more predictable exchange rule (see, for example, Cardoso and Fishlow 1990). According to the government, tax and credit subsidies were compensation for the overvalued exchange rate, the consequence of high tariffs that repressed the demand for foreign exchange. It was a kind of “second best” policy to correct the distortion of reduced imports. Export subsidies were a very explicit and transparent feature of export promotion, and a typical case of “unfair trade policy” that aroused increasing opposition from developed countries, mainly the United States. This policy was discontinued during the 1980s due to fiscal deficits and international trade disputes, and was replaced with two “maxi-​ devaluations” of the Brazilian currency in 1979 and 1983. The results were impressive: in 1964, manufactured exports represented 6.2% of the total, and in 1980, 44.8%, as shown in Figure 19.2. Another feature of trade policy during this period was the timid and short-​lived 1967 tariff liberalization reform, followed by a radical reversal in 1974, after the first oil shock. Between 1974 and 1976, import duties were doubled for consumer goods and a sharp increase occurred for intermediate and capital goods. In 1976, imports of 1,300 superfluous goods were “temporarily prohibited,” but this restriction remained effective for 14 years. There was also an extensive use of the “law of the similar” to ban imports. Import restraints were so binding that by the end of the 1980s the Brazilian economy was close to autarchy, with an import/​GDP ratio of 4.4%. Instead of using the conventional devaluation cum contractionary fiscal and monetary policy to face balance of payment deficits, the Brazilian government decided to postpone the adjustment and rely on external debt (with floating interest rates) to finance the adverse supply shocks and maintain a higher GDP growth rate, with an additional round of ISI in intermediate and capital goods. The literature points out that one of the consequences of this policy was the increase of the relative price of capital goods that became an important drag on further growth (see, for example, Bacha and Bonelli 2012).

Trade Policy from the 1930s to the Present    401 Table 19.1 Brazilian Economic Performance, 1967–​1980 Yearly Average Growth

1967/​1973

1974/​1980

GDP

11.1%

6.8%

Industrial GDP

13.2%

6.6%

Inflation (GDP deflator)

21.9%

49.7%

Current account (%GDP)

0.8%

–​4.5%

Foreign Debt, US$ billion,1973 and 1980

14.9

64.3

Source: Elaborated by the author from data provided by IBGE.

Prior to the oil and interest rate shocks, Brazilian economic performance had been outstanding, attaining the nickname of the “economic miracle”: between 1967 and 1973, the yearly average GDP growth was 11.1%, and industrial growth was at a record rate of 13.2% per year. Even with the oil shocks, the growth cum foreign debt strategy did work up to the end of the 1970s. GDP growth was 6.8% and industrial growth 6.6% per year, but balance of payment deficits reached an annual level of 4.5% of GDP in the period 1974 and 1980. Foreign debt skyrocketed from US$14.9 billion in 1973 to US$64.3 billion in 1980, as can be seen in Table 19.1. The real exchange rate was overvalued, and in 1979 the government abandoned the crawling peg rule with a maxi-​ devaluation, followed by prefixing the exchange rate correction in 1980, as a nominal anchor to control inflation. It did not work, and the country entered into a three-​digit level of annual inflation. After a second large devaluation in 1983, inflation jumped from 100% to 200% per year, and during the next decade the country was haunted by the risk of hyperinflation. The Cruzado Plan (1986) used a general price freeze, including the exchange rate, to curb inflation, and as a consequence foreign exchange reserves were exhausted and the government declared a unilateral default on foreign debt in 1987. In 1979, inflation reached double-​digit levels in developed countries, and then came the final act ending the period of cheap money for developing countries:  Volcker’s disinflation tight monetary policy—​followed by other developed countries’ central banks—​caused a deep world recession in the early 1980s, worsened terms of trade for commodities exporters, led to higher interest outlays on foreign debt, and, in September 1982, came the Mexican default. Brazil entered a phase of balance of payments adjustment under the International Monetary Fund (IMF) conditionality, economic growth vanished, and the country entered into the “lost decade.”

19.2.4. Import Liberalization Reform and a Comeback for Protectionism: 1988–​2015 By the end of the 1980s, high tariffs and NTBs became dysfunctional and thus one important impediment to investment, productivity growth, and competition in the

402   Simão Davi Silber Brazilian economy. Brazil was so isolated from the world market that clearly the first move had to be a unilateral opening up of the domestic market to foreign competition. The costs of near autarchy were very high, distorting resource allocation, reducing economic welfare, and hindering economic growth. Starting in 1988, a series of measures were undertaken reducing the level and the variance of import controls. Import duties were reduced, and the government agency for NTBs was discontinued (Kume, Piani, and Souza 1988). During six years tariffs fell sharply, but after 1995 no additional progress was made in terms of trade liberalization, and a reversal occurred after the mid-​1990s. The opening up of the economy was implemented in three steps: between 1988 and 1989 the reform eliminated tariff redundancy (water in the tariff), leaving untouched the NTBs controls by CACEX (the state agency responsible for administrative import controls); the period 1990–​1993 saw the elimination of the 42 special import regimes, quotas, and a list of 1,300 forbidden import items, while CACEX was discontinued and a new Tariff Law was approved, reducing import duties scheduled in a period of four years; finally, in 1994, after the inauguration of the Real Plan, there was one additional round of tariff reduction, anticipating the tariff levels of the Ouro Preto Protocol that instituted Mercosur’s common external tariff (CET). By the mid-​1990s, trade policy reform moved import controls to market instruments (tariffs and exchange rate) instead of using the unpredictable, informal, and nontransparent bureaucratic controls of previous decades prone to rent-​seeking activities. The results were impressive. Prior to reform, the most favored nation (MFN) effectively applied import tariff rates on manufactured goods was 54.9% (UNCTADSTAT), one of the highest in the world; in 2014, the tariff was one-​quarter of the initial level, reaching 12.7%, accompanied by the removal of important NTBs. Even though this represents one important step in trade reform, the resultant scheme still resembles the structure of the previous regime: high tariff rates by international standards and a large variation in nominal and effective protection, discriminating against imports of capital goods, essential for productivity growth (Moreira 2009). The same sectors that had the highest protection at the peak of import substitution remain sheltered from international competition. There is plenty of empirical evidence on the effects of the Brazilian trade reform on productivity growth and competitiveness. Hay (2001), using panel data for the period 1986–​1994 for large manufacturing firms, found important total factor productivity (TFP) gains, and reductions in market share and profits. The supply shock from import liberalization forced firms to improve their efficiency. Cavalcante and Guillén (2002) using panel dada for 16 industrial sectors in the period 1985–​1997, found that trade reform increased average TFP in the majority of Brazilian manufacturing industry sectors, although they did not find significant changes in domestic market power associated with increased import competition. Productivity growth was attributed to access to imported inputs and technologies, not to competitive pressures from abroad. Córdoba and Moreira (2003) concluded that total factor productivity increased in the

Trade Policy from the 1930s to the Present    403 manufacturing sectors of Brazil and Mexico after the unilateral and regional trade integration. With opposite results, Muendler (2004), using firm level statistics for the period 1986–​1998, found that competition in product markets had an important role in boosting TFP growth, but access to inputs in world markets and the elimination of inefficient firms had no significant effect on productivity growth. Tyler and Gurgel (2009), using a computable general equilibrium model to simulate the effect of the trade liberalization of the 1990s, found favorable effects: an increase of 1.9% in welfare, measured by the increase in consumption, a 9.1% real exchange rate devaluation, and income redistribution in favor of Brazilian relative abundant natural resources. The model also simulated increases in trade flows and changes in production composition in the direction of agriculture, mineral products, and light manufacturing, sectors where the country’s comparative advantage lies. But Brazilian trade liberalization is rather limited when compared with the average world pattern, and the relative price of capital goods is still high. The following statistical data illustrate the challenges facing trade policy for the future. Compared with a large sample of countries, Brazil stands as a country of high nominal tariffs. The weighted manufactured goods tariff of 12.7% in 2014 appears in the upper quartile of the distribution, well above (72%) the average tariff (UNCTADSTAT). It is worth mentioning that the present Brazilian tariff is above the 2006 level, and the last 20 years have seen no further effort to liberalize trade and reap the advantages of better resource allocation and productivity growth. Macroeconomic imbalances since the mid-​1990s, linked with overvalued exchange rates and international financial turbulence, gave new strength to protectionist demands and tariff hikes, and NTBs were implemented. The Brazilian anti-​dumping and unfair trade practices agency became very active. According to the latest DECOM (2015) report, in the period ranging from 2005 to 2015, 274 anti-​dumping investigations have been conducted, and in 58% of the cases countervailing duties have been imposed. As pointed out by Tavares and Miranda (2008), in most cases the decisions protected domestic monopolies or oligopolies. According to the World Trade Organization (WTO), there are only two countries in the world with a higher number of anti-​dumping investigations in 2016:  India and the United States. Similar results appear for unfair trade practices (subsidies) and safeguard measures, protecting domestic firms with monopoly power. There is a large variation in nominal and effective protection, according to the latest empirical calculations as given by Castilho (2015). The nominal tariff ranges from 0% for oil and natural gas to the level of 32% for trucks and buses, but the great variance occurs with the effective tariff rate. It ranges from –​3.1% for oil and natural gas to 132.7% for trucks and buses. The average nominal tariff in 2014 was estimated at 12.2%, and the effective rate at 26.3%. Compared with the tariff levels of 1988, where the average nominal tariff was 38.5% and the effective rate was 50.4%, there was a significant reduction in tariff levels, but there is room for further trade liberalization.4 There is no economic rationale for the present tariff structure. Why protect the truck and bus

404   Simão Davi Silber industries with an effective rate of 132.7% and penalize the production of alcohol and oil with negative rates? The only answer is the vested interests of protected sectors consolidated in the tariff structure and in NTBs. But the costs for the country are high in terms of distorted resource allocation, employment, welfare loss, and reduced rates of growth. The high protection given to the machinery and equipment sectors is one important burden for investment in Brazil. There is convincing empirical evidence showing that high relative prices of capital goods have a negative impact on a country’s investment and growth rates. Using a selected sample of developing countries (UNCTAD 2016), it is possible to point out that the Brazilian applied weighted tariff on machinery is relatively high. The Brazilian tariff of 13.84% in 2014 is the highest of representative countries of Latin America, Asia, and the BRICs (Brazil, Russia, India, and China).5 One stylized fact of the Brazilian economy is the low rate of investment. In the last five years the average gross investment/​GDP rate was 19.2%, and there is no sound economic argument to penalize investment that is badly needed to increase productivity and growth. Brazil was one of the latest countries to embark on trade liberalization and did not go far enough from its autarchic model, taking as a yardstick the average degree of trade openness of developing countries. After the unilateral trade reform and the MERCOSUR initiatives of the beginning of the 1990s, no significant move can be identified in terms of trade policy liberalization. Domestic and international turbulence, inconsistent domestic macroeconomic policies with recurrent exchange rate overvaluations, the strengthening of protectionist lobbies, and a different government strategic view after 2003 can all be invoked to explain the paralysis in trade policy reform and the increase in protectionism. After the 2008 financial crisis there was a strange increase in technical barriers to trade6 and the definition of a “new industrial policy” with preferences for domestic suppliers in government procurement (overprice up to 25%); tax rebates for the auto industry to increase national content; ambitious national content requirements (65%) in “Pre-​Salt” investments; BNDES credit subsidies for capital goods and selected “national winners” in industrial targeting. Particularly, there was a revival of South-​South integration initiatives. The abandonment of trade negotiations in the Western Hemisphere and with the European Union eliminated the possibility of preferential access to high per capita income markets. The trade agreements with India, the Southern African Customs Union, and Latin American countries were very limited in scope and did not significantly increase bilateral trade flows. Even though it is not possible to blame the retreat of liberalization as the only factor explaining poor economic performance, recent years have witnessed a GDP growth slowdown without parallel in 115 years. In the period 2011–​2015, average annual GDP growth was down to 1% and industrial output fell by 3.2%, on average, per year. Balance of payment deficits increased to 4.3% of GDP in 2014, and a huge nominal devaluation in 2015 (50%), coupled with a deep recession, reduced foreign deficit.

Trade Policy from the 1930s to the Present    405

19.3.  Brazilian Trade Performance and the Exchange Rate It is a stylized fact that the degree of openness of the Brazilian economy has been historically low, and during the 1970s and 1980s the country followed the opposite path of the rest of the world, reducing the share of its trade in GDP (UNCTAD 2016b). Following trade liberalization there was a gradual increase in export and import flows. After the first oil shock of 1973, the share of imports declined steadily up to the end of the 1980s. Tight import controls and an additional round of ISI in intermediate goods, petrochemicals, and capital goods explain the path to near autarchy. Imports fell to 6% of GDP, one of the lowest of the world and incompatible with any criteria of comparative advantage. Since 1990, imports have had an increased share in domestic supply, and this is explained by the opening up of the economy. With a decade lag, exports followed the same downturn trajectory, being reversed only in the first decade of the twenty-​first century. Exports fell from 14% to 7% of GDP as a consequence of overvalued exchange rates, associated with various attempts to control inflation, the discontinuity of tax subsidies to exports, and isolation of world-​class technologies embodied in imports and FDI flows. In 1999, Brazil adopted a floating exchange rate regime followed by three large devaluations of the real (1999, 2001, and 2002), and this was the first factor behind export expansion. Compared with the pattern globally, Brazilian insertion in world markets is still modest, calling for additional steps in trade liberalization policies. In the last half-​century, international economic integration advanced steadily. In the early 1960s, exports represented 12% of world GDP. With trade liberalization through multilateral and regional agreements, exports now represent 30% of world GDP. Brazil, however, did not follow the same track. In 2015, the Brazilian export/​GDP ratio was similar to the world ratio of 1973. Taking a long-​run perspective, the price competitiveness of Brazilian exports has fluctuated widely, and this had been a hindrance on exports, mainly of manufactured goods. During the period 1985–​2015, the real exchange rate became several times highly overvalued as a consequence of various attempts to use the exchange rate as an anchor against inflation and large inflows of foreign capital attracted by high domestic interest rates. With a 33% appreciation of the real exchange rate during the first period of the Real Plan, balance of payment deficits increased sharply and the country was on the verge of an additional default on foreign debt in 1998. In January 1999, after an acute speculative attack against the real, the Central Bank had no choice but to move to the floating exchange rate regime. After 1999, there were four large real devaluations of the Brazilian currency, but after 2004 the exchange rate appreciated again up to 2014. There are two major factors explaining this exchange rate behavior in the short run. The first one is Brazilian sovereign risk. Since the 1980s, inconsistent macroeconomic policies led to increased foreign and government debts, and risk fluctuated wildly,

406   Simão Davi Silber determining the path of the exchange rate. The huge devaluations of 1999, 2001, 2003, and 2014 are explained by risk averse-​behavior with respect to Brazilian financial assets. With the consolidation of three pillars of the macroeconomic policy (inflation targeting, floating exchange rates, and fiscal responsibility) Brazil reached in 2008 the status of “investment grade” and risk and the exchange rate embarked on a consistent downturn up to 2014. After 2011 macroeconomic policy became increasing inconsistent with the so called New Macroeconomic Matrix, based on expansionary fiscal and monetary policies, and followed by higher inflation, price and exchange rate controls, and huge balance of payment and fiscal deficits. This macroeconomic disarray led to the “Great Recession” of 2015–​2016, when GDP fell by 7%, as opposed to the 5.3% reduction of 1930–​1931, during the Great Depression. Brazil lost investment grade status in 2015, and the government is proposing fiscal adjustment and reforms to avoid an explosive path of government debt. There is a major chance that the country is entering a new “lost decade.” The second factor to explain lower exchange rates is the inappropriate mix of macroeconomic policy. It is a well-​known result in macroeconomic models that—​with floating exchange rates and integrated financial markets—​the combination of expansionary fiscal policy with restrictive monetary policy produces higher equilibrium interest rates, lower exchange rates, and balance of payment deficits. Price competitiveness in world markets depends crucially on the exchange rate, and the present macroeconomic policy represents an important constraint to export performance. The main factors behind the export recovery after 1999 were exchange rate devaluations, rapid world demand expansion, and the sharp increase in commodity prices associated with the emergence of China as a major player in world markets. Finally, structural reforms and support policies by government agencies must be mentioned as favorable factors affecting export performance. Import liberalization and privatization were responsible for important productivity gains, increasing the competitiveness of Brazilian companies. Cases like Embraer and Vale are vivid examples of success in world markets after privatization. The role of the state agricultural research body EMBRAPA (see, in this volume, Bacha, Chapter 13; Martha and Alves, Chapter 15) is essential to understanding the increased share in world agricultural markets through innovations in seed varieties and production techniques. But the Brazilian share in the world’s exports increased only modestly. According to UNCTAD (2016) data, in 1981 the Brazilian share in world exports was 1.15%; in 2015 it reached 1.16%. Increasing market share is a measure of dynamic export competitiveness, and by this criteria Brazil offered a dismal performance. The following analysis covers the last three decades, the period of import liberalization, stabilization of inflation at lower levels, changes in the exchange rate regime, and increased world demand for commodities that lasted up to 2011. Over this period, both the product composition and direction of Brazil’s trade have experienced considerable changes, and the long-​run response of exports and imports to the exchange rate has varied according to product types. The stylized facts that emerge are as follows.

Trade Policy from the 1930s to the Present    407 Table 19.2 Composition of World Exports (Percentage), 2014

Africa

Agricultural Products

Fuels and Mining Products

Manufactures

11.5

62.9

21.3

Asia

6.7

11.3

80.0

Brazil

39.0

24.4

33.3

Europe

10.6

11.1

74.8

Middle East

2.3

64.9

20.7

North America

11.1

16.7

67.6

South and Central America

30.6

39.9

25.5

9.5

20.5

66.2

World Source: WTO (2016).

The volume of world exports grew at an annual rate of 5.1% and Brazilian exports at 4.8%. In nominal terms, world exports expanded 8.1% per year and the Brazilian exports by 7.8%. The composition of world exports has evolved gradually in the direction of manufactured goods, but the Brazilian trade structure moved in the opposite direction during the last 15 years. The present structure of world exports can be depicted by the data of Table 19.2. The share of Brazilian agricultural exports are almost four times the world average; on the other side, manufactured exports are 50% below the world average. There is no region of the world with a higher share of agricultural exports than Brazil. The proportion of manufactured goods exports has declined during the last decade, with an increased share of primary products. The composition of Brazilian exports has changed in the direction of primary products (see Silva e Silva, Chapter 31 in this volume). At the end of the 1990s, the share of commodities was 22.8%; in 2015, it had increased to 45.6%. The main exports of this group are iron ore, petroleum, meat, soybeans, sugar, and coffee. The opposite move was observed in manufactured goods, with its share diminishing from 59% to 38.1%. In this group the main exports are vehicles, iron and steel, airplanes, organic chemicals, and machinery. The changes in Brazilian exports are shown in Figure 19.2. With respect to imports, the opening up of the economy substantially increased the share of manufacturing, from 57% in 1989 to 84.4% in 2015. The more important changes in the composition of Brazilian trade were, on the export side, the increased share of commodities and the reduced share of manufacturing in export flows. In the case of commodities, increased world demand and comparative advantage of the Ricardian type are the factors behind the observed tendency in the last two decades. The declining share of manufacturing is explained by the increasing competition from Asian countries and overvalued exchange rates. According to UNCTAD (2002), primary products and resource-​based manufacturers have lost share in world

408   Simão Davi Silber exports during several decades, and in those sectors Brazil had increased export concentration. For developing countries as a whole, the fastest-​growing export was of high technology sectors, where Brazil had a minor share of its exports. The UNCTAD study defineda “winner” in world markets as a country that increased its share in world markets for manufactured goods. The striking result during the period 1985–​2000 is that Brazil does not appear once among the top 20 “winning” countries for all categories of manufacturing industries. The positive results for commodities are associated with the rapid growth of trade during the last decade, the effect of China in particular, and domestic supply factors like the rapid increase in productivity associated with import liberalization, privatization (Embraer and Vale), and new technologies developed by EMBRAPA. Brazil has not been able to change its comparative advantage toward the new technological sectors. To increase competition of manufactured goods, it is necessary to reduce the level of import protection to enhance modernization and efficiency gains, particularly diminishing tariffs on imported capital goods. Another important task is the elimination of the domestic tax burden on exports. Brazil has a complex tax system with a high tax burden (33% of GDP in 2015), imposing large administrative costs and indirect taxation on exports, and this is a significant constraint on Brazilian firms’ competitiveness in world markets. The current policy of rebates for state and federal value added tax is not enough to eliminate taxes levied on exports being part of the so-​called Brazilian cost. It is also fundamental to control government current expenditure to commit higher government resources to investment in education, research, and infrastructure essential to the competitiveness of the industrial sector in the long run. Brazil has one of the more diversified industrial bases among developing countries, but is not able to reap this advantage in world markets due to the heavy burden of inappropriate government policies. Two final subjects on trade issues: the first one is regional market diversification, and the second is the outstanding agribusiness performance. One distinguished feature of recent export performance has been the increasing sales to new markets, as pointed out by Bonelli and Pinheiro (2007). All developing regions increased purchases from Brazil, as can be seen in Table 19.3. The share of MERCOSUR more than doubled between 1989 and 2015, as a consequence of preferential regional agreements and lower transport costs associated with lower distances and infrastructure improvements.7 Developed countries declined in importance as major trade partners, the United States in particular. In 1989, 65.3% of Brazilian exports went to the European Union, the United States, and Japan; this share fell to 32.7% by 2015. The major change is the emergence of China as the major trade partner. Brazil is a “global trader.” Brazilian exports are distributed among all regions of the world. This does not come as a surprise. The size of the country, different climates (ranging from tropical to temperate), a large endowment of natural resources and labor, and a highly diversified industry all influence the composition and regional distribution of Brazilian exports. Another distinctive feature of the present Brazilian trade profile is the regional specialization of exports. Manufactured exports are dominant in only two regions: Latin

Table 19.3 Destination of Brazilian Exports (Percentage) 1989

2015

Basic

Semi-​ manufactured

Manufactured

16.5%

13.8%

69.7%

6.4%

6.5%

87.1%

14.2%

4.0%

China

11.7%

Japan

40.4%

United States Middle East European Union

Africa ALADI excluding MERCOSUR MERCOSUR

Russia total

Basic

Semimanufactured

Manufactured

2.8%

36.7%

23.1%

40.2%

4.4%

6.4%

17.0%

4.3%

78.7%

7.7%

81.8%

4.0%

12.2%

2.1%

85.6%

9.6%

27.5%

60.8%

1.8%

80.4%

13.2%

6.4%

19.1%

34.6%

24.9%

7.2%

66.8%

18.4%

14.8%

2.6%

10.7%

11.6%

77.7%

24.2%

19.2%

19.1%

61.7%

12.0%

26.7%

11.1%

62.2%

3.3%

61.6%

9.9%

28.5%

0.5%

46.9%

15.4%

37.8%

33.9%

48.7%

16.2%

35.1%

18.1%

0.0%

0.0%

0.0%

0.0%

73.7%

15.0%

11.3%

1.3%

28.1%

17.1%

54.8%

46.8%

14.2%

39.0%

Source: Ministry of Development, Industry, and Trade statistics.

% Total Exports

% Total Exports

410   Simão Davi Silber America and the United States. With the exception of MERCOSUR, in all other markets the country’s exports followed the uniform pattern of reduced importance of manufactured goods. This is the result of one important pitfall in the Brazilian trade policy strategy. When multilateral trade negotiations waned, the country moved to South-​South agreements with minor effects on trade flows. The international trade specialization with developing countries is limited by market size and similarity in factor endowments. A smaller market does not permit exploration of the comparative advantage of a larger array of manufactured products and the benefits of economies of scale. Similar factor endowments impose very high adjustment costs for a full-​blown free trade agreement. The adjustment costs of a bilateral agreement with India or China would not be trivial in labor-​and capital-​intensive industries.8 On the other hand, the absence of preferential access to developed countries’ markets explains the falling share of Brazilian exports in those markets. Developed countries have been very active in regional trade agreements. The North American Free Trade Agreement (NAFTA), the European Union enlargement, and their agreements with Latin American countries explain, at least partially, the eroded Brazilian position in developed countries’ markets. The competition in the manufactured goods world market increased sharply in the last three decades with the entry of new and major players. China, with its aggressive trade and exchange rate policies, is the great novelty, becoming in one generation the largest exporter in the world. This is a challenge facing Brazil imminently: how to increase market access in a segmented world through regional agreements, new competitive players, and laggard multilateral trade negotiation to open agricultural markets. One distinctive aspect of the Brazilian insertion in world trade is the outstanding performance of agribusiness. From the early 1960s, the share of agribusiness in total exports fell steadily as a consequence of the import-​substitution policy and government export incentives for manufactured goods exports. In 1990, agribusiness exports were reduced to 28% of total exports and remained at this level up to 2005; in the past decade it has increased to 46.2% (2015). Agribusiness has always been an export-​oriented sector, and presently its share in exports is five times bigger than its share in GDP. There are several factors behind export growth during the past decade: the high growth rate of world demand; price increases of commodities in world markets; increased competitiveness of agribusiness; regional diversification of exports; and the elimination of the anti-​export bias of trade policy after import liberalization. Agribusiness’s exports upsurge has had two basic characteristics: changed composition of exported products, and regional diversification in the direction of developing country markets. With the increased importance of meat and sugar, today agribusiness exports are diversified within tropical products, temperate products, and meat. During this decade there was a relative decline of exports to developed countries, and developing country markets are becoming increasingly important. Within developed countries, the reduction was across the board: diminishing share to all relevant markets; for the developing countries, China appears as the most dynamic market for agribusiness exports (see

Trade Policy from the 1930s to the Present    411 Table 19.2). There is not a unique factor to explain regional diversification, but clearly rapid per capita growth in Southeast Asia and highly protected markets in the developed countries can be pointed to as some of the key factors explaining the observed tendency in agribusiness trade. The “China effect” and Vale’s privatization also explain Brazil becoming a large exporter of iron ore. One stylized fact of the world economy after World War II was the diminishing share of agriculture in world production and trade. The opposite movement happened with Brazilian agriculture. In 1990, Brazil had 2.4% of the world’s agriculture trade; by 2014, the Brazilian share was double that at 5%. Brazil is the third-​largest exporter in the world, behind the European Union and the United States. Brazil is now a “major player” in several markets, ranking as the largest exporter for orange juice, sugar, poultry, coffee, and second place in soybean products, corn, and beef. The degree of openness of the agribusiness sector to exports is the largest of the country. According to recent FIESP (2016) data, Brazilian industry exported 17.3% of its production in 2015, clearly inferior to the agriculture sector figures of 37.7% in the same year. The increased competitiveness of Brazilian agriculture has depended on a series of factors. The most important to explain output expansion was productivity growth. There are several empirical studies showing that total factor productivity growth was impressive. Avila and Evenson (2005) data show that TFP in Brazilian agriculture, in the period 1981–​2001, grew on average 3.2% per year—​the largest among all developing countries. Gasques (2006) estimates that between 1975 and 2005 total farm productivity doubled. Productivity growth is linked to government policy decisions to invest in EMBRAPA’s research in new agricultural technologies, and also to investment in human capital. Public universities graduated a large number of forestry engineers and veterinarians necessary for high-​tech production in agriculture. The main competitors in agriculture world markets are the developed countries, and their exports depended on subsidies, modern agriculture technologies, and large-​scale production. As the world’s third-​largest exporter, Brazil was able to overcome unfair trade practices and became a major player in the world agriculture market. Trade liberalization had also a positive impact in agribusiness export performance. In the period of import substitution, the agriculture sector was penalized by trade policies through overvalued exchange rates, tariffs in final products and imported inputs, quotas, licensing, and domestic taxation of exports. Trade reform during 1987–​2004 eliminated export licensing and quotas, reduced tariffs and NTBs for agriculture imports, and in 1996 the value added tax on exports of agriculture and semi-​manufactured goods was abolished. During this period the domestic government program of credit subsidies and price support was phased out, forcing the agriculture sector to compete in an open world market environment. Finally, the rapid increase in international demand should be mentioned to explain the success of agricultural trade. The results have been impressive, and the agribusiness sector was responsible for changing the Brazilian trade balance during the last 25 years. In 1989, agribusiness trade surplus reached US$11 billion, or 67% of total trade surplus; in 2015, with a trade surplus of US$88.2 billion, the agribusiness sector was the only

412   Simão Davi Silber sector with trade surplus and compensated the trade deficits of the manufacturing and services sectors. A summing up of Brazilian trade performance:  the country continues as a global trader, with an increasing relative share of new developing markets, a decreasing share of manufactured goods exports, and an increasing share of agribusiness and mining. The main factors to explain export performance are productivity growth, economic reforms, diversification of markets, and natural resources. Brazil experienced an export boom after 2002, with two main characteristics: new markets, and a change in export composition.

19.4.  The Balance of Payments and the Real Exchange Rate in the Long Run With the advantage of hindsight and almost a century of economic literature and statistical data, it is possible to give a comprehensive explanation of the links between the current account, capital flows, and the real exchange rate. For the balance of payments, the picture that emerges is recurrent and long-​lasting current account deficits, as shown in Figure 19.3.

8.00% 6.49%

6.41%

6.00%

6.38% 4.94%

3.96% 3.23% 3.48%

4.00%

2.52% 3.78%

4.38%

–4.00%

1.73%

–0.34%

–0.52% –1.30% –1.55% –2.46% –1.58% –2.82%

–2.02%

–1.74% –2.85%

–1.13%

–2.07%

–2.80%

–1.67%

–3.77% –4.72%

–6.00% –6.47%

–6.01%

1930 1932 1934 1936 1938 1940 1942 1944 1946 1948 1950 1952 1954 1956 1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

–8.00%

3.02%

1.59% –1.30%

0.00%

4.27% 4.52%

3.47%

1.94% 1.43% 2.00%

–2.00%

4.84%

Current account balance

Net capital inflow

Figure 19.3.  Brazilian balance of payments, 1930–​2015 as percentage of GDP. Sources: IBGE and Brazilian Central Bank.

Trade Policy from the 1930s to the Present    413 Given the structural low level of domestic savings, Brazil used extensively foreign capital to finance domestic investments. Balance of payment surplus occurred only in exceptional periods such as the Great Depression and World War II, when capital inflows collapsed, or during the early 1970s and the first decade of the 2000s, when commodities prices boomed with a very favorable external environment. But the dominant feature was four long cycles of balance of payment deficits, followed by increased foreign debt and crises. The largest came after the oil shocks of the 1970s, followed by the debt crisis and default of the 1980s: current account deficits peaked at 6% of GDP, financed by a similar size of inflows of petrodollars. During the Real Plan, with a fixed nominal exchange rate for four and a half years, came another period of large deficits, and Brazil almost went into foreign debt default again. Starting in 2008 came the most recent period of increased balance of payment deficits, followed by a large devaluation of the Brazilian currency in 2015. This chapter has stressed that the Brazilian government has always been cautious with exchange rate devaluations due to the impacts on terms of trade, inflation, and the government budget. How can the large real devaluations be explained in the long run? The answer is the foreign budget constraint. Current account deficits over time increase foreign liabilities that have to be served in foreign currency. This is the famous “original sin” in international economics theory. This cash flow problem is depicted in Figure 19.4 through the gross foreign debt/​export ratio and the real exchange rate. 5.0

180.0 160.0

5.0

140.0 4.0

120.0 100.0

3.0

80.0 2.0

60.0 40.0

1.0

20.0 0.0

1920 1922 1924 1926 1928 1930 1932 1934 1936 1938 1940 1942 1946 1948 1950 1952 1954 1956 1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014

0.0

BRL$/US$ Real exchange rate

Gross external debt/export ratio

Figure  19.4.  Index of the real exchange rate and Brazilian gross foreign debt/​export ratio, 1920–​2015. Source: IBGE and Brazilian Central Bank.

414   Simão Davi Silber As foreign debt increased, a higher proportion of export revenues was committed to interest payments and other foreign exchange remittances, like profits and dividends. In the Brazilian case, when the foreign debt/​export ratio exceeded 3 (right-​side scale in Figure 19.4) it became increasingly difficult to serve foreign liabilities, and the country became more vulnerable to terms of trade and capital inflow reversals that triggered large exchange rate devaluations to correct balance of payments disequilibria. The statistical correlation between the two statistical series is 0.53% over a period of 95 years.

19.5.  Concluding Remarks Historically, Brazilian trade and investment policies have been characterized as highly protectionist to foster inward industrialization growth. A large array of tax and credit subsidies for industrial investments and high tariffs and NTBs protected the domestic market from import competition, leaving it for export-​promotion policies to mitigate the anti-​export bias of trade policies. Even the liberalization attempt of the late 1980s and early 1990s was partial and insufficient to change the Brazilian development model’s focus on the domestic market. Brazil is almost an autarchy in terms of trade flows. During this period, Brazil implemented important trade policy changes, going from ISI to export promotion, trade liberalization, and ultimately a return of protectionism in the last two decades. Booms and busts in the world economy are important factors shaping Brazilian trade policy in the twentieth and twenty-​first centuries. Protection to foster industrial growth has been of paramount importance. Most of the time, trade policy was inward looking with a strong anti-​export bias due to overvalued exchange rates and strong incentives to divert production into domestic absorption and industrial growth. The results were impressive during the half a century from 1930 to 1980, when industrial average annual growth was 8.1%. But in the last 35 years, industrial growth has almost vanished, to a dismal annual rate of 0.7%. There was no room for additional stages of inward-​looking ISI based on shifting underemployed supply of cheap labor from agriculture to the industrial sector, typical of the Lewis development model. Brazil is a typical case of a country caught in the “middle-​income trap,” with an annual per capita growth rate of 1% since 1980. By this pace, per capita income will double after 70 years. Trade policy alone cannot be blamed for this poor performance, but the country is still very isolated from the world market, and trade liberalization is one of the big challenges for the near future. The benefits of the trade and investment liberalization of the 1990s were blunted by international crisis and inconsistencies in domestic macroeconomic policies. After 2002, the world economy boomed for six years, with important consequences for the Brazilian foreign sector. The country’s exports quadrupled in nominal terms, with important composition and regional changes. The benefits of more trade and FDI

Trade Policy from the 1930s to the Present    415 were clear, and the country moved to a higher growth rate plateau. In a certain way, the twenty-​first century resembles the nineteenth century. The similarity is that the most dynamic economies had a scarce endowment of natural resources, and their growth irradiated to the world by increasing demand for staples, oil, and mining products, with very favorable effects on Brazilian exports. This was the case with Great Britain in the nineteenth century, and is the case with China today. This period also witnessed the resurgence of world FDI. By the end of the 1980s, FDI was less than 1% of world GDP and rose to 4.7% in the period 2012–​2014. In 2014, Brazil ranked sixth in world FDI stock. During this period, Brazil implemented important economic reforms, ranging from trade liberalization, the stabilization of inflation, an aggressive privatization program, to moving to the floating exchange rate regime. Trade and foreign investment policies have always been decisive in Brazil, given the frequent balance of payments crises and their restrictive impacts on economic growth. In which direction has trade reform and international demand changed the composition of trade? In which sectors has Brazil been more competitive in world markets? What were the consequences of “getting the fundamentals right” on trade flows? What is the direction of trade policy reform? These are the some key questions to be answered. By the middle of the second decade of the twenty-​first century, trade policy is still inward looking: the country presents an export/​GDP ratio of 13% and a modest share of 1.16% of world trade, similar to 40 years ago. According to OECD (2015) statistics on foreign value added share in gross exports, Brazil stands with the second-​smallest share in the world of foreign value added in exports, higher only than Saudi Arabia. Brazil implemented an important trade reform during the 1990s, compared with its historically inward-​looking policies. Tariffs and NTBs were removed, new opportunities were open for FDI, and the country became more integrated in the world economy. A more open economy was able to reap the advantages of international specialization, and the gains in efficiency, productivity, and competitiveness in the Brazilian case are well documented empirically. The structure of foreign trade moved in the direction of the increasing share of commodities in exports and high technology manufactured goods in the import side, reflecting present comparative advantages. Three decades after of the beginning of economic reforms, Brazil still faces important unsettled issues in trade. The first challenge for policymakers is to choose the appropriate mix of macroeconomic policies. Price competitiveness in world markets depends crucially on the exchange rate. The present combination of loose fiscal policy with restrictive monetary policy produces higher equilibrium interest rates, lower exchange rates, and balance of payment deficits. A more restrictive fiscal policy and an expansionary monetary policy will be suitable to promote exports through a devalued exchange rate and lower interest rates without jeopardizing inflation control. It is also fundamental to control government current expenditure to commit higher government resources to investment in education, research, and infrastructure, essential to competitiveness in the long run. This has been an important constraint on export performance during the three previous

416   Simão Davi Silber decades, and is one important part of the “Brazil cost” (overvalued exchange rate, high tax burden and interest rates, and costly infrastructure). One additional side effect of a sound fiscal policy is to improve the business environment for investment, attracting both the domestic and foreign capital essential for brighter growth prospects. Higher domestic savings are fundamental to avoid the recurrent balance of payment crisis that constrained Brazilian growth in the post–​World War II period. Important fiscal controls have been approved by Congress in 2016: an amendment to the Constitution will restrict government expenditure growth over the next 20 years and the social security reform is being seriously discussed in the Congress today. New rules for concessions in infrastructure and Pre-​Salt could boost investment in the near future. With a more restrictive fiscal policy and inflation close to the target, there is room for reducing the policy interest rate and a more devalued currency. If the United States implements an expansionist fiscal policy with the Federal Reserve increasing interest rates, there will be an additional pressure for US dollar strength. Even though nominal and effective tariffs had been reduced, by international standards, they are comparatively high and with a large variance. There is no economic rationale for the present tariff structure. Nominal tariff rates range from zero to 32%, and the effective exchange rate from –​3.1% to 132.7%. The only explanation for this outcome is the strength of special interest groups consolidated in the tariff structure. High tariffs on capital goods penalize investments, badly needed in a country where the investment rate is low. The costs of the present tariff structure for the country are high in terms of distorted resource allocation, less employment, welfare loss, and reduced rates of growth. The direction of tariff reform is clear: reduce the level and the variance of the present tariff schedule. Ideally, a Chilean-​style homogeneous tariff would eliminate privileges and rent-​seeking activities. An additional advantage of a reduced tariff rate is to increase the demand for imports contributing to the devaluation of the exchange rate, reinforcing the results of the previous proposal. It should be noted that the proposed round of trade liberalization is a completely different case compared with the opening process of the 1990s. Then, the country was so isolated from the world market that the first task depended only on domestic will and the move for more trade was unilateral. Now, market access is of paramount importance. This leads to the final challenge: market access at the regional level. During this decade Brazil moved to South-​South agreements with minor effects on trade flows. The international trade specialization with developing countries is desirable but is limited by market size and similarity in factor endowments. On the other hand, the absence of preferential access to developed countries’ markets explains the falling share of Brazilian manufactured exports to those markets. Developed countries have been very active in regional trade agreements. NAFTA, the European Union enlargement, and their agreements with Latin American countries eroded the Brazilian position in developed countries’ markets. This is one of the challenges facing Brazil in the near future: how to increase market access in a world that is segmented by regional agreements, new competitive players, and laggard multilateral trade negotiation to open agricultural markets.

Trade Policy from the 1930s to the Present    417

Notes 1. The BRL$/​US$ real exchange rate was calculated using data from IBGE (1990 and 2006) and FED (2016). The deflators chosen were GDP deflator for the Brazilian economy and CPI for the United States. 2. Cr$ was the symbol of the Brazilian currency, the cruzeiro, used in the period 1942–​1967. 3. Acronym for the government agency for money, credit, and exchange rate guidelines created in 1945, a precursor of the Brazilian Central Bank opened in 1964. 4. According to Kume’s (1996) calculations for 1988. 5. For capital goods without a domestic counterpart, there is the possibility to reduce the tax burden through a bureaucratic and nontransparent import regime called ex-​tarifário. 6. Probably the most bizarre recent technical barrier to trade is the globally very rare three-​pin socket for electrical appliances in force since 2010. 7. MERCOSUR signed trade agreements with Chile and Bolivia (1996), the Andean Community (2003), India (2004), Southern African Customs Union (2008), and Lebanon and Tunisia (2014). Brazil and Mexico (2002) signed a bilateral agreement within the LAIA initiative granting free trade status for 800 tariff lines, even though trade is heavily concentrated in autos and auto parts. 8. Silber (2004) finds that trade liberalization with India would increase Brazilian imports of apparel, textiles, chemicals, and pharmaceutical products. DECOM (2015) reports that—​in the two previous decades—​26% of all anti-​dumping countervailing duties were applied on Chinese exports of manufactured goods.

References Abreu, M. P., and A. S. Bevilaqua. 1996a. “Brazil as an Export Economy 1880–​1930.” Texto para discussão # 363. Departamento de Economia PUC-​Rio. Available at http://​www.econ.puc-​ rio.br/​biblioteca.php/​trabalhos/​show/​1233. Abreu, M. P., A. S. Bevilaqua, and D. M. Pinho. 1996b. “Import Substitution and Growth in the Brazilian Economy: 1890s–​1979s.” Texto para discussão # 366. Departamento de Economia PUC-​Rio. Available at http://​www.econ.puc-​rio.br/​biblioteca.php/​trabalhos/​show/​1236. Bacha, E., and R. Bonelli. 2012. “Accounting for the Rise and Fall of Post-​WW-​II Brazil’s Growth.” Institute for Economic Policy Studies/​Casa das Garças. Available at http://​iepecdg. com.br/​wp-​content/​uploads/​2016/​03/​120630BachaBonelli.pdf. Bonelli, R., and A. C. Pinheiro. 2007. “New Export Activities in Brazil: Comparative Advantage, Policy or Self-​Discovery?” IPEA, texto para discussão No. 1269a. Available at http://​www. ipea.gov.br/​sites/​000/​2/​publicacoes/​tds/​td_​1269a.pdf. Brazilian Central Bank (BCB) Statistics. 2016. Available at https://​www3.bcb.gov.br/​sgspub/​ localizarseries/​localizarSeries.do?method=prepararTelaLocalizarSeries. Cardoso, E., and A. Fishlow. 1990. “Trade Policies and Consequences.” Chapter 6 of Developing Country Debt and Economic Performance, Vol. 2: The Country Studies: Argentina, Bolivia, Brasil, Mexico, edited by Jeffrey D. Sachs, 335–​351. Chicago: University of Chicago Press. Available at http://​www.nber.org/​chapters/​c8949. Castilho, M. F. 2015. “A estrutura recente de proteção nominal e efetiva no Brasil.” Federação das Indústrias do Estado de São Paulo e Instituto de Estudos para o Desenvolvimento Industrial.

418   Simão Davi Silber Available at http://​www.fiesp.com.br/​indices-​pesquisas-​e-​publicacoes/​a-​estrutura-​recente-​ de-​protecao-​nominal-​e-​efetiva-​no-​brasil/​. Federação das Indústrias do Estado de São Paulo. 2016. “CEI: Coeficientes de exportação e importação da indústria brasileira.” Available at http://​www.fiesp.com.br/​indices-​pesquisas-​ e-​publicacoes/​coeficiente-​de-​exportacao-​e-​importacao/​. Federal Reserve System. 2016. “Consumer Price Index for All Urban Consumers—​Not Seasonally Adjusted 1913–​2016.” Available at https://​fred.stlouisfed.org/​series/​CPIAUCNS. Hay, D. A. 2001. “The Post-​1990 Brazilian Trade Liberalization and the Performance of Large Manufacturing Firms:  Productivity, Market Share and Profits.” The Economic Journal 111 (July): 620–​641. Available at http://​www.jstor.org/​stable/​2667950?cookieSet=1/​. International Monetary Fund World Economic Outlook Database. 2016. Available at https://​ www.imf.org/​external/​pubs/​ft/​weo/​2016/​01/​weodata/​index.aspx. Instituto Brasileiro de Geografia e Estatística (IBGE). 1990. “Estatísticas históricas do Brasil—​de 1550 a 1988.” Available at http://​biblioteca.ibge.gov.br/​visualizacao/​monografias/​GEBIS%20-​ %20RJ/​seriesestatisticasrestrospectivas/​Volume%203_​Estatisticas%20historicas%20do%20 Brasil_​series%20economicas_​demograficas%20e%20sociais%20de%201550%20a%201988.pdf. Instituto Brasileiro de Geografia e Estatística (IBGE). 2006. “Estatísticas do Século XX.” Available at http://​seculoxx.ibge.gov.br/​economicas/​contas-​nacionais. Kume, H., G. Piani, and C. F. B. Souza. 1988. “A política brasileira de importação no período 1987–​1998:  Descrição e avaliação.” Available at http://​www.ipea.gov.br/​sites/​000/​2/​livros/​ aberturacomercial_​/​Capitulo%201_​politica.pdf/​. Lopez-​Cordoba, E. and M. M. Moreira. 2003. “Regional Integration and Productivity:  The Experiences of Brazil and México.” INTAL ITD-​STA Working Paper No. 14. Available at http://​www.iadb.org/​intal/​. Mata, Maria Eugénia, and J. L. Love. 2008. “A Reversal in the Historical Role of Tariffs in Economic Growth? The Cases of Brazil and Portugal.” Estudos Econômicos 38 (3): 461–​492. Mayer, J., A. Butkevicius, and A. Kadri. 2002. “Dynamic Products in World Exports.” UNCTAD Discussion Papers No. 159. Available at http://​ideas.repec.org/​p/​unc/​dispap/​159.html. Ministério do Desenvolvimento, Indústria e Comércio Exterior. Departamento de Defesa Comercial. 2015. “Relatório do DECOM—​ 2015.” Available at http://​www.mdic.gov.br/​ images/​REPOSITORIO/​secex/​decom/​Relat%C3%B3rios_​DECOM/​2015.pdf. Ministry of Development, Industry and Trade Statistics. 2016. Available at http://​www. desenvolvimento.gov.br/​sitio/​interna/​interna.php?area=5&menu=576. Moreira, M. M. 2009. “Brazil’s Trade Policy: Old and New Issues.” In Brazil as an Economic Superpower?, edited by L. Brainard and L. Martinez-​ Diaz, 137–​ 158. Washington, DC: Brookings Institution Press. Muendler, M. 2004. “Trade, Technology and Productivity: A Study on Brazilian Manufactures, 1986–​1998.” CEFifo Working Paper No. 1148. Available at http://​www.cesifo-​group.de/​ portal/​page/​portal/​ifoHome/​. OECD. 2015. “Trade in Value Added:  TiVa.” Available at http://​stats.oecd.org/​Index. aspx?DataSetCode=TIVA2015_​C1. Pasteels, J. M. 1998. “Foreign Trade Statistics:  A Guide to Their Use In Market Research.” Geneva: International Trade Center, UNCTAD/​WTO. Silber, Simão Davi. 2004. “Setores exportadores e importadores da Índia.” Convênio CEPAL/​ IPEA. Fipe –​Fundacao Instituto de Pesquisas Economicas, Texto para Discussão n. 4. Trade Centre. UNCTAD/​WTO. Available at https://​www.wto.org/​english/​thewto_​e/​coher_​e/​ wto_​itc_​e.htm.

Trade Policy from the 1930s to the Present    419 Tavares de Araújo, J., and P. Miranda, P. 2008. “Antidumping and antitruste: peculiaridades do caso brasileiro.” Série Breves Cindes No. 8. UNCTADSTAT. 2016. “Most Favoured Nation Average Applied Import Tariff Rates on Non-​ Agricultural and Non-​ Fuel Products.” Available at http://​unctadstat.unctad.org/​wds/​ TableViewer/​tableView.aspx. UNCTADSTAT. 2016b. “International Trade in Goods and Services.” Available at http://​ unctadstat.unctad.org/​wds/​ReportFolders/​reportFolders.aspx. World Trade Organization. 2016. “I-​TIP Goods: Integrated Analysis and Retrieval of Notified Non-​Tariff Measures.” Available at https://​i-​tip.wto.org/​goods/​Default.aspx.

Pa rt  I V

B R A Z I L’ S R E G ION S

Chapter 20

Re gional Di spa ri t i e s Carlos R. Azzoni and Eduardo A. Haddad

20.1. Introduction An important initial fact to highlight is the sheer territorial size of Brazil. With an area of 8.5 million square kilometers, Brazil is larger than Australia, the continental parts of the United States and Europe, and smaller only than Russia, China, the United States (all 50 states), Canada, and the enlarged Europe (53 countries). At its longest extent, the distance between the coast in the east and the neighboring countries in the west reaches 2,700 miles. Possible connections with other economies to the north and to the west are limited by, in the former case, the Amazon forest, and in the latter, the Andean range of mountains. The economies of the West and North regions of the country are either isolated or connected to the coastal areas many miles away, either with respect to foreign markets or the main population and internal market areas located in the Southeast and South of the country. Long distances and the quality of roads are important economic issues impairing the competitiveness of these regions. There are good connections only in the South and Southeast with Uruguay, Argentina, and Paraguay, and, with limitations related to topography and distance, with Chile. This amplifies the attraction of the internal market concentrated in the Southeast and South of the country. The other relevant aspect of Brazilian economic geography relates to the 2,700-​mile north-​to-​south spread of its territory (see Figure 20.1). Small parts are located either in the Northern Hemisphere, close to the line of the equator (around 6% of the area and 5.3% of the population) or in the temperate zone (7% of the area and 14.5% of the population). The remaining parts are subtropical, with a concentration of population in coastal areas. This prompts important differences in the resource base of the regions, including levels and variations in temperature, rain regimes, the presence of at least 10 different types of vegetation, soil types, landscapes, and so on, that comprise the natural

424    Carlos R. Azzoni and Eduardo A. Haddad

Figure 20.1.  Brazilian states and macro regions.

basis of the country.1 Given this heterogeneity, it is not surprising to find huge regional disparities in economic and social conditions. The urban system is also concentrated in the coastal area and is markedly skewed in terms of the size distribution of cities. Half of the national population of the country’s 5,570 municipalities is located in the 200 largest cities. In 2015, the main metropolitan area, São Paulo, hosted over 21 million inhabitants in its 37 municipalities, but the surrounding area (constituting the same labor market) reached over 173 municipalities and 32 million inhabitants.2 The second largest metropolitan area was Rio de Janeiro, with 12  million inhabitants, followed by Belo Horizonte, with almost 6  million. Fourteen other cities contain populations larger than 1 million. Given their location in the coastal area or close to it, density varies significantly. The North and Center-​ West areas present demographic densities of 2.66 and 5.86 inhabitants per square kilometers; the poor Northeast region is denser, with 27.33 inhabitants per square kilometer; the South shows 38.38, and the rich Southeast, 67.77 inhabitants per square kilometer (Figure 20.2).

Regional Disparities   425

Population density, munucipalities, 2016 (people per sq. km) ≤ 5.0 5.1–10.0 10.1–25.0 25.1–50.0 50.1–100.0 100.1–250.0 250.1–500.0 500.1–1000.0 1,000.1–5,000.0 > 5,000.0

N 0

250 500 750 1.000 km

Figure 20.2.  Population density by municipality, 2016.

20.2.  Regional Disparity Indicators In the first decades after the arrival of the Portuguese, very few activities were developed since the country offered no easily accessible source of gold or silver. Exploration for a specific type of tree3 used to color clothes was the first surge of activity. Once the easy-​to-​ find trees were exhausted, interest faded and the country was left alone. Some attempts were made to populate the newfound land, but it was only with the beginning of the production of sugar and, to a lesser degree, tobacco that this goal was partially reached. This happened in the Northeast region, mainly in the present state of Pernambuco,4 in the mid-​sixteenth century and lasted until the mid-​eighteenth century. The region was

426    Carlos R. Azzoni and Eduardo A. Haddad the economic core of the country at the time, with the remaining area left essentially untouched. Gold was finally found in 1697 in the Southeast region, although initially not in the expected volume, and was followed by diamonds in the same region in 1714. In the years that followed, a gold rush started, thus bringing about huge transformations in the area. In parallel with events occurring in Europe, this prompted massive migratory flows from Portugal to the colony, mainly to its Southeast portion. This included even the movement of the capital city from Salvador (in the Northeast region) to Rio de Janeiro in 1763. Though sugar and tobacco were produced mainly to supply the international market, the large population contingents in the Southeast constituted sizable internal markets for food and some services. This is the first distinction between the regions, although both used slaves to produce their products. A new era came with coffee, planted and developed in the Southeast. The crop was brought to Brazil in the eighteenth century, but only in the early nineteenth century did it finally boom. This started in the coastal areas between Rio de Janeiro and São Paulo, but attained full steam as it reached adequate soil inland to the west in the states of São Paulo and Paraná. This almost 100-​year cycle produced several transformations in the economy, including the substitution of European migrants for slaves—​thus amplifying the local market—​the construction of important railways, the development of ports, and so on. This sizable internal market, the qualification of the labor force, the presence of potential entrepreneurs, and the development of a transportation (railways) infrastructure provided the conditions for the present concentration of activities in the country. By the time of the Great Depression and World War I, the economy of the Southeast was well ahead of the other parts of the country. The capital city of Rio de Janeiro and the city of São Paulo were important urban centers concentrating the new middle class. Import limitations during World War I and the Great Depression gave space for the development of manufacturing activities to substitute for imported goods. Exporting coffee demanded some supporting manufacturing, such as tools and textiles for the sacks, which were developed in these urban centers, mainly in São Paulo. A natural step was to expand these activities to other manufacturing areas, which gave way to the industrialization of the country, with a focus in the Southeast. Estimates of regional data for the distant past illustrate the shape of regional disparities in the country. Rands Barros (Chapter 21 in this volume) presents a long-​ term series of per capita GDP for the Northeast region, indicating that it was above the national average until the mid-​nineteenth century, the time at which the coffee boom began in the Southeast. Reis and Monasterio (2008) have shown that regional disparities at the time, based on the 1872 population census, were not much different from the present levels. Official statistics are only available after 1939, and they are summarized in Table 20.1. As Rands Barros’s chapter on the Northeast analyses, the region hosted almost 28% of the national population in the most recent census, and has lost shares in population and production since 1940. Being populous, it attracts attention in terms of regional policy, thus constituting the most significant “regional problem” in Brazil. In

Regional Disparities   427 Table 20.1 Regional Disparity Indicators Shares (% of National Totals) Population Regions

Area

North

45.3

Northeast

18.3

3.9

10.6

12.5

4.4

5.0

0.75

0.63

7.8

8.8

14.7

0.48

0.50

2.0

8.3

2.9 10.9

São Paulo State

2.9

17.4 21.6

South

6.8

0.7

0.0

0.2

1.0

0.43

0.41

77.6

56.4

55.6

1.41

1.31

11.0

3.8

38.6

33.6

1.8

1.48

13.9 14.4 15.3 16.4 29.5

1.0

24.5

15.2

1.11

1.15

1.1

5.9

9.6

0.7

1.25

7.4

0.9

5.3

44.5 42.1 63.0 54.9 23.4

2.7

1.6

2.7

2014

35.0 27.8 16.9 13.9 16.5

Southeast

18.9

Per Capita GDP

Value Added, 2013

1940 2010 1939 2014 Agric. Mining Manuf. Tertiary 1939

Piauí State

Midwest

GDP

31.3

2.1

0.6 32.1

9.4

19.2

Source: IBGE, Censo Populacional and Contas Regionais.

contrast, the rich Southeast, with 11% of the area, hosted 42% of population in 2010, and 55% of the national GDP in 2014. Although this region lost almost 8 percentage points over the period, it is still the locomotive behind the national economy, especially in manufacturing and services. The South is small in terms of area, is second in economic importance, and has increased its share in population and GDP over time. A comparison of the states of Piauí in the Northeast and São Paulo in the Southeast, with identical area sizes, is illustrative. The former lost 20% of its share in population, and one-​third of its share in GDP; the latter lost 5% of its share in population, but gained 3% of share in GDP. The gainers in terms of population and production are the North and the Midwest. The North is another poor area, but with less importance in terms of population. It has more than doubled in its share of the population from 1940 to 2010, and doubled its share of national GDP. Its growth is based on natural resources, with important logging and mining activities and some cattle ranching. It also hosts a very active tax-​free import zone (Zona Franca de Manaus), established in the 1960s. The Midwest more than tripled its share of the population, and multiplied its share of GDP more than fourfold. Besides the transference of the capital city to Brasília in 1961—​now a metropolis with more than 3 million inhabitants, and 3% of the national value added—​the advance of the agricultural frontier toward the west explains its growth in importance. Table 20.1 also provides information on the importance of the regions in broad sectors of activities. In 2013, the North region hosted 12.5% of all mining activities in Brazil. This is the sector with the highest importance, though agriculture is also important in the region. In the Northeast, agriculture and tertiary activities are highlights. The Midwest excels in terms of agriculture, and the South, in agriculture and manufacturing. In the

428    Carlos R. Azzoni and Eduardo A. Haddad richest region, the Southeast, mining is of the highest importance (oil in Rio de Janeiro and iron ore in Minas Gerais), along with manufacturing and tertiary activities. The preceding indicators highlight the concentration of people and production, but disparities also can be expressed in terms of inequality. The final two columns of Table 20.1 provide information on per capita income levels, in relation to the national average. The Southeast region is well above the national average, followed by the Midwest and the South. The Midwest has almost doubled its relative level of per capita income, powered by the penetration of agriculture into the region, and the movement of the national capital in 1961 to the region. The new capital has exerted a large influence on that outcome since the Federal District exhibits the highest level of per capita income in the country, influenced by the high percentage of well-​paid public workers. At any rate, the increase in its share in production (7.3 percentage points [p.p.]) was larger than

< 0.500 0.500–0.549 0.550–0.599 0.600–0.649 0.650–0.699 0.700–0.749 0.750–0.799 ≥ 0.800

Figure 20.3.  Human Development Index by municipality, 2010.

N 0

250 500 750 1.000 km

Regional Disparities   429 the increase in its share in population (2.7 p.p.). As for the North, its increase in production share (2.6 p.p.) was smaller than the increase in its population share (4.4 p.p.), thus leading to a decrease in distance from the national per capita income average. The poor Northeast region was never able to reach half the per capita income level of the country. Education is another important way of presenting inequality. The share of illiterates in the population over 20 years of age in 2014 was less than 27% in the South and Southeast, 31% in the Midwest, 39% in the North, and 42% in the Northeast. The share of the population over 10 years of age with six or more years of schooling was 60% in the North, 55% in the Northeast, 66% in the Midwest, 67% in the South, and 70% in the Southeast. Public health services are very poor in Brazil, and people who can afford it rely on private health insurance. In the Southeast, 36% of the population was insured, 30% in the South, 25% in the Midwest, and only 13% in the Northeast (PNAD 2008, Special Issue on Health). Water services were present only in 80% of houses in the Northeast in 2014, in contrast with 91% in the Southeast. Sewage collection was available for 88% of houses in the Southeast, and only 41% in the Northeast (PNAD 2013; SNIS 2016). Indicators of regional inequality can be summarized in the Human Development Index of municipalities, which is a simple average of indicators5 of life expectancy, schooling, and per capita income. Results for the most recent census are displayed in Figure 20.3.

20.3. Competitiveness The preceding section describes just how pronounced are regional disparities in Brazil, whether in terms of concentration or inequality. It is interesting to investigate the causes behind these differences and to provide information on trends. The most significant region in terms of policy interest is evidently the Northeast, since it hosts almost 28% of the population (and votes). That is why there is a specific chapter in this Handbook devoted to that region. As mentioned earlier, given the size of the country, especially the spread in latitude, it is natural to find huge differences in terms of weather conditions and endowment of natural resources. From the hot and dry areas in the Northeast, subject to cyclical droughts, to the rainy and hot Amazon area, to the cooler Southern states, Brazil’s weather is heterogeneous. In addition, soil quality varies, as mentioned earlier with respect to coffee. In general, the best soil for agriculture is located in the southern half of the country, where the rain regime is also more favorable for agriculture in general. Available measures of productivity confirm these endowment differences, since total factor productivity indicators are the highest in these regions (Imori 2012). An important development in this context is the advancement of agriculture toward previously unfit land in the Midwest. As the chapters on agriculture in this volume describe, an impressive government-​led research effort has made it possible to explore large savannah areas with new varieties of soybeans, corn, beans, and even cotton. The region is highly competitive in terms of physical productivity at the farm gate, but

430    Carlos R. Azzoni and Eduardo A. Haddad high transportation costs attenuate this advantage as the products approach the main markets and the ports. Nonetheless, Brazil is one of the leading exporters of grains, with most of them coming from the savannah areas. Before coffee, manufacturing activities—​as precarious as they were at the time—​were widespread all over the main urban centers in the country, and essentially supplied the surrounding markets. The introduction of coffee provided many stimuli for the development of manufacturing activity. The high demand for textile products related to shipping the product promoted a captive market for the rudimentary manufacturing bases of the urban centers next to the planting areas. The massive labor market associated with the crop, with migrants from Europe substituting for slaves, provided a local demand for wage goods, thus reinforcing the stimulus for manufacturing development. The construction of railways enlarged the market for plants located in the main urban centers in the coffee-​growing areas, whose products had not only better quality but also lower costs than those produced elsewhere. Through this process, a concentration started in the coffee-​related urban areas, producing the same pattern of concentration observed at present. The two world wars and the Great Depression increased the cost of imports—​if not precluded them—​which produced an increased demand for local producers. As competitiveness already existed with the established manufacturing centers, further concentration was an expected byproduct. Thus, by the end of World War II, the main manufacturing centers were in São Paulo and Rio de Janeiro. After a short period of abundant international reserves, when imports of consumer goods were massive, local producers became more competitive and accessed the increasing local market. An intense import-​substitution program followed, which even further boosted local production, now already concentrated in spatial terms. This pattern of concentration has continued to the present, given the economies of agglomeration typically required by manufacturing activities. Studies indicate that increased concentration leads to increased productivity through a reduction in transaction costs related to the proximity of connected sectors, the sorting of firms and workers, large pools of well-​qualified workers, and so on. Barufi (2015) has shown that an increase in urban size, even after controlling for the characteristics of workers, increases productivity, as represented by wage levels. Gonzaga and Azzoni (2016) have estimated that even after controlling for the sorting of firms and individuals, the urban premium is still present: a 1% in urban size increases productivity by 3%–​4%; that is, other than sorting, it has a place in explaining large centers’ productivity edge. In reality, relative measures of productivity calculated by Azzoni and Ferreira (1998) have shown that productivity advantages of the traditional industrial centers outweigh their larger wage levels. Schettini and Azzoni (2013) have estimated relative productivity levels for the first decade of the twenty-​first century, and their numbers indicate that the traditional centers are still ahead in terms of competitiveness. Infrastructure is clearly an important factor in economic growth, especially so at the regional level, and notably in a country with wide heterogeneity in the supply of such a factor. Schettini and Azzoni (2015) estimate stochastic frontiers at the regional level to assess the role of infrastructure on regional growth between 2000 and 2010. The

Regional Disparities   431 availability of transportation infrastructure is markedly heterogeneous, with deep concentration in the states of São Paulo, Paraná, and Rio de Janeiro, the economic core of the country. Similar situations are found for the urban infrastructure (water and sewage, and electricity), and for communications (telephones and Internet). Their results indicate that increases in the regional coverage of transportation infrastructure leads to an increase in regional efficiency. A similar result was found for the urban infrastructure and for communications. Based on their results, a simulation was made in which regional efficiency indexes were computed with and without infrastructure variables. The difference between these results indicates the role of infrastructure in regional growth. The results show that the agricultural and mineral frontiers, localized in the North and Midwest, are the regions most negatively affected by the lack of reasonable infrastructure. However, their results do not show impacts of great amounts, at least in the short run. As Grimes (2014) indicates, infrastructure investments give the private sector a future option, not an obligation. On the other hand, Crescenzi and Rodríguez-​Pose (2012) emphasize that investments directed toward bottlenecks have larger positive impacts, indicating that investments should be prioritized according to their impacts. A rich way to look at competitiveness relates to the quality of human capital in the regions. This variable can be approached both by labor supply (quality of labor available in terms of education, experience, discipline, cultural attitude toward work, etc.) and demand (what sort of occupations are being offered by the establishments located in the region). Using the adaptation of a comprehensive American study6 performed by Maciente (2013), who associated skill levels with each occupation in Brazil, it is possible to evaluate the relative position of the regions. Moreno and Azzoni (2016) used these skills to determine the innovative potential of Brazilian regions in manufacturing, and discovered a high degree of concentration. As for the cognitive skills required by the occupations, it is clear that only the states of São Paulo, Rio de Janeiro, and the Federal District demand above-​average skills; the same states use below-​average physical-​strength skills. This means that these areas, comprising the traditional economic center of the country, display a productive structure more intensive in sophisticated skills, which can thus give them a competitive edge against the other regions. In terms of growth in skill requirements in the twenty-​first century, however, there seems to be a convergence; that is, states offering low-​skill occupations, on average, presented faster growth in the incidence of such sorts of skills between 2002 and 2014. Thus, although the absolute levels indicate a divide between the more advanced regions and the backward ones, it seems that this situation has the potential to change in the future. However, the speed of change is modest, and it would take a very long time before the labor-​demand quality would be even across regions. The tertiary sector is becoming more and more important in economic terms, accounting for about 70% of the national GDP and even more in terms of employment, particularly in the informal sector (Azzoni and Guilhoto 2011). This sort of activity is urban by its very nature and is widespread in the territory, and follows, in general, the geographical distribution of the population. However, there are important aspects of this

432    Carlos R. Azzoni and Eduardo A. Haddad area of activity that should be mentioned. Government-​related activities and more basic sectors, such as services to families, tend to follow the population more closely than private activities. More sophisticated commerce (wholesale, for example) and services (consulting, financial, etc.) tend to be more concentrated. Azzoni and Andrade (2005) found that the rich areas were losing competitiveness in commerce and in traditional services in the 1980s and 1990s, but were becoming more competitive in the modern subsectors within the tertiary sector, such as in services to firms, computing, and so on. As for spatial concentration, the majority of subsectors presented decreasing concentration in the period, although six subsectors presented increasing concentration.

20.4.  Per Capita Income Convergence The productivity levels mentioned in the previous section indicate that the regions responsible for the great part of the national GDP still have a competitive edge, meaning that the concentration pattern observed tends to be maintained in the future. This conclusion is reinforced by the analysis of recent changes in productivity levels across regions also present in those studies, which have shown that lagging regions are also those with the least productivity growth. The combination of these two lines of results points to the continuation of the highly concentrated pattern of the geographical distribution of economic activities in the country. At least, it precludes any optimism in terms of radical changes in the observed situation. Another strand of literature deals with the inequality in per capita income levels across regions, and how it evolves over time. Following Barro and Sala-​i-​Martin (1992, 1995), many studies have dealt with this subject in Brazil (Azzoni 1994, 1997, 2001; Azzoni and Servo 2002; Azzoni and Silveira-​Neto 2005; Cravo and Resende 2013; Ferreira 1996, 2000; Laurini et al. 2005; Lima and Resende 2007; Magalhaes et al. 2005; Mossi et al. 2003). The results indicate that there are no signs of absolute convergence, which would mean that all regions would tend to the same absolute per capita income level at some point in time. Thus, the present situation, not only in terms of concentration but also in terms of regional income inequality, would tend to remain. The existence of a persistent regional dualism, in which spatial spillovers are very important, is one of the most robust results of these studies. Geography has also been shown to play an important role in driving regional growth. Regional efficiency of manufacturing sectors, for instance, independent of technology intensity, is heavily affected by neighborhood effects (Schettini et al. 2011). Spatial spillovers are also important in determining regional innovative capacity and technology diffusion in Brazil (Gonçalves and Almeida 2009). Demography influences the economic environment either through income inequality across individuals, or through the possible influence of demographic characteristics on economic growth, with important effects on the process of convergence across Brazilian regions (Menezes et al. 2012). The existing uneven spatial distribution of demographic indicators that are deemed important to understand

Regional Disparities   433 the current and future levels of welfare of a region (e.g., infant mortality) may, however, be improved by investments in health infrastructure, with more likely impacts in the long run (Barufi et al. 2012). Once the hypothesis of absolute convergence has been excluded, the next question concerns what the equilibrium level of regional inequality, which is analyzed by looking at conditional convergence, would be. Contrary to absolute convergence, in this case each region could converge to a region-​specific per capita income level, compatible with their endowments of capital and labor (and the evolution of that endowment). In this case, there would be a particular level of regional economic inequality that would be an equilibrium situation, thus meaning that there would be no mechanisms leading to a change in that situation. All studies that have dealt with convergence have found evidence of conditional convergence. This result is robust for different periods of time (the longest is 1939–​2013), types of geographical areas (states, meso or micro regions, municipalities), ways of measuring income (per capita income—​GDP/​inhabitant or labor productivity—​GDP/​ worker), econometric techniques applied (panel data, spatial econometrics), and so on. Some aspects are worth mentioning, such as the fact that convergence is faster for labor productivity than for per capita income. It seems that the labor market, which considers only employed people—​a subset of the adult population—​is working toward conditional convergence faster than the population at large. This means that differentials in the age composition of the population make it more difficult for poor regions to catch up with rich regions. Another negative aspect is that convergence of per capita income is faster for people over 60  years of age, mainly caused by pensions and retirement payments. Again, age composition differentials play an important role. One interesting finding is the presence of convergence clubs, that is, clusters of regions with similar per capita income levels. A  clearly identified rich-​region club involves the states of the Southeast and the South, and another is the poor-​region club, encompassing states in the Northeast (and some states in the North). The conclusion of this sort of analysis is that the equilibrium level of regional inequality is one that will register a clearly marked disparity between the North and the South. Since it is the steady-​ state solution for the models behind the studies, it is the situation the economic system tends to produce. Even considering this caveat, the time necessary to reach the steady-​ state solution is long, indicating that the adjustment process is taking place at a very slow pace.

20.5.  Economic Integration of the Regions Given the pronounced disparities among Brazilian regions, it is important to observe how integrated the regions are. Based on the national input-​output tables developed

434    Carlos R. Azzoni and Eduardo A. Haddad by the national statistics office (IBGE), some multiregional I-​O models have been estimated (Guilhoto and Sesso-​Filho 2005, 2010). Of the 27 states, only five present net exports to other states. The isolated state of Amazonas, home of the free-​trade area Zona Franca de Manaus, exports 64% of its production to other states, essentially in the wealthy parts of the country. Espírito Santo state in the Southeast exports 14% of its production, but this is mostly transference to other states of imports from abroad. The state of São Paulo exports only 12% of its production to other states, but the absolute amount is four-​fifths of the total internal trade surplus. Two other states in the Southeast (Paraná and Santa Catarina) also have trade surplus, but at a much lesser value. In summary, except for the artificial economy of Zona Franca de Manaus in the North, the main exporter is the state of São Paulo, where economic activity is concentrated. This indicates how the Brazilian economy gravitates around its traditional economic center. The largest import shares (imports/​local production) take place in the Northeast as a macro region (24%), which also shows the largest absolute amount. The region is followed by the Federal District with 22%, and the North region with 17%. The states of Minas Gerais and Rio de Janeiro, in the Southeast, present lower shares but important absolute amounts. While interregional exports are more concentrated, interregional imports are more widespread across all states. Considering trade with other countries, the interregional surplus states of São Paulo, Amazonas, Paraná, and Santa Catarina present negative balances. The state of Pará (mining and timber) in the North, the grain and beef producing states of the Midwest, Minas Gerais (mining products) and Rio de Janeiro in the Southeast, and Rio Grande do Sul in the South show positive trade balances with other countries. This positions São Paulo state, and the traditional economic center of the Brazilian economy, as a net exporter to other states. Thus, economic activity radiates out from that traditional area to reach other parts of the national economy, which, in turn, export primary products to other countries (grains, beef, cotton, mining products).

20.6.  Explicit Regional Policies Given the significant share of the Northeast region in terms of population, and therefore political power, regional policies have been designed with the aim of overcoming its backwardness. In the 1950s, development agencies for the backward regions of the Northeast and North were established, and a series of incentives for investment were put into place with the aim of attracting capital, including the creation of regional development banks. Ferreira (2004) describes the instruments utilized by the federal government and provides an assessment of these experiences. With the change in the national constitution in 1988, the previous instruments, which, as indicated by the numbers presented at the start of the chapter, had proven to be ineffective, were replaced

Regional Disparities   435 by investment funds. A recent review of these initiatives can be found in Soares et al. (2009), Resende (2012, 2013), and Silva et al. (2007; 2009). Resende (2013) provides a summary of the evaluation literature relating to regional policy. From the late 1980s until recently, the agreed agenda for Brazil included the competitive integration of the country into the global trade network, with additional domestic concerns focused on sustainable stabilization and social cohesion. This implied attracting foreign investment and a responsible (balanced) budget policy for all levels of government, reinforced by the promulgation of the Lei de Responsabilidade Fiscal (Fiscal Responsibility Law) in 2000. The latter restricted regional policies based primarily on redistributional expenditures, as was the case in the 1970s (Haddad 1999). The research undertaken in the last two decades by regional scientists in Brazil refers to this context of policies in such areas as macroeconomic stabilization, economic opening with respect to both trade and investment, and the expansion of market forces within the domestic economy, common to many developing countries that have embraced the “Washington Consensus.” Studies have shown that the interplay of market forces in the Brazilian economy tends to favor the more developed regions of the country. An increase in regional inequality (Guilhoto and Fonseca 1998; Haddad et al. 2002, 2005) is a very likely regional repercussion of trade liberalization policies, including the creation of MERCOSUR. The loss of importance of MERCOSUR, in turn, brings challenges for Brazilian regional economies as specific rules for the free-​trade area (FTA) seem to overcome the usual effects of relative competitiveness associated with movements in relative prices (Vieira et al. 2014). Regional governments in Brazil continue to adopt tax incentive programs to attract private investment to their jurisdictions. The case of the automobile industry deserves the special attention of regional scientists. New investments were attracted in the late 1990s and the first decade of the 2000s by policies implemented by the Brazilian government, which has played an active role in negotiations with foreign investors in the country. Up to today, state governments have engaged in strong competition for the incoming capital through fiscal incentives. It has been documented that the regional dualism in Brazil is also associated with a strong productive dependence of the less developed regions on the more developed regions (Guilhoto et al. 2002). This poses a very real danger that the benefits of new inward investments are not fully internalized by the states that seek to instigate them. In this regard, from a regional perspective, it is important that effective strategies are devised that minimize this risk (Amann et al. 2007; Perobelli et al. 2007). Moreover, investments in the poorer regions, which tend to be more beneficial to the improvement of regional imbalances in the country, do not generate the same level of national growth as investments in the more developed, denser areas that benefit from agglomeration economies (Haddad and Hewings 1999). Specialized structures of production are also important features of regional economies, even in more developed states, that help in understanding the local impact of regional tax incentives (Porsse et al. 2007).

436    Carlos R. Azzoni and Eduardo A. Haddad Even though fiscal incentives continue to play a role in attracting capital to the regions, for private investors the search is dominated by attention to maximal financial returns with little concern for regional equity; location is defined on a purely economic basis. A stream of research has looked at the implication of investors’ rationale for the location of investments in Brazil. While Silva and Hewings (2012) attempted to understand, from a theoretical perspective, the role played by the internal organization of the firms, other authors have addressed important empirical features of the Brazilian economy, common to many developing countries, such as financial constraints and volatility of the business cycle (Kalatzis et al. 2008, 2011). The results reveal that, in the case of Brazil, there are significant differences across regions in the importance of investment determinants, thus bringing significant insights with respect to the design of regional policies in the country (Azzoni and Kalatzis 2008).

20.7.  Sectoral Policies with Regional Effects As the chapters on agriculture in this volume demonstrate, investments in technology to boost the primary sector have proven to be a trend-​changing tool in reducing regional disparities in Brazil. At the same time, it can be argued that most sectoral policies could act in favor of the more developed regions (export promotion favoring states in the Southeast, incentives to ethanol as fuel benefiting São Paulo state, etc.). Thus, it is important to consider how tools developed to solve particular problems, be they sectoral or social, have influenced the pattern of regional inequality in the country. One important area of regional science research in Brazil relates to the development of large-​scale integrated modeling systems for impact analysis. By inserting a core Computable General Equilibrium Model (CGE) in a broader modeling framework, Brazilian scholars have been able to suppress some of the shortcomings of isolated models. Applications for transportation policies have dealt with market imperfections in the Brazilian spatial economy by introducing non-​constant returns and non-​iceberg transportation costs in an interregional CGE model integrated with a geographic information system (GIS) transportation network model (Haddad and Hewings 2005). In the case of Brazil, as well as in many developing countries where transportation costs are high and accessibility low compared to European or North American standards, handling market imperfections becomes imperative, as does the need to address internal spatial issues from the perspective of Brazil’s increasing involvement with external markets. Projects of spatially connective infrastructure have been assessed using this methodology, which provides insightful results concerning the various trade-​offs that emerge. Haddad et al. (2011) make it clear that for policies of domestic integration in the country, given different policy options, decision-​makers face nontrivial choices; different projects perform differently in different dimensions, usually presenting outcomes with different

Regional Disparities   437 hierarchies related to multidimensional policy goals. This is also true for other types of infrastructure investment. For instance, the choice of ports for government investment would have, potentially, significant implications for the hinterlands serving those ports, as well as for other areas that may be able to access them once the investments have been completed, along with very strong regional development policy implications (Haddad et al. 2010). Despite being still fettered to the reins of the perfectly competitive modeling paradigm, Almeida et al. (2010) add to the previous results by revealing that, methodological differences aside, the evidence concerning the nature of the relationship between the provision of transport infrastructure and regional equity is controversial due to a fundamental characteristic associated with this issue. In other words, even with the same theory or model, method, and its specification, one may continue to obtain different results concerning this relationship. This outcome arises because this relationship crucially depends on where the transport infrastructure is located. In addition to methodological considerations, there seems to exist authentic spatial reasons that might yield controversial results. Indeed, transport infrastructure is strongly region-​dependent. The spatial structure of the provision of transport infrastructure matters in this question, since it plays a fundamental role in determining the effects on the economic system, as shown in a model developed for the state of Minas Gerais, Brazil. Large-​scale integrated modeling systems have also been developed for regional impact analysis of energy policies in Brazil (Santos et al. 2013). Simulations of the long-​run regional impacts of electric power tariff policy in Brazil have shown that the heterogeneity of energy intensity and the differentials of energy substitution drive the spatial impacts of changes in electric power prices. On the other hand, the recent trend of spatial dispersion of electric power prices might contribute to a decrease in long-​run economic growth and to an increase in regional inequalities in Brazil. Since the 1990s, the energy sector has been the subject of a variety of reform initiatives that are changing the market structure and energy price levels. These reforms have also been triggered by the implementation of neoliberal policies in the Brazilian economy. Energy policy in the country has stimulated energy diversification to increase the inter-​ fuel substitution. Some studies have attempted to understand the new patterns of sectoral and regional consumption of energy that have emerged in the country (Carvalho et al. 2013; Perobelli and Oliveira 2013). Emphasis on renewable energy has implications for food security in the country, since biofuel production in Brazil relies heavily on processing sugarcane. There is an ongoing debate regarding the risks associated with diverting farmland or crops for biofuel production to the detriment of the food supply. The expansion of sugarcane cultivation in Brazil, spurred particularly by increased demand for ethanol, has triggered the need to evaluate the economic, social, and environmental impacts of this process, both on the country as a whole and on the growing regions. Despite some evidence that the presence of sugarcane cultivation in these areas is not relevant to determining their social conditions (Chagas et al. 2012), positive demand shocks upon the sugarcane agro-​industry have a greater impact with respect to

438    Carlos R. Azzoni and Eduardo A. Haddad income upon the less developed regions of the country (North) compared to the Center-​ South (Costa et al. 2006; Martínez et al. 2013). In the context of the fiscal adjustment process of the 1990s, the role of the central government in directly stimulating productive activities has been replaced by strategies of socioeconomic inclusion. Seemingly non-​spatial government policies in the form of spatially blind social programs have played an important role in the recent decline in regional income inequality in Brazil (Silveira-​Neto and Azzoni 2011, 2012). However, regional inequality continues to be very high and this issue will continue to be on the research agenda for many years to come. Important components of income and particularly wealth inequality are still unknown in the case of Brazil. The distribution of prop­ erty rights and rents on natural resources need to be better understood (Goeschl and Igliori 2006). Furthermore, a more complete picture of income and wealth distribution is still needed. As has been shown by Piketty (2014), there can now be no doubt that the phenomenon of inequality is not predominantly about the inadequacy of the skills of lagging workers. Understanding the process of wealth accumulation across regions may change drastically our prescriptions of regional policies. Brazil, as is the case with many developing countries, has experienced a rapid process of urban expansion around the CBD7 of its main cities that was not followed by the implementation of an adequate infrastructure, thus causing important urban problems (Haddad and Nedović‐Budić 2006; Menezes et al. 2013; Silveira-​Neto et al. 2015). Recent experimentation with integrated modeling of metropolitan systems has proven relevant for assessing the consequences of apparently local phenomena related to the city of São Paulo: floods (Haddad and Teixeira 2015) and local transportation infrastructure (Haddad et al. 2015). The key message is that one needs to consider interactions both inside and outside a prime metropolitan system to recognize the role they play in an integrated interregional system. The lack of redundancy in the economic infrastructure of developing countries (i.e., the inability to have alternatives to solve problems of logistics, communication, or energy in the advent of unexpected events) poses interesting research questions for regional scientists in Brazil. As an example of a global phenomenon, ongoing global climate change will have potential consequences for the competitiveness of regions in the future. Resource-​oriented activities, such as agriculture, mining, timber, and so on, and the related processing industries, deal with different restrictions as compared to footloose activities. The immediate impact of climate change will very likely be more intense on activities more dependent on nature. However, the repercussions of these effects will be felt in other sectors as well, thus affecting the composition of regional income and household consumption, with an influence on the tertiary sector of the main cities in the region, and finally reaching the industrial sectors supplying regional demand. Thus, it is expected that the initial stimuli from natural resource-​based industries will eventually result in major changes in the economy of the region as a whole (Azzoni and Haddad 2012). The most vulnerable regions to climate change are traditionally the less-​developed areas of the country: the Amazon and the Northeast (Barbieri et al. 2010). This is a challenging interdisciplinary research area, bringing various challenges to regional scientists in the

Regional Disparities   439 country. It could provide opportunities in the form of increased integration between institutions, more accurate data through information sharing and interdisciplinary approaches, and a greater understanding of the potential impacts of climate change on Brazil to ensure the most effective responses by the relevant political, economic, and social sectors. As detailed by Hoffmann (Chapter 22 in this volume), income inequality has been diminishing lately, though it remains at elevated levels. Together with the stabilization of the economy and the appreciation of the minimum wage, part of the explanation for this is the implementation of social policies, including cash transfers to poor families. As Silveira-​Neto and Azzoni (2011) demonstrate, although the largest part of the reduction in regional inequality is associated with the functioning of the market, the social programs implemented account for more than 24% of the reduction in regional inequality. This is impressive, since the amounts transferred account for less than 1.7% of disposable income. Thus, the social programs have unintended regional impacts of a larger magnitude, as compared to the explicit regional programs.

20.8.  Concluding Remarks As a country with a large territory, Brazil presents pronounced regional disparities. Economic activity and the population are concentrated in a small part of the territory. Even within this reduced area, the geographical distribution is highly uneven. Besides concentration, regional inequalities are marked in the country in terms of per capita income, education, access to public services, and so on. This scenario of concentration and inequality is quite persistent, as the data available indicate. This chapter has explored some aspects of this phenomenon. Besides providing a description of the observed disparities and their evolution, we have presented some insights into regional competitiveness and how it has evolved over time, convergence of per capita income, and regional integration. We conclude with a discussion of regional policy, both intended and unintended. The need for regional policies is controversial. First, it is necessary to determine if the observed disparities are the equilibrium situation of the economic system or not. In the latter case, implementing policy measures could correct market failures that lead to excessive concentration of inequality and, therefore, constitute an improvement in overall economic efficiency. However, if the observed level of disparities were the natural result of the operation of the economy—​an equilibrium situation—​then any form of intervention would result in some loss of efficiency. Which is the right answer is still debatable (Garcilazo et al. 2010; Gil 2010; Pessôa 2001; Neumark et al. 2014). This leads to the discussion of whether policies should be place-​based or people-​based; that is, if mechanisms such as those implemented in Brazil are the right tools to use. Although the debate is still inconclusive, the evidence is eloquent in pointing to the failure of the traditional place-​based regional policies implemented in the country.

440    Carlos R. Azzoni and Eduardo A. Haddad The present levels of inequality are a testimony to that. However, this is not tantamount to concluding that policies should be people-​based, such as the social policies implemented recently have been. It is true that people-​based policies have been the most effective way of reducing regional inequality (although not as good at reducing regional concentration). However, it can also be that the place-​based policies of the past suffered from implementation problems; that is, the policies were not well designed and/​or there were inefficiencies in their implementation, hence their failure. In either case, the fact is that regional disparities are important and persistent in the country. Whether for political or economic reasons, society has for better or worse devoted attention to this issue. However, success in reducing these disparities has yet to be achieved. The most effective ways of changing the observed regional concentration have been sectoral policies, such as the effort to make Brazilian agriculture more competitive, or social policies. The former opened opportunities for scarcely populated regions (the Midwest and lower North regions), far away from the populated poor Northeast. However, cash transfer programs have reduced inequality in a perceivable way. Economists must yet try harder to come up with the right answers on how to solve the “regional problem” in Brazil.

Notes 1. 2. 3. 4. 5. 6. 7.

http://​www.ibge.gov.br/​home/​geociencias/​cartogramas/​ctb.html. IBGE, Arranjos Populacionais (2016). Whose name, Pau Brasil (Brazilwood), in turn gives the country its name. See Rands Barros, Chapter 21 in this volume, for specific discussion of the Northeast region. The index varies between zero and 1; the higher the index, the better. ONet, https://​www.onetonline.org/​find/​descriptor/​browse/​Abilities/​. Central Business District.

References Almeida, E. S., E. A. Haddad, and G. J. D. Hewings. 2010. “Transport-​Regional Equity Issue Revisited.” Regional Studies 44 (10): 1387–​1400. Amann, E., E. A. Haddad, F. S. Perobelli, and J. J. M. Guilhoto. 2007. “Structural Change in the Brazilian Automotive Industry and Its Regional Impacts.” Latin American Business Review 7 (3–​4): 97–​119. Azzoni, C. R. 1994. “Crescimento econômico e convergência das rendas regionais:  O caso brasileiro à luz da nova teoria do crescimento.” In XXI Encontro Nacional de Economia, 1994, Vol. 1, 185–​205. Anais Do XXII Encontro Nacional De Economia. Florianopolis. Azzoni, C. R. 1997. “Concentração regional e dispersão das rendas per capita estaduais: Análise a partir de séries históricas estaduais de PIB, 1939–​1995.” Estudos Econômicos 27 (3): 341–​393. Instituto de Pesquisas Econômicas, São Paulo. Azzoni, C. R. 2001. “Economic Growth and Regional Income Inequality in Brazil.” The Annals of Regional Science 35 (1): 133–​152.

Regional Disparities   441 Azzoni, C. R., and A. S. Andrade. 2005. “The Tertiary Sector and Regional Inequality in Brazil.” Région et Développement 21: 155–​172. Azzoni, C. R., and D. A. Ferreira. 1998. “Competitividad regional y reconcentración industrial:  El futuro de las desigualdades regionales en Brasil.” EURE (Santiago) XXIV (73): 81–​111. Azzoni, C. R., and Guilhoto, J. 2011. “Size and Importance of Tertiary Activities in Brazil.” In The Economies of Argentina and Brazil: A Comparative Perspective, edited by W. Baer and D. Fleischer, 151–​163. Northampton, MA: Edward Elgar. Azzoni, C. R., and E. A. Haddad. 2012. “Climate Changes and the Future of Regions.” In Networks, Space and Competitiveness, edited by R. Capelo and T. Dentinho, 137–​148. London: Edward Elgar. Azzoni, C. R., and A. E. G. Kalatzis. 2008. “Incorporating Demand-​side Aspects into Regional Policy: Variations in the Importance of Private Investment Decision Factors across Regions.” The Annals of Regional Science 44: 69–​82. Azzoni, C. R., and L. M.  S. Servo. 2002. “Education, Cost of Living and Regional Wage Inequality in Brazil.” Papers in Regional Science 81: 157–​175. Azzoni, C. R., and R. Silveira-​Neto. 2005. “Decomposing Regional Growth:  Labor Force Participation Rates, Structural Changes, and Sectoral Factor Reallocation.” The Annals of Regional Science 39: 221–​239. Barbieri, A. F., E. P. Domingues, B. L. Queiroz, R. M. Ruiz, J. I. Rigotti, J. A. M. Carvalho, and M. F.  C. Resende. 2010. “Climate Change and Population Migration in Brazil’s Northeast: Scenarios for 2025–​2050.” Population and Environment 31: 344–​370. Barro, R., and X. Sala-​ i-​ Martin. 1992. “Convergence.” Journal of Political Economy 100 (2): 223–​251. Barro, R., and X. Sala-​i-​Martin. 1995. Economic Growth. New York: McGraw-​Hill. Barufi, A. M.  B. 2015. “Agglomeration Economies and Labour Markets in Brazil.” PhD Dissertation, Department of Economics, University of São Paulo. Available at http://​www. teses.usp.br/​teses/​disponiveis/​12/​12138/​tde-​04022016-​162856/​pt-​br.php. Barufi, A. M. B., E. A. Haddad, and A. Paez. 2012. “Infant Mortality in Brazil, 1980–​2000: A Spatial Panel Data Analysis.” BMC Public Health 12 (181): 1–​15. Carvalho, T. S., F. S. Santiago, and F. S. Perobelli. 2013. “International Trade and Emissions: The Case of the Minas Gerais State—​2005.” Energy Economics 40: 383–​395. Chagas, A. L.  S., R. Toneto, and C. R. Azzoni. 2012. “A Spatial Propensity Score Matching Evaluation of the Social Impacts of Sugar Cane Growing on Municipalities in Brazil.” International Regional Science Review 35 (1): 48–​69. Costa, C. C., H. L. Burnquist, and J. J. M. Guilhoto. 2006. “Relations of the Regional Brazilian Cane Agro-​industry with the National Economy: Analysis Applied to the Centre-​South and North-​Northeast.” Applied Economics 38 (5): 519–​531. Cravo, T. A., and G. M. Resende. 2013. “Economic Growth in Brazil:  A Spatial Filtering Approach.” The Annals of Regional Science 50 (2): 555–​575. Crescenzi, R., and A. Rodríguez-​Pose. 2012. “Infrastructure and Regional Growth in the European Union.” Papers in Regional Science 91 (3): 487–​513. Ferreira, A. H. B. 1996. “Evolução recente das rendas per-​capita estaduais no Brasil: O que a nova evidência mostra.” Belo Horizonte CEDEPLAR/​UFMG. [Texto para discussão 102]. Available at http://​www.cedeplar.ufmg.br/​pesquisas/​td/​TD%20102.pdf Ferreira, A. H.  B. 2000. “Convergence in Brazil:  Recent Trends and Long-​Run Prospects.” Applied Economics 32 (4): 479–​489.

442    Carlos R. Azzoni and Eduardo A. Haddad Ferreira, P. C. 2004. “Regional Policy in Brazil:  A Review.” Available at https://​www. researchgate.net/​publication/​228464935_​Regional_​policy_​in_​Brazil_​a_​review. Garcilazo, J. E., J. O. Martins, and W. Tompson. 2010. “Why policies May Need to Be Place-​Based in Order to Be People-​Centred.” VOX CEPR’s Policy Portal, November 20, 2010. Available at https://​www.researchgate.net/​publication/​228464935_​Regional_​policy_​in_​Brazil_​a_​review. Gil, I. 2010. “Regional Development Policies:  Place-​ Based or People-​ Centred?” VOX CEPR’s Policy Portal, October 9, 2010. Available at http://​voxeu.org/​article/​ regional-​development-​policies-​place-​based-​or-​people-​centred. Goeschl, T., and D. C. Igliori. 2006. “Property Rights for Biodiversity Conservation and Development: Extractive Reserves in the Brazilian Amazon.” Development and Change 37 (2): 427–​451. Gonçalves, E., and E. Almeida. 2009. “Innovation and Spatial Knowledge Spillovers: Evidence from Brazilian Patent Data.” Regional Studies 43 (4): 513–​528. Gonzaga, D., and C. Azzoni. 2016. “Location and Wages: The Contribution of Firm and Worker Effects in Brazil.” 44o Encontro Nacional de Economia, ANPEC. Available at https://​www. anpec.org.br/​encontro/​2016/​submissao/​files_​I. Grimes, A. 2014. “Infrastructure and Regional Economic Growth.” In Handbook of Regional Science, edited by M. M. Fischer and P. Nijkamp, 331–​352. Berlin-​Heidelberg: Springer-​Verlag. Guilhoto, J. J. M., and M. A. R. Fonseca. 1998. “The Northeast and the Rest of Brazil Economies in a Mercosur Context, 1992–​2014: An Econometric Interregional Input-​Output Approach.” Studies in Regional Science 29 (1): 171–​185. Guilhoto, J. J. M., G. J. D. Hewings, and M. Sonis. 2002. “Productive Relations in the Northeast and the Rest of Brazil Regions in 1995:  Decomposition and Synergy in Input-​Output Systems.” Geographical Analysis 34 (1): 62–​75. Guilhoto, J. J. M., and U. A. Sesso Filho. 2005. “Estimação da matriz insumo-​produto a partir de dados preliminares das contas nacionais.” Economia Aplicada 9 (2): 277–​299. Guilhoto, J. J. M., and U. A. Sesso Filho. 2010. “Estimação da matriz insumo-​produto utilizando dados preliminares das contas nacionais: Aplicação e análise de indicadores econômicos para o Brasil em 2005.” Economia & Tecnologia. UFPR/​TECPAR. Ano 6, 23: 53–​62. Haddad, E. A. 1999. Regional Inequality and Structural Changes: Lessons from the Brazilian Experience. Aldershot, UK: Ashgate. Haddad, E. A., E. P. Domingues, and F. S. Perobelli. 2002. “Regional Effects of Economic Integration: The Case of Brazil.” Journal of Policy Modeling 24: 453–​482. Haddad, E. A., E. P. Domingues, and F. S. Perobelli. 2005. “Brazil-​Argentina Trade and Its Impacts in Brazilian States.” Investigaciones Regionales 7: 113–​137. Haddad, E. A., and G. J. D. Hewings. 1999. “The Short-​Run Regional Effects of New Investments and Technological Upgrade in the Brazilian Automobile Industry:  An Interregional Computable General Equilibrium Analysis.” Oxford Development Studies 27 (3): 359–​383. Haddad, E. A., and G. J. D. Hewings. 2005. “Market Imperfections in a Spatial Economy: Some Experimental Results.” Quarterly Review of Economics and Finance 45: 476–​496. Haddad, E. A., G. J. D. Hewings, F. S. Perobelli, and R. A. C. Santos. 2010. “Regional Effects of Port Infrastructure: A Spatial CGE Application to Brazil.” International Regional Science Review 33 (3): 239–​263. Haddad, E. A., G. J. D. Hewings, A. A. Porsse, E. Van Leeuwen, and R. S. Vieira. 2015. “The Underground Economy: Tracking the Higher-​Order Economic Impacts of the São Paulo Subway System.” Transportation Research Part A: Policy and Practice 73: 18–​30.

Regional Disparities   443 Haddad, E. A., F. S. Perobelli, E. P. Domingues, and M. Aguiar. 2011. “Assessing the Ex Ante Economic Impacts of Transportation Infrastructure Policies in Brazil.” Journal of Development Effectiveness 3 (1): 44–​61. Haddad, E. A., A. A. Porsse, and W. A. Rabahy. 2013. “Domestic Tourism and Regional Inequality in Brazil.” Tourism Economics 19: 173–​186. Haddad, E. A., and E. Teixeira. 2015. “Economic Impacts of Natural Disasters in Megacities: The Case of Floods in São Paulo, Brazil.” Habitat International 45: 106–​113. Haddad, M. A., and Z. Nedović‐Budić. 2006. “Using Spatial Statistics to Analyze Intra‐Urban Inequalities and Public Intervention in São Paulo, Brazil.” Journal of Human Development 7 (1): 85–​109. Imori, D. 2012. “Eficiência produtiva da agropecuária familiar e patronal nas regiões brasileiras.” MS thesis, FEAUSP. Available at http://​www.teses.usp.br/​teses/​disponiveis/​12/​ 12138/​tde-​19032012-​160907/​pt-​br.php. Kalatzis, A. E. G., C. R. Azzoni, and J. A. Achcar. 2008. “Financial Constraints and Investment Decisions: Evidence from a Highly Unstable Emerging Economy.” Applied Economics 40 (11): 1425–​1434. Kalatzis, A. E.  G., C. F. Bassetto, and C. R. Azzoni. 2011. “Multicollinearity and Financial Constraint in Investment Decisions: A Bayesian Generalized Ridge Regression.” Journal of Applied Statistics 38 (2): 287–​299. Laurini, M., E. Andrade, and P. Pereira. 2005. “Income Convergence Clubs for Brazilian Municipalities: A Non-​Parametric Analysis.” Applied Economics 37 (18): 2099–​2118. Lima, M. A.  M., and M. Resende. 2007. “Convergence of Per Capita GDP in Brazil:  An Empirical Note.” Applied Economics Letters 14 (5): 333–​335. Maciente, A. 2013. “The determinants of agglomeration in Brazil:  input-​output, labor and knowledge externalities.” PhD dissertation, University of Illinois. Magalhaes, A., G. J. D. Hewings, and C. R. Azzoni. 2005. “Spatial Dependence and Regional Convergence in Brazil.” Investigaciones Regionales 6: 5–​20. Martínez, S. H., J. Van Eijck, M. P. Cunha, J. J.  M. Guilhoto, A. Walter, and A. Faaij. 2013. “Analysis of Socio-​economic Impacts of Sustainable Sugarcane-​Ethanol Production by Means of Inter-​Regional Input-​Output Analysis:  Demonstrated for Northeast Brazil.” Renewable and Sustainable Energy Reviews 28: 290–​316. Menezes, T., R. Silveira-​Neto, and C. R. Azzoni. 2012. “Demography and Evolution of Regional Inequality.” The Annals of Regional Science 49: 643–​655. Menezes, T. A., R. Silveira-​Neto, J. L. Ratton, and C. Monteiro. 2013. “Spatial Correlation between Homicide Rates and Inequality: Evidence from Urban Neighborhoods.” Economics Letters 120: 97–​99. Moreno, E., and C. Azzoni, C. 2016. “Potencial inovativo da indústria nas regiões brasileiras. Revista Brasileira de Inovação, Unicamp 15 (2): 275–​304. Mossi, M., P. Aroca, I. Fernandez, and C. R. Azzoni. 2003. “Growth Dynamics and Space in Brazil.” International Regional Science Review 26 (3): 393–​418. Neumark, D., and H. Simpson. 2014. “Place-​Based Policies.” NBER Working Paper No. 20049, National Bureau for Economic Research, April. Available at http://​www.nber.org/​papers/​ w20049. Oliveira, H., and E. P. Domingues. 2005. “Considerações sobre o impacto dos FNO e FCO na redução da desigualdade regional no Brasil.” In XXXIII Encontro Nacional de Economia, 2005.

444    Carlos R. Azzoni and Eduardo A. Haddad Perobelli, F. S., E. A. Haddad, S. Q. Bastos, and E. Pimentel. 2007. “Fiscal Incentives and Regional Development Projects: Mercedes-​Benz in Juiz de Fora (MG), Brazil 1996/​1999.” Latin American Business Review 7 (3–​4): 49–​75. Perobelli, F. S., and C. C. C. Oliveira. 2013. “Energy Development Potential: An Analysis of Brazil.” Energy Policy 59: 683–​701. Pessôa, S. A. 2001. “Existe um Problema de Desigualdade Regional no Brasil? Anais do Encontro Nacional da Anpec, 2001.” Available at http://​www.anpec.org.br/​encontro2001/​ artigos/​200105174.pdf. Piketty, T. 2014. Capital in the Twenty-​First Century. Translated by Arthur Goldhammer. Cambridge, MA: Belknap/​Harvard University Press. PNAD—​Pesquisa Nacional por Amostra de Domicílios, IBGE. (Various years). Available at http://​w ww.ibge.gov.br/​home/​estatistica/​indicadores/​t rabalhoerendimento/​pnad_​ continua/​default.shtm. Porsse, A., E. A. Haddad, and E. P. Ribeiro. 2007. “Economic Effects of Regional Tax Incentives:  A General Equilibrium Approach.” Latin American Business Review 7 (3–​4): 195–​216. Reis, E., and L. Monastério. 2008. “Mudanças na concentração espacial das ocupações nas atividades manufatureiras no Brasil—​1872z-​1920.” IPEA, RJ, Texto para discussão 1361. Available at http://​www.ipea.gov.br/​agencia/​images/​stories/​PDFs/​TDs/​td_​1361.pdf. Resende, G. M. 2012. “Measuring Micro-​and Macro-​Impacts of Regional Development Policies: The Case of the FNE Industrial Loans in Brazil, 2000–​2006.” Regional Studies 48 (4): 646–​664. Resende, G. M. 2013. “Regional Development Policy in Brazil:  A Review of Evaluation Literature.” REDES—​Rev. Des. Regional, Santa Cruz do Sul 18 (3): 202–​225. Santos, G. F., E. A. Haddad, and G. J.  D. Hewings. 2013. “Energy Policy and Regional Inequalities in the Brazilian Economy.” Energy Economics 36: 241–​255. Schettini, D., and C. R. Azzoni. 2015. “Determinantes regionais da produtividade industrial: O papel da infraestrutura.” In Produtividade no Brasil, Desempenho e Determinantes, edited by Fernanda De Negri and Luiz Ricardo Cavalcante, Vol. 2: Determinantes, 391–​414. Brasília: IPEA. Schettini, D., and C. R. Azzoni. 2013. “Diferenciais regionais de competitividade industrial do Brasil no século 21.” Economia (Brasília) 14: 361–​387. Schettini, D., C. R. Azzoni, and A. Paez. 2011. “Neighborhood and Efficiency in Manufacturing in Brazilian Regions:  A Spatial Markov Chain Analysis.” International Regional Science Review 34 (4): 397–​418. Silva, A. M., G. M. Resende, and R. Silveira-​Neto. 2007. “Uma avaliação da eficácia do FNE, no período 1995–​2000.” Análise Econômica (UFRGS) 25: 233–​261. Silva, A. M., G. M. Resende, and R. Silveira-​Neto. 2009. “Eficácia do gasto público:  uma avaliação do FNE, FNO e FCO.” Estudos Econômicos 39 (1): 89–​125. Silva, C. E.  L., and G. J.  D. Hewings. 2012. “Locational and Managerial Decisions as Interdependent Choices in the Headquarter-​ Manufacturing Plant Relationship:  A Theoretical Approach.” The Annals of Regional Science 48: 703–​7 17. Silveira-​Neto, R., and C. R. Azzoni. 2011. “Non-​Spatial Government Policies and Regional Income Inequality in Brazil.” Regional Studies 45 (4): 453–​461. Silveira-​Neto, R., and C. R. Azzoni. 2012. “Social Policy as Regional Policy:  Market and Nonmarket Factors Determining Regional Inequality.” Journal of Regional Science 52 (3): 433–​450.

Regional Disparities   445 Silveira-​Neto, R., G. Duarte, and A. Paez. 2015. “Gender and Commuting Time in São Paulo Metropolitan Region.” Urban Studies 52 (2): 298–​313. SNIS. 2016. Sistema Nacional de Informações sobre Saneamento: Diagnóstico dos serviços de água e esgotos—​2014. Brasília: SNSA/​MCIDADES. Soares R. B., J. N. P. Sousa, and A. Pereira Neto. 2009. “Avaliação de impactos do FNE no emprego, na massa salarial e no salário médio em empreendimentos financiados.” Revista Econômica do Nordeste 40 (1):217–​234. Vieira, F. V., E. A. Haddad, and C. R. Azzoni. 2014. “Export Performance of Brazilian States to Mercosul and Non-​Mercosul Partners.” Latin American Business Review 15 (3–​4): 253–​267.

Chapter 21

Brazil’s Nort h e ast Alexandre Rands Barros

21.1. Introduction The Northeast is a macro-​region with 56.56 million inhabitants, representing 27.7% of the total population of Brazil, according to 2015 statistics from IBGE (Brazilian Institute of Geography and Statistics). This population is distributed across an area of 1,554,291.74 square kilometers, some 18.3% of the country’s total area. If the Northeast were a country itself, it would be the fourth largest country by area in Latin America, and second largest if it were in Europe, behind only Russia. Its population would be the third largest in Latin America, behind only the rest of Brazil and Mexico, and the sixth largest if it were in Europe.1 This population is spread over 1,794 municipalities, and includes 102 cities with populations over 50,000. The most important metropolitan areas are Recife (population 3.94  million), Salvador (population 3.98  million), and Fortaleza (population 4.02 million). The region has a very low per capita gross domestic product (GDP) (US$7,648.99 in 2015), when nominal figures are corrected for purchasing power parity. This would make it the second poorest country in South America and the sixth poorest in Latin America. The Northeast’s per capita GDP reached only 51% of the Brazilian average in 2015. Thus, this region is of particular interest not only because it is large and populous, but also because it is particularly poor in a country that allows free trade and population flows among regions. The Northeast was also the region that was first heavily populated by European settlers, as well as a significant place of forced settlement for those of African origin as a result of the Atlantic slave trade. For example, it is estimated that in 1600 the Northeast held around 94.5% of people of African descent in Brazil. Only by the eighteenth century did the picture begin to change, as the Northeast’s share of Brazil’s overall population started to fall. The region’s early colonization, together with its current relative underdevelopment and large size (in population, geographical area, or even total GDP), merits special attention. The following section (21.2) of the chapter reviews some general economic

Brazil’s Northeast   447 and geographical data, while section 21.3 analyzes general historical developments. Section 21.4 discusses the causes of underdevelopment, while section 21.5 provides an overview of regional policies relied upon in order to overcome economic disadvantage. Section 21.6 analyzes the prospects for the future convergence of per capita GDP with the rest of the country. Section 21.7 summarizes the major conclusions of the chapter.

21.2.  Economics and Geography The Northeast is a region whose per capita GDP has been about half the national average over the last 50 years. Figure 21.1 charts the evolutions of its proportional contribution to GDP since 1939. It can be seen that while the region’s share has oscillated, there is no actual fall in regional disparities according to this measure. This relative backwardness makes the region the poorest among all Brazilian regions. When taken together with the size of the local population, such statistics make the Northeast the major focus of the regional question in Brazil. The economic problems of the region are not restricted to aggregate per capita GDP. Within the Northeast, a higher share of population lives under the national poverty line, and it is also the region with the lowest averaging schooling years and life expectancy of 0.60 0.50 0.40 0.30 0.20 0.10

NE/BR

2014

2011

2008

2005

2002

1999

1996

1993

1990

1987

1984

1981

1978

1975

1972

1969

1966

1963

1960

1957

1954

1951

1948

1945

1942

1939

0.00

Long-term trend

Figure 21.1.  Per capita GDP proportion: Northeast/​Brazilian average, 1939–​2015. Note: Long-​term trend is estimated from model of structural brakes identification before 2000; average growth rate between 2001 and 2006; simple stability of average between 2006 and 2011, and the average growth rate was introduced again after this latter year. Sources: IBGE Regional accounts for 1939 to 2014 and author’s estimations from data from IBGE, PNAD for 2015.

448   Alexandre Rands Barros Table 21.1 Basic Economic and Social Indicators for the Northeast and for Brazil Overall Statistics

Northeast

Brazilian Average

Share of population living in households with per capita income up to half minimum wage (2015).

35.0%

18.9%

Average schooling years of population over 25 years old (2015)

6.6

7.9

Illiteracy rates of population over 15 years old (2015)

16.2%

8.0%

Share of urban population in total (2015)

73.1%

84.7%

Life expectancy (years) (2014)

72.5

75.1

Source: IBGE, PNAD.

its population, as seen in Table 21.1. Most social indicators paint the same picture: the Northeast is the region with the lowest social development in Brazil and it is where Brazilian underdevelopment is most pronounced. The Northeast is also the region with the second-​highest income concentration among all Brazilian macro-​regions, behind only the Center-​West. The concentration of other social and economic indicators, such as land, education, and access to water supply, also shows that the Northeast is the Brazilian region with the strongest social distortions. In spite of its already advanced urbanization, it is still the region with the lowest share of the population living in urban areas, as seen in Table 21.1. Together with the low educational level of this rural population, these are factors that help to explain the region’s low GDP per capita and social development. Another important statistic through which to understand the economic problems of the region concerns the ethnic composition of its population. In 2014, 73% of the population in the Northeast was black, mixed race, or indigenous Brazilian, while 47% of the population of the rest of Brazil was included in these ethnic groups (data from IBGE and the PNAD, National Household Survey). One of the perennial highly negative legacies of Brazil’s development process has been the systematic restriction of access for these groups to quality public services, notably education. Barros (2011, 2016) argues that this discriminatory process has adversely affected human capital formation among these groups and, by extension, the development of the region. Other important geographical features of the Northeast lie in its weather conditions. The region has mainly a tropical climate, with 57.53% of its area comprising semiarid lands. This region, called the Sertão, is often subject to periods of drought, which brings severe hardship to all farming activities in the region. Nevertheless, the Sertão is still home to more than 50% of the population of the Northeast. As these are lands of poor productivity, the idea that their prevalence in the Northeast is one of the major causes of its relative underdevelopment is common outside academic research, such as in journalism or policy discussion. This hypothesis is discussed further in section 21.4.

Brazil’s Northeast   449

21.2.1. The Nature of Regional Disparities There are two possible pure sources of per capita income differences among individuals, although many combinations of them can also hold in reality. Inequality can arise because an individual has a set of productive attributes that generates higher income, although each particular attribute yields the same marginal income for all individuals, whatever the amount they have. Inequality arises in this case from uneven distribution per capita of the set of productive attributes. The second source of income inequality arises from the possibility that two individuals with the same set of productive attributes may receive different incomes. In other words, there are unequal returns to similar productive attributes when they are owned by different individuals, whatever the source of that difference—​perhaps gender, ethnic origin, or even region of residence. This is a case of discrimination in productive factors markets. These two sources of differences in individual incomes can be extrapolated to regional average per capita incomes, if the sets of individuals living in each of the two regions are subject to these two income differences when compared to the set of individuals living in other regions. An important question for understanding regional disparities in Brazil is what combination of these two possibilities explains existing regional inequalities. Barros (2011) presents estimations on income for counties with different average performance in education. These suggest that regional inequalities in the distribution of this productive asset were the major and only source of regional disparities in Brazil. If the Northeast’s human capital indicator were altered such that the regional average matches the Southeast and South averages, there would be no regional disparity. Similarly, Barros (2012) presents statistical exercises for the same purpose, but relying on individual income determination equations. The results are similar. Individuals living either in the Northeast or the Southeast and South have the same income whenever they have the same human capital. These two estimations indicate that regional inequalities in Brazil arise from differences in the per capita distribution of productive attributes, more specifically human capital. They do not have a discriminatory source.

21.3.  Historical Background Brazil’s initial connection with Europe commenced in the sixteenth century and was rooted in the extraction of timber as the key economic activity (Furtado 1959b; Simonsen 1957). In fact, this was the first economic activity that could sustain long and expensive transoceanic trade; precious metals were not found in the country in the first initial exploratory expeditions of the Portuguese. Timber extraction spread over most of the Brazilian coast, from São Paulo to Maranhão, along a stretch of rainforest called the Mata Atlântica. Thus, the first European settlements in Brazil were widespread along the

450   Alexandre Rands Barros whole coast, although the Northeast was already a priority because of its greater proximity to Europe. Nonetheless, the sixteenth century saw the beginnings of sugarcane plantations, as sugar was a commodity that could, through its productive process, add value sufficient to justify the expensive inter-​ocean trade, if undertaken on a large scale. The tropical lands of the Northeast coast were found very suitable for such agrarian exploitation. Local temperatures were high and there were abundant rains and rich soils, as demanded by a grass species such as sugarcane. Thus, European economic occupation of Brazil began in this region (Baer 2008). This is why in 1700, according to estimates, more than 70% of the non-​native population was living in the Northeast. Sugarcane production created a market for the expansion of local activities in the surrounding region. Livestock, including mainly cattle, horses, and donkeys, and foodstuff production, such as manioc flour, beans, and maize, were the major activities, which expanded in regions further from the coast. The sugar industry also demanded many urban activities, such as wholesale and retail trade, and services such as health and education, in addition to all state control over international trade and associated tax collection. These activities justified the settlement of some cities in the region, which became important colonial centers, such as Olinda, Recife, and Salvador. Available data indicate that the Northeast enjoyed a higher per capita GDP than the Brazilian average over the initial two and a half centuries of colonization. Settlements around sugarcane production provided the impetus for such development leadership. Only after 1750, once the gold cycle had assumed leadership of Brazilian exporting output, did Northeastern per capita GDP fall below the Brazilian average. Nevertheless, the Northeast recovered its developmental leadership (or at least equivalence) in 1830, with the new upsurge in the role of sugar in Brazilian exports; the Northeast was still the major locus of its production in Brazil. The region only started to fall behind, more strongly this time, after 1870. Coffee exports had begun to lead the Brazilian economy, and most incoming immigrants settled in the South, in particular in the coffee-​growing regions in the Southeast. As a consequence, the Northeast’s relative per capita GDP fell to such an extent that it reached only 48% of the Brazilian average in 1939. Since then it has been oscillating around that level with little significant change.

21.4.  Causes of Underdevelopment Competing hypotheses have been advanced in the literature to explain the Northeast’s relative economic backwardness. Some of them are similar in nature to hypotheses used to explain Brazil’s comparative underdevelopment as a whole. The most important are briefly reviewed here.

Brazil’s Northeast   451

21.4.1. Climate Adversity Hypothesis This hypothesis argues that the Northeast is poor because of the large share of drought-​ prone land that it contains. Thus, the existence of the Sertão and its area and population sizes would be the ultimate cause of the Northeast’s relative backwardness. According to this hypothesis, geographical adversity would lead to the low productivity of a reasonable share of the local working population engaged in primary activities and of a large share of exploited land areas. Consequently, average output per worker would be low in this area, bringing down the region’s average. What this view does not explain is why still more than 50% of the total population lives in these drought-​prone areas and why other areas with a more regular rain distribution—​such as the hinterlands of Maranhão and the rural parts of the coastal areas—​have an even lower per capita GDP than the Sertão. This point is stressed by Barros (2011, Chapter 4).

21.4.2. Structuralist and Dependency Hypothesis The climate adversity hypothesis was challenged by Furtado (1959a), who transplanted the logic of Latin American structuralism to explain the relative backwardness of the region. His argument was that Brazilian industrialization was boosted in the late nineteenth and early twentieth centuries, periods when coffee was the major primary commodity exported from Brazil (see also Furtado 1959b). As a consequence, the local market was larger in the Southeast, where this commodity was produced. This brought to life a strong industrial base in the South and left the Northeast relatively more dependent on primary activities. Later, national protectionist policies for infant industries generated a deterioration of the net barter terms of trade of Northeast exports, as they increased the prices of industrial goods imported by the region, which were increasingly purchased from the Southeast. This reduced local capital accumulation and slowed its growth. This was, according to Furtado, one of the major causes of its relative backwardness compared to the Southeast.2 Additionally, the Northeast also lost relative dynamism, when compared to the Southeast, because industry in the latter region was the most dynamic sector in overall national productivity gains. Moreover, industry experienced more dynamic demand. According to this hypothesis, such industrial concentration in the Southeast was not reversed over the years because of the internal economies of scale associated with these productive structures. Thus only a public policy to promote industrial development in the Northeast would prove able to reverse regional disparities in Brazil. Although Furtado’s (1959a, 1959b) contributions are the primary presentation of this idea, it gained ground in the literature thanks to further influential studies. Thus, versions of this hypothesis can be found in works by Baer (2008, Chapter 11), Hirschman (1963, Chapter 1), and Goodman and Albuquerque (1974).

452   Alexandre Rands Barros Building on these structuralist economic ideas, dependency theorists also attempted to explain why such phenomena as specialization in primary commodities persisted, why the region featured late industrialization, and the reasons behind the absence of policy to tackle such unbalanced industrial growth. Dependency theorists argued that alliances of local dominant social classes with segments of others from outside the region assured that the institutional framework and macroeconomic policies fostered the existing economic specializations. This unhealthy class alliance was ultimately responsible for the relative underdevelopment of the region. According to these views, the major political alliances in the Northeast were led by the large landowners, who were associated with wholesale traders—​who benefited from local imports of industrial goods and exports of primary commodities—​and with industrial leaders in the Southeast to foster economic policy arrangements that would reproduce local specializations. Such political alliances bestowed great political power upon large landowners in the region and even nationally. Moreira (1979), Oliveira (1981), and Kon (1976) are authors who have developed analyses within this dependency tradition.

21.4.3. The Comparative Advantage Hypothesis Nathaniel Leff (1972) challenged the hypothesis that industrialization was the cause of regional inequality and argued that differences in export growth in the nineteenth century were the major source of Brazilian regional disparities. The Southeast expanded under the impetus of coffee exports in that period, while the Northeast’s exports were mainly concentrated in sugar and cotton. Coffee exports, however, had grown much faster than those of sugar and cotton (although world markets for these two last commodities had expanded much faster than Brazilian exports of them). This was, for Leff, the explanation for the widening of per capita income differences. Leff (1972) argues that this difference in exporting growth arises because the Northeast was trapped in a unified currency with the Southeast. The higher relative comparative advantage of Brazil in coffee production forced the local exchange rate to reach an equilibrium level that would be dominated by coffee’s relatively high productivity. Thus, sugar and cotton growers ended up facing an overvalued currency for their relative costs when compared to other producers around the world. The consequence was that their profits were squeezed and they did not expand output to keep pace with world market increases. This determined the lower growth rate of GDP in the Northeast over the nineteenth century, compared to that of the Southeast. Leff argues that had the Northeast been independent and had its own currency, its devaluation could have assured higher competitiveness of sugar and cotton exports and generated a higher growth rate for GDP. It is also worth mentioning Leff ’s introduction of some political economy ideas to explain why, in the early 1920s, when there was overproduction of coffee in the world market, there was not a devaluation of the exchange rate that could have reversed

Brazil’s Northeast   453 the profit squeeze on sugar and cotton producers. In his view, successive Brazilian governments preferred to expend resources on attempting to avert coffee price falls. This prevented the erosion of unequal development between the two regions.

21.4.4. Rands Barros’s Hypothesis In earlier work (Barros 2011), I challenge the structuralist and dependency hypotheses, stressing that there is a free flow of capital among Brazilian regions and that, consequently, the generation and appropriation of surplus locally is not relevant to investments in any region. Thus, if there are more profitable opportunities for investments in one region, capital will flow for this reason, regardless of where the savings are generated. Such flows exist even on an international level (Barros 2016, Chapter 3), but they are stronger still between regions within Brazil. Furthermore, I argue that the flow of financial capital is faster than that of people and, therefore, human capital. Consequently, physical capital also adjusts faster than the other factors of production. Thus, the adjustment of physical capital to the relative regional availability of human capital and labor is made through the interregional flow of the former. Of course, there are also flows of labor and human capital among regions, but finance and physical capital adjust faster. Thus it is the latter two varieties of capital that are responsible for most of the adjustments when imbalances concerning factor proportions have to be corrected. This means that the lack of savings and physical capital need not be a cause of regional inequalities (as would be held by both the structuralist and the comparative advantage hypotheses). Additionally, I stress that there exists reasonable arbitrage in all markets for factors of production. Consequently, the returns to each are similar across regions. Certainly such arbitrage is not complete, but it is strong enough to keep differences in returns to the factors of production among regions in Brazil quite small. A consequence of such a hypothesis is that two workers with the same qualifications will earn a very similar wage independently of the region where they are settled. Certainly, the measure of such differences has to correct for the costs of living. It should also correct for fringe benefits, but they are not relevant in a homogeneous country such as Brazil. According to this view, regions with a higher per capita stock of human capital will have a higher per capita GDP, as there is complementarity between these two types of capital. Hence, total output per worker will be higher when there is higher per capita human capital, as it implies that the per capita generation of income will correspond to returns to a higher stock of human and physical capitals, plus the return to workers—​ wages—​which are similar in the many regions. Altogether the generated income will be higher in the region with the higher per capita stock of human capital. Certainly, even under the full arbitrage in the factors markets hypothesis, differences could still arise from the proportion of the active population to the total population, but these can be assumed to be small. Hence, all of these arbitrages tend to generate higher per capita income in the region with a higher per capita stock of human capital.

454   Alexandre Rands Barros In addition, I postulate that there is a tendency for two families, each living in one region in a two-​regions model, to replicate their disparities in human capital over generations. There are no market incentives working to narrow this gap. Thus, regional disparities in human capital tend to be reproduced over generations. Only if there were sufficiently strong public educational policies would there be a decrease in this disparity. As there have not been such policies in Brazil, since the state has always been controlled by local elites representing a minority of the total population, there has been therefore a long-​term reproduction of regional disparities in human capital. I also show how such an equilibrium, with different per capita stocks of human capital among regions, is stable; its existence does not generate any endogenous market forces to push it to any other equilibrium, since all factors of production will receive the same returns in whatever region they are employed. Thus, if inequalities in the per capita stock of human capital arise because of some historical eventuality, this will generate regional inequalities that will be reproduced over generations. As such, patterns of regional inequality have tended to remain ingrained. Particular circumstances surrounding the settlement of many Brazilian regions up to 1939 generated an original disparity in the per capita stock of human capital, which was particularly fostered by immigration in the second half of the nineteenth century and early years of the twentieth century. This inequality was reproduced over the years thereafter. This, then, is why there is no convergence of per capita GDP, as postulated by models with only one good, and no flow of factors of production among regions. Given historical circumstances and the ongoing evolution of public policy, the South and Southeast would eventually have more developed human capital than the Northeast. This would generate the regional disparities. The arbitrage assumption, that two people with the same attribute would have the same earnings, is testable, as well as the assumption that differences in the availability of human capital among regions is the cause of regional disparities. Barros (2012) tests the assumption, while Barros (2011) empirically verifies the conclusions that flow from it. The test of the conclusion (presented in Barros 2011, Appendix 1) is facilitated with county-​level data in Brazil. A relationship is estimated between per capita GDP in the country as a function of indicators of human capital, productive effort, and cost of living differences. Next, per capita GDP in each county is simulated under the hypothesis that the regional average of these variables in the Northeast was the same as in the South and Southeast together. The outcome is that average per capita GDP in the Northeast would be slightly higher than for these other regions. Furthermore, differences in human capital indicators account for such differences, as the impact of productive effort and cost of living offset each other. Barros (2012) tests the hypothesis of two persons living in different regions earning the same income. This test is made through estimations of Mincer-​type equations determining individual incomes as a function of regional dummies, among other explanatory variables. Although this type of test is common in the empirical literature on the Brazilian economy, and normally attributes a very large role to regional dummies to explain income differences, I argue that such results are a consequence of

Brazil’s Northeast   455 missing variables, which have been proxied by regional dummies. To offset this mistake, I created several groups of workers that have more similar features, and estimated Mincer-​type individual income equations within these groups. The major results, shown in Table 21.2, indicate that inequalities in individual incomes, when the impacts of these missing variables are reduced, are quite small. Thus, the assumption of arbitrage among regions in the markets for factors of production, particularly labor and human capital in this case, is a good approximation, which has been neglected by the other hypotheses aiming to explain regional disparities in Brazil. Following the tradition of recent modern growth theory, my hypothesis identifies both immediate and more fundamental causes of regional disparities in Brazil. The former is the difference in per capita human capital, when local per capita stock is compared to other regions, especially the Southeast and South. As for the fundamental cause, I stress class conflicts in the region, whose result has been the almost complete negligence within public policy of efforts to enhance access to better education for the Table 21.2 Proportion of Average Individual Incomes between Residents of Northeast and Southeast/​South (after Corrections for Individual and Some Regional Attributes) Correction Made Worker Category

Original

Only Individual Education

Total Education

Non-​metropolitan mixed men with labor market registration

0.85

0.99

1.15

Non-​metropolitan white men with labor market registration

0.76

0.96

1.03

Metropolitan white men with labor market registration (excluding Rio de Janeiro and São Paulo)

1.11

0.96

0.95

Non-​metropolitan mixed men with non-​labor market registration

0.78

0.85

0.92

Metropolitan white men with non-​labor market registration (excluding Rio de Janeiro and São Paulo)

0.93

0.87

0.93

Metropolitan mixed men with labor market registration (excluding Rio de Janeiro and São Paulo)

1.05

0.93

0.87

Non-​metropolitan white men with non-​labor market registration

0.71

0.81

0.81

Metropolitan mixed men with non-​labor market registration (excluding Rio de Janeiro and São Paulo)

1.19

1.15

1.17

Source: Barros (2012). Note: Individual indication corrects only for the educational attainment of the individual, while total indicator corrects also for the educational attainment of the region, capturing Lucas’s (1988) hypothesized existence of externalities from education.

456   Alexandre Rands Barros poorer social strata. Those of European descent controlled the state and maintained public education as a very low priority in the Northeast. As a consequence, there was no narrowing in the educational gap between regions. The low emphasis on granting educational access to the poorer social strata, mainly African and native descendants, was a consequence of adverse national political priorities. Regions with a higher proportion of these social groups, the Northeast in particular, were more detrimentally affected by these political priorities (Oliveira 2017).

21.4.5. Institutionalist Hypothesis There is no formal presentation of the institutionalist hypothesis to explain the Northeast’s relative backwardness, although many economists have argued that institutions are the major cause.3 The idea is that there was not an efficient allocation of resources to human capital building in the region because of its institutional framework. The colonial heritage built proportionally more extractive, rather than inclusive, institutions in the region. For example, slavery was stronger in the Northeast, as well as the role of large farms in the agrarian structure, which was beyond the optimal size for economic efficiency and which fostered social relationships that blocked the possibility of improvements to the educational systems. These extractive institutions dumped human capital accumulation and the local emergence of efficient firms. In the work mentioned in the previous section (e.g., Barros 2011), I did not disregard the role of institutions. In my view, however, they are generated and continuously evolve as a consequence of the social conflict or class struggle in society. Hence, they are more volatile than is typically held by the institutionalist view, and have class struggle as their ultimate determinant. Social classes or groups always try to solidify their gains in social disputes through adjustments to the institutional framework. Nevertheless, when they face setbacks, there are changes in institutions that undo their temporary gains. Thus, institutions adjust to social disputes, although they also set the stage for them. The human capital building path is determined by these disputes, whose results are temporarily frozen in institutions. Thus, if the basic public educational institutional framework in Brazil has been ineffective at reducing social and regional disparities, it is not a consequence of bad luck. Rather, it reflects the interests of those social groups that are able to influence its structure.

21.4.6. Summing Up Despite the differences in these interpretations of the fundamental causes of the Northeast’s underdevelopment, some clear conclusions arise. First, the expansion of coffee production in the Southeast was an important factor in the development of regional disparities. It might even be considered their cause, as in the comparative advantage hypothesis, or an underlying determinant that supported differences in the

Brazil’s Northeast   457 industrialization path, as in the structuralist hypothesis. Alternatively, it may simply have provided a stimulus for a change in the evolution of human capital in the South and Southeast, as in the Rands Barros hypothesis. All hypotheses, then, acknowledge some role for the growth of the coffee industry in Southeast Brazil in explaining the widening of regional disparities. A second conclusion relates to the timing argument presented in section 21.3 that the origin of the sharp widening in disparities was in the nineteenth century. Most major explanations of the relative underdevelopment of the Northeast—​the structuralist, dependency, comparative advantage, and Rands Barros views—​all share this conclusion. A third conclusion also shared by most (although not all) of the explanatory hypotheses is that social relationships played a major role in determining regional inequalities. This conclusion is found in the dependency, institutionalist, and Rands Barros views. Thus, it seems reasonable to suggest that some focus on political economy issues is necessary to understand the persistence of regional inequality in Brazil.

21.5.  Regional Policies in Brazil Regional policies directed toward Northeastern development have to some extent reflected the competing explanations for the region’s underdevelopment.4 Whichever hypothesis appeared ideologically and politically stronger at a given moment tended to shape policy. Often political conflicts involved different social alliances, adopting different hypotheses to explain the relative backwardness of the region. Typically this ideological defense arose from the actual benefits that the various social groups involved stood to gain from the policies. Nevertheless, political alliances could force more than one hypothesis to drive forward regional development policies at the same time. As an example, it was not difficult to find policies to promote industrial development coexisting alongside public efforts to reduce the adverse impact of weather factors. Prior to Furtado’s structuralist hypothesis, the dominant idea was that climate hazards were the major impediment to the development of the Northeast. As such, policy emphasis tended to be on programs to reduce the impact of droughts in the region. DNOCS (Departamento Nacional de Obras contra a Seca) was the major institution, created in 1909, to reduce the economic problems faced by the region. The primary goal of this institution was to build infrastructure that could alleviate the adverse impact of droughts through the construction of ponds, dams, and collection wells. During droughts the effort to build such infrastructure also provided a source of income for those small farmers and rural workers who were severely hit by the droughts and could not count on their main source of income from primary activities. CODEVASF was another important regional institution, created in 1948. Its goal was to develop economic activities along the São Francisco River, activities that would be resilient due to the fact that this river did not dry up during droughts. CODEVASF’S policy instruments were mainly aimed at building up infrastructure to facilitate the

458   Alexandre Rands Barros access of the population to areas where they could be employed productively near the river, thereby encouraging more resistant agriculture and livestock farming to flourish in the Northeast. It is worth mentioning that both CODEVASF and DNOCS still exist, although the policy of combating the effects of drought lost its ideological and political prominence in the 1960s. Still under the dominance of the climate adversity hypothesis, in 1952 Banco do Nordeste do Brasil (Bank of the Northeast of Brazil) was created. It is a regional development bank, whose major original goal was to finance economic projects that could help alleviate the adverse impact of droughts. The 1946 Constitution had established a drought fund, which was regulated in 1949, but experiences surrounding the management of these resources were less than satisfactory until the creation of the Bank in 1951. Despite its primary focus on financing development in areas subject to droughts, it was already active in the more general promotion of economic activities. After 1960, the priorities of the Banco do Nordeste changed, with more emphasis on industrial projects. In the 1990s its priorities changed again, the major focus moving to small firms, including primary activities projects (although large restructuring projects still featured significantly within the overall allocation of resources, since they were seen as important in changing the local productive structure). Over the years, the Bank financed private projects at subsidized interest rates and over repayment periods longer than those available in the Brazilian private financial market. Thus, in times of sufficient funding, the Bank has been an important source of credit in the Northeast; funding availability, however, has varied over the years depending on federal government priorities.5 Even before Furtado’s major contributions, which guided regional development policies in Brazil after 1960, the view that the promotion of industrial growth is an important pillar of regional development had gained ground. These ideas had already emerged in Brazil in the nineteenth century, with the support of influential policymakers such as Sezerdelo Correa and Rui Barbosa. There was also theoretical support internationally from contributions such as those of Rosenstein-​Rodan (1943, 1944); see also Lebret (1955) and Singer (1956) for examples. Thus, some support for industrial promotion already appeared in earlier policy proposals. Banco do Nordeste do Brasil and CODEVASF already included the possibility of industrial promotion in their development strategies, although this element was not dominant until the 1960s (following Furtado’s rise to prominence). In 1960, the federal government, strongly influenced by Celso Furtado, created SUDENE (Superintendência de Desenvolvimento do Nordeste), a regional planning institution whose major goal was to plan the Northeast’s development, in addition to supplying a special source of financing for local industrial development (Hirschman 1963, Chapter 1). SUDENE collected much important data on the region and managed a captive capital market in which local businesses could raise some investment resources to facilitate the expansion or establishment of new activities. The resources would come from a rebate option that enterprises could take on their income taxes, so that they would still be able to own part of these resources to apply to investments in the region.

Brazil’s Northeast   459 This became an important source of liquidity for local investors. As they could purchase the investment bonds applied in their enterprises at discount prices in the captive capital market, this ended up as a sizable subsidy to capital in investments in the region. Local industrial development would then occur through this mechanism, according to the technocratic view of the time. This policy was fully in line with the structuralist interpretation of the source of regional inequalities. Nevertheless, with extensions of the sectors to which the resources of this fund from income taxes of enterprises could be employed, the policy increasingly became more of a clever instrument to subsidize capital in the region than necessarily a stimulus to industrial development. More recently, all the preceding policies have become marginalized in the face of two major national developments. The first was a shift in the criteria governing the distribution of federal government tax proceeds among state and municipal authorities. A substantial transference of resources through these instruments emerged from changes in these distributional rules. The second shift was the rise in the importance of social policies in total federal government expenditures. The expenditure bias here moved toward transfers or subsidies directed at the poorest social strata. Naturally, then, less wealthy regions received a substantial share of the proceeds of these policies. Indeed, this latter shift has become de facto the second most important instrument of regional policy in Brazil; only the measure of taxing poorer regions comparatively lightly has more impact (Barros 2014).6 It is important to stress that policies carried out by SUDENE and Banco do Nordeste after 1960, which had Furtado’s structuralist ideas as their underlying theoretical support, may have helped to promote the industrialization of the Northeast. Nevertheless, they did not bring any significant reversal of the region’s relative backwardness, as the data in Figure 21.1 indicate.7 As a consequence, the motivations of policies started to change. A sharpened focus on the poorest social strata assumed primacy from the 1990s on, even if the underlying policy instruments were still the same as before. In parallel, transfers through the public sector to local authorities, either from states or municipalities, have recently become increasingly important. This has been a consequence of the political weight of the Northeast region, which, as noted earlier, is quite heavily populated. As a consequence, regional policies have become increasingly determined by the interests of political groups, whose social support comes from populist logic in which they capture votes directly from the population and then drive the state according to their own interests. Federal social policies work to capture the electorate, and transfers to the states and municipalities are partly used to finance the interests of the groups in power, often civil servants included among them. No hypoth­ esis explaining regional disparities lends support to these policies as instruments to overcome relative underdevelopment, although the dependency, Rands Barros, and institutionalist hypotheses all can account for their existence. Policies to promote human capital accumulation more intensively are supported by the Rands Barros hypothesis, yet have still not been implemented as regional policy in Brazil. However, the recent boost in resources to public education has had a higher

460   Alexandre Rands Barros impact on the supply of this service to the poorest regions. Nevertheless, the boost was still below that required to narrow the gap at the necessary speed. Brazil is still falling short in its obligation to properly train and educate people in the Northeast, compromising the ability of citizens there to compete effectively in labor markets.

21.6.  Perspectives of Convergence Although there has been no effective long-​term convergence of per capita GDP between the Northeast and the rest of Brazil in the last 75 years, it is possible that there will be some partial convergence in the future. There are two major reasons why one might anticipate the possibility of a future, perhaps partial, convergence of per capita GDP. First, there has lately been an increase in the political strength of those social classes that stand to benefit most from a good public education and further access to tertiary education. This has increased the resources flowing to education in the country and should ensure better access for the poorer social strata. Certainly, such changes will reduce the gap in per capita human capital among regions in the future. Second, Brazil is also a country in which there are high returns to education, especially to tertiary education. Nevertheless, these returns are already falling (Barbosa and Pessoa 2008), as there are policies to reduce income disparities, which press for relative rises in low wages and labor income. Such changes certainly reduce regional inequalities, as the population of the poorest regions benefits proportionally more from the relative fall in the return to education. This is because in poorer regions a higher share of the population has lower educational attainments. These changes will also generate a long-​term impact on regional disparities, forcing some convergence. The beginning of such a convergence has already been detectable in the past 13 years. The data in Figure 21.1 already indicate a change in the long-​term trend. From 2002 to 2006 there was some convergence, followed by stagnation in the trend up to 2011. However, from 2011 on, convergence is once more evident. As this trend has educational investment as one of its determinants, it may tend to persist and perhaps will still evolve positively for some years. Some convergence in schooling years is also evident (see Figure 21.2), a process that started in 1985. This reflects the re-​democratization process and a concomitant rise in power of the poorer sections of the population. Thus, there are grounds for hope that interregional per capita GDP will converge as a consequence of this convergence in educational investment. The convergence generated by these long-​term trends, however, will not be absolute. Per capita income will not be the same in all regions; only the gap will narrow a little, as differences in per capita income among people with distinct stocks of human capital will fall, at the same time that differences in the per capita stock of human capital among regions will also fall. Nevertheless, neither of these regional differences will be fully eliminated, as there is still more investment per student in the Southeast and South than in the Northeast.

Brazil’s Northeast   461

1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Figure  21.2.  Proportion of schooling years Northeast/​Brazil (population over 25  years old, 1981–​2015). Source: IBGE, PNAD. Extracted from IPEADATA.

21.7. Conclusions The Northeast is a large and poor region, where Brazil’s most acute social problems are overrepresented. Especially high rates of poverty, low per capita income levels, highly skewed income concentration, below-​average educational attainment, and deficient public medical care are just some of the challenging issues faced by this region. These features, together with the sheer size of the region—​in population, area, and GDP—​ unsurprisingly attracts the interest of researchers and policymakers. Despite its current situation, the Northeast was the richest region in Brazil until around 1750; it only fell behind following the surge of the South and Southeast in the late nineteenth century. The rise of the latter was driven by the coffee sector and subsequent industrialization. At the same time, in tandem with these structural changes, the South benefited from a higher level of investment in human capital. Investment here would, in the years to follow, continually outstrip that of the Northeast. The low initial human capital level of poorer groups in the Northeast created the conditions for lower productivity growth over time in the region. Furthermore, the low access to political power among poorer groups prevented the launch of an inclusive educational policy that would have increased human capital endowments and, through this, productivity and income levels. In the absence of such policies, reproduction of low human capital over generations has led to the current situation of low per capita GDP and development. Regarding the future, as argued earlier, a degree of convergence is likely to happen, but this will be only partial and limited. Improvements in educational policy will narrow the human capital gap among regions, but this will take a long time and will be limited, since expenditure per student is still much lower in the Northeast than further south. Nevertheless, the recent fall in the returns to education in Brazil has also shifted the equilibrium of disparities. It has already contributed to a recent small decrease in regional inequality. However, only a radical change in the region’s educational policy will speed up convergence.

462   Alexandre Rands Barros

Notes 1. Data in this paragraph taken from the World Bank’s World Development Indicators, available at http://​data.worldbank.org/​data-​catalog/​world-​development-​indicators/​. The Northeast would be the fourth largest country by area in Latin America behind Brazil, Argentina, and Mexico. It would be second in Europe, and has almost three times the area of France. (Turkey is considered to be in Asia for the purposes of this ranking.) 2. These arguments of Celso Furtado’s were also used in a publication by the Working Group for the Development of the Northeast (GTDN 1959), which was largely written by Furtado. 3. Naritomi, Soares, and Assunção (2007) give a version of this hypothesis to explain the backwardness of a particular region in the State of Minas Gerais, which is included in the Northeast, but this represents only a small part of the region. 4. Maia Gomes (2009) and Baer (2008) offer surveys of regional policies in Brazil. Hirschman (1963, Chapter 1) gives a history of policies until 1950. Barros (2011, Chapter 7) also offers an analytical presentation of these policies. 5. For a detailed presentation of the history of Banco do Nordeste, see Oliveira and Vianna (2005). 6. This comparison between these two policies assumed that social security is not a social transference policy, otherwise social transference policies would come first. 7. See also Ferreira (2005) for an evaluation of regional policies toward the Northeast. His basic conclusion is that all were either irrelevant or had negligible impacts.

References Baer, W. 2008. The Brazilian Economy, 6th edition. Boulder, CO: Lynne Rienner. Barbosa Filho, F. H., and S. Pessoa. 2008. “Retorno da educação no Brasil.” Pesquisa e Planejamento Econômico 38 (1): 97–​125. Barros, A. R. 2011. Desigualdades regionais no Brasil. Rio de Janeiro: Elsevier. Barros, A. R. 2012. “As desigualdades regionais de renda entre indivíduos no Brasil.” Recife, UFPE. Mimeo. Barros, A. R. 2016. Roots of Brazilian Relative Economic Backwardness. London: Elsevier. Bolt, J., and J. L. van Zanden. 2014. “The Maddison Project:  Collaborative Research on Historical National Accounts.” The Economic History Review 67 (3): 627–​651. de Castro, A. B. 1971. 7 Ensaios sobre a economia brasileira. Rio de Janeiro: Forense. Ferreira, P. C. 2005. Regional Policy in Brazil: A Review. Manus: Fundação Getúlio Vargas. Furtado, C. 1959a. A Operação Nordeste. Rio de Janeiro: Instituto Superior de Estudos Brasileiros. Furtado, C. 1959b. Formação econômica do Brasil. Rio de Janeiro: Fundo de Cultura. Goodman, D., and R. C. de Albuquerque. 1974. Incentivos à industrialização e desenvolvimento do Nordeste. Rio de Janeiro: IPEA/​INPES. GTDN. 1959. Uma política de desenvolvimento para o Nordeste. Mimeo, produced for the Working Group for the Development of the Northeast (GTDN), Recife. Hirschman, A. 1963. Journeys towards Progress. New York: Twentieth Century. Kon, A. 1976. Crise regional e planejamento. São Paulo: Perspectiva. Lebret, L. 1955. Estudo sobre desenvolvimento e implantação de indústrias, interessando a Pernambuco e ao Nordeste. Recife: Comissão de Desenvolvimento Econômico de Pernambuco.

Brazil’s Northeast   463 Leff, N. 1972. “Economic Development and Regional Inequality: Origins of the Brazilian Case.” Quarterly Journal of Economics 86 (2): 243–​262. Lucas, R. 1988. “On the Mechanics of Economic Development.” Journal of Monetary Economics 22 (1): 3–​42. Maia Gomes, G. 2010. Conflito e conciliação: Políticas de desenvolvimento regional no mundo contemporâneo. Fortaleza: Banco do Nordeste. Moreira, R. 1979. O Nordeste brasileiro. Rio de Janeiro: Paz e Terra. Naritomi, J., R. Soares, and J. Assunção. 2007. “Rent Seeking and the Unveiling of ‘de facto’ Institutions: Development and Colonial Heritage within Brazil.” NBER Working Paper No. 13545. Cambridge, MA. Oliveira, C., and P. Vianna. 2005. Desenvolvimento regional: 50 anos do BNB. Fortaleza: Banco do Nordeste. Oliveira, F. 1981. Elegia para uma re(li)gião, 3rd edition. Rio de Janeiro: Paz e Terra. Oliveira, V. 2017. Autonomia regional e financiamento da educação básica: Pernambuco, São Paulo e Rio Grande do Sul, 1850–​1930. São Paulo: Companhia das Letras. Rosenstein-​Rodan, P. 1943. “Problems of Industrialization of Eastern and South-​Eastern Europe.” Economic Journal 53 (210–​211): 202–​211. Rosenstein-​Rodan, P. 1944. “The International Development of Economically Backward Areas.” International Affairs 20 (2): 157–​165. Simonsen, R. 1957. História econômica do Brasil. São Paulo: Companhia Editora Nacional. Singer, Hans W. 1956. Economic Development of Northeast Brazil. Report No. Taa/​Bra/​2. New York: United Nations.

Pa rt  V

S O C IOE C ON OM IC DI M E N SION S

Chapter 22

Changes in I nc ome Distribu tion i n Bra z i l Rodolfo Hoffmann

22.1. Introduction Latin America is widely recognized as being a region of great income inequality, but even within this regional context Brazil stands out for its exceptionally high inequality. The country’s continental dimensions and significant geographical heterogeneity play their part, but the major causes of this inequality lie in Brazil’s history. This chapter is limited to events since 1964, and in particular changes to income distribution in the period 1995–​2015. The debate about the increase in inequality during the 1960s is examined in section 22.2, in particular the importance of policy changes after the military coup of 1964. Section 22.3 deals with the main changes up to 1994: the increased labor participation of women; the conservative modernization of agriculture that reduced differences be­ tween average incomes in agriculture and other sectors, but increased inequality within the agricultural sector; and the impact of inflation on income distribution. In section 22.4, Pesquisa Nacional por Amostra de Domicílios (PNAD) data are used to chart the decline in inequality from 1995 until 2014. Sections 22.5 and 22.6 analyze the immediate determinants of this decline, looking at income decomposed into its components and evaluating the way in which each component contributes to inequality or the reduction thereof. Sections 22.7 and 22.8 discuss the roles of, respectively, rises in the minimum wage and the degree of schooling. Section 22.9 examines the role of other factors, such as gender and skin color. Section 22.10 discusses the trend of reduced differences between regions in the period 1995–​2014, while section 22.11 shows that there was no increased polarization in the distribution of income; in fact, contrary to the findings of some researchers, polarization diminished alongside the reduction in inequality. Section 22.12 shows how the economic crisis that hit the country beginning in 2014 interrupted the

468   Rodolfo Hoffmann process of declining inequality in per capita family income; from 2014 to 2015, income inequality among economically active persons increased. Section 22.13 concludes the chapter.

22.2.  1960–​1970 The military dictatorship that seized power in 1964 became more authoritarian in 1968 with the Institutional Act No. 5 decree. From 1968 to 1973, Brazil’s economy enjoyed very high rates of growth, in a period often called the “Brazilian economic miracle.” With the publication of the 1970 census results, and their comparison with 1960 census data, it be­ came clear that Brazil had seen a pronounced growth in inequality in income distribu­ tion in the 1960s. Papers by Hoffmann and Duarte (1972) and Fishlow (1972) identified the economic policy of the dictatorial government as one of the causes of increased inequality. Given that normal channels of political expression were closed, academic studies became an important way to express criticism of the government of the time (Lopes 1973). The international repercussions of Fishlow’s article (1972) were such that Brazil became known as an example of the extraordinary growth of inequality within the space of a decade. The first reaction of those who defended the dictatorship was to question the credibility of the census data. Citing the arguments of Hoffmann (1971) and Duarte (1971), Simonsen (1972, 50) states that “the debate about the increase in income concentration between 1960 and 1970 can only be sustained with a good deal of statis­ tical frivolity.” The growth of inequality was later recognized as a fact, following publication of a book by Langoni (1973) that was written with the encouragement and support of Finance Minister Antônio Delfim Netto. However, the increased inequality was seen as something arising from the normal functioning of the market, rather than being in any way related to the nature of the government or the drastic political change represented by the 1964 coup. Langoni had privileged access to microdata from the 1960 and 1970 censuses and produced a very competent econometric analysis, although with obvious political bias, proceeding as if neither a military coup that jailed union leaders across the country nor all the subsequent repressive economic policy had any effect on income dis­ tribution. Polarization within the ongoing debate became associated with the question of whether or not the author was an opponent of the dictatorial government. There is ex­ tensive literature on this debate; see, for instance, Tolipan and Tinelli (1975) and Taylor et al. (1980). Recent work by Souza and Medeiros (2015) offers new data to evaluate the causes of the growth of inequality between 1960 and 1970. Using income tax data, they find that “top income shares have fluctuated sometimes in tandem with major political changes. The military coup of 1964, in particular, was followed by a rapid rise of the top 1% in­ come share, reversing the previous trend” (129).

Changes in Income Distribution in Brazil    469

22.3.  1970–​1980: Democratization without the Immediate Reduction of Inequality There is a historical trend of growing female participation in the labor market, and this can be seen in the period 1970–​1980. Since women typically receive lower pay, rising fe­ male participation contributed to further widening the unequal distribution of income among the economically active population. Nevertheless, the contribution of women’s income to family income in the period 1970–​1980 was such that census data reveal a slight reduction of inequality in the distribution of per capita family income (Hoffmann and Kageyama 1986). Agricultural income grew more strongly in the 1970s, reducing the gap between the primary sector and other sectors of the economy. However, agricultural modernization, conducted in a manner that favored large establishments and mechanization (conserv­ ative modernization), helped to reinforce the high level of inequality within the sector (Hoffmann 1990). President João Figueiredo (1979–​1985) headed the last administration of the military government. In 1984 there were widespread popular demonstrations calling for imme­ diate direct presidential elections—​diretas já—​but the campaign met with only partial success:  Tancredo Neves, a conservative civilian politician, won an indirect election for president in 1985, but fell ill before taking office and died shortly thereafter; the vice president, José Sarney, became the first civilian president since 1964. Direct presidential elections were to take place only in 1989, following the promulgation of the new 1988 Constitution. The process of re-​democratization did not lead, at least in the short term, to a re­ duction in the inequality of income distribution (Henriques 2000). This was to occur only in the new millennium, during and after the government of President Fernando Henrique Cardoso. The Sarney government faced an urgent problem in the shape of very high inflation. The Cruzado Plan, which included a price freeze, was introduced at the end of February 1986. However, the plan failed and its beneficial effects on income distribution were ephemeral. Soon after the November 1986 gubernatorial and congressional elections, inflation once again exceeded 10% per month. Inflation had two effects on measures of inequality of income distribution, when cal­ culated based on household survey data. First, there was a real effect associated with the fact that inflation was more harmful to the relatively less well off, because the wages of the poorest workers tended to be readjusted with more delay. In Brazil’s period of highest inflation, the relatively well-​off were better able to protect themselves against rising prices because they had bank accounts where the current-​account balance was indexed against inflation on a daily basis. However, high inflation also caused statistical noise in

470   Rodolfo Hoffmann reported income values, helping to increase their dispersion and hence the measures of inequality. In September 1989, the reference month for the National Household Sample Survey (Pesquisa Nacional por Amostra de Domicílios—​PNAD), inflation was close to 40% per month. In such a situation the simple confusion between the wage in September and that of the following month, when the interview was conducted, generated huge random variation in the data, causing the Gini index of per capita household income calculated from this data to exceed 0.63. A regression analysis using data for the period 1979–​1997 indicated that each addi­ tional 10 percentage points of inflation in the data reference month increased the Gini index of income distribution for the economically active population, excluding those with zero income, by more than 1 percentage point (Hoffmann 1998).

22.4.  Controlling Inflation and the Reduction of Inequality It was only with the advent of the Real Plan in 1994 that inflation was finally brought under control. There were immediate effects on income distribution, as seen in the sig­ nificant growth in sales of certain items, in particular consumer discretionary products such as yogurt. But most important was the creation of a reasonably stable national cur­ rency, which made possible the subsequent implementation of government programs to reduce inequality and poverty. The PNAD was not conducted in 1994, and it can be argued that the 1992 and 1993 surveys were seriously contaminated by information error. Thus, it is reasonable to doc­ ument the fall in inequality using PNAD surveys conducted in the period 1995–​2014.1 We look first at the distribution of per capita household income. Inequality fell only slightly from 1995 to 2001, with the Gini index dropping from 0.599 to 0.594. The fall then became more intense and systematic, with the Gini index reaching 0.513 in 2014. Other measures confirm the substantial reduction in inequality in that period: Theil’s T passes from 0.727 in 1995 to 0.719 in 2001 and then to 0.532 by 2014. The percentage of total income received by the richest tenth falls from 47.7% in 1995 to 47.2% by 2001 and 40.6% by 2014. Figures 22.1 and 22.2 illustrate the reduction of inequality in the distri­ bution of per capita household income in the period 1995–​2014 (note that the PNAD survey was not conducted in 2000 and 2010). Figure 22.2 shows that, prior to 2002, the percentage of total income received by the richest hundredth in the distribution of per capita household income exceeded the share received by the poorest half. In 2014 the share of the richest hundredth had fallen to about 11%, while the share of the poorest half increased to more than 17%. Similarly, the distribution of income among employed persons with income greater than zero2 shows that the fall in inequality intensifies after 2001, as seen in Figure 22.3. The Gini index drops from 0.581 in 1995 to 0.560 in 2001 and 0.483 in 2014. Theil’s T for

Changes in Income Distribution in Brazil    471 0.75 0.70 0.65 0.60 0.55

Gini index

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

0.50

Theil T index

Figure 22.1.  Gini and Theil’s T indices for the distribution of per capita household income in Brazil, 1995–​2014. 18 17 16

%

15 14 13 12 11 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

10

Poorest 50% share

Richest 1% share

Figure 22.2.  Shares of total household income received by the poorest half and the richest 1% of household members in Brazil, 1995–​2014.

the same three years is 0.687, 0.652 and 0.495 respectively, and the percentage of income received by the highest-​paid tenth is 46.8%, 45.8% and 40.0%. Despite the substantial reduction from 1995 to 2014, inequality remains high in Brazil. Data from the 2014 PNAD for the distribution of per capita household income show that, for the whole of Brazil (including the rural area of the former North region), the Gini index is 0.515, Theil’s T is 0.535, and the bottom half has a share of total in­ come (17.2%) that is almost equal to that of the richest two-​hundredths (17.1%), while

472   Rodolfo Hoffmann 0.70

0.65

0.60

0.55

0.50

0.45

Gini index

Theil T index

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

0.40

Theil L index

Figure 22.3.  Gini index and Theil’s T and L for the inequality of the distribution of earnings among employed persons with earnings greater than zero in Brazil, 1995–​2014.

the richest hundredth receives 11.3%, the richest 5% receive 28.6%, and the richest tenth receive 40.7%. Comparisons with the National Accounts show that income is underreported by 30%–​40% in the PNAD (Hoffmann 1988, 2001). The PNAD does not take into account the value of production for self-​consumption, which can be an important component of income for a family farmer. This income is contemplated in the Household Budget Survey (Pesquisa de Orçamentos Familiares—​POF), but without significantly altering the overall picture of income distribution in Brazil. Taking into account comparisons of total income reported in the PNAD and POF with the results of National Accounts and income tax returns, the conclusion is that both the PNAD and the POF include greater underreporting of higher incomes, thus causing the degree of inequality to be underestimated. Combining income tax data with the results of the PNAD, Medeiros et al. (2015a) arrive at a much higher estimate of inequality, with almost half of all income concen­ trated in the richest 5%. In addition, they conclude that inequality remained practically stable from 2006 to 2012, rather than decreasing as suggested by the PNAD data. The authors recognize the risks inherent in combining the two databases. On the one hand, there is no doubt that the PNAD data underestimates inequality because incomes typ­ ical of the relatively rich, such as interest, dividends, and profits, are subject to greater under-​declaration. On the other hand, the variation in inequality revealed by the PNAD

Changes in Income Distribution in Brazil    473 can be seen as important and worthy of consideration, even if it is necessary to enter a caveat that it is based on “income detected by the PNAD.”

22.5.  Evaluation of the Determinants of Changes in Inequality One way to analyze what the determinants of inequality are is to see how different components of income contribute to the overall Gini index (G). If incomes xi are ranked in ascending order (x1 ≤ x2 ≤ …≤ xn), the Gini index can be defined as a function of the covariance between incomes and their position in the order: G=



2 cov (i, xi ) , (1) nµ

where µ is the average income of the n people. If income xi is decomposed into k parts (xhi, with h = 1, 2, . . . , k) and µ h is the average of part h, then its concentration ratio (Ch) is defined in terms of the covariance between the value of the part for each person and its position in the ranking based on their total income:

Ch =

1 cov ( xhi , i ) (2) 2µ h

As shown by Rao (1969), the Gini index is an average of the concentration ratios weighted by the participation (ϕh ) of each part in total income: k

G = ∑ϕhCh



h =1

(3)

From this expression we can obtain the decomposition of the change in the Gini index between year 1 and year 2:∆G = G2 − G1. Using the macron over a symbol to indicate the average of the values in the two years, we have (Hoffmann 2006, 2013a; Soares 2006): k



∆G = ∑ (Ch − G ) ∆ϕh + ϕh ∆Ch  h =1

(4)

Note that each of the k parts of the change in the Gini index comprises a composition effect, associated with the change in the part’s share in total income (∆ϕh), and a concen­ tration effect that is associated with the change in its respective concentration ratio (∆Ch). This dynamic decomposition of the Gini index makes it possible to calculate the per­ centage contribution of each part of income to the change in the Gini index in a specific period.3

474   Rodolfo Hoffmann Another way to decompose the changes in inequality is to conduct counterfactual simulations. Artificial distributions are constructed that substitute, in the initial distri­ bution, one or more parts of the final distribution. Comparing the measure of initial inequality with that of the simulated distribution, changes can be attributed to specific parts or factors (Barros et al. 2006). This technique was used by Barros et al. in a large number of excellent studies on the recent changes in income distribution in Brazil.

22.6.  The Importance for the Reduction of Inequality of Income Transfers from the Federal Government and other Income Components The 1988 Constitution established the right to the Continued Provision Benefit (Benefício da Prestação Continuada—​BPC), whereby every person unable to work or classified as elderly (over 65 years of age) living in a very poor household should receive a minimum wage from the government. The number of people receiving this benefit grew as potential beneficiaries learned about their rights, and duly applied. A federal conditional cash transfer program was created in 2001. Called the School Stipend (Bolsa Escola), this provided a monthly payment to a parent, preferably the mother, of a child in a poor family who maintained a satisfactory level of school attend­ ance. In 2003 the School Stipend was incorporated into the Family Stipend Program (the well-​known Bolsa Família—​PBF), which grew to benefit some 14 million households in 2015. The concession and value of the benefit depend on the family’s poverty level and the number of pregnant women, children, and adolescents, but it is generally less than one-​third of the minimum wage, which as of mid-​2016 was equivalent to US$274 per month at the official exchange rate. PBF is more tightly focused on very poor families than the BPC. Using data from the Household Budget Survey (Pesquisa de Orçamentos Familiares—​ POF) for 2008–​2009, it can be seen that income from Bolsa Família is strongly progres­ sive as a share of per capita family income, with a concentration ratio of −0.562. Income from the BPC has a concentration ratio of −0.104 (Hoffmann 2010) and thus makes a lesser contribution to reducing inequality. Figure 22.4 shows the Lorenz curve for the distribution of per capita family income and the concentration curves for various parts of this income, according to 2008–​2009 POF data. The horizontal axis measures the cumulative proportion of people ranked according to their per capita family income, while the vertical axis plots the corre­ sponding cumulative proportion of total income (generating the Lorenz curve) or parts of income (generating the concentration curves). The earnings concentration curve is almost superimposed on the Lorenz curve.

Changes in Income Distribution in Brazil    475 1 0.9

Accumulated income proportion

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Accumulated population proportion Line of perfect equality

Earnings

Lorenz curve

Federal transfers

PBF

BPC

INSS

Public servants pensions

Figure 22.4.  The Lorenz curve and concentration curves for some components of per capita family income in Brazil, 2008–​2009.

In the PNAD, unfortunately, Bolsa Família transfers are aggregated with other in­ come sources such as interest and dividends. This makes it necessary to use statistical techniques to disaggregate the reported value. Applying the decomposition of expres­ sion (4), we can estimate that the combined income of the PBF and BPC accounted for around 18% of the fall in the Gini index from 0.599 in 1995 to 0.527 in 2011. During this period the share of this part of the per capita household income in the total rose from 0.1% to 1.3%. This is a small part of total income, but one that has a substantial impact in reducing inequality, given its focus on the very poor. The earnings of private-​sector employees,4 whose share of total income rose from 40% in 1995 to 43% in 2011, accounted for about 44% of the reduction in the Gini index in that period (Hoffmann, 2013b). The Gini coefficient for the distribution of per capita household income decreased from 0.599 in 1995 to 0.513 in 2014. Using the same methodology for this period, it can be seen that the part that estimates the income from PBF and BPC accounts, once again, for 18% of the reduction of the Gini index, even though it represents less than 1.6% of total

476   Rodolfo Hoffmann declared income in 2014. The earnings of private-​sector employees, who enjoyed a 41.9% share of total income in 2014, once again accounted for 44% of the reduction in the Gini index. Earnings of military personnel and civil servants directly employed by the public sector increased as a share of total declared income from 10.1% to 10.7% in the period 1995–​2014. These earnings constitute a regressive part of the per capita household in­ come (with a concentration ratio above 0.73 in both years), and helped to reinforce in­ equality. Separating military and civil service incomes, we can see that the share of the latter in total income increased from 9.5% to 10.2% and was the only component that contributed to increasing inequality, rather than reducing it. It is important to remember that the relative contribution of any given part of the per capita household income to the overall change in inequality will depend on the chosen measure of inequality. While the Mehran and Piesch coefficients are both based on the Lorenz curve, as is the Gini index, the Mehran index gives more weight to changes in the left tail of the distribution and the Piesch index gives more weight to changes in the right tail. When per capita household income is divided into parts, these two indices can be decomposed in the same way as the Gini index (Hoffmann 2004, 2007b). The contribu­ tion of estimated PBF and BPC income to reducing inequality in the period 1995–​2014 is 18% of the variation of the Gini index, 24% of the variation of the Mehran index, and 15% of the reduction in the Piesch index. It was only from 2004 that the PNAD started to cover the rural areas of the North. This means that we can analyze the entire national territory on a regional basis for the period 2004–​2014. Brazil is usually considered to be divided into five regions: North, Northeast, Southeast, South, and Center-​West. In the following analysis, however, it was deemed appropriate to separate out the state of São Paulo from the rest of the Southeast region, given the state’s demographic and economic importance. The country was thus divided into six regions, as shown in Table 22.1. Looking at the national territory as a whole, the Gini index for the distribution of per capita household income fell from 0.570 in 2004 to 0.515 in 2014. Using expression (4), the main contributions to the reduction of 0.055 in this index were as follows: (a) 38.7% due to changes in remuneration of private-​sector employees. This, in turn, resulted from the reduction of educational inequality, a reduction in the return on increased schooling, and growth in the real value of the minimum wage; (b) 26.9% due to changes in pensions paid by the government. This is a regressive part of total income where the concentration ratio decreased mainly because of the increase in the minimum wage, which is also the minimum for such payments; (c) 14.3% due to expansion of Bolsa Família. The contribution of Bolsa Família to the reduction of the Gini index is clearly greater in the two poorest regions:  44.9% in the North and 25.8% in the Northeast. This same contribution is 11.0% in the Southeast region minus São Paulo (Minas Gerais + Espírito Santo + Rio de Janeiro), 8.0% in São Paulo, 4.0% in the South, and 3.3% in the Center-​West.

Changes in Income Distribution in Brazil    477 Table 22.1 Participation of six regions in population and reported income, according to the PNAD, 2014 Region North

% of Population

% of Income

Relative Income

8.5

5.8

0.68

Northeast

28.0

17.7

0.63

MG + ES + RJ

20.4

22.2

1.09

SP

20.9

27.2

1.30

South

14.5

17.7

1.22

7.7

9.3

1.22

Center-​West

Note: MG + ES + RJ = Minas Gerais + Espírito Santo + Rio de Janeiro.

22.7.  Real Growth of the Minimum Wage One of the factors determining how the distribution of wages affects overall inequality is the level of the minimum wage. The rise in inequality that followed the 1964 mili­ tary coup is associated with a major reduction in the real value of the minimum wage from 1964 to 1974. This is a variable that reflects the repressive economic policy of the period. Applying regression analysis to data for 1967–​1971, it was possible to show the intensifying effect of the level of the real minimum wage on inequality in the distribu­ tion of industrial wages (Hoffmann 1973). Studies indicate that in the nine years from September/​October 1996 to September/​ October 2005, the real value of the minimum wage increased by 43%, and in the following nine years, from 2005 to 2014, it increased by a further 50%. Estimated density functions of income distribution of employed people show, visually, how the value and growth of the minimum wage affect the shape of the distribution (Bacha and Hoffmann 2015; Hoffmann 2008; Lemos 2009; Soares 2004).5 It is important to bear in mind that Brazil’s minimum wage (salário mínimo—​SM) is not just the lower limit for wages in the formal labor market, it is also the unitary value of the BPC transfer and the floor value for pensions paid by the government. In the 2014 PNAD, 9.9 million people reported receiving exactly 1 minimum wage (then R$724) as compensation for their main job, 17.7 million reported receiving a government pension of exactly 1 minimum wage, and 2.5 million reported other income sources, usually gov­ ernment transfers such as the BPC, equal to 1 minimum wage.6 Even if we consider just the effect on the labor market, it is clear that there are limits to reducing inequality via increases in the real value of the minimum wage. Exaggerated increases can contribute to increasing unemployment. But under the current legislation, which guarantees an annual increase to fully compensate for consumer price inflation of

478   Rodolfo Hoffmann the previous year, plus a real increase in line with GDP growth of the penultimate year, minimum wage increases have had a major impact on the government deficit. One im­ portant step would be to break the mandatory link between the minimum wage and the BPC and the minimum state pension, to ensure that these benefits maintain their real value. Only after such a step had been taken would it be appropriate to examine the pos­ sibility of eventual real increases in their value.

22.8.  The Importance of Schooling Education enjoys unanimous recognition as one of the most important determinants of a person’s income. Given that education is also valuable in and of itself, this means that every researcher can finish his or her analysis recommending more schooling and improvements to the quality of education, especially if that researcher is actually a university professor. By adjusting earnings equations to household survey data in Brazil, it is common to obtain a coefficient of determination around 50%.7 This can be interpreted by saying that around 50% of the variations in the logarithm of individual incomes of employed persons can be explained by the explanatory variables considered (these essentially in­ clude age, gender, schooling, color, region, and sector), with particular emphasis on the contribution of schooling. In the case of persons employed in agriculture, inclusion of the area owned by the producer as an explanatory variable in the earnings equation substantially reduces the marginal contribution of education (Fishlow 1973, Ney and Hoffmann 2003a, 2003b). The size of an agricultural property is a variable collected in the PNAD, and it can be taken as a proxy for the physical capital owned by the producer. Unfortunately there is no equivalent variable for other sectors. The rate of return to each additional year above 10 years of schooling has fallen since 2002, thus helping to reduce inequality. At the same time, dispersion of schooling decreased among employed persons in non-​ agricultural sectors, and this also contributes to reducing income inequality associated with schooling (Hoffmann and Oliveira 2014). Barros et al. (2010) use counterfactual simulations to estimate that be­ tween 2001 and 2007 the price effect (the relation between labor earnings and schooling) and the quantity effect (the distribution of schooling) contributed, respectively, 35% and 11% to the reduction in the Gini index of the earnings distribution.

22.9.  Earnings Differences due to Color, Gender, and Other Factors In the PNAD, people are classified into four categories of skin color plus one ethnic cat­ egory. The classification is self-​assigned by the person responding to the survey, rather

Changes in Income Distribution in Brazil    479 than being assigned by the interviewer. These five groups, and their participation in the population of employed persons in 2014, using literal translation of the PNAD Portuguese-​language classification, were as follows: white (46.5%); black (9.7%); brown (43.0%); yellow (0.5%); and Indian (indigenous) (0.3%). Relative income (income compared to the overall average) was 0.74 for black; 0.76 for brown; 0.77 for indigenous; 1.27 for white; and 1.97 for yellow. The participation of the inequality between the five groups in the overall inequality of the distribution of labor earnings is 6.9% for Theil’s T and 8.1% for Theil’s L. The income gap between genders (relative income of 0.83 for women and 1.12 for men) represents 2.0% of total inequality for Theil’s T and 2.4% for Theil’s L. Considering the division of the country into 27 components (26 states and the Federal District), the participa­ tion is 6.3% for Theil’s T and 7.7% for Theil’s L. Dividing the working population into six categories of occupation (military personnel and civil servants directly employed by the public sector; domestic workers; formally employed private-​sector workers; informally employed private-​sector workers; self-​employed people; and employers), the partici­ pation of inequality among these categories in the total overall inequality is 19.9% and 20.5% for Theil’s T and Theil’s L, respectively. Adjusting an earnings equation, it is possible to estimate the effect of the various categories by controlling the effect of the other variables included in the model. The effect of being a woman has declined slowly over time, but still represents −28% of ex­ pected earnings in an equation estimated with 2014 data, in which the dependent vari­ able is the logarithm of monthly income of all work and the explanatory variables are the following: gender; age and its square; two variables for education to capture the different rates of return up to 10 years of schooling and above that level;8 the logarithm of the number of working hours per week; binary variables for color, sector of activity, region, rural or urban residence; and five employment categories (domestic workers; military personnel and civil servants directly employed by the public sector; other employees; self-​employed people; and employers). Persons self-​describing as black and brown tend to earn 9% less than those self-​ describing as white, but the expected income of those self-​describing as yellow is 7% higher than that of whites, after controlling the effect of all other variables included in the model. The effects of skin color when controlling other factors are substantially lower than the gross differences, mainly because the average level of schooling is lower for black and brown groups than for white, while for yellow it is greater. In the case of gender, the opposite occurs. The controlled effect is greater than the gross effect, be­ cause among employed people women have longer average schooling than men. The estimated equation also shows that the earnings of domestic workers tend to be 17% lower than that of other employees in the private sector (adopted as a basis for compar­ ison). On the other hand, the earnings of military personnel and civil servants directly employed by the public sector tend to be 35% higher, always after controlling for the effect of other factors. Jesus (2015) uses the Oaxaca-​Blinder method to decompose the difference between the income of the set of those self-​describing as black and brown and the income of

480   Rodolfo Hoffmann the group self-​describing as white in that part associated with differences in personal characteristics (the averages effect) and another part related to differences in the way these characteristics are reflected in income (the parameters effect). His study shows that from 1995 to 2014 the averages effect was always greater, and was falling, while the parameters effect remained stable. The part of income relating to schooling within the averages effect was the most im­ portant. Another important factor contributing to reducing the difference was the geo­ graphical redistribution of the population groups self-​describing as black and brown. It is therefore just as well that there are no barriers to migration within the country.

22.10.  Regional Differences In this section, in addition to separating out São Paulo (SP) from the rest of the Southeast region because of the state’s demographic and economic importance, the Federal District has been separated out from the Center-​West, given the special characteristics of the nation’s capital. We therefore analyze per capita household income distribution within and between seven regions using PNAD data from 1995–​2014. A first conclusion is that inequality between the regions has fallen. Even with the decline in overall ine­ quality, the participation of inequality between the seven regions in the overall Theil’s T falls from more than 9.5% in the years 1995–​1997 to about 8% or less in 2008–​2014. The fall in the relative position of São Paulo’s per capita household income, from 1.5 in 1995–​ 1997 to 1.3 or less in 2008–​2014, has been an important part of reducing interregional inequality. The highest average per capita household income is always found in the Federal District, followed by São Paulo and the South. The lowest is the Northeast, followed by the North. The Federal District stands out for the high participation of civil service remuneration in the total declared income (greater than 25% in most years of the series). Participation of the estimated income from Bolsa Família and the BPC in total income is, as would be expected, highest in the two poorest regions, the Northeast and North, and exceeds 4% in the Northeast in the years 2012–​2014.9 The unequal distribution of per capita household income declined in six of the seven regions, the exception being the Federal District, which had a Gini index of 0.58 in 2014, compared with a value below 0.52 in all other six regions. The Gini index of inequality within the Northeast region is systematically higher than the Gini index for Brazil as a whole. The lowest Gini indices are generally seen in the South and in São Paulo. Interregional inequality also falls when analyzing income distribution among employed persons. However, when considering only those people employed in the agricultural

Changes in Income Distribution in Brazil    481 sector, it would appear, on the contrary, that interregional inequality and also inequality among the 27 federal components (26 states and the Federal District) shows an upward trend in the period 1992–​2014 (Hoffmann and Jesus 2016).

22.11. Polarization If the population is divided into rich and poor, total inequality increases when the gap between the two strata increases, and/​or when inequality increases within the two strata. For polarization between rich and poor to be conceptually distinct from total inequality, the measure used to calculate polarization should grow with the gap between rich and poor, but should also increase when there is a reduction of the differences within the strata. This is the basic idea that guided the early work proposing synthetic measures of polarization. But while Wolfson (1994) was limited to bipolarization, Esteban and Ray (1994) and Duclos et al. (2004) proposed a measure of multi-​polarization. It is still common for “polarization” to be used rhetorically as a synonym for “ine­ quality” or “high inequality.” An initial and essentially correct analysis of the polari­ zation of income distribution in Brazil was provided by Castro and Scorzafave (2007). However, Figueirêdo et  al. (2007) calculated a measure of multi-​polarization and interpreted it as a measure of bipolarization between rich and poor, claiming, wrongly, that this polarization was growing. In fact, bipolarization fell in Brazil in recent years, accompanying the fall in inequality. The value calculated by Figueirêdo et al. (2007) identified the formation of various peaks in the distribution, located at round values and/​or values associated with the minimum wage. This is a purely statistical problem, but the authors confused it with an entirely different phenomenon, which is the polari­ zation between rich and poor (Hoffmann 2008). Figure 22.5 shows two estimates of the probability density function for the distribution of income from the main job of em­ ployed persons in Brazil in 2014, drawing on PNAD data. Using a wider window, the estimated curve is smooth, but the use of a narrower window highlights the spikes as­ sociated with the minimum wage (MW) and round numbers indicated by vertical lines (with values in tens of reais). The highest peak is associated with the minimum wage (R$724 in 2014). More recently, Clementi and Schettino (2013) provide another example of finding a false increase of income polarization in Brazil. They used a measurement of “polariza­ tion” proposed by Handcock and Morris (1999) that is indistinguishable from a measure of inequality between rich and poor. Worse still, Clement and Schettino did not realize that when transposing the income distribution to the right (the sum of a constant to the incomes), a reduction of inequality occurred. The transposition required for calculating Handcock and Morris’s measurement of inequality must necessarily be made in the dis­ tribution of the income logarithms.

482   Rodolfo Hoffmann 40

1

2

3

35

4

MW

10

15

20

30

30

50

25

20

15

10

5

0 3.8

4.2

4.6

5.0

5.4

5.8

6.2

6.6

7.0

7.4

7.8

8.2

8.6

9.0

9.4

9.8

Y Curves:

Kernel (c = 1)

Kernel (c = 2.5)

Figure 22.5.  Kernel density curves for the distribution of income from the main job of occu­ pied persons in Brazil in 2014, with the highest peak at the minimum wage (MW).

22.12.  Crisis Interrupts Falling Inequality The Brazilian economy entered into crisis in 2014, principally as a result of the federal government’s severe fiscal problems. Quarterly data from the “continuous PNAD” show that the unemployment rate, which was below 8% in 2012–​2014, had risen to 9% by the end of 2015 and reached 11% in the first quarter of 2016. Given that the unemployed are members of the economically active population (EAP) with zero labor income, the Gini index of inequality of the distribution of labor income among the EAP inverted its downward trend in the fourth quarter of 2014. Taking the PNAD data for 2015, we obtain a number for the Gini index of the per capita family income that is practically the same as that for 2014 (0.513, as we saw in section 22.4). Thus, the systematic decline in inequality seen since 2001 has halted.

Changes in Income Distribution in Brazil    483 If data of the income distribution of occupied persons with positive earnings are taken, there is a decline in inequality from 2014 to 2015. However, since rising unem­ ployment is an essential feature of the crisis, it is more adequate to analyze the distri­ bution of earnings among all economically active persons, including the unemployed, whose labor earnings are zero. This distribution also shows a clear tendency of declining inequality from 1998 to 2014, with the Gini index reducing from 0.620 to 0.529, while in 2015 it rises to 0.539 (Hoffmann 2016). Considering the components of per capita family income, it is possible to analyze the contribution of each component to the value of the Gini index in each year (according to expression (3), presented in section 22.5). From 2014 to 2015, the overall average income declined considerably. However, the incomes of civil servants and the value of pensions and other retirement benefits paid by the government were less affected by the crisis. Thus, the contributions of these two components to the value of the Gini index became higher (Hoffmann 2017).

22.13. Conclusion In light of the findings presented in the previous sections, we may now draw some gen­ eral conclusions and highlight several opportunities and difficulties to be faced in re­ ducing the inequality of income distribution in Brazil. Political change and corresponding public policies have a fundamental impact on income distribution, as illustrated by the effect of the 1964 military coup on increased inequality in the 1960–​1970 decade. It should be remembered that the impact of dicta­ torship on the distribution of income was almost immediate, while, conversely, the po­ litical opening in the 1980s yielded only gradual results. The shrinking of the real value of the minimum wage between 1964 and 1974 reflects repression under dictatorship, while the fact that the real value of the minimum wage almost doubled from 1996 to 2014 arose from a very different political environment favoring a reduction in inequality. It should be stressed that the control of inflation after the Real Plan of 1994 was es­ sential both in avoiding the regressive impact of the inflationary tax and in allowing the establishment and gradual expansion of income transfer policies (such as Bolsa Família) that played a significant role in reducing inequality and poverty. The economic crisis that surfaced in Brazil in 2014 halted the reduction in income inequality, and it is neces­ sary to recognize that income transfer programs have clear limitations as instruments to promote further reductions in inequality. Nobody disagrees about the need to improve education and the qualification of the labor force. However, equally, specific public policies are needed. From a technocratic standpoint, it is not difficult to say what should be done to fur­ ther reduce the inequality of income distribution in Brazil. The proposal would include a pension reform eliminating all privileges of civil servants (as proposed by Nicholson 2007), reform of the tax system to make it much more progressive, rationalization of

484   Rodolfo Hoffmann compensation norms for public employees, and regulation of public-​sector strike leg­ islation. It is much more difficult to build the political will to implement such changes. The scope and progressivity of income tax should be increased, although we know that to this end it would be necessary to have better information and control over the global movement of capital, as discussed, for instance, by Piketty (2014).

Notes 1. Before 2004 the PNAD did not include rural areas of the old North region, and so data for these areas are excluded from 2004–​2014 PNADs when making comparisons for the period 1995–​2014. 2. Deleting observations with missing data on variables used to fit earnings equations. 3. This technique can be extended to other measures of inequality (Hoffmann 2007). 4. Private-​ sector workers, including formally and informally employed but excluding self-​employed. 5. It should be noted that these studies considered the minimum wage in September and October, because the PNAD reference month in the analyzed period was September, and the monthly salary for September is usually received in early October. The studies also used the National Index of Consumer Prices (Índice Nacional de Preços ao Consumidor—​INPC) as deflator. 6. The number of people reporting income in the range R$700–​R$750 was: 12.5 million as in­ come from their main job; 18.9 million for government pensions; and 2.5 million for other sources. 7. See, for instance, Langoni (1973), Ney and Hoffmann (2003b) and Hoffmann and Simão (2005). 8. See Hoffmann and Simão (2005). 9. Silveira Neto and Azzoni (2013) analyze the impact of social programs on regional inequality.

References Bacha, Edmar L., and Rodolfo Hoffmann. 2015. “Uma interpretação estatística do PIB, da PNAD e do salário mínimo.” Revista de Economia Política 35 (1): 64–​74. Bacha, Edmar L., and Lance Taylor. 1980. “Brazilian Income Distribution in the 1960s: ‘Facts’, Model Results and the Controversy.” In Models of Growth and Distribution for Brazil, ed­ ited by Lance Taylor, Edmar L. Bacha, Eliana A. Cardoso, and Frank J. Lysy, 296–​342. New York: Oxford University Press for the World Bank. Barros, Ricardo P. de, Mirela Carvalho, and Samuel Franco. 2007. “O papel das transferências públicas na queda recente da desigualdade de renda brasileira.” In Desigualdade de renda no Brasil: Uma análise da queda recente, Vol. 2, edited by Ricardo P. de Barros, Miguel N. Foguel and Gabriel Ulyssea, 41–​86. Brasília: IPEA. Barros, Ricardo P. de, Mirela Carvalho, Samuel Franco, and Rosane Mendonça. 2006. “Uma análise das principais causas da queda recente na desigualdade de renda brasileira.” Econômica 8 (1): 117–​147.

Changes in Income Distribution in Brazil    485 Barros, Ricardo P. de, Mirela Carvalho, Rosane Mendonça, and Samuel Franco. 2010. “Markets, the State and the Dynamics of Inequality: The Case of Brazil.” UNDP Discussion Paper. Barros, Ricardo P. de, and Miguel N. Foguel. 2000. “Focalização dos gastos públicos sociais e erradicação da pobreza no Brasil.” In Desigualdade e pobreza no Brasil, edited by Ricardo Henriques. Rio de Janeiro: IPEA. Pp. 719–​739 Barros, Ricardo P. de, Miguel N. Foguel, and Gabriel Ulyssea, eds. 2006–​2007. Desigualdade de renda no Brasil: Uma análise da queda recente, Vols. 1, 2. Brasília: IPEA. Barros, Ricardo P.  de, Ricardo Henriques, and Rosane Mendonça. 2000a. “A estabilidade inaceitável: Desigualdade e pobreza no Brasil.” In Desigualdade e pobreza no Brasil, edited by Ricardo Henriques, 21–​48. Rio de Janeiro: IPEA. Barros, Ricardo P.  de, Ricardo Henriques, and Rosane Mendonça. 2000b. “Education and Equitable Economic Development.” Economia 1 (1): 111–​144. Barros, Ricardo P. de, and Rosane Mendonça. 1993. “Geração e reprodução da desigualdade de renda no Brasil.” In Perspectivas da economia brasileira, Vol. 2, 471–​490. Rio de Janeiro: IPEA. Bonelli, Regis, and Lauro Ramos. 1993. “Distribuição de renda no Brasil:  Avaliação das tendências de longo prazo e mudanças na desigualdade desde meados dos anos 70.” Revista de Economia Política 13 (2, 50): 76–​97. Campello, Tereza, and Marcelo C. Neri. 2013. Programa Bolsa Família: Uma década de inclusão e cidadania. Brasília: IPEA. Castro, Sueli Aparecida Correa, and Luiz Scorzafave. 2007. “Ricos? Pobres? Uma análise da polarização da renda para o caso brasileiro.” Pesquisa e Planejamento Econômico 37 (2): 283–​297. Clementi, Fabio, and Francesco Schettino. 2013. “Income Polarization in Brazil, 2001–​2011: A Distributional Analysis Using PNAD Data.” Economics Bulletin 33 (3): 1–​16. Duarte, João Carlos. 1971. “Aspectos da distribuição da renda no Brasil em 1970.” Master’s dis­ sertation, ESALQ-​USP, Piracicaba. Duclos, Jean‐Yves, Joan Esteban, and Debraj Ray. 2004. “Polarization: Concepts, Measurement, Estimation.” Econometrica 72 (6): 1737–​1772. Esteban, Joan-​ Maria, and Debraj Ray. 1994. “On the Measurement of Polarization.” Econometrica 62 (4): 819–​851. Figueirêdo, Erik Alencar de, José Luis da Silva Netto Jr., Sabrino da Silva Porto Jr.. 2007. “Distribuição, mobilidade e polarização de renda no Brasil: 1987 a 2003.” Revista Brasileira de Economia 61 (1): 7–​32. Fishlow, Albert. 1972. “Brazilian Size Distribution of Income.” American Economic Review 62 (2): 391–​402. Fishlow, Albert. 1973. “Distribuição da renda no Brasil: Um novo exame.” Dados 11: 10–​80. Instituto Universitário de Pesquisas do Rio de Janeiro. Morris, Martina, and Mark S. Handcock. 1999. Relative Distribution Methods in the Social Sciences. New York: Springer. Henriques, Ricardo, ed. 2000. Desigualdade e pobreza no Brasil. Rio de Janeiro: IPEA. Hoffmann, Rodolfo. 1971. Contribuição à análise da distribuição da renda e da posse da terra no Brasil. Tese de Livre-​Docência, Piracicaba: ESALQ-​USP. Hoffmann, Rodolfo. 1973. “Considerações sobre a evolução recente da distribuição da renda no Brasil.” Revista de Administração de Empresas 13 (4): 7–​17. Hoffmann, Rodolfo. 1988. “A subdeclaração dos rendimentos.” São Paulo em Perspectiva 2 (1): 50–​54.

486   Rodolfo Hoffmann Hoffmann, Rodolfo. 1990. “Distribuição da renda e pobreza na agricultura brasileira.” In Agricultura e políticas públicas, edited by Walter Belik, Guilherme Costa Delgado, José Garcia Gasques, and Carlos Monteiro Villa Verde, 3–​111. Brasília: IPEA. Hoffmann, Rodolfo. 1995. “Desigualdade e pobreza no Brasil no período 1979–​90.” Revista Brasileira de Economia 49 (2): 277–​294. Hoffmann, Rodolfo. 1998. “Desigualdade e pobreza no Brasil no período 1979–​97 e a influência da inflação e do salário mínimo.” Economia e Sociedade 11: 199–​221. Hoffmann, Rodolfo. 2001. “Income Distribution in Brazil and the Regional and Sectoral Contrasts.” In Structure and Structural Change in the Brazilian Economy, edited by Joaquim J. M. Guilhoto and Geoffrey J. D. Hewings, 85–​106. Aldershot, UK: Ashgate. Hoffmann, Rodolfo. 2004. “Decomposition of Mehran and Piesch Inequality Measures by Factor Components and Their Application to the Distribution of Per Capita Household Income in Brazil.” Brazilian Review of Econometrics 24 (1): 149–​171. Hoffmann, Rodolfo. 2006. “Transferências de renda e a redução da desigualdade no Brasil e cinco regiões entre 1997 e 2004.” Econômica 8 (1): 55–​81. Hoffmann, Rodolfo. 2007a. “Distribuição da renda e da posse da terra no Brasil.” In Dimensões do agronegócio brasileiro: Políticas, instituições e perspectivas, edited by Pedro Ramos and Antônio Márcio Buainain. Brasília: MDA, NEAD Estudos no. 15. Hoffmann, Rodolfo. 2007b. “Transferências de renda e redução da desigualdade no Brasil e em cinco regiões, entre 1997 e 2005.” In Desigualdade de renda no Brasil: Uma análise da queda recente, Vol. 2, edited by Ricardo P. de Barros, Miguel N. Foguel, and Gabriel Ulyssea, 17–​40. Brasília: IPEA. Hoffmann, Rodolfo. 2008. “Polarização da distribuição da renda no Brasil.” Econômica 10 (2): 169–​186. Hoffmann, Rodolfo. 2009. “Desigualdade da distribuição da renda no Brasil: a contribuição de aposentadorias e pensões e de outras parcelas do rendimento domiciliar per capita.” Economia e Sociedade 18 (1–​35): 213–​231. Hoffmann, Rodolfo. 2010. “Desigualdade da renda e das despesas per capita no Brasil, em 2002–​2003 e 2008–​2009, e avaliação do grau de progressividade ou regressividade de parcelas da renda familiar.” Economia e Sociedade 19 (3–​35): 647–​661. Hoffmann, Rodolfo. 2011. “Distribuição da renda agrícola e sua contribuição para a desigualdade de renda no Brasil.” Política Agrícola 20 (2): 5–​22. See also “Author’s Correction” in Política Agrícola 21 (3) (2012): 131. Hoffmann, Rodolfo. 2011. “The Evolution of Income Distribution in Brazil.” In The Economies of Argentina and Brazil:  A Comparative Perspective, edited by Werner Baer, and David Fleischer, 187–​218. Cheltenham, UK: Edward Elgar. Hoffmann, Rodolfo. 2013a. “How to Measure the Progressivity of an Income Component.” Applied Economics Letters 20 (4): 328–​331. Hoffmann, Rodolfo. 2013b. “Transferências de renda e desigualdade no Brasil (1995–​2011).” In Programa Bolsa Família: Uma década de inclusão e cidadania, edited by Tereza Campello and Marcelo C. Neri, Chapter 12, 207–​216. Brasília: IPEA. Hoffmann, Rodolfo. 2016. “A desigualdade relevante não caiu de 2014 a 2015.” Texto para Discussão No. 37 do IEPE/​Casa das Garças. Hoffmann, Rodolfo. 2017. “Desigualdade da distribuição da renda no Brasil: O que mudou em 2015?” Texto para Discussão No. 38 do IEPE/​Casa das Garças. Hoffmann, Rodolfo, and João Carlos Duarte. 1972. “A distribuição da renda no Brasil.” Revista de Administração de Empresas 12 (2): 46–​66.

Changes in Income Distribution in Brazil    487 Hoffmann, Rodolfo, and Josimar G. Jesus. 2016. “Distribuição do rendimento das pessoas ocupadas no Brasil de 1992 a 2014, destacando as atividades agrícolas.” SOBER, 54º Congresso da Sociedade Brasileira de Economia, Administração e Sociologia Rural. Hoffmann, Rodolfo, and Angela Kageyama. 1986. “Distribuição da renda no Brasil, entre famílias e entre pessoas, em 1970 e 1980.” Estudos Econômicos 16 (1): 25–​51. Hoffmann, Rodolfo, and Marlon G. Ney. 2004. “Desigualdade, escolaridade e rendimentos na agricultura, indústria e serviços, de 1992 a 2002.” Economia e Sociedade 13 (2–​23): 51–​79. Hoffmann, Rodolfo, and Régis B. de Oliveira. 2014. “The Evolution of Income Distribution in Brazil in the Agricultural and the Non-​Agricultural Sectors.” World Journal of Agricultural Research 2 (5): 192–​204. Hoffmann, Rodolfo, and Rosycler C.  S. Simão. 2005. “Determinantes do rendimento das pessoas ocupadas em Minas Gerais em 2000:  O limiar no efeito da escolaridade e as diferenças entre mesorregiões.” Nova Economia 15 (2): 35–​62. Jesus, Josimar G. 2015. “Diferenças de rendimento entre negros e brancos no Brasil: evolução e determinantes.” Masters Thesis, University of São Paulo. Langoni, Carlos Geraldo. 1973. Distribuição da renda e desenvolvimento econômico do Brasil. Rio de Janeiro: Editora Expressão e Cultura. Lemos, Sara. 2009. “Minimum Wage Effects in a Developing Country.” Labor Economics 16: 224–​237. Lopes, José Sérgio Leite. 1973. Sobre o debate da distribuição da renda: Leitura crítica de um artigo de Fishlow, Vol 13, no. 3. Sao Paulo: Revista de Administração de Empresas. Medeiros, Marcelo, Pedro H.  G. F. Souza, and Fabio Avila de Castro. 2015a. “O topo da distribuição de renda no Brasil: Primeiras estimativas com dados tributários e comparação com pesquisas domiciliares, 2006–​2012.” Dados—​Revistas de Ciências Sociais 58 (1): 7–​36. Rio de Janeiro. Medeiros, Marcelo, Pedro H. G. F. Souza, and Fabio Avila de Castro. 2015b. “A estabilidade da desigualdade de renda no Brasil, 2006 a 2012: Estimativa com dados do imposto de renda e pesquisas domiciliares.” Ciência & Saúde Coletiva 20 (4): 971–​986. Ney, Marlon Gomes, and Rodolfo Hoffmann. 2003a. “Desigualdade de renda na agricultura: O efeito da posse da terra.” Economia 4 (1): 113–​152. Ney, Marlon Gomes, and Rodolfo Hoffmann. 2003b. “Origem familiar e desigualdade de renda na agricultura.” Pesquisa e Planejamento Econômico 33 (3): 541–​572. de Oliveira, Régis Borges, and Rodolfo Hoffmann. 2013. “Desigualdade de rendimentos entre os empregados da agricultura brasileira de 1992 a 2009: O efeito do salário mínimo.” Revista Econômica do Nordeste 44 (1): 125–​144. Piketty, Thomas. 2014. Capital in the Twenty-​ first Century. Cambridge, MA:  Harvard University Press. Ramos, Lauro R. A. 1990. “The Distribution of Earnings in Brazil: 1976–​1985.” PhD disserta­ tion, University of California, Berkeley. Rao, Vasala M. 1969. “Two Decompositions of Concentration Ratio.” Journal of the Royal Statistical Society. Series A (General) 132 (3): 418–​425. Rocha, Sonia. 2008. “Transferências de renda focalizadas nos pobres: O BPC versus o Bolsa Família.” Sinais Sociais 3 (8): 150–​186. Silveira Neto, Raul M., and Carlos R. Azzoni. 2013. “Os programas sociais e a recente queda da desigualdade regional no Brasil.” In Programa Bolsa Família: Uma década de inclusão e cidadania, edited by Tereza Campello and Marcelo C. Neri, Chapter  13, 217–​ 232. Brasília: IPEA.

488   Rodolfo Hoffmann Simonsen, Mário Henrique. 1972. Brasil 2002. Rio de Janeiro: APEC/​Bloch. Soares, Fabio Veras, Sergei Soares, Marcelo Medeiros, and Rafael Guerreiro Osório. 2007. “Programas de transferência de renda no Brasil:  Impactos sobre a desigualdade.” In Desigualdade de renda no Brasil: Uma análise da queda recente, Vol. 2, edited by Ricardo P. de Barros, Miguel N. Foguel, and Gabriel Ulyssea, 87–​129. Brasília: IPEA. Soares, Sergei S. D. 2004. “O impacto distributivo do salário mínimo: A distribuição individual dos rendimentos do trabalho.” Economia Aplicada 8 (1): 47–​76. Soares, Sergei S.  D. 2006. “Análise de bem-​estar e decomposição por fatores da queda na desigualdade entre 1995 e 2004.” Econômica 8 (1): 83–​115. Souza, Pedro H. G., and Marcelo Medeiros. 2015. “Top Income Shares and Inequality in Brazil, 1928–​2012.” Journal of the Brazilian Sociological Society 1 (1): 119–​132. Tolipan, Ricardo, and Arthur C. Tinelli, eds. (1975). A controvérsia sobre distribuição da renda e desenvolvimento. Rio de Janeiro: ZAHAR Editores. Taylor, Lance, Edmar L. Bacha, Eliana A. Cardoso, and Frank J. Lysy, eds. 1980. Models of growth and distribution for Brazil. World Bank. Vaz, Daniela V., and Rodolfo Hoffmann. 2007. “Remuneração nos serviços no Brasil:  O contraste entre funcionários públicos e privados.” Economia e Sociedade 16 (2–​30): 199–​232. Wolfson, Michael C. 1994. “When Inequalities Diverge.” The American Economic Review 84 (2): 353–​358.

Chapter 23

T he Devel opme nt of Br azilian Edu c at i on A Tale of Lost Opportunities? Claudio de Moura Castro

This chapter reviews Brazilian education, returning to its origins in order to make sense of the present situation.1 Overall, poor performance today is an inevitable consequence that follows from the trajectory of educational backwardness throughout the country’s history. In fact, it is surprising how rapidly the country has been able to grow given the scale of deficiencies in its educational system. Indeed, Brazil would appear to represent a clear exception to the theory that education is a necessary condition for development. Maybe this is so; but what seems clear is that the pathway to ongoing development today is not the same as that which allowed Brazil to enjoy the highest growth rate in the world from 1870 and 1989 (Maddison 1995). Today, proceeding with such a poor record on education has become far more difficult. As such, improving education should form a priority goal for Brazilian society. If we wish to realize this ambition, understanding the past is a necessary first step.

23.1.  Late Start, Acceleration, Deceleration: A Historical Overview of Brazilian Education The first colonists to arrive in what is now Brazil were soldiers and explorers from Portugal, who were mostly illiterate and in search of riches. Given the incompatibility of the local Indians with sedentary and organized life, African slaves were imported in

490   Claudio de Moura Castro great numbers, almost all of them also illiterate. Only a handful of administrators were able to read and write (Marcílio 2005). It is estimated that in the eighteenth century, the literacy rate was around 3% of the population. This contrasts with around 15% literacy in Europe at the time, shortly after Gutenberg’s invention of the printing press (Roser and Ortiz-​Ospina 2016). The arrival of the Portuguese royal family during the Napoleonic wars significantly improved the situation. At the end of the nineteenth century, illiteracy was estimated to be around 75%, similar to levels in Portugal at the time. Thus from the outset Brazil lacked a strong baseline for education (Roser and Ortiz-​Ospina 2016). The spectacular economic growth that began at the end of the nineteenth century did not incorporate significant investment in schooling. After the end of World War II, en­ rollment reached around half of primary-​school-​age children. But from then on, growth in rates of enrollment accelerated, and by the late 1990s almost 99% of children from age 7–​14 were enrolled in schools (Todos Pela Educação n.d.). At the turn of the millennium, there was a significant slowdown. Secondary educa­ tion stopped growing; in fact, in the last decade there has been a small but significant re­ duction in the number of secondary school graduates. Today, around 71% finish ensino fundamental and 51% graduate from ensino médio (secondary education).2 Comparisons with other Latin American countries can help put Brazilian educational underdevelopment in perspective. Until a few decades ago, the literacy rates of coun­ tries such as Paraguay, Ecuador, Panama, and Colombia were higher than that of Brazil. While Brazilian secondary enrollment rates have stagnated, Chile is approaching uni­ versal enrollment at this level. Leaders such as Presidents Sarmiento and Varella began pushing for universal lit­ eracy in Argentina and Uruguay, respectively, from the mid-​nineteenth century. As such, in comparison to Brazil these neighboring countries had a head start of almost a century. Thanks to the acceleration in enrollment in Brazil, the gap has been reduced. This is no minor achievement, given how sizable this gap was. However, since by world standards Latin America is a poor performer in education, reaching regional levels does not necessarily place Brazil in an enviable position compared to the rest of the world. Turning to higher education, it is worth noting that the first Spanish-​American university was founded in 1538, in Santo Domingo. Several others followed. Brazil, however, only gained its first professional colleges after the arrival of the Portuguese court in the early nineteenth century, and the first university proper was only created in 1932. At this level at least, progress has been remarkable. The growth of higher education in the last few decades has been impressive. Given the sophistication of the systems created, Brazil has overtaken many earlier starters. Nevertheless, having started so late, higher education enrollment rates are still lagging. The main bottleneck is in the supply of graduates of secondary schools.

The Development of Brazilian Education    491

23.2.  Basic Education This section reviews basic education, focusing on measures of achievement, problems, and successes.

23.2.1. The Predicament of a Federal Government That Operates No Schools Municipalities operate ensino fundamental (primary schools) while states are in charge of ensino médio (secondary schools).3 The federal Ministry of Education has some hundred universities and a network of technical schools, but no primary or secondary schools. Municipalities and states hire teachers and run schools with money they collect directly from taxpayers. As such, the nature of funding in education gives the federal government little leverage over local and state governments. This engenders the somewhat odd situation that the Ministry of Education is respon­ sible for education but neither runs schools nor pays the bills. It deals with policy and legislation, runs an evaluation agency, collects statistics, and supplements basic edu­ cation funds. Equally importantly, it dominates policy and media discussion of educa­ tion (ANDI et al. 2000). But the Ministry has little direct power over municipal or state schools, neither in theory nor in practice. There are still some federal funding mechanisms that have an influence. FUNDEB4 redistributes funds from rich to poor states. If the reallocated funds are insufficient to meet a given threshold for a less prosperous state, federal monies are expected to make up the shortfall. This amounts to around 10% of expenditures in basic education. In addition, there is a plurality of small and low-​budget programs with minimal impacts. The Ministry also imposes a minimum salary level for teachers; arguably this is of doubtful constitutional validity and is problematic for many local governments. Considering the mediocre tradition of education in Brazil, one might wonder what to expect in the weakest of municipalities in the country. The lowest performing systems are typically in small towns; in some, political intrigues hamper educational provision and consequently the quality is extremely poor. However, in other places free from such conditions, some enterprising mayors have raised education standards to European levels.5 One auspicious consequence of so many largely autonomous education systems is a wide range of experimentation. There are, in a sense, thousands of educational laboratories throughout Brazil; some strike gold and generate interesting and valuable ideas (Bruns, Evans, and Luque 2012, Chapter 3). However, the flipside is that, practically left to their own devices, many individual schools suffer from multiple maladies. These include perverse mechanisms for

492   Claudio de Moura Castro choosing principals (e.g., election by peers), lack of incentives to improve performance, promotion by seniority, lack of discipline, archaic civil service regulations, and many other factors.

23.2.2.  The Brazilian Education System in Numbers Acceleration of enrollment came very late in Brazilian history. Nevertheless, it allowed the country to catch up and surpass most Latin American countries. Unfortunately, other than in higher education, this growth came to a halt at the turn of the millennium. Figures for enrollment from 1920 show one million students in primary schools. This figure rose to 4.3 million by 1950 and 50 million by 2013. In 1920, the system covered 9% of school-​age children. This proportion had risen to 26% by 1950. More recently, 98% of 6–​14 year old children have been covered.6 The acceleration in growth, particularly since the 1950s, has been an impressive achievement. According to the most recent Educational Census, 71% of the relevant age cohort now finishes basic education. Secondary education is completed by 51%. Compared to the past, this is quite an achievement. Compared to world or even Latin American standards, these results are mediocre at best. Another challenge reflected in statistics has been repetition. Up to the 1970s, around half of first grade students would fail and subsequently retake the year. Surprisingly, less than 1% evaded at this level (Fletcher and Castro 1993). Overall, repetition has slowly receded and the aggregate figure today is around 30% (Todos Pela Educação n.d., 49). Overall then, the trajectory of education in Brazil has seen a very late start, slow growth for a long time, and then significant acceleration in the second half of the twen­ tieth century. Brazil has risen from being a somewhat laggardly performer to almost the top of Latin America rankings.

23.2.3.  An Impressive Evaluation System The country has a comprehensive and sophisticated set of evaluation programs. However, as discussed in the following, in most cases the quality measured by these tests is very weak. In 1995, Minister Paulo Renato de Souza launched SAEB (Sistema de Avaliação da Educação Básica). Since then, samples from each state, at grades four, eight, and eleven, have provided Brazil with a reliable barometer of the state of its schools. This was followed in 2005 with a new test, Prova Brasil. Instead of a sample, Prova Brasil is a uni­ versal test of students in grades 4 and 8. The results are widely published, including the mean scores of each public school. ENEM (Exame Nacional do Ensino Médio) is another test, aimed at graduating sec­ ondary students and applied voluntarily to both public and private students. This test

The Development of Brazilian Education    493 recently became the official criterion of access to federal universities. One problematic aspect of this decision is that, by aggregating all subjects, it helps to perpetuate the situa­ tion of a single homogenous secondary program for all students. A test to evaluate literacy levels in the early grades of primary schools was experimen­ tally launched under education minister Fernando Haddad. Due to technical problems, this test was discontinued, but it prepared the way for similar initiatives in various Brazilian states. Since all of these tests are only biannual, several states have created their own tests in order to have yearly monitoring of their schools. We can sum up these developments under three broad propositions: (1) Brazil has joined the ranks of developed countries, judged by the breadth of its evaluation initiatives. It has built a comprehensive evaluation system, from early education to PhD programs. (2) The technical quality of the tests and the professionalism of its application are more than adequate, commanding respect from relevant experts. (3) The use of the tests by education administrators and families is still below op­ timal. Although ENEM has become the accepted yardstick by which schools are evaluated, the other measures and tests remain underutilized.

23.2.4.  What the Tests Say about Quality However, the contrast between the impressive sophistication of the tests and the still-​ weak performance of students is stark. Unlike the test for graduating students,7 SAEB and its cousin Prova Brasil are based on a matrix of competencies, similar to that of the internationally used Programme for International Student Assessment (PISA) measure. In other words, objectively measurable levels of proficiency of students can be ascertained. Similarly, it is possible to measure the progress of students from grade to grade, and of the system over time, thanks to the use of Item Response Theory testing methodology. On the whole, the results from these tests since their introduction have not been flattering for Brazil. In 1997, 35% of fifth-​grade students had the expected level of pro­ ficiency in Portuguese; this then dipped slightly before climbing from 2005 on and reaching 45%. In ninth grade and at the end of secondary education, a slight dip was also seen, but with no significant recuperation. In mathematics, the same dip is observed, but (except for strong improvement in fifth grade), there is no sign of recovery. In fact, there is a continuous degradation of math­ ematics scores, and minimum attainment falls from 18% to 9% (Todos pela Educação 2015, 70–​7 1). Overall, the pattern is of a slight degradation of the results through time, followed by a modest improvement in the past few years, except at the secondary education level. (Although arguably the observed variations may be within the margin of error.) On the

494   Claudio de Moura Castro positive side, scores remained reasonably stable through the significant expansion in enrollment that brought to school cohorts that were previously left out. Clearly, the overall picture represents a considerable challenge for Brazil. At the fifth-​grade level, around 60% of students remain functionally illiterate (Todos pela Educação 2015, 70–​7 1). This factor is at the root of all other problems in Brazilian educa­ tion. Reading is simply an essential prerequisite for any further learning, and if students fail to master it, everything that follows is jeopardized. More or less the same is true of mathematics. Thus far, we have been looking at evaluations based on Brazilian tests. A brief com­ parison with PISA results complements this analysis. Brazil has participated in this evaluation since 2000, the first year it was organized. Of 28 countries volunteering to participate, Brazil was the weakest performer (Castro 2002). Subsequently, more countries joined PISA; the latest test in 2015 involved 70 coun­ tries. Of these, less than half were developing countries, not too different from Brazil in terms of per capita income and other indicators. Depending on subject or year, Brazil tends to rank ahead of around 10 countries. Examining the PISA numbers (OECD 2015) yields some interesting findings. (1) Contrary to popular belief, the distance between poor and rich students is con­ siderably below the average for other countries (the top 10% are 27 points from the lower end, compared to 47 points for the OECD as a group). However, this is less to do with the poor performing well than with the rich faring less well compared to their counterparts in rich countries. (2) Brazilian students perform at a level that equates to four years less schooling than students from Organsiation for Economic Co-​operation and Development (OECD) countries. (3) The proportion of Brazilian students performing below the lowest level initially defined by PISA is substantial. It necessitated the creation of a new lower category. (4) Poorer countries such as Albania, Colombia, Moldova, Malaysia, and Thailand all perform better than Brazil, while Chile, with around the same income per capita, performs considerably better. (5) The degree of Brazil’s improvement in its PISA score was higher than almost any other country between 2003 and 2012. In mathematics, scores increased from 356 to 391 points. Tempering this good news is the fact that the 2015 test suggests stagnation. (6) Brazil currently sits on the boundary between the third and fourth quartiles in terms of PISA results. However, bearing in mind that the countries participating in PISA testing tend to be the top performing in the world, we can say that roughly 50 countries display higher educational attainment than Brazil, and 150 lower attainment. In summary, educational quality has remained largely stagnant in the last 20  years. By many standards, Brazil’s performance should be considered disappointing.

The Development of Brazilian Education    495 However, some doomsayers’ claims of a worryingly deteriorating education system is also not backed up by the evidence; if anything, there has been a slight improvement. Additionally, Brazil is not among the worst performers in the globe, as is sometimes claimed by some who ignore the fact that less developed countries do not participate in PISA.

23.2.5. Secondary Education: The Structural Inadequacies behind Disappointing Results It would probably be accurate to make the general statement that most countries are not entirely happy with their secondary schools and persistently tinker with their structure. However, it is also hard to deny that Brazil has a particularly strong set of reasons for concern. As a broad generalization, there are two models for secondary education. One is the European model, offering different tracks for each career and individual profile. To illus­ trate, in addition to the dual system, Germany has three versions of secondary schools. Furthermore, there are countries, such as France, that offer different paths within different models of schools (i.e., a combination of the two systems). The second type is the American model, essentially comprising a single school for all students but offering a menu of courses that can be chosen from by each student with a considerable degree of freedom. Whatever their merits and shortcomings, both models offer one critically impor­ tant feature: choice. By contrast, the Brazilian model consists of one single type of sec­ ondary school, but also with just a single curriculum for all students. This model is ill-​equipped to cater to the numerous variations among students in previous academic performance, preferences, and abilities, and ends up being unnecessarily restrictive. Gifted mathematics students continue with the same heavy load of literature and bi­ ology as everyone else; future poets struggle their way through derivatives and integrals. Students from schools catering to favelas take the same curricula as those in the very top elite schools. All of this asks too much from students, and the weak results in sec­ ondary education—​both in enrollment and level of learning—​are partly thanks to this one-​size-​fits-​all model. The case of the recent new national curriculum (Base Curricular) demonstrates the difficulties involved in escaping this situation. In addition to other shortcomings, it continues to prescribe an overambitious curriculum. There are no rules preventing schools from diversifying and offering more of certain subjects, but in general Brazilian education is in a counterproductive habit of teaching too much. Teachers typically have to rush through overly long syllabi, leaving insufficient time to learn anything well, resulting in so-​called surface learning. A further problem is that the most sought-​after higher education institutions are the tuition-​free public universities, whose entrance tests weight all subjects equally. This means that any school diversifying its offerings would jeopardize its students’ chances

496   Claudio de Moura Castro of getting into those universities due to their comparative lack of mastery of this or that subject. Some of these problems are compounded by a lack of properly prepared teachers in mathematics and science. The proportion of teachers without requisite qualifications reaches 61% in physics and 40% in chemistry (Nascimento et al. 2014, 44). As in many other countries, rote learning prevails, with little application of the subjects presented in class. A review of mathematics textbooks published in 2001 demonstrates that none of the books examined contextualized the concepts taught (Lima, ed. 2001). It is unsurprising, then, that the numbers are highly disappointing. Opinion polls of students indicate that typically they are highly critical of their schools and struggle to generate interest in what is taught. As a result, enrollment shrinks rather than grows, and dropout rates also increase, currently standing at around 30% (Todos Pela Educação 2015). Arguably, the first years of schooling are the real challenge in Brazilian education—​ but throughout the world the model for primary education is similar. By contrast, in Brazil the structure of secondary education is (internationally speaking) unusual and hampers education, but this means that viable alternatives are available and the struc­ ture could be fixed relatively quickly. There has been awareness of this predicament for some time, but only now has a potential solution arrived, with Congress approving a new law aiming to diversify secondary schools. This suggests a step in the right direc­ tion, but it is still to be implemented.

23.2.6.  Teacher Training: A Vexed Question To what extent are the shortcomings of the Brazilian education system due to poor teacher preparation? Consensus in answer to this question seems unlikely. For instance, an opinion poll conducted by CNT/​Sensus in 2008 found that teachers on the whole considered the education received by Brazilian students to be good.8 This suggests a dis­ tinct gap between perceptions and hard data such as PISA scores. In the 1970s, the commonly identified culprit was the low level of schooling of teachers, some of whom had not even completed primary education. However, now three-​quarters of teachers hold a higher education diploma.9 In other words, there has been a revolution in the schooling level of teachers—​but the quality of education has barely improved. Since years of schooling are a poor indicator of competency, we cannot reject the hypothesis that inadequate teacher preparation is to be blamed for the bad results of students. Meanwhile, spending in secondary schools per student has trebled in the last 10 years, suggesting that money is also a poor remedy for ailing education. Looking for the culprit, then, other possibilities include that (1)  the occupation attracts weak candidates, (2) teacher colleges typically are low-​performing, and (3) there is a pronounced lack of supervised teaching experience in classrooms during training. On the first point, data on admission scores of entrants to federal universities sorted by majors chosen indicate that those opting for education careers are among the

The Development of Brazilian Education    497 lowest-​scoring entrants.10 Meanwhile, the lack of attractiveness of teaching is partly due to a much greater choice of careers now being open and attainable for women. Additionally, in Brazil public schools can be a challenging work environment, with fre­ quent discipline problems, further discouraging this choice of career. On the second and third points, a new version of teacher training, inspired by the French system, was developed and approved in the 1990s. However, under pressure from teachers from federal graduate schools of education, the Ministry of Education scrapped this model. As a result, future teachers have to enroll in programs originally intended to prepare school counselors. These programs offer a plethora of theoretical subjects and ideological discussion and little that is concrete and practical in terms of learning how to teach and the subject matter to be taught. Teachers graduate with scant experience of having actually taught under a competent mentor. Arguably this is compounded by the professors in schools of education themselves having had little ex­ perience of teaching.

23.2.7.  Technical Education Technical education is defined as a program that combines a regular secondary degree with some professional training. This is a formula widely used in Europe. It contrasts with German-​ style apprenticeships and with the American comprehensive high schools, offering a wide range of courses, including some that prepare students for the trades or provide them with some practical skills. Brazil also has a large training system—​the Sistema S—​under the auspices of the employers’ association. As such, federal technical schools are relatively minor players in preparing workers and technicians. Technical schools operate on a model that is rather unstable due to an attempt to perform at least three separate functions. The first is academic, leading to a secondary diploma and access to higher education. The second is to offer some preparation in tech­ nology, in a particular field. The third is to teach the manual skills necessary for a given occupation. The problem with this is as follows. If the academic degree that a student achieves is too weak, it makes access to higher education difficult. For that reason, such students often in turn cannot find a suitable job, and hence are pushed toward manual work. However, schools lack sufficient time for and emphasis on the hands-​on training needed for these jobs. As such, many graduates from technical schools are essentially underpre­ pared for just about any future possibility. This is a common problem throughout Latin American technical schools. In Brazil, the excellence of academic teaching in the federal network of technical schools attracts high-​level students able to pass the very competitive university en­ trance tests. From their point of view, this tuition-​free program is a good deal. However, taxpayers are essentially sponsoring expensive schools with advanced laboratories and workshops, but many graduates of these schools end up going to universities anyway,

498   Claudio de Moura Castro rather than deploying the technical skills learned. Ultimately, these schools are failing to prepare the technicians needed by industry. I pressed essentially this point regarding the subversion of the intentions of the tech­ nical school model as long ago as the 1970s (Castro 2004). During the tenure of Minister Paulo Renato, I suggested breaking these schools into two tracks: academic and tech­ nical. The idea was that, since the upper classes and the more talented youth from lower classes are set on the higher education path, they would not enroll in the technical track. This would ipso facto free up vacancies in the technical side for more modest students genuinely interested in the corresponding careers. The proposal was accepted and the programs were split. Any student could apply to one or the other track, or both. The solution was implemented, despite some confusion and uncertainty at the beginning. There was also a considerable critical reaction from left-​leaning educators influenced by Gramscian ideas who argued that education must be integrated (Castro 2004, 7–​11). When the Workers’ Party came to power in 2003, new regulations mandated the rein­ tegration of the two tracks. But since the law made it possible to offer technical curricula after graduation from high school, this alternative became the predominant model in the private sector and in SENAI schools (vocational and industry training schools) (Leite 2016, 8, 19). What might be the reasons for taking the technical track after graduation? According to the legislation, schools must offer everything that is mandated in the full academic curricula—​more than three thousand hours. As such, adding another 800 to 1,200 hours of technical courses makes the integrated technical education even more heavily burdened than the regular secondary (academic courses cannot be replaced by technical ones). Since the technical track typically caters to less academically able students (in Brazil as elsewhere), mandating a workload that is around one thousand hours heavier discourages enrollment. Hence, students prefer to graduate from secondary school and later enroll in a technical program lasting one year. After the reintegration decree, the situation in the federal schools remains murky. Despite the mandated integration, there are twice as many students taking the tech­ nical track after graduation from secondary school, compared to the integrated track.11 For all these reasons, the proportion of the secondary cohort enrolled in technical education remains very modest. A recent law on secondary education addresses this problem, proposing to make the workload of technical students equivalent to that of the regular academic track; this may help. In recent years, the federal government created PRONATEC (Programa Nacional de Acesso ao Ensino Técnico e Emprego), a program to financially support technical and vocational education. This is a promising idea, but its implementation is faulty at best, and the initiative arguably has more to do with a desire to achieve positive-​looking sta­ tistics; the government has exerted little control over implementation. The program was meant to offer training opportunities to the poorer segments of society, and as such selected the students to be funded. As often happens with such

The Development of Brazilian Education    499 programs, particularly those involving huge numbers, unavoidable mismatches occur. Additionally, little was done to monitor performance and evaluate results. Evaluations of PRONATEC by the Finance Ministry (Filho n.d.) and the World Bank have been rather negative. At any rate, private technical education has expanded, now covering 88% of the en­ rollment. Considering the low purchasing power of the students, in general they get what they pay for (i.e., not very much). Most careers are in business or health care.

23.2.8.  Private Education Brazil was a latecomer to the world of education. For a long time, most of what existed came from nongovernmental initiatives. Religious orders were the most frequent sponsors, in particular the Catholic church. But from the late nineteenth century, American Protestants started coming to Brazil; Presbyterians, Methodists, Seventh-​Day Adventists, and Baptists. Subsequently, some of these founded prestigious institutions, such as Mackenzie University (Nascimento 2005). A  few individual educators also created their own schools. Often, the institutions were named after their creators, as also happened in the United States. In the latter part of the twentieth century, public schools began to grow. While private schools never decreased in absolute terms, they became an increasingly smaller propor­ tion of the total. Today, only around 15% of enrollment in basic education is private. As for the quality of private schools, IDEB (Índice de Desenvolvimento da Educação Básica) numbers show that they are considerably superior to their public counterparts. In 2015, scores in the private sector were 51% higher. Part of that is due to the higher status of the students and their voluntary participation. However, some research has suggested that even controlling for intake quality, they still perform better, probably due to better management and better teachers (Filho 2007).12 Throughout the second half of the twentieth century, private schools had to specialize; quality education has been the predominant option, and today the very best schools are private. In the ENEM list of the top hundred schools, only five are public. The religious schools fill a second niche. Traditionally, these were the Catholic schools that had come very early to the country, when the public school sector barely existed. Last but not least, some institutions are associated with certain pedagogical traditions. This is the case of the Piaget, Montessori, and Waldorf schools. “Constructivism” is fash­ ionable, although it is unclear to what extent this is actually practiced in classrooms.

23.2.9.  Brazilian Innovations in Education As sometimes happens in societies trying to catch up, some creative innovations can appear. This seems to be the case in Brazil. A few such innovations are mentioned in the following.

500   Claudio de Moura Castro Sistemas de ensino (“teaching systems”) are the most idiosyncratic feature of Brazilian education. Leading private institutions progressively became suppliers of services for other schools. These include books, teacher training, evaluation, technical support, and many other resources, such as marketing materials. They became, as it were, a “soft” franchising service. A  few years ago, municipalities also started contracting their services. Business philanthropy is another area in which Brazil has shown considerable crea­ tivity. These initiatives are directed to help public schools from the inside. Telecurso 2000 is a television program in the K–​12 area. It was created to help enterprises educate their staff, but it eventually migrated to regular schools. Around 7  million students have obtained their degrees through this means. It transports into education the long tradition of Brazilian soap operas, with expensive production and first-​rate actors. The Ayrton Senna Institute has a very effective acceleration programs for repeaters, having already offered courses to several million students. The evaluations show very impressive results (Veloso 2011, 239). Instituto Unibanco has a well-​evaluated program to fund improvements in secondary education. ISMART is a program to select highly talented students of very low social status and prepare them to be enrolled in the very best secondary schools in the country (Veloso 2011, 239–​240).

23.3.  Higher Education As a broad generalization, it is probably accurate to state that Brazilian higher education is considered to function significantly better than the lower levels. Much effort has been devoted to creating an ambitious set of public universities. It seems that much of the dynamism of the economy from the 1960s onward derived from the existence of a significant group of highly educated leaders and professionals. However, the controversies remain. Was that the best strategy? Should the country first have invested in quality basic education?

23.3.1.  The Expansion of Public Higher Education In 1930, total enrollment was only 33,723, and up to the 1960s Brazilian higher educa­ tion lagged behind most of the rest of Latin America. With the postwar prosperity, a decision was made to establish at least one federal university in each of the 27 states. The plans were grandiose, with the universal adoption of the Humboldt model of research universities, full-​time professors, and huge campuses.

The Development of Brazilian Education    501 In 1960, Brazil enrolled 59,624 students. In 1970, enrollment reached 425,478. Today, it is around 8.5 million students—​quite an achievement, by any standards (Tachibana, Filho, and Komatsu 2015). This investment, funded by the Inter-​American Development Bank, allowed the growth of high-​quality institutions (at least compared with what had existed before). Close to world-​class universities were created in the wealthiest states. In the less pros­ perous ones, the new universities were not as impressive, but they may have had equally powerful impacts.

23.3.2.  The Predicament of Public Universities In most cases, the grandiose Humboldt model of teaching and research turned out to be unrealistic. Full-​time faculty and PhD students are a necessary but not sufficient con­ dition for research. Hence, for a number of reasons, serious research bloomed only in about a dozen federal universities, all of them in the most advanced states. If research remained concentrated in a few centers, full-​time contracts for professors were widespread and resulted in inevitably high costs per student in all of these universities. Average per student costs of Brazilian federal universities are equivalent to those of the OECD average.13 In other words, while the model only successfully created a few research centers, the fact that the funding formula was the same for all meant that all of the others cost the equivalent of research universities, despite being little more than teaching-​only colleges. In addition, the original 1960s blueprint for federal universities was poorly designed in terms of governance. Subsequent developments made the regulations even less func­ tional, as follows: (1) The universities proclaim their autonomy, backed by legislation; they claim to take no orders from the Ministry. However, in practice they have very little freedom in finances and recruitment (i.e., the worst of all worlds). (2) There are no rewards for efficiency or excellence, and, conversely, very few penalties for poor performance, absenteeism, waste, and so on. (3) Rectors are elected by all faculty, students, and, in some cases, all university staff. This populist method can be counterproductive for some universities. (4) Student movements have a long tradition in Latin America, but more recently in Brazil the really powerful political actors in higher education have been the teacher unions. They resist proposals to make faculty more accountable and, for a variety of reasons, often go on strike. (5) In addition to the high costs, all these factors militate against the expansion of the federal universities. In fact, good students are increasingly voting with their feet and enrolling in private institutions.

502   Claudio de Moura Castro

23.3.3. Private Higher Education: From Religious Institutions to Initial Public Offerings From the 1980s onward, it became financially impossible to keep expanding public universities at the same pace. But the number of graduates from secondary schools began to grow, creating political pressure to expand enrollment. Given the alternatives of being criticized for not offering enough vacancies or being criticized for allowing the private sector to expand, the Ministry of Education tacitly preferred the latter alternative. The strategy was to relax the bureaucratic hurdles faced by private colleges. Additionally, higher education institutions were permitted to be for-​profit, if desired. Today, around half of private institutions declare their intention to make profits. The Catholic and Protestant universities have been around for many decades. But the recent expansion came from non-​religious private colleges, in many cases former high schools. Around 75% of higher education enrollment is now in the private sector. Beginning in the mid-​1990s, the private sector expanded very rapidly. Today, pri­ vate colleges and universities enroll almost 7.8 million students. Distance education has grown faster than face-​to-​face; between 2001 and 2010, it expanded by 110%, reaching then 15% of the total enrollment, with 1.34 million students. This compares with 80% for face-​to-​face education. All in all, higher education displayed very impressive growth. Gross enrollment went from 14.5% in 2002 to 28% in 2012 (Tachibana, Filho, and Komatsu 2015, 4). By any standards, this is very impressive growth. But note that net en­ rollment is much smaller than that, at 15% in 2012. The difference is due to the large pro­ portion of students who enter higher education at over 25 years of age. Fierce price competition is the norm in the most crowded areas, such as business ad­ ministration. As a result, tuition is falling (in real terms). Of course, the growth of pri­ vate education depends on potential candidates having the resources to pay for tuition. As the pool of paying students reaches its limits, distance education is the formula to further lower the levels of tuition. It is interesting to note that since 2012 the number of students entering higher edu­ cation has been larger than the number of graduates in secondary schools (1.95 million enrolling, compared to 1.84 million graduating, in 2013) (Tachibana, Filho, and Komatsu 2015, 19). This surprising situation reflects both the stagnation of secondary graduations and the strong presence of more mature students who decide to enroll in higher educa­ tion once their professional life is stable (in 2012, 40% were over 25 years old). As this pool of relatively well-​off candidates is exhausted, further expansion of the private system depends on public funding, since distance education is not always a good substitute for face-​to-​face classes. There has been a reduction in the growth rate in more recent years. The year 2016 ended with a negative figure. As a response, two programs have taken shape and have expanded considerably in the last several years. Under PROUNI (Programa Universidade para Todos), tax rebates of for-​profit institutions are exchanged for fellowships for their students. The combination

The Development of Brazilian Education    503 of economic and social filters makes this a good program. Meanwhile, FIES (Fundo de Financiamento Estudantil) is the student loan program of the Ministry of Education. It has been around for many decades, but has been expanded greatly in the last sev­ eral years. It is not a well-​conceived program. Its abrupt contraction following the 2014 elections created much confusion and ill feeling. At the turn of the millennium, foreign (for-​profit) universities took note of the dyna­ mism of the Brazilian market for higher education. A few of them invested in existing local institutions. This rush provoked consternation among those of a nationalistic in­ clination, fearing that higher education would meet the same fate as the Brazilian phar­ maceutical firms that gigantic multinational laboratories swallowed up. However, the denationalization of higher education did not happen. There are very few for-​profit higher education initiatives around the world. The overall sce­ nario is not at all similar to the pharmaceutical or automobile industries, with their mammoth multinational firms. What happened was quite different from the predictions. The more aggressive colleges began to expand, by means of the acquisi­ tion of other institutions. Indeed, the proprietors of K–​12 sistemas de ensino, banks, investment institutions, and individual entrepreneurs saw higher education as an in­ teresting business. Growth has been spectacular. Several of them enroll more than 100,000 students. One of them, Pitágoras/​Kroton, boasts more than one million students. Such actors have brought to education sophisticated marketing techniques and pro­ fessional management and auditing practices. However, they have shown little interest in the content and delivery of education. Over-​concentration is becoming an issue, with predatory competition beginning to appear. Most private colleges offer excellent physical locations and facilities, even in schools that are not so impressive in academic terms; in a very competitive market, they believe that attractive architecture helps sales. Given the legal requirement to have a certain proportion of faculty with graduate degrees, Brazilian colleges have professors with respectable academic credentials. Probably no other Latin American country employs so many holders of master’s and PhD degrees. What is missing? Simplifying somewhat, we could say that this is a level in which the schools are good and the students weak. The main reason is the poor quality of the K–​12 schooling that supplies universities with students. But institutions are reluctant to adjust their courses to the level of the students who enroll. Indeed, there is an unfortunate ten­ dency to teach for the ideal students the teachers would like to have, rather than the ones who are in front of them. Regardless of disagreements and ideological disputes, getting an education still makes economic sense. In 2013, graduates of higher education institutions earned 2.41 times more than those with a secondary diploma. This income differential is shrinking but still justifies the efforts of the youth to get their diplomas and the energy and push of the private sector to offer the courses.

504   Claudio de Moura Castro

23.3.4.  Provão and ENADE: What the Evaluations Say In the late 1990s, a national test was created to evaluate how much of the official curricula had been mastered by graduating students of higher education. This test, nicknamed Provão, became a powerful tool to measure quality. In subsequent years it was renamed ENADE (Exame Nacional de Desempenho de Estudantes). Similarly to the evaluation of graduate schools, this was a pioneering initiative globally, in an area that remains unex­ plored in just about all countries. If mastery of the official curricula is an indicator of quality, ENADE provides the country with a means to compare institutions of higher learning. The adequacy of the tests depends on the areas. In the hard sciences, tests are not controversial. In areas subject to ideological disputes, such as education or philosophy, they may be less reliable or more biased. Elsewhere I  have used ENADE data to compare institutions (Castro, Giuntini, and Lima 2011).14 Surprisingly, these data show that the difference in quality between private and public institutions is quite small. Also instructive is to notice that, among the 100 top performing institutions, 65 are private. These results are even more significant when we consider that private colleges typically cost around a third of what federal universities cost. Perhaps even more surprising is that the quality of for-​profit institutions was no different from that of not-​for-​profits (Castro, Giuntini, and Lima 2011, 29–​38). Whatever determines quality is not correlated with whether or not the institution is profit-​making. Another counterintuitive result comes from those institutions that have had initial public offerings (IPOs) and their shares are on the stock market. Surprisingly, the mean performance of the students is the same as in the other private institutions. There is one difference, however. The variance in scores is much narrower, meaning that they do not permit their schools to lower their performance very much. But none performs signifi­ cantly above the mean either. Also surprising is that, contrary to the conventional wisdom, the weakest-​performing courses in public universities are just as bad as the worst in the private institutions.

23.4.  Graduate Schools and Technology Graduate schools are the crown jewels of Brazilian education. Up to the late 1960s, the few existing programs were of the old European model. A doctorate consisted of a stu­ dent preparing a thesis under the supervision of an advisor, if at all.

23.4.1.  Graduate Schools Beginning in the 1960s, with the initial support of the Ford and Rockefeller foundations, and later of the US Agency for International Development (USAID), Brazil set in motion

The Development of Brazilian Education    505 a serious program to send students abroad for graduate studies. Eventually, a few thou­ sand enrolled in American institutions and returned with masters or doctorates. Progressively, CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) (within the Ministry of Education) and CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) (then, within Planning) took over from the American sponsors and continued the funding of scholarships. This succession of sponsors is remarkable for the seamless continuity of policies and the strictly merito­ cratic formulae adopted. In a society so used to nepotism, the machinery to choose and send students abroad is striking. When the flow of returning graduates began to increase, funds from the Planning Ministry became available to build laboratories and create graduate schools. FINEP (Financiadora de Estudos e Projetos) was the key agency in this process.15 Another source of funds was a World Bank loan, under the PADCT (Programa de Apoio ao DesenvolvimentoCientífico e Tecnológico) project. The first cohorts of returning students shaped the graduate schools in the image of the American models with which they were familiar. Both Brazilian agencies created ambi­ tious scholarship programs to support the students of the new graduate schools. In the late 1970s, CAPES created an evaluation system for the growing number of graduate schools. Interestingly, this started as an improvised peer-​review assessment, in order to allocate quotas of scholarships to the better programs. Eventually, it became public, with the grades of every graduate school being published on the Internet. This system remains powerful and prestigious (despite a few hitches). A very significant feature was connecting grades obtained in evaluations to the number of scholarships granted to each institution—​which in turn would be allocated to their students. In my view, this link has been the most powerful engine driving the continuous improvement in graduate schools.16 The governance of federal universities is riddled with problems. But the factor of ex­ ternal agencies funding research on a competitive basis created a productive environ­ ment for the graduate programs. There is a world of difference between graduate and undergraduate programs. Undergraduate programs are lost in the entropy of a weak and conflictive bureaucracy. Graduate programs thrive, competing for projects, research funds, publications, and evaluations. Today, there are close to 3,000 programs offering a master’s degree or a doctorate.17 In 2013, enrollments in master’s programs reached around 160,000 students. PhD programs enrolled around 110,000. Of the total, only 17% are in private institutions. In 2013, around 60,000 students completed their master’s degree and 23,000 their doc­ torate (Posgraduando.com 2016). This is one of the highest figures anywhere in the world. It is thanks to this vast output of graduate degrees that colleges can expand while still finding enough professors with the requisite academic profiles. In the 1950s, one could not find a single paper by Brazilian authors in the Current Content periodicals. Between 2009 and 2013, Brazil published 35,663 papers credited in the same databases. This places Brazil as the 13th largest producer of research.18

506   Claudio de Moura Castro In the 1970s, Argentina—​in addition to its four Nobel laureates—​was ahead of Brazil in number of publications. Today, Argentina produces one-​fourth of the Brazilian output. Brazil is ahead of countries such as Holland, Switzerland, and Russia. There is ongoing controversy about the quality of Brazilian science. Serious observers such as Simon Schwartzman claim that it is not good research, much of it being trivial (Schwartzman 2005). Perhaps, but is that not an accusation that could be leveled against the global output of research? The frequency of citations of Brazilian papers is modest, at 0.65. This number is higher than Russia and almost as high as India, both of which are long-​established forces in science.19 However, it is below the world average of 1.0, and 1.45 for the United States. Perhaps the greatest weakness of Brazilian research is thematic fragmentation. Too many researchers merely continue their PhD work, or the research of their supervisors. As a result, in many areas there is a lack of critical mass to make a difference in particular research themes. To sum up, the single most impressive Brazilian achievement in education is the de­ velopment of a large network of graduate schools, with its corresponding research. By any standard, this was a remarkable feat. But it shines even more in a country with so few achievements in the field of education.

23.4.2.  Technology Made in Brazil? Overall, the tendency is to denigrate Brazilian achievements in technology and inno­ vation. Indeed, in 2013, Brazil published 35,000 papers, while it only registered around 3,000 patents. The accumulated number of patents reached 41,453, in contrast to the 2.2 million patents of the United States.20 Could we say that the stagnation of productivity growth in the last several years is solid evidence of a failure to innovate? This remains an open question. Brazil ranked a modest 19th place in registered patents in 2012. However, that number still puts it ahead of all of Latin America. In other words, the reasons for the modesty of the numbers are probably the same around the world, rather than a particular weakness of the country. According to the Global Innovation Index (2014), no developing country is ahead of Brazil. In other words, it is not a star, but not a laggard either.21 Moving from science to technology is a very arduous transition. Few countries have done truly well in this move. Science is an enterprise within the confines of the academy. Technology, by contrast, requires selling some idea or instrument to profit-​calculating businessmen. If it is not ready and functioning, it is not salable. If it takes too long, someone else will have already supplied the market.

23.5.  What about the Future? This concluding section focuses on two different topics. The first is how and what to teach. The second is the perennial debate over why quality remains poor.

The Development of Brazilian Education    507

23.5.1. How to Teach for the Twenty-​first Century if Children Cannot Read? The newspapers abound with pleas to teach for the twenty-​first century, not for the workers of the industrial revolution. This seems fair enough. How can Brazil become the global leader it dreams of being if its educational system remains caught in the past? However, let us recall that engineers or scientists did not create the industrial revolution. Instead, it was mechanics who could read and deployed their skills to absorb the growing technical litera­ ture and to correspond with colleagues pursuing similar projects (Rosen 2012). If a large segment of the labor force is not functionally literate and achievement in mathematics remains so poor, how far can Brazil go? Furthermore, how can Brazilians put to use the new technologies that are so critical for further progress? Imagination and creativity significantly depend on a qualitatively different education. Let us examine one aspect of this controversy. Critics of the system tell us that students have to learn to deal with problems that have more than one answer. This seems reasonable—​but the disagreements concern when this jump to more open-​ended intellectual pursuits has to take place. My view is that it must happen much later and to a more limited extent than proclaimed by many. We see T-​shirts and posters reproducing Einstein’s dictum that imagination is more important than knowledge. This certainly may have been true for Einstein—​after obtaining a serious doctorate. In earlier stages, students need to tackle the basics, seriously and with full rigor. Open questions in mathematics and physics only appear in graduate school. Before that, one has to learn the exact meaning of well-​constructed sentences, theories, and authoritative interpretations of the endless sequence of what is subsumed under the term Western civilization. Since the early grades, students should make, explore, and become engaged in projects. There, they should begin imagining solutions and building contraptions. But it is too early to start migrating to open-​ended theoretical issues and to deny the prece­ dence of “yes or no” answers.

23.5.2.  Why Are Improvements in Education So Slow? Many educated Brazilians wonder why, after so much effort, so many perorations, and so many plans, progress in the quality of education is so slow, if it even exists all. There are many reasons. Let us comment on two of the most salient. The first reason is the absence of good models of quality education. Our schools were late to emerge and were transplanted from a country with scant traditions in education. Brazilians just do not know what a good school is like. The second issue is more insidious. Politicians respond to demands from their constituencies and can ignore what is not perceived as a keen preference of voters. Conversely, they will try to respond to what voters think is essential.

508   Claudio de Moura Castro In 2008, an opinion poll by CNT/​Sensus found that parents think the education of their children is adequate or good. Several other polls found equivalent results.22 Hence, they do not think that improving education should be a priority. Of the other 30% that do, half of them have children in private schools, and thus are not affected by bad public schools. To politicians, improving education requires fighting unsavory battles. Students have to work harder, teachers have to prepare better for their classes and grade homework, and principals must impose their authority to make sure that everything works prop­ erly. From a political perspective, the numbers appear to speak for themselves: 70% are happy with what they have. Politicians act pragmatically. It seems that the path to reform requires convincing Brazilian society at large that it is heavily handicapped by the poor quality of education. Frequent mentions of Prova Brasil and PISA may be helping. But, evidently, it has not been enough.

Notes 1. The author is thankful for comments by Biorn Maybury-​Lewis, Helio G. Barros, James Ito-​ Adler, João Batista de Araujo e Oliveira, and José Tarquínio Prisco. However, they are not responsible for the present text. 2. Ensino fundamental corresponds to eight years of schooling and middle school in the United States; ensino médio is secondary or high school, corresponding to 11 or, more re­ cently, 12 years of schooling. 3. Some states still run a considerable proportion of primary schools, despite laws mandating the transfer of this level to municipalities. 4. FUNDEB (Fundo de Desenvolvimento do Ensino Básico) is a transfer mechanism between states, complemented by federal funding. Richer states transfer funds to those unable to reach the mandated minimum budgets. 5. Among the best known are Sobral and Foz do Iguaçu. For a broader treatment of the sub­ ject, see Fundação Santillana (2016). 6. Early figures from Fundação IBGE, Séries Estatísticas Retrospectivas (1970) cited in the publication Revista de Estudos Pedagógicos no.  101. Latest numbers from Todos Pela Educação (n.d., 20). 7. At the higher education level, the test for graduating students (ENADE) can only tell which schools are ahead or behind others (i.e., it is a comparative and relational measure, showing which schools are better or worse). That is a different question from that of whether schools are actually good or bad in themselves, something frequently overlooked by the media. The tests also lack comparability over time. 8. “Pais aprovam escolas ruins,” Veja, August 16, 2008. 9. Anuário brasileiro da educação, 102. 10. https:// ​ q uerobolsa.com.br/ ​ f aculdade/ ​ u fmg/​ n otas-​de-​corte-​sisu?gclid=CM70_​ _​ args8CFQQHkQodT5gMJg. 11. A thesis about the federal technical school in Piauí illustrates this vexing instance of a non-​ enforceable law (Pereira 2012). 12. Despite this, PISA results show that less than 1% of Brazilian students in science (most of whom come from private institutions) are among those in the top category, contrasting with 8.9% of male OECD students (OECD 2015, 2).

The Development of Brazilian Education    509 13. http://​nces.ed.gov/​programs/​coe/​indicator_​cmd.asp. 14. The official indicator lumps together ENADE scores, percentage faculty with PhDs, with master’s, or with full-​time contracts, plus other indicators. None of them correlates with scores in ENADE; they merely pad the indicator and bias the results in favor of public universities that can afford these extra costs. 15. Finep is an independent parastatal agency to fund research and development. 16. I should disclose that I was director general of CAPES from 1979 to 1982 and thus inti­ mately involved in the institutionalization of the evaluation system. 17. There is some double counting since most programs offer both master’s and doctorates. 18. Pesquisa Fapesp magazine, June 2015, 8. 19. Pesquisa Fapesp magazine, June 2015. 20. http://​anprotec.org.br/​site/​2014/​04/​brasil-​ocupa-​penultima-​posicao-​em-​ranking-​de-​ patentes/​. 21. http://​www.wipo.int/​econ_​stat/​en/​economics/​gii/​. 22. “Pais aprovam escolas ruins,” Veja, August 16, 2008. Similar results were obtained by INEP (2004).

References ANDI, MEC, IAS, UNICEF, NEMP, FUNDESCOLA, and CONSED. 2000. Mídia e educação: Perspectivas para a qualidade da informação. Brasília. Available at https://​www. unicef.org/​brazil/​pt/​midiaedu.pdf. ANPROTEC. 2014. “Brasil é o penúltimo em ranking de patentes.” Available at http://​anprotec. org.br/​site/​2014/​04/​brasil-​ocupa-​penultima-​posicao-​em-​ranking-​de-​patentes/​. Bruns, Barbara, David Evans, and Javier Luque. 2012. Achieving World-​Class Education in Brazil: The Next Agenda. Washington, DC: World Bank. Castro, Claudio de Moura. 2002. A penosa evolução do ensino e seu encontro com o PISA. Brasília: INEP. Castro, Claudio de Moura. 2004. “Technical Education:  The Chronicle of a Turbulent Marriage.” In The Challenges of Education in Brazil, edited by Colin Brock and Simon Schwartzman, 115–​134. Oxford: Symposium Books. Castro, Claudio de Moura, Aldo Giuntini, and Luciana Lima. 2011. Avaliação no ensino superior: Acertos e derrapagens. Brasília: ABMES. Dutta, Soumitra, Bruno Lanvin, and Sacha Wunsch-​ Vincent, eds. 2014. The Global Innovation Index 2014:  The Human Factor in Innovation. Available at https://​www. globalinnovationindex.org/​userfiles/​file/​reportpdf/​gii-​2014-​v5.pdf. Filho, Fernando de Hollanda Barbosa. n.d. “Pronatec Bolsa Formação.” Secretaria de Política Econômica/​Ministério da Fazenda. Available at http://​portal.mec.gov.br/​docman/​abril-​2010-​ pdf/​22061-​24092015-​lancamento-​estudos-​pronatec-​ministerio-​fazenda-​pdf/​file. Filho, Naércio Menezes. 2007. Os determinantes do desempenho educacional do Brasil. São Paulo: Instituto Futuro Brasil. Fletcher, Philip R., and Cláudio de Moura Castro. 1993. “Mitos, estratégias e prioridades para o ensino de 1º grau.” Estudos en Avaliação Educacional (8): 39–​56. Fundação Carlos Chagas. Fundação Santillana. 2016. Destino educação: Escolas inovadoras. Futura/​Santillana, Moderna. Available at https://​pt.calameo.com/​read/​002899327713298808b0d. INEP. 2004. Pesquisa nacional qualidade da educação: A escola pública na opinião dos pais. Brasília: INEP.

510   Claudio de Moura Castro Leite, Elenice. 2016. “Panorama do Ensino Vocacional no Brasil.” In Educação Média Profissional no Brasil:  Situação e Caminhos, edited by Simon Schwartzman, 123–​149. São Paulo: Fundação Santillana. Lima, Elon Lages, ed. 2001. Análise de livros de matemática para o ensino médio. Rio de Janeiro: Sociedade Brasileira de Matemática. Maddison, Angus. 1995. Monitoring the World Economy: 1820–​1992. Paris: OECD. Marcílio, Maria Luiza. 2005. História da escola em São Paulo e no Brasil. São Paulo: Instituto Braudel, Segunda Parte. Nascimento, Paulo Meyer, Carolina Andrade Silva, and Paulo Henrique Dourado da Silva. 2014. “Subsídios e proposições preliminares para um debate sobre o magistério da Educação Básica no Brasil.” Radar (32):  37–​51. Available at http://​www.ipea.gov.br/​portal/​images/​ stories/​PDFs/​radar/​140508_​radar32.pdf. Nascimento, Esther Fraga Vilas-​Bôas Carvalho do. 2005. Educar, curar, salvar: Uma ilha de civilização no Brasil tropical. Maceió: Edufal. OECD. 2015. Results from PISA 2015: Brazil. OECD Country Note. Paris: OECD. Pereira, Samara C. S. 2012. “A política de educação professional media integrada: Entre o legal e o real.” Master’s thesis, Universidade Federal do Piauí. Posgraduando.com. 2016. “GeoCapes: Os números da pós-​graduação no Brasil em detalhes.” Available at http://​posgraduando.com/​geocapes-​os-​numeros-​da-​pos-​graduacao-​no-​brasil-​ em-​detalhes/​. Rosen, William. 2012. The Most Powerful Idea in the World. New York: Random House. Roser, Max, and Esteban Ortiz-​Ospina. 2016. “Literacy.” Available at https://​ourworldindata. org/​literacy/​. Schwartzman, Simon. 2005. “Doutorados no mundo, o Modo II e a dama adormecida.” Available at http://​www.schwartzman.org.br/​sitesimon/​?p=32&lang=en-​us. Tachibana, Thiago Y., Naércio de Menezes Filho, and Bruno Komatsu. 2015. Ensino superior no Brasil. São Paulo: Insper. Todos Pela Educação. n.d. De olho nas metas. São Paulo: Todos pela Educação, Meta 1. Todos Pela Educação. 2015. Anuário brasileiro da educação:  2015. São Paulo:  Todos Pela Educação. Torres, Haroldo da Gama, Danilo França, Jacqueline Teixeira, Rafael Camelo, and Edgard Fusaro. 2013. O que pensam os jovens de baixa renda sobre a escola. São Paulo: CEBRAP. Veloso, Fernando. 2011. “A evolução recente e propostas para a melhoria da Educação no Brasil.” In Brasil: A Nova Agenda Social, edited by Edmar Lisboa Bacha and Simon Schwartzman, 215–​253. Rio de Janeiro: GEN.

Chapter 24

An ti-​P overt y T ra nsfe rs and P overt y Re du c t i on Armando Barrientos

24.1. Introduction In the new century, Brazil has attracted considerable attention among low-​and middle-​ income countries for the way it has successfully combined economic and social policies to reduce poverty, inequality, and social exclusion.1 The emergence of large-​scale so­ cial assistance institutions addressing poverty and social exclusion is central to these achievements. Among them, Brazil’s Bolsa Família, an anti-​poverty program reaching 14 million households, including one-​third of all children in the country, has been par­ ticularly influential in countries in Asia and Africa. Less well known, but no less impor­ tant, social pension schemes have pushed pension coverage of people aged 65 and over to just over 86%, among the highest in the region. Previdência Social Rural provides around 7.5  million transfers annually, largely old age pensions, to informal workers in rural areas. The Benefício de Prestação Continuada provides income transfers to 3.7 million older people and people with disabilities in extreme poverty. This chapter provides a comprehensive examination of social assistance in Brazil—​ its main components, evolution, effectiveness, and sustainability—​and assesses its contribution to poverty reduction. The emergence of social assistance2 in Brazil has been swift, but far from linear. The 1988 Constitution, following 20 years of dictatorship, is the marker for the rapid expansion of social assistance programs and policies in the years that followed. However, the policy instruments the Constitution supported, Previdência Social Rural and Benefício de Prestação Continuada, were not especially innovative or farsighted. Their orientation was firmly rooted in Bismarckian welfare policy, specifically on the distinction between individuals with or without the ability to work (Jaccoud, Hadjab, and Chaibub 2009). They focused on old age poverty and on disability, but failed to

512   Armando Barrientos address child poverty (Barros and Carvalho 2003), and favored pure income transfers that replicated the “compensatory” approach of the golden age of European social assistance. Bolsa Família developed instead out of municipal experimentation with Bolsa Escola, rooted in a mix of guaranteed income proposals, multidimensional perspectives on poverty, and education interventions. Bolsa Escola provided transfers to households in extreme poverty, conditional on children attending school. The spread of these programs was sanctioned at the federal level in the 1997–​2001 period. In parallel, a pro­ gram for the eradication of child labor, Programa de Erradicação do Trabalho Infantil (PETI), emerged in 1996 as a human capital conditional income transfer program addressing hazardous child labor in areas with high incidence of this type of work. Other income transfer programs emerged in 2001. The incoming Lula administration eventu­ ally consolidated the various income transfer programs into Bolsa Família in 2003 and established the Ministry of Social Development in 2004 to provide overall leadership and management in social assistance.3 The Unified Social Assistance System has ex­ panded to add ground-​level intermediation and referrals. Recent policy developments include Brasil Sem Miséria, focused on strengthening economic inclusion, and Brasil Carinhoso, which guarantees a floor to transfers with the objective of eradicating extreme poverty. This chapter provides a critical review of the growing literature on social assistance in Brazil, focusing on tracing the main institutional developments and assessing their poverty-​reduction effectiveness from an economic perspective. The discussion in the chapter makes a contribution in three respects. First, the huge interest in conditional cash transfers globally has concentrated attention on Bolsa Família,4 as the flagship of social assistance in Brazil, but other programs are important as well. The two social pensions in Brazil reach over 10 million people with a budget over two times greater than that of Bolsa Família. It is important to consider all the components in the round. Second, the impact evaluation literature often focuses on inputs and outcomes, paying scarce consideration to institutional factors. The chapter pays close attention to the design, evolution, and implementation of social assistance programs in Brazil. Third, the chapter provides a comprehensive review of the substantive literature highlighting the role of social assistance transfers in ensuring a sustained reduction of poverty. The chapter highlights throughout the challenges for the future of social assistance in Brazil. The rest of the chapter is organized as follows. Section 24.2 provides an overview of poverty trends in Brazil. Section 24.3 provides an overview of the emergence of social assistance in Brazil, paying attention to the design, implementation and evolution of its different components. Section 24.4 presents a brief review of the economics of social as­ sistance, with a view to setting and organizing expectations on their outcomes. Section 24.5 assesses the outcomes from these programs as identified in available impact evalua­ tion studies. A final section concludes.

Anti-Poverty Transfers and Poverty Reduction    513

24.2.  Poverty Trends in Brazil

Poverty and extreme poverty headcourt rates and Gini

As shown in Figure 24.1, Brazil experienced a sustained fall in the headcount rate of poverty after 2003, whether the focus is on extreme poverty or poverty. Extreme pov­ erty describes households whose income is insufficient to purchase a basic basket of food. Adding nonfood essentials—​the costs of basic services like health, education, and transport—​to the food poverty line sets a poverty line. At the turn of the century, 35% of the population in Brazil was estimated to be in poverty, while 15% of the population was in extreme poverty. By 2014, the rates were 13% and 4.2%, respectively. Over this decade and a half, the share of the population in poverty and in extreme poverty shrank to around one-​third of the baseline. This is a remarkable achievement in the context of Brazil, especially as poverty rates had remained stable in the 1990s. It is more remarkable that the sustained decline in poverty was accompanied by an equally large and sustained fall in income inequality. As Figure 24.1 demonstrates, the Gini measure of inequality fell from 59.4% at the turn of the century to 51.8% in 2014, a reduction of around 10% from an admittedly high baseline.5 There is a measure of consensus regarding the main factors behind the sustained fall in poverty and inequality in Brazil. Rapid economic growth led to marked improvements in labor market opportunities, as shown by a parallel fall in the rate of unemployment 70 60 50 40 30 20 10 0

1992 1993 1995 1996 1997 1998 1999 2001 2002 2003 2004 2005 2006 2007 2008 2009 2011 2012 2013 2014

Extreme poverty % pop.

Poverty % pop.

Gini

Figure 24.1.  Poverty and inequality trends in Brazil, 1992–​2014. Data source: IPEA Data, accessed June 19, 2016.

514   Armando Barrientos until after the 2008 global financial crisis. Social and labor market policies strengthened the income gains for low-​income households. The minimum wage rose in real terms, with implications for wage setting in formal employment, but also in informal employ­ ment. The minimum wage indexes social assistance transfers and the minimum pension. In addition, the rapid expansion of Bolsa Família from around 6 million households in 2003 to 14 million in 2014 greatly expanded the reach of the program. Soares et al. (2010), relying on a decomposition of income sources reported in the national house­ hold survey Pesquisa Nacional por Amostra de Domicílios (PNAD),6 conclude that Bolsa Família and non-​contributory pensions accounted for 31% of the reduction in inequality in the period 1999–​2009, around 15% each. They also conclude that, in the absence of Bolsa Família, poverty would have been 16% higher and extreme poverty a third higher.7 The literature suggests that anti-​poverty transfers or social assistance have made an important contribution to the reduction of poverty and inequality in Brazil. The next section traces the emergence of social assistance following the 1988 Constitution.

24.3.  The Emergence of Social Assistance in Brazil The 1988 Constitution represents a watershed in the development of social assistance in Brazil. It placed social inclusion at the top of the public policy agenda. The policy and institutional development that followed pursued three main inclusion strategies responsible for the current configuration of social assistance: (1) integrating informal workers in social insurance institutions through a special regime; (2) providing income transfers to older and disabled people in poverty; and (3) setting up human capital in­ come transfer programs targeting the population in extreme poverty. They led to the three main components of social assistance in place in Brazil today. Figure 24.2 provides a bird’s eye view of these components. The following sections will examine in some de­ tail their design and evolution.

24.3.1.  Previdência Social Rural as a Special Regime for Informal Rural Workers The 1988 Constitution provided the basis for the rapid expansion of social assistance. Following 20 years of dictatorship, Brazil emerged in 1985 with a huge “social debt,” a commitment to address this debt, and a ferment of ideas on the way forward. The discussions of the constituent assembly led to far-​reaching change in the principles underpinning social policy. It established social assistance as a distinct policy dimen­ sion, based on citizenship and social rights. Responsibility for the provision of so­ cial assistance was placed with the state. This was a break with the past, where social

Anti-Poverty Transfers and Poverty Reduction    515 Previdência Social Rural [A] 1988

Benefício de Prestação Continuada [B]

Constitution Bolsa Escola [C] Municipal

Federal Bolsa Família [C]

BrasilCarinhoso

PETI (Child Labor) [C] 4 other ccts

1988

1991

1995 1996

2001

2003

2012

Three strategies of social assistance expansion in Brazil: [A] Preferential inclusion of rural informal workers into the private sector social insurance fund [B] Social pensions for older and disabled people in households with per capita household income below 1/4 of minimum wage. [C] Human development conditional income transfers like Bolsa Escola (1995–2003), Bolsa Família since 2003. Brasil Carinhoso (2012) provides a guaranteed minimum income floor at the extreme poverty line. Number of Transfers/Public Subsidy in 2010: [A] 7.8 m/1.4% GDP [C] 13.2 m/0.6% GDP

[B] 3.2 m/0.7% GDP

Figure 24.2.  Evolution of social assistance components in Brazil. Source: Author.

protection entitlements were grounded squarely on a Bismarckian contributory prin­ ciple, while state responsibility for social assistance was limited to partial and residual support for provision of services by private and charitable bodies (Jaccoud, Hadjab, and Chaibub 2009). The aspiration in the Constitution to address the large inequalities existing between urban and rural areas in Brazil encouraged discussion on extending social insurance coverage to workers in agriculture. This led to Law 8212/​8213 in 1991, establishing Previdência Social Rural. The legislation created a special regime by which waged and own account workers (in family farms) could access social insurance benefits provided by the National Institute of Social Security (Instituto Nacional de Previdência Social; INPS) with a minimum of 10 years contribution record. This is in recognition of the informal and irregular nature of employment in this sector. Crucially, the legislation delayed implementation of the contribution requirement for 10 years to enable workers to meet it and also to provide them with an incentive to start contributing (Delgado and Cardoso 2000). Under the Previdência Social Rural, women aged 55 and over and men aged 60 and over who could demonstrate having worked in mining, agriculture, or fishing qualify for a pension equivalent to one minimum wage.8 Previdência Social Rural involved a

9 8 7 6 5 4 3 2 1 0 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

Number of transfers in millions December of each year

516   Armando Barrientos

Figure 24.3.  Previdência Social Rural transfers, 1980–​2009. Data source: Informe de Previdencia Social (2010).

significant scaling up of entitlements to workers in agriculture (Barbosa 2010), as can be seen in the rise in the number of benefits during 1992 and 1993, shown in Figure 24.3. With the introduction of Previdência Social Rural in 1992, the number of transfers jumped from 4.1 million in 1992 to 6 million in 1994. In 2009, Previdência Social Rural paid 5.3 million old age and 0.4 million disability pensions (Barbosa 2010). The remainder is accounted for by maternity, sickness, and other benefits. The introduction of Previdência Social Rural is hugely significant because it provided an initial break with the Bismarckian contributory principle and a shift toward citizenship-​ based old age pension entitlements (Delgado and Cardoso 2000). On paper, the fi­ nancing modality is hybrid, financed through a tax on sales of agricultural produce and contributions from waged and own account workers. In practice, the tax on sales of ag­ ricultural produce amounts to around 10% of benefit expenditure, and the contribution requirement has only recently been implemented. Transfers are largely funded out of the public subsidies to the private sector social insurance fund (Schwarzer and Querino 2002), and are described in official documentation as “semi-​contributory” (Mesquita, Jaccoud, and dos Santos 2010). Previdência Social Rural is part of the Regime Geral de Previdência Social, the private-​sector pension fund managed by the Ministério de Previdência Social.

24.3.2.  Benefício de Prestação Continuada and Conventional Social Assistance The new Constitution explicitly acknowledged the right of older people and people with disabilities to a minimum guaranteed income. This was consolidated in the 1993

Anti-Poverty Transfers and Poverty Reduction    517 law Lei Orgânica da Assistência Social (LOAS) defining the role of public agencies in the delivery of social assistance programs under a Sistema Unico de Assistência Social (SUAS). LOAS also led to the implementation of the Benefício de Prestação Continuada, a non-​contributory pension covering older and disabled people in extreme poverty. In practice, the Benefício de Prestação Continuada meant an extension of the Renda Mensal Vitalícia, a social assistance pension introduced in the 1970s by the military rulers.9 The Benefício de Prestação Continuada began to be implemented in 1996, and provided a transfer of one minimum wage (as established in the Constitution) to people aged 70 (subsequently reduced to 67 and later to 65 in 2003) and over, or with disabilities, living in households with per capita income at a quarter of the minimum wage or less. The regulations include a review of entitlements every two years. Figure 24.4 shows the number of transfers under the old age and disability components of Renda Mensal Vitalícia and Benefício de Prestação Continuada on an annual basis. The trend is a steady reduction in Renda Mensal Vitalícia transfers and a steady rise in Benefício de Prestação Continuada transfers. The jump in the number of old age beneficiaries of Benefício de Prestação Continuada in 2003 reflects the lowering of the minimum age of entitlement from 67 to 65 years. By 2010 there were 3.7 million beneficiaries, split evenly between old age and disability. The Benefício de Prestação Continuada reflects a more conventional approach to so­ cial assistance, based on a distinction between individuals who are able to work and earn an income, and those who are not in that position due to old age or disability. It is focused on extreme poverty, as it is a requisite for entitlement that household income

4,000,000 3,500,000

Number of transfers

3,000,000 2,500,000 2,000,000 1,500,000 1,000,000 500,000 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 BPC Old

RMV Old

BPC Dis

RMV Dis

Total

Figure 24.4.  Social assistance pension transfers Benefício de Prestação Continuada and Renda Mensal Vitalícia (for old age and disability). Data source: IPEA and MDS-​SAGE.

518   Armando Barrientos per capita is less than one-​quarter of the minimum wage, but lacks an explicit inactivity test. Benefício de Prestação Continuada and Previdência Social Rural have transfers set at one minimum wage. Providing individuals meet eligibility requirements, transfers are a right or entitlement acknowledged in the Constitution. This has the implication that entitlements are not subject to budgetary restrictions. Whereas Previdência Social Rural is restricted to rural residents, Benefício de Prestação Continuada is available nationwide. The two social pension schemes, combined with social insurance pensions, have greatly extended the coverage of social protection among households with older people in urban and rural areas. In 2008, 86.2% of the population aged 65 and over were re­ ceiving a pension transfer. Among Latin American countries, Brazil has one of the highest rates of pension coverage among older people (Gasparini et al. 2007). The social pension schemes ensure that the rate of pension coverage of older people significantly exceeds the coverage of social insurance among the working population, estimated at 65.9% in 2008.

24.3.3. Human Development, from Bolsa Escola to Bolsa Família Bolsa Família has a different origin. Its roots are in Bolsa Escola, a program introduced in a handful of municipalities in parallel in 1995 as a means of addressing the impact of crises on poor households. Its intellectual origins are to be found in guaranteed income proposals and a growing realization among policymakers and researchers that poverty is multidimensional and persistent. A strand of proponents advocated Bolsa Escola–​ type programs as an education intervention.10 The Constitution did not change the basic political structures in Brazil, but gave an enormous impetus to decentralization. In Brazil, municipalities are federal entities, with considerable room for experimentation. A  handful of municipalities began experimenting with guaranteed-​income schemes linked to children’s schooling and other interventions. Bolsa Escola emerged from municipal activism on poverty reduc­ tion. The experimental programs soon began to be replicated in other municipalities. In 1997 the federal government offered financial incentives to municipalities to support the adoption of Bolsa Escola.11 Bolsa Escola became a federal program in April 2001 under the responsibility of the Ministry of Education. Federal initiatives included PETI, first introduced in 1996. Initially located in municipalities with a high incidence of child labor in hazardous employment, the program provided direct transfers to households as well as remedial education in after-​school sessions. The program was especially successful, in part because of the supplementary education provided (Brazilian Court of Audit 2003).12 The apparent success of Bolsa Escola and PETI, and especially their core idea of pro­ viding direct transfers to households in poverty, stimulated similar policy initiatives in

Anti-Poverty Transfers and Poverty Reduction    519 other areas. The Ministry of Health introduced a Bolsa Alimentação in September 2001, aimed at expectant mothers and infants and with the objective of reducing malnutrition and infant mortality. In 2003, the Ministry of Mines and Energy began to implement a gas subsidy, Auxilio Gás, to compensate households in poverty for the phasing out of gas subsidies. The arrival in government of Lula in 2002 did not seem auspicious for this policy agenda at first; his campaign emphasized the fight against hunger, and the Fome Zero initiative popularized the three-​meals-​a-​day guarantee (Hall 2006). In office he created an Extraordinary Ministry for Zero Hunger, which floated a raft of new interventions, including a new family subsidy, the Cartão Alimentação, providing in-​kind and cash transfers. Very soon, however, experts, policymakers, and beneficiaries themselves advocated a change in policy. The fact that plans for the consolidation of transfers in cash had been discussed by his campaign team facilitated a change of course. He announced the implementation of Bolsa Família as a single program aiming to provide transfers to households in extreme poverty, integrating all the existing subsidy programs, to begin in 2003. A new Ministry for Social Development and Zero Hunger was established to manage Bolsa Família in 2004.13 Bolsa Família greatly expanded the coverage of Bolsa Escola and the other income transfer programs, as can be seen in Figure 24.5. The figure shows the number of households participating in Bolsa Família, which increased from 6.5 million in 2004 to 14 million in 2013.

16 14 12

Millions

10 8 6 4 2 0

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

Number of households

Figure 24.5.  Bolsa Família: number of participating households. Data source: IPEA data.

520   Armando Barrientos

24.3.4. Implementation Table 24.1 summarizes the main components of social assistance in Brazil, including the Previdência Social Rural. Benefício de Prestação Continuada and Previdência Social Rural are pure income transfers, with the consequence that their implementation and management is rela­ tively uncomplicated. With Previdência Social Rural, once entitlements are established, the main administrative burden is associated with the payment of the transfers. The means test required for entitlement to the Benefício de Prestação Continuada poses an Table 24.1 Social Assistance in Brazil: Main Programs Previdência Social Rural

Beneficio de Prestação Continuada

Bolsa Família

Target population

Rural informal workers Older and disabled people with low contribution in households in extreme capacity poverty

Households in extreme poverty and households in moderate poverty with children

Eligibility

Long-​term rural informal workers (> 15 years), subsistence agriculture, fishing, mining, limited land

Households with per capita income ≤ R$70 (US$40) and households with children with per capita income ≤ R$140 (US$82)

Monthly benefits

Old age pension is one One minimum wage minimum wage R$545 R$545 (US$318) (US$318)

Basic transfers = R$70 (US$40). Variable transfer = R$32 (US$18) per child (0–​15) up to five; and R$ 38 (US$22) for each youth (16–​17) up to two. Households with per capita income > R$70 and ≤ R$140 receive child transfers only. From 2012, the Beneficio de Superação da Extrema Pobreza provides a “top up” to households with incomes below R$70 after transfers.

Reach

7.8 million beneficiaries

3.7 million beneficiaries

14 million households

Budget as % GDP

1.4

0.7

0.6

Agencies responsible

Regime Geral de Previdência Social Ministério de Previdência Social

Ministério de Ministério de Desenvolvimento Social Desenvolvimento Social Ministério de Previdência Caixa Econômica Federal Social

Aged ≥ 65 in households with per capita income ≤ one-​quarter minimum wage

All values are for 2010 (US$ PPP 1 = RS$1.713). Source: Barrientos (2013b).

Anti-Poverty Transfers and Poverty Reduction    521 additional complication requiring management at the local level. On paper, entitlement is subject to review every three years, but in practice, review is less frequent. Bolsa Família is by comparison a complex program to implement and manage (Cotta and Paíva 2010; Paíva, Falcão, and Bartholo 2013). The selection of beneficiaries is the outcome of several processes (Barros et al. 2008). The federal government sets a target number of participants and allocates quotas to the different municipalities. The quotas are identified from poverty profiles based on household survey data. The municipalities use their knowledge of the distribution of households in poverty, sometimes based on local poverty mappings, to disseminate information on the program to potential beneficiaries. Potential beneficiary households are then registered in the CadUnico da­ tabase; this involves households filling out a standard questionnaire. Municipalities have an incentive to reach potential beneficiary households and for this purpose en­ gage with local institutions and organizations. The majority of registrations take place in schools, clinics, and so on. The information provided on registration is then sent to the federal agency and is used to perform a means test to determine eligibility and therefore selection. The federal agency notifies municipalities of selected households.14 Once selected for participation in the program, households are provided with a monthly transfer. The transfer amounts depend on the income of the households and their composition (see Table 24.1).15 In 2010, households with per capita incomes up to R$70 were entitled to a fixed household transfer of R$70 plus an additional amount of R$32 for each child 0–​15 years of age, up to a maximum of five, and an additional amount of R$38 for each youth aged 16–​17, up to a maximum of two. The maximum benefit amount for households in extreme poverty was therefore R$242. Households with per capita incomes between R$71 and R$140 were entitled to the child and youth transfers, but not to household transfer, and their maximum transfer amount was R$172.16 As a comparison, the minimum wage in April 2011 was R$545.17 Since 2012, the Beneficio de Superação da Extrema Pobreza has provided a “top up” to households with incomes below R$70 after transfers. This ensures that no participant is in (income) ex­ treme poverty. Bolsa Família attaches several conditions to the continuation of transfer payments, covering schooling, health, and social services. Participating households commit to ensuring that children up to 15 years of age have an 85% record of school attendance, while youths aged 16–​17 are required to complete 75% of school attendance. Health conditions include a full immunization schedule, child development monitoring for children under seven years of age, prenatal monitoring for pregnant women, and monitoring for nursing mothers. Attendance of remedial education for children and youths at risk of child labor, which was a requirement under the PETI program, is retained in Bolsa Família. The conditions have a diagnostic role. Failure to comply with conditions should lead to consideration of whether additional interventions are needed. Non-​compliance can lead to suspension of payments and exclusion from the program.18 Compliance following suspension of a single payment can lead to restitution of payment with arrears. In a comparative context, conditions in Bolsa Família are described as “soft conditions” (Cecchini and Madariaga 2011).

522   Armando Barrientos In March 2011, following the election of Dilma Rousseff to the presidency, a new Plano Brasil Sem Miséria was announced. The aim of the Plan is to articulate public policy directed at eradicating extreme poverty.19 It is a Plan and not a Program, like Bolsa Família, and in many respects it constitutes another stage in the development of social assistance in Brazil (Paíva, Falcão, and Bartholo 2013). Drawing a parallel be­ tween Fome Zero and Plano Brasil Sem Miséria—​as an aspirational strategy involving multiple programs, policies, and interventions—​is superficially attractive but hopefully wide of the mark. Like Fome Zero, the Plano Brasil Sem Miséria involves around 120 interventions and programs distributed in 22 ministries and multiple agencies (Paes-​ Souza 2013). It shares the ambition of eradicating extreme poverty, as well as a multidi­ mensional perspective on poverty. But in its design and implementation it appears to have assimilated the lessons from the evolution of anti-​poverty policy in the country. In sum, the evolution of social assistance in Brazil suggests a shift in focus over time from (1) a complement to social insurance aimed at incorporating excluded sectors on a preferential basis; to (2) conventional social assistance directed at vulnerable groups; to (3) human development–​focused income transfers. Plano Brasil Sem Miséria gathers to­ gether the main lessons from this evolution: the eradication of extreme poverty requires a coordinated effort directed at the population in extreme poverty, aimed at improving income/​consumption and improving access to basic services and facilitating productive employment.

24.4.  The Economics of Social Assistance This section assesses the contribution of social assistance to the reduction of poverty in the context of economic and financial perspectives on the role of transfers in growth-​ redistribution trade-​offs. The discussion on the emergence and evolution of social assistance in Brazil earlier in the chapter identified three main strategies: the universalization of social insurance (Previdência Social Rural); the protection of vulnerable groups (Benefício de Prestação Continuada); and human development–​focused guaranteed minimum income (Bolsa Família and other transfer programs) (Barrientos 2013b). They have led to two main types of policy instrument:  pure income transfers and human development income transfers. Pure income transfers are based on the assumption that poverty is primarily asso­ ciated with insufficient income. In line with “canonical” social assistance, subsidies support citizens who are unable to earn an income (Atkinson 1995). Old age transfers or so-​called non-​contributory pensions20 are pure income transfers. Here, the source of insufficient income is assumed to be age-​related declines in labor productivity. Supplementing the income of households in poverty works to reduce their consumption

Anti-Poverty Transfers and Poverty Reduction    523 deficits. Figure 24.6 provides a basic representation of a guaranteed minimum income at the poverty line z. In standard pension schemes, transfers replace income y* at retire­ ment. They are intended to generate withdrawal from the labor market. Social assistance old age transfers do not require withdrawal from the labor market explicitly, but they do so in practice. Previdência Social Rural transfers are perceived as retirement income by recipients. Benefício de Prestação Continuada transfers are dependent on beneficiaries meeting the condition that per capita household income does not exceed a quarter of the minimum wage. Income supplements to older people are not expected to generate large and adverse incentive effects on work or saving (Barrientos 2013a; Moffitt 2002). Previdência Social Rural sought to engage rural informal workers in the contributory system through subsidized entitlements expected to be replaced gradually by workers’ savings. Income supplements focused on older workers and took account of the spe­ cific characteristics of their employment. Benefício de Prestação Continuada extended provision to older people and people with disabilities with income below the extreme poverty line. Old age anti-​poverty transfers raise several concerns. First, their poverty reduction effectiveness will depend in practice on the extent to which transfers are distributed within beneficiary households. Secondary effects on household members can be neg­ ative or positive. Pension receipt could in principle lead to other household members lengthening periods of unemployment, or children and youths remaining in school. Alternatively, pension receipt could facilitate labor migration by other household members. Second, demographic and economic factors could in the future significantly increase demand for social pensions. In Brazil, population aging and a deterioration

Income after transfers

z

y*

Income before transfers

Figure 24.6.  Pure income transfers, social pensions. Note: A transfer set at z, the minimum wage in Brazil’s case, ensures that recipients have a guaranteed minimum income at z. Benefício de Prestação Continuada entitlements are dependent on the per capita household income of recipients not exceeding a quarter of the minimum wage. Above this qualifying level, transfers are suspended. As incomes in excess of y* carry an implicit 100% tax, this is consistent with weakened work incentives. Source: Author.

524   Armando Barrientos of formal-​sector pension scheme coverage of older people will require larger social pension budgets to meet the demand for old age transfers. Bolsa Família transfers, on the other hand, combine supplementation of household consumption with support for household investment in human development of chil­ dren in households in poverty (Barrientos and DeJong 2006). The objective is to redis­ tribute consumption as well as productive capacity among households in poverty. The structure of the transfers, taking account of the number of children and their ages, and attached conditions aim to ensure that participant households invest in the schooling and health of children. Figure 24.7 provides a representation of the expected effect of Bolsa Família transfers on the allocation of household disposable income to schooling. The household budget pre-​transfers is captured by the line a–​b. The transfer shifts the household budget line to c–​d. For a household at point e before the transfer, points f–​g on the improved budget line are feasible. A condition that children must attend school is represented in expenditure terms by point h as the minimum schooling expenditure. To meet the condition, the participant household will move from point e to point g on the new budget line. Assessing poverty reduction effectiveness in human development conditional in­ come transfer programs requires that we pay attention to household income and con­ sumption, and potential secondary effects in labor supply, as in pure income transfers;

Other goods and services

c

a f

g

e

h

b

d

Schooling

Figure 24.7.  The role of conditions in Bolsa Família. Note: Household choice between schooling and other goods (child labor): Initially a–​b is the budget constraint and a representative household is at point e. Social assistance transfers move the budget constraint to c–​d. The household can now “consume” more of other goods keeping schooling constant f or, alternatively, more schooling at g keeping other goods constant, or any point between f and g. A condition of minimum schooling at h ensures the household chooses g to meet the condition. Source: Author.

Anti-Poverty Transfers and Poverty Reduction    525 but in addition, effectiveness requires improvements in household productive capacity, especially children’s human development. Assessing the evolution of social assistance in Brazil, the view that its role and scope are increasingly being defined around human development objectives has some support (Barrientos 2013c; Barrientos and Telias Simunovic 2015). The low visibility of Previdência Social Rural as a social assistance instrument has drawn attention away from the project of universalizing social insurance. Economic change has undermined this project, including the liberalization of the labor market, persistent deficits in so­ cial insurance funds, and export-​led growth strategies (Guimaraes 2007). Social assis­ tance is no longer seen as a competitor to social insurance but increasingly, instead, as a complement. The expansion of social assistance in Brazil has been made possible in part by a sig­ nificant rise in the tax/​gross domestic product (GDP) ratio, which is now at a level com­ parable to that of some Organisation for Economic Co-​operation and Development (OECD) countries (Barrientos 2014). Studies on the distributional effects of social pro­ tection transfers demonstrate that social assistance transfers are strongly progressive, while social insurance transfers are essentially regressive. A full picture of the distribu­ tional effects of social assistance transfers needs to take account of the contribution of the target population of social assistance to government revenues. In theory, if the value of the transfers from social assistance is equal to the tax collected from households in poverty, the tax-​transfer scheme will not have redistributed income toward households in poverty. The handful of tax-​transfer studies for Brazil suggest that, overall, the distri­ butional effects of taxes and transfers together is muted (Immervoll et al. 2006; Lustig, Pessino, and Scott 2013; Silveira 2008). Given the progressivity of social assistance, its relatively low level of public subsidies, and the contribution of low-​income groups to tax revenues, there should be few concerns with the sustainability of current financial arrangements.

24.5.  Poverty Reduction Effectiveness This section reviews the main findings from the literature on the outcomes of the different components of social assistance in Brazil. The focus is on program outcomes on poverty and well-​being and on labor supply.

24.5.1.  Previdência Social Rural Using a dedicated rural household survey, Delgado and Cardoso (2000) track the effects of the Previdência Social Rural on the well-​being and livelihoods of rural elders and their households. They establish that the incorporation of rural workers and their access to social insurance old age and disability pensions reached the target population and

526   Armando Barrientos strengthened economic activity in rural areas. The reform is framed as a step toward the universalization of the right to social insurance, with a focus on poverty reduction (Schwarzer and Querino 2002). Barbosa (2010) compared rural head-​count poverty in Brazil in 2008 with and without Previdência Social Rural transfers using PNAD data.21 Excluding the transfers from household income resulted in a 68.1% head-​count poverty rate and a 41.3% indi­ gence head-​count rate, compared to the actual poverty rates estimated at 53.5% in 2008, while the rural indigence head-​count rate was estimated at 26.1%. He concludes that Previdência Social Rural transfers lifted 4 million rural Brazilians out of extreme pov­ erty. The impact on poverty gaps is likely to be even more significant. The implementation of Previdência Social Rural led to a large increase in the number of transfers in 1992–​1993, approximating the conditions of an exogenous policy change for the relevant population. Several studies have exploited this shift in transfer policy using cross-​section household survey data from before and after 1992–​1993 in order to throw light on the impact of the change on outcomes. Carvalho (2008a) finds that pension receipt was associated with a large drop in participation and hours of work among newly qualifying rural pensioners, a fall of around 8% in participation and total hours of work by 22.5 compared to urban workers. He also investigates the income effect of the pension on co-​resident children’s work and school enrollments.22 He finds a sig­ nificant increase in school enrollments, especially among girls, of around 20% of the rate of enrollment gap (Carvalho 2008b). This is consistent with a drop in child labor. The sex of the pensioner matters; girls benefit from co-​residing with a female pensioner, while boys benefit from co-​residing with a male pensioner. Ponczek (2011) focuses on the role of household bargaining in decisions on schooling. He finds, in contrast to Carvalho, that the presence of an eligible male pensioner has significant effects on literacy and schooling. He finds no significant effects in households with eligible female pensioners.

24.5.2.  Benefício de Prestação Continuada Soares et  al. (2006) use 2004 PNAD data to study the effects of targeted transfer programs in Brazil on poverty and inequality. Their analysis also includes non-​ contributory pension transfers and minimum contributory pensions, as they approxi­ mate Previdência Social Rural transfers. The study compares poverty rates for incomes with and without these transfers. They find that Benefício de Prestação Continuada and Bolsa Família together account for a 2 percentage point reduction in the poverty head count, while minimum pensions are responsible for a 5 percentage point reduction. The strong effects of the Benefício de Prestação Continuada are explained by the fact that minimum pensions are significantly more generous than Bolsa Família transfers, plus the fact that the latter was just beginning its process of expansion. Kassouf et  al. (2011) report on an impact evaluation of the Benefício de Prestação Continuada based on PNAD household survey data for the years 2004–​2006. The pro­ gram had been in place since 1996, but they exploit the reduction in the age of entitlement

Anti-Poverty Transfers and Poverty Reduction    527 from 67 to 64 in 2003 to identify the impact of transfers on this population. They find no changes in household composition following entitlement. Regarding labor supply effects, they find a reduction in the labor force participation of direct beneficiaries of around 2%–​3%, no significant effects on young co-​residents aged 19–​29, but small neg­ ative labor force participation effects on adult co-​residents 30–​49. The study identifies a reduction in child labor, but no significant effects on school attendance. The survey evidence confirms that non-​contributory pension benefits are shared within households, thus having an impact on a larger population than simply the di­ rect beneficiaries (Barrientos 2008; Barrientos and Mase 2012). In addition, non-​ contributory pensions encourage livelihoods and productive investment, especially in rural areas (Delgado and Cardoso 2000).

24.5.3.  Bolsa Família Bolsa Família was designed as an extension of existing human development condi­ tional cash transfer programs. Evaluation surveys were only collected in 2005 (AIBF1) and 2009 (AIBF2), and the results from the analysis of the 2009 data were placed in the public domain only in 2012 (de Brauw et al. 2012). The impact evaluation compared outcomes across participants and non-​participants.23 The main findings include the following: improvements in children’s weight-​for-​height and body mass, as well as in some immunizations; improvements in school attendance by 4 percentage points (C2 and C3), larger for girls and for the Northeast; improvements in progression and a re­ duction in grade repetition; children’s entry into the labor market reduced by a year; increased prenatal visits by participant expectant mothers (1.6 additional visits); improved influence of mothers in decisions over household budget and contraception. The analysis found no significant effects on labor supply, but it did note a reduction in formal sector hours by males and an increase in hours worked in the informal sector. The vast majority of studies examining the impact of Bolsa Família rely on cross-​ section PNAD data. The PNAD survey data lacks direct identification of Bolsa Família beneficiaries, except for supplements in 2004 and 2006. For all other waves of the data, the identification of Bolsa Família participants is done through the unique mon­ etary values reported under a question on residual income (Foguel and Barros 2010; F. V. Soares et al. 2006).24 Data from the income and expenditure survey Pesquisa de Orçamentos Familiares (POF) does include a question permitting direct identification of Bolsa Família participants, but it is collected only every five years. Surprisingly, perhaps, there are few studies assessing the impact of Bolsa Família on poverty. This is in part explained by the difficulties involved in identifying beneficiaries of social programs in Brazil’s household survey data, as well as the fact that the improved growth of the Brazilian economy and the expansion of a range of social policies make it harder to isolate the impact of Bolsa Família.25 Soares et al. (2010) estimate poverty and extreme poverty head counts with and without Bolsa Família transfers as a counterfac­ tual. This approach does not account for behavioral responses to the transfers. Their

528   Armando Barrientos analysis of changes in poverty in the decade 1999–​2009 attributes to Bolsa Família one-​ sixth of the reduction in poverty (2 percentage points of a reduction from 26% to 14%) and around one-​third of the reduction in extreme poverty (1.6 percentage points of a reduction from 9.9% to 4.8%). Some studies focus on the impact of Bolsa Família on schooling and health. Magalhães and Lima (2013) summarize the findings from impact evaluation studies on the impact of the program on basic education, while Craveiro and Ximenes (2013) do the same for health. Some of these impacts were explored in the impact evalua­ tion discussed at the start of this section. In addition, Glewwe and Kassouf (2012) used school census panel data for 1998–​2005 to examine the effects from the expan­ sion of Bolsa Escola in 2001 as a natural experiment. They find that Bolsa Escola raised enrolments by 5.5% in Grades 1–​4 and by 6.5% in Grades 5–​8. They also find a reduction in dropout rates and improvements in grade progression among Bolsa Escola/​Bolsa Família participants. They simulate the longer term effects of the program on the pro­ ductive capacity of participant children and suggest that an 11% rise in labor earnings associated with a predicted additional 1.5 years of schooling is greater than the costs of the program. In this simulation, Bolsa Escola and Bolsa Família pay for themselves in terms of improved productivity. One study focuses on the combined impact of the Family Health program and Bolsa Família on child morbidity and mortality at the mu­ nicipal level, using data collected by the Health Ministry (Rosella et al. 2013). The study finds strong and positive impacts from both programs, especially in reducing child mortality and morbidity. Finally, the impact of Bolsa Família on the labor supply of participants has been studied extensively. Oliveira and Soares (2012) summarize this literature. For the population as a whole, studies find a reduction in child labor consistent with a rise in school attendance (Ferro, Kassouf, and Levison 2010). Few studies find any significant effects of the program on adult labor at the extensive margin (Foguel and Barros 2010). Teixeira (2010) finds a small positive increase in the labor force participation of women, but no effect for men. As regards the intensive margin, studies find small but signifi­ cant effects but, depending on the data and econometric model employed, the findings are mixed with positive estimated effects as likely as negative ones. Whereas labor supply effects are likely to be marginal for the program population, disaggregating the effects by gender and region often lead to clearer and stronger findings. Ribas and Soares (2011) find stronger effects in urban areas, especially metropolitan areas, in­ cluding a reduction of labor force participation by women of around 4.4 percentage points and an increase in hours of work in informal employment by males. The latter finding is likely to reflect the influence of the income test at the margins of eligibility (Firpo et al. 2013). The important fact to keep in mind is that labor force participation rates among adults in households eligible for, or participating in, Bolsa Família are high, at least as high as for the population as a whole (Castro et al. 2010). Taking on board the design and reach of Bolsa Família and the conditions of Brazil’s labor market suggests that concerns

Anti-Poverty Transfers and Poverty Reduction    529 with labor market incentives are less significant than in high-​income countries. The economics literature on labor supply (dis)incentives is dominated by the income-​ maintenance perspective of “classical” social assistance, which assumes beneficiaries as either unemployed or inactive in a full employment (formal) economy (Moffitt 2002). This approach requires adaptation to labor market conditions in Brazil. Nevertheless, the fact that currently adverse labor supply effects from social assistance programs are marginal does not imply that future expansion of benefit coverage or levels might not have larger labor supply effects. In sum, a review of the outcomes of social assistance programs in Brazil suggests that they have had a strong measurable impact on the reduction of poverty and in­ equality. They also have some impact on strengthening the productive capacity of households in poverty, especially children. Labor supply effects are significant for older people (social pensions) and children (Bolsa Família), but muted for the popu­ lation of working age.

24.6. Conclusions Brazil showed a large and sustained reduction in poverty and inequality in the first decade of the new century. Rapid economic growth, allied to social and labor market policy, were responsible for these trends. The chapter has examined the contribution of emergent social assistance institutions, their evolution and outcomes. The 1988 Constitution is the marker for the rapid expansion of social assistance programs and policies since, but this expansion also shows significant innovation and diversity. It is possible to identify three main strategies for inclusion: (1) the incorpo­ ration of informal workers as a special regime under social insurance, the main objec­ tive of Previdência Social Rural; (2) conventional social assistance transfers to older and disabled people in poverty, the focus of Benefício de Prestação Continuada; (3) human development income transfer programs focused on extreme poverty, as demonstrated by Bolsa Escola and Bolsa Família. The evolution of these strategies, and the respective institutions, has over time largely maintained their focus and scope. From an economic perspective, social pensions or old age transfers—​Previdência Social Rural and Benefício de Prestação Continuada—​are pure income transfers. Their main objective is to provide a guaranteed minimum income to older people and people with disabilities deemed unable to work. Transfers are equivalent to one minimum wage in order to ensure minimum living standards for the direct recipients. The focus on these groups mitigates potential work and saving disincentives. Demographic factors, population aging, and household size all influence the poverty-​reduction effectiveness of these transfers. Bolsa Família transfers also aim to raise household consumption, but in addition they facilitate household investment in productive capacity, especially children’s human capital. Whereas income effects are dominant in the context of social

530   Armando Barrientos pensions, substitution effects are important in the context of human development in­ come transfers, through the effects of conditions on the schooling of children and the work incentives, positive and negative, faced by their parents. Anti-​poverty programs re­ duce labor force participation among children and older adults—​especially in the case of old age transfers—​but have marginal labor supply effects among adults of working age. Social assistance programs have been effective in reducing poverty, and, perhaps against all expectations, inequality. Studies attribute one-​fifth and one-​third of the re­ duction in poverty and extreme poverty, respectively, to social assistance transfers. They have also contributed to human capital, in helping to universalize basic education and primary health care. In addition, they have had some success in achieving specific outcomes relating to child development: immunization, nutrition, reduction in mor­ tality and child labor, and increased school enrollments and attendance. The growth slowdown experienced by the Brazilian economy in the aftermath of the 2008 global financial crisis has fueled continued discussion of the design, reach, and fi­ nancing of social assistance programs in Brazil. Public discussion of budget restrictions and potential reforms will be intense, but there is no question that these new institutions are here to stay.

Notes 1. During the first decade of the new century, the share of the Brazilian population in ex­ treme poverty declined from 22% to 11%, while the Gini coefficient of per capita household income fell by 10%. 2. Social assistance describes tax-​financed public programs and policies addressing poverty and vulnerability. Social insurance describes contributory schemes addressing life-​course and work-​related contingencies. Together, social assistance and social insurance are the main components of social protection (Barrientos 2013a). 3. In 2016 the Ministry of Social Development was merged with the Ministry of Agrarian Development. 4. Medeiros et al. (2008) examine Bolsa Família and Benefício de Prestação Continuada but not Prêvidencia Social Rural. 5. Reductions in poverty and inequality can be observed in most countries in Latin America in this period, but comparison with the BRICS countries highlights how remarkable these trends are. China, India, Russia, and South Africa managed poverty reduction but suffered from a rise in inequality (Cevik and Correa-​Caro 2015). 6. Studies on inequality also make use of the income and expenditure survey Pesquisa de Orçamentos Familiares (POF). 7. The contribution of social assistance to poverty reduction is be discussed in more detail in section 24.5. 8. Initially, entitlement was restricted to workers in agriculture, an occupational restriction, but changes in the regulations in 2008 added a rural residence requirement. 9. It provided one-​half of the minimum wage to people aged 70 and over with at least 12 contributions to social insurance. 10. Senator Cristovam Buarque in particular.

Anti-Poverty Transfers and Poverty Reduction    531 11. In 1998 60 municipalities had adopted the program. Their number mushroomed to 1,115 by 2000. 12. Both federal Bolsa Escola and PETI played a role in the social development strategy of President Cardoso and his social policy advisor Vilmar Faria. The latter describes the role of direct income transfer programs, including the BPC, PRS, and Bolsa Escola, in the overall development strategy (Faría 2002). 13. Until 2004, social assistance was the responsibility of the Ministry of Labor and Social Assistance. 14. Bolsa Família’s target was 12.7  million households in 2010. Is this the “right” number? S.  Soares, Ribas, and Soares (2009) compare eligible and recipient households using PNAD data for 2006. They find a large proportion of eligible households not receiving transfers:  6.6  million. They argue that volatility in incomes in Brazil means that the government’s calculation that at a point in time in 2006 there were 11  million eligible households missed out on the fact that over a three-​month period, say, the number of eli­ gible households is much larger. They estimate that to reach all households with per capita income below one-​quarter of the minimum wage at some point during 2006, the govern­ ment should have included 15 million households in Bolsa Família. 15. The transfer values in the text are those applying in April 2011 in reais. The exchange rate with the US$ has been volatile recently; 2011 implied PPP exchange rate with international US$ is 1.713. 16. There is also a special variable transfer amount, calculated on a case-​by-​case basis, to households that were recipients of the component transfers of Bolsa Família and are af­ fected financially by the migration to the new program. 17. Bolsa Família transfers have been upgraded by government on five occasions, but there is no legal requirement for benefits to be indexed. 18. To 2010, just over 2 million households received a warning; 0.7 million had a payment blocked; 0.3 million had a payment suspended; and 0.09 million were excluded (Curralero et al. 2010). 19. Information from the 2010 census estimates the population living in extreme poverty, R$70 per capita or less, at 15 million, or 7.8 percent of the population. 20. Tax as a proportion of household income in Brazil is constant for all income deciles. If any­ thing, the poorest decile faces a slightly higher tax incidence (Silveira 2008). 21. The poverty line employed was one-​half of the minimum wage, and the indigence line one-​quarter of the minimum wage. 22. Cortez Reis and Camargo (2007), using PNAD data from 2003, find that 15–​21-​year-​ olds co-​resident with a pensioner (social insurance and social assistance pensions) are more likely to be at school and less likely to be in work, than those in households without pension income. 23. The impact evaluation report identifies three main comparison groups:  comparison 1 (C1) compares new participants in 2009 versus nonparticipants in 2009; comparison 2 (C2) compares all new participants regardless of whether registered in the Cadastro Único in 2005 against nonparticipants in 2009 who had either registered in the CU or received benefits in 2005 but no longer in 2009; comparison 3 (C3) compares all participants in 2009 against nonparticipants in 2009 who had either registered in the Cadastro Único or received benefits in 2005 but no longer in 2009 (same as in C2). 24. For a detailed discussion of all data issues, see Soares et al. (2006). 25. Researchers have focused on studying the incidence of social programs, as opposed to their impact.

532   Armando Barrientos

References Atkinson, A. B. 1995. “On Targeting and Family Benefits.” In Incomes and the Welfare State: Essays in Britain and Europe, edited by A. B. Atkinson, 223–​261. Cambridge: Cambridge University Press. Barbosa, Edvaldo Duarte. 2010. “A Previdência Social Rural Brasileira.” Informe de Previdência Social—​Ministério Da Previdência Social 22 (8): 1–​10. Barrientos, Armando. 2008. “Cash Transfers for Older People Reduce Poverty and Inequality.” In Institutional Pathways to Equity: Addressing Inequality Traps, edited by Anthony J. Bebbington, Anis A. Dani, Arjan De Haan, and Michael Walton, 169–​192. Washington, DC: World Bank. Barrientos, Armando. 2013a. Social Assistance in Developing Countries. Cambridge: Cambridge University Press. Barrientos, Armando. 2013b. “The Rise of Social Assistance in Brazil.” Development and Change 44 (4): 887–​910. Barrientos, Armando. 2013c. “Transferências de renda para o desenvolvimento humano no longo prazo.” In Programa Bolsa Família: Uma década de inclusão e cidadania, edited by Tereza Campello and Marcelo Cortes Neri, 417–​434. Brasília: IPEA. Barrientos, Armando. 2014. “On the Distributional Implications of Social Protection Reforms in Latin America.” In Falling Inequality in Latin America: Policy Changes and Lessons, edited by Giovanni Andrea Cornia, 240–​260. Oxford: Oxford University Press. Barrientos, Armando, and Jocelyn DeJong. 2006. “Reducing Child Poverty with Cash Transfers: A Sure Thing?” Development Policy Review 24 (5): 537–​552. Barrientos, Armando, and Julia Mase. 2012. “Poverty Transitions among Older Households in South Africa and Brazil.” European Journal of Development Research 24 (4): 570–​588. Barrientos, Armando, and Amanda Telias Simunovic. 2015. “Child Poverty Reduction in Brazil: Reversing Bismarck?” In Theoretical and Empirical Insights in to Child and Family Poverty: Cross National Perspectives, edited by Elizabeth Fernandez, Cinzia Canali, Tiziano Vecchiato, and Anat Zeira, 159–​179. Sydney: Springer. Barros, Ricardo Paes de, Mirela de Carvalho, Samuel Franco, and Rosane Mendonça. 2008. “A importância das cotas para a focalização do programa Bolsa Família.” Texto para Discussão TD 238. Universidade Federal Fluminense. Brauw, Alan de, Daniel O. Gilligan, John Hoddinott, and Shalini Roy. 2012. “The Impact of Bolsa Família on Child, Maternal, and Household Welfare.” Report. International Food Policy Research Institute. Brazilian Court of Audit. 2003. “TCU Evaluation of the Child Labor Eradication Programme.” SEPROG. Carvalho, Irineu Evangelista de. 2008a. “Old-​Age Benefits and the Labour Supply of Rural Elderly in Brazil.” Journal of Development Economics 86: 129–​146. Carvalho, Irineu Evangelista de. 2008b. “Household Income as a Determinant of Child Labour and School Enrollment in Brazil: Evidence from a Social Security Reform.” IMF Working Paper No. WP/​08/​241. Washington, DC: International Monetary Fund. Castro, Jorge Abrahão de, Natália Sátyro, José Aparecido Ribeiro, and Fabio V. Soares. 2010. “Desafios para a inclusão produtiva das famílias vulneráveis: Uma Análise exploratória.” Texto para Discussão 1486. IPEA. Cecchini, Simone, and Aldo Madariaga. 2011. “La trayectoria de los Programas de Transferencia Con Corresponsabilidad (PCT) en América Latina y El Caribe.” Report. CEPAL.

Anti-Poverty Transfers and Poverty Reduction    533 Cevik, Serhan, and Carolina Correa-​Caro. 2015. “Growing (Un)equal: Fiscal Policy and Incoem Inequality in China and the BRICS+.” IMF Working Paper No. WP/​15/​68. Washington, DC: International Monetary Fund. Cortez Reis, Maurício, and José Márcio Camargo. 2007. “Rendimientos domiciliáres com aposentadorias e pensôes e as decisôes dos jovens quanto à educacão e a participação na forca de trabalho.” Pesquisa e Planejamento Economico 37 (2): 221–​246. Cotta, Tereza Cristina, and Luis Henrique Paíva. 2010. “O Programa Bolsa Família e a proteção social no Brasil.” In Bolsa Família 2003–​2010: Avanços e desafios, edited by Jorge Abrahão de Castro and Lucia Modesto, Vol. 1, 57–​99. Brasilia: IPEA. Craveiro, Clélia Brandão Alvarenga, and Daniel de Aquino Ximenes. 2013. “Dez anos do Programa Bolsa Família: Desafios e perspectivas para a universalização da educação básica no Brasil.” In Programa Bolsa Família: Uma década de inclusão e cidadania, edited by Tereza Campello and Marcelo Cortes Neri, 109–​124. Brasilia: IPEA. Curralero, Cláudia Baddini, Ana Amélia da Silva, Daniel de Aquino Ximenes, Beatriz Pinto de Almeida Vasconcellos, Kelva Karina Nogueira, Kathleen Sousa Oliveira, Ana Carolina Feldenheimer, and Eduardo Augusto Fernandes Nilson. 2010. “As condicionalidades do Programa Bolsa Família.” In Bolsa Família 2003–​2010: Avanços e desafios, edited by Jorge Abrahão de Castro and Lucia Modesto, Vol. 1, 151–​178. Brasilia: IPEA. Delgado, Guilherme Costa, and José Celso Cardoso. 2000. A universalização de direitos sociais no Brazil: A Prêvidencia Rural nos anos 90. Brasilia: IPEA. Faría, Vilmar. 2002. “Institutional Reform and Government Coordination in Brazil’s Social Protection Policy.” CEPAL Review 77: 7–​24. Ferro, Andrea, Ana Lúcia Kassouf, and Deborah Levison. 2010. “The Impact of Conditional Cash Trasfer Programs on Household Work Decisions in Brazil.” Research in Labor Economics 31: 193–​218. Firpo, Sergio, Renan Pieri, Euclides Pedroso, and André Portela Fernandes. 2013. “Evidence of Eligibility Manipulation for Conditional Cash Transfer Programs.” Working Paper C-​Micro 26. FGV São Paulo School of Economics. Foguel, Miguel Nathan, and Ricardo Paes de Barros. 2010. “The Effects of Conditional Cash Transfer Programmes on Adult Labour Supply: An Empirical Analysis Using a Time-​Series-​ Cross-​Section Sample of Brazilian Municipalities.” Estudos Economicos 40 (2): 259–​293. Gasparini, Leonardo, Javier Alejo, Francisco Haimovich, Sergio Olivieri, and Leopoldo Tornarolli. 2007. “Poverty and the Elderly in Latin America and the Caribbean.” Discussion Paper No. 55. CEDLAS. Guimaraes, Ricardo J. R. 2007. “Searching for the Vulnerable: A Review of the Concepts and Assessments of Vulnerability Related to Poverty.” European Journal of Development Research 19 (2): 234–​250. Hall, Anthony. 2006. “From Fome Zero to Bolsa Familia:  Social Policies and Poverty Alleviation under Lula.” Journal of Latin American Studies 38 (4): 689–​709. Immervoll, Herwig, Horacio Levy, José Ricardo Nogueira, Cathal O’Donoghue, and Rozane Bezerra de Siqueira. 2006. “The Impact of Brazil’s Tax-​Benefit System on Inequality and Poverty.” Discussion Paper No. 2114. IZA. Jaccoud, Luciana, Patricia Dario El-​ Moor Hadjab, and Juliana Rochet Chaibub. 2009. “Assistência Social E Securança Alimentar:  Entre Novas Trajetórias, Vehlas Agendas E Recentes Desafíos (1988–​2008).” In Políticas sociais: Acompanhamento e análise 17, edited by Diretoría de Etudos e Políticas Sociais, 175–​250. Brasilia: IPEA.

534   Armando Barrientos Lustig, Nora, Carola Pessino, and John Scott. 2013. “The Impact of Taxes and Social Spending on Inequality and Poverty in Argentina, Bolivia, Brazil, Mexico, Peru and Uruguay.” CEQ Working Paper No. 13. Tulane University. Magalhães, Helvécio Miranda, Jr., Patricia Constante Jaime, and Ana Maria Cavalcante de Lima. 2013. “O papel do setor saúde no Programa Bolsa Família:  Histórico, resultados e desafíos para o sistema único de saúde.” In Programa Bolsa Família: Uma década de inclusão e cidadania, edited by Tereza Campello and Marcelo Cortes Neri, 93–​108. Brasilia: IPEA. Medeiros, M., T. Britto, and F. Soares. 2008. Targeted Cash Transfer Programmes in Brazil. Brasília: International Poverty Centre. Mesquita, Ana Cleusa Serra, Luciana Jaccoud, and Maria Paula Gomes dos Santos. 2010. “Garantia de renda na política social brasileira: Entre a proteçao aos riscos sociais e o alivio à pobreza.” Mimeo. IPEA. Moffitt, Robert A. 2002. “Welfare Programs and Labour Supply.” In Handbook of Public Economics, edited by Alan J. Auerbach and Martin Feldstein, Vol. 4, 2394–​ 2430. London: Elsevier Science B. V. Oliveira, Luis Felipe Batista de, and Sergei S. D. Soares. 2012. “O que se sabe sobre os efeitos das transferencias de renda sobre a oferta de trabalho.” Texto para Discussão 738. IPEA. Paes-​Souza, Rômulo. 2013. “Plano Brasil sem miséria: Incremento e cudanca na política de protecao e promocao social no Brasil.” Working Paper No. 113. International Policy Centre for Inclusive Growth. Paíva, Luis Henrique, Tiago Falcão, and Leticia Bartholo. 2013. “Do Bolsa Família o Brasil sem miséria: Um resumo de percurso brasileiro recent na busca da superação da pobreza ex­ trema.” In Programa Bolsa Família: Uma década de inclusão e cidadania, edited by Tereza Campello and Marcelo Cortes Neri, 25–​46. Brasilia: IPEA. Rosella, Davide, Rosana Aquino, Carlos A. T. Santos, Romulo Paes-​Sousa, and Mauricio L. Barreto. 2013. “Effect of a Conditional Cash Transfer Programme on Childhood Mortality: A Nationwide Analysis of Brazilian Municipalities.” The Lancet 382: 57–​64. Schwarzer, Helmut, and Ana Carolina Querino. 2002. “Non-​ Contributory Pensions in Brazil:  The Impact on Poverty Reduction.” ESS Paper No. 11. Social Security Policy and Development Branch. ILO. Silveira, Fernando Gaiger. 2008. “Tributação, Prêvidencia E Assistência Social.” Mimeo. Tesouro Nacional. Soares, Fabio Veras, Sergei Soares, Marcelo Medeiros, and Rafael Guerreiro Osório. 2006. “Cash Transfer Programmes in Brazil: Impacts on Inequality and Poverty.” International Poverty Centre. Soares, Sergei, Rafael Perez Ribas, and Fábio Veras Soares. 2009. “Focalização e cobertura do Programa Bolsa-​Família:  Qual o significado dos 11 milhões de famílias?” Texto para Discussão 1396. IPEA. Soares, Sergei, Pedro Herculano Ferreira de Souza, Rafael Guerreiro Osório, and Fernando Gaiger Silveira. 2010. “Os impactos do benefício do Programa Bolsa Família sobre a desigualdade e a pobreza.” In Bolsa Família 2003–​2010: Avanços e desafios, edited by Jorge Abrahão de Castro and Lucia Modesto, Vol. 2, 25–​52. Brasilia: IPEA. Teixeira, Clarissa Gondim. 2010. “A Heterogeneity Analysis of the Bolsa Familia Programme Effect on Men and Women’s Work Supply.” Working Paper No. 61. International Policy Centre for Inclusive Growth.

Chapter 25

Sou th-​S ou t h C o operati on for So cial Devel opme nt Brazil and Africa Examined Anthony Hall

25.1.  Brazil and the Growth of South-​S outh Cooperation The concept of “South-​South” cooperation (SSC) emerged in the 1950s as a distinc­ tive approach to foreign aid, gaining momentum in the post–​Cold War period. It encompasses not just economic, social, and political dimensions, but also security and defense concerns. This chapter will concentrate on social development, but it is worth noting that a major dimension of SSC relates to collaboration among countries bordering the South Atlantic, which includes Amazonia, under Brazil’s national defense strategy (Aguilar 2013, Wiesebron 2013). As well as the “softer” social policies on which this chapter will focus, therefore, it should not be forgotten that “harder” strategic issues also figure strongly in wider foreign policy considerations. When President Lula took office in 2003, he saw Brazil’s foreign policy as an op­ portunity to consolidate South-​South relations and expand the country’s influence into new areas. This included building ties with the BRIC countries (Brazil, Russia, India, and China), pursuing diplomatic space through the Union of South American Nations (União de Nações Sul-​Americanas; UNASUR), and striving for a leadership role in major UN organizations such as the World Trade Organization (WTO) and the Food and Agricultural Organization (FAO). For historical, cultural, and geographical reasons, Sub-​Saharan Africa was singled out as a major focus for extending Brazil’s “soft power” and the politics of solidarity. In particular, the Community of Portuguese

536   Anthony Hall Language Countries (Angola, Cape Verde, Guinea Bissau, São Tomé and Príncipe, and Mozambique) was prioritized for cooperation purposes. A  major objective was to increase the volume and influence of Brazil’s overseas development assistance, underpinned by the rapid expansion of its foreign reserves from less than US$37 billion in 2002 to over US$252 billion by 2010. Brazil is officially classified as a “South-​South cooperation provider” (OECD 2015). The country has been notoriously slow in publishing up-​to-​date aid statistics. Until recently, figures from the Development Assistance Committee/​ Organisation for Economic Co-​operation and Development (DAC/​OECD) for Brazil were only avail­ able for 2010. These put Brazil’s annual aid bill at US$923 million, about US$500 million of which would meet the criteria for Official Development Assistance (ODA), up from US$362 million in 2009. The OECD treats the higher figure with some skepticism on the grounds that it may include elements that would not officially count as development cooperation, such as peacekeeping operations. Recently released figures (Brazil 2016a) show that during 2005–​2013, Brazil spent a total of US$4,146 million on aid, peaking in 2010 at US$923 million. However, by 2011–​2013 the annual aid bill fell significantly to just under US$397 million, or well under half that for 2010. Around 56% was channeled through international organizations via multilateral arrangements. Brazil had an ambitious political agenda during the Lula years. It sought to hold out the perhaps optimistic “promise that it was possible to grow, reduce poverty, and in­ crease social inclusion while also remaining a democracy committed to human rights” (Mares and Trinkunas 2016, 69). Through its program of technical and financial assis­ tance, Brazil also aimed to “build broad support for its candidacy in global institutions in pursuit of its ultimate objective: reforming the global multilateral institutions so as to give greater weight to its interests and policy preferences” (Mares and Trinkunas 2016, 74). The Ministry of External Relations (Itamaraty) oversees Brazil’s development co­ operation and coordinates its humanitarian aid and technical cooperation through the Brazilian Cooperation Agency (Agência Brasileira de Cooperação; ABC), along with financial cooperation such as debt relief, some concessional loans, and multilateral allocations. According to the latest official report, Brazil’s SSC activities in Africa involve an im­ pressive variety of activities (Brazil 2016a). These include social protection and nutri­ tional security, urban development and financial inclusion, human rights, agricultural research in several countries (Angola, Bolivia, Cape Verde, Cuba, Mozambique, Peru, and Senegal), public health, HIV/​AIDS initiatives, education and scholarships, scien­ tific research and technical cooperation, and peacekeeping and humanitarian assis­ tance, among others. A small but growing portion of Brazil’s aid is channeled into “triangular coopera­ tion,” in which two countries form a partnership to provide technical assistance to a developing country. These tripartite agreements involve partnerships with interna­ tional organizations such as the United Nations Development Programme (UNDP), FAO, World Food Program (WFP), United Nations Educational, Scientific and Cultural Organization (UNESCO), and DAC members such as Germany, the United States,

South-South Cooperation for Social Development    537 Japan, and the United Kingdom to support developing countries in South America, Africa, East Timor, and Haiti. Trilateral cooperation forms an intermediary mode be­ tween the more traditional bilateral and multilateral categories. In this mode, typically, a donor and a recipient country collaborate with a third party (such as a UN agency) to deliver technical assistance. The most active donors in triangular cooperation have been Japan, Chile, Brazil, Norway, Germany, Mexico, Guatemala, and Colombia, with between 20 to 160 projects lasting an average of 33 months each (OECD 2016). Over the past decade or so, Brazil’s development assistance in Africa has been con­ centrated on support for social development, namely social protection policy (such as conditional cash transfer programs (CCTs), as well as support for agriculture, health, education, food security, and humanitarian assistance. This focus has helped to carve out a particular niche and to build Brazil’s reputation as a provider of SSC, in contrast to the standard North-​South paradigm that governs traditional aid relationships. The latter is regularly portrayed as somehow “exploitative” and overwhelmingly serving donor interests. South-​South cooperation as a distinctive set of aid policies has been around for longer than many people imagine. At present, SSC accounts for just 10 percent of the total aid budget globally, and is far outweighed by traditional aid under DAC rules. However, it is considered qualitatively superior in many respects in terms of aid ef­ fectiveness. The concept emerged from the creation of the Non-​Aligned Movement at the Bandung conference in 1955, when 29 Asian and African countries, as well as some 30 national liberation movements, agreed on a set of “fresh” principles that would govern international development cooperation. In an attempt to influence the global aid agenda and break links with the old colonial powers, these new principles included respect for human rights and national sovereignty, along with the promotion of mutual interests and justice. During the 1970s, the United Nations was instrumental in promoting SSC, in par­ ticular the UNDP. This movement culminated in the Buenos Aires conference (1978) in which 138 nations participated and which established the technical guidelines to be followed under the Buenos Aires Plan of Action, or Plano de Ação de Buenos Aires (PABA) (Pino 2014). Later, following the downturn in SSC that came with the advent of neoliberal influences in the 1980s and 1990s, the United Nations was able again to use its political neutrality and decentralized structure to refocus attention on the basic princi­ ples of SSC. These fundamental (some would say idealistic) principles are reiterated in Brasília through the corridors of power in technical cooperation bodies, including the ABC, as well as bilateral and multilateral agencies. The High Level UN Conference on South-​ South Cooperation, held in Nairobi on the 30th anniversary of Buenos Aires, reaffirmed basic SSC principles. South-​South cooperation, it was argued, should not be seen as offi­ cial development assistance (ODA), but rather as a “partnership among equals based on solidarity” in which there is “respect for national sovereignty, national ownership and independence, equality, non-​conditionality, non-​interference in domestic affairs and mutual benefit” (UN 2010, 2).

538   Anthony Hall The overriding sentiment expressed in official interviews is that, above all, SSC should be “demand-​driven.” That is, aid donors should be responsive to national requests and not imposed or “ventriloquized” by outsiders seeking to impose their will for commer­ cial or political reasons. Essentially, the SSC aid machine, according to its proponents, could be used as a mean of “empowering” recipients. Furthermore, techniques such as triangular cooperation could produce a multiplier effect by combining the roles of var­ ious stakeholders to jointly maximize their impacts.

25.2.  Brazilian Solutions for African Problems? Under Lula’s presidencies, Brazil “sold” itself on the world stage as an effective ex­ ample of how to help eliminate world poverty. The government research agency Instituto de Pesquisa Econômica Aplicada (Institute for Applied Economic Research; IPEA) proudly announced in 2013 that “Brazil is helping the world to reduce pov­ erty” (Lisboa 2013). Initially based on the impact of the country’s CCT program, Bolsa Família, it was assumed that the direct transfer of technologies such as the Single Registry1 could also contribute to the speedy reduction of poverty and inequality else­ where. As one analyst noted, “The Brazilian solution for ending poverty was a refer­ ence point for other countries, to the extent that it became an export product” (Lisboa 2013, 23). In fairness, the same study also recognized that many of these achievements were also due to the specific contextual circumstances of Brazil, namely expanding employment opportunities, a progressive minimum wage policy, and an efficient banking system. When President Lula came to power, he and his government were able expand their influence overseas by forging closer ties with Lusophone countries in Africa. An inten­ sive and ambitious program of official visits across the Atlantic enabled Brazil to test and apply its social experiments from Brazil in the search for solutions to poverty and food insecurity in Africa. Brazil’s diplomatic presence and its development cooperation ac­ tivities were expanded significantly. During the Lula administration, 40 new embassies were opened, including 19 in Africa (Leite et al. 2014). In a peculiar variant of the blue­ print bias, no doubt buoyed up by Brazil’s growing self-​confidence in the first decade of the 2000s and emboldened by Lula’s re-​election, the popular perception among some technical experts was that “for every African problem there is a Brazilian solution” (cited in Cabral et al. 2016, 47). During the Lula administrations (2003–​2006 and 2007–​2010), South-​South coop­ eration “was brought to the center of the foreign policy agenda” (Leite et al. 2014, 48). Lula paid official visits to 12 African countries, while the minister of foreign affairs un­ dertook 67 visits to 37 African countries (Leite et al. 2014). These contacts inspired a series of fresh development initiatives and encouraged ABC budgetary expansion.

South-South Cooperation for Social Development    539 During 2011–​2013, over two-​thirds of the value of technical cooperation was allocated to Africa (Brazil 2016a). Not only was the country able to present a kinder development face through its SSC philosophy, this aim was also thought to be consistent with Brazil’s pursuit of a permanent seat on the UN Security Council by reinforcing its role as an instrument of soft foreign policy. There seems to have been a general consensus that, during this period, “[t]‌echnical cooperation with developing countries is, therefore, emerging as an important operational instrument of Brazilian foreign policy” (Cabral and Weinstock 2010a, 2). As Cabral and Weinstock (2010b, 2) argue, taken together, Brazil’s policy success in combating poverty, its appropriate level of technical expertise, its political neutrality, and its lack of conditionalities “give Brazil some comparative advantages in the aid scene.” This seemed to be underlined by the growing demand for Brazilian cooperation from developing countries. The number of technical cooperation projects initiated rose from 23 in 2003 to 413 by 2009. Brazilian cooperation largely benefited Latin America and the Caribbean between 2005–​2010, at 49%, followed by Africa, at 35%. However, to put aid matters in perspective, it can be said that Dilma Rousseff ’s ad­ ministration marked a significant turning point away from the “social development” approach to SSC. Since her second administration and fall from grace, following her impeachment and the assumption of interim president Michel Temer, there has been a greater focus on Latin America, together with a reduction in ABC’s technical coop­ eration budget. Even before the political turmoil, this was slashed from R$52 million in 2011 to R$36 million in 2012 (Leite et al. 2012). Development cooperation with other countries seems to have been deprioritized, with a view to promoting trade and private investments. Commercial ties have gradually replaced social grants as Brazil seeks more industrial engagement with Africa. It seems that there is a new quest for integration into more sophisticated export markets of Europe and Asia. This recent emphasis on promoting Brazil’s economic interests coincides with, for example, diminished support for overseas vocational education and training in Africa, spearheaded by Serviço Nacional de Aprendizagem Industrial (Brazilian National Service for Industrial Apprenticeship; SENAI) as efforts are refocused domestically. These policy choices are consistent with the declaration of the new minister of foreign affairs, José Serra, who pointedly stated in his inaugural address on May 18, 2016, that “[d]‌iplomacy will again reflect the legitimate values and economic interests of Brazilian society at the service of Brazil as a whole rather than the convenience and ideological preferences of a single political party and its allies overseas” (Brazil 2016b). Some of these issues are taken up in the concluding section of this chapter.

25.3.  Institutional Diversity The Brazilian Cooperation Agency (ABC) was set up in 1987, with the support of UNDP, as a department of the Ministry of Foreign Affairs (Itamaraty) to oversee and coordinate

540   Anthony Hall technical cooperation. However, it has no financial or human resource management au­ tonomy and is subordinated to foreign policy. One of the most notable features of Brazil’s SSC program has been its fragmentation and institutional complexity. Altogether, it has been estimated that up to 120 agencies have been involved in implementing SSC projects, coordinated by ABC at various times, although how effectively remains a moot point. Currently, according to ABC, around 95 official bodies are involved in executing Brazil’s SSC program (Brazil 2016a). It is an open secret that, due to high staff turnover, typical of the diplomatic service, it has been difficult for ABC to provide effective coordination. According to critics, ABC is “handicapped by its incapacity to perform basic development assistance functions such as procuring goods and services for developing countries” (Cabral and Weinstock 2010, 2; see also Leite et al. 2014). The lack of technical and analytical capacity limit ABC’s role largely to diplomatic representation. In these situations, bilateral and multilateral bodies, as well as civil society organiza­ tions, have filled in the skills gap. Chief among these has been the UNDP, which has be­ come the main agency for operationalizing technical cooperation in Brazil, and which was instrumental in setting up Brazil’s cooperation agency. This strategy has helped by­ pass ABC-​MRE’s (Ministério das Relações Exteriores; Ministry of External Relations) limited staff and technical capacity and the constraints imposed by national legislation, although the short-​term nature (1–​2 years) of many UNDP contracts remains a serious limitation. The government research body IPEA, attached to the Secretariat for Strategic Affairs of the Presidency of the Republic, provides data gathering and research facilities. Together with the International Poverty Centre for Inclusive Growth (IPC-​IG), IPEA fosters studies that contribute to monitoring and evaluation activities within the SSC program. However, the lackadaisical record on this front has made assessing the achievements of the aid machine problematic and has led to recent staffing changes within IPEA. Other major institutional players from the outset of Lula’s Africa campaign have been the Ministry of Social Development (Ministério do Desenvolvimento Social; MDS) and Ministry of Agrarian Development (Ministério do Desenvolvimento Agrário; MDA), which played prominent roles in launching social protection initiatives and, latterly, in establishing food security programs, as detailed in the following. International politics, however, continue to reflect changing domestic priorities. In May 2016, the MDA was unceremoniously abolished as part of a min­ isterial reduction by President Temer, and its greatly reduced role was folded into a Special Secretariat, attached to the presidency, physically housed on the top two floors of the MDS building. The renamed Ministry of Social and Agrarian Development (Ministério do Desenvolvimento Social e Agrário; MDSA), led by minister Osmar Terra, reflects the altered priorities in Brazil’s aid program. Its reduced political status seems to echo a move away from the original stress on small-​scale farmer programs

South-South Cooperation for Social Development    541 under the Lula administration, toward a new emphasis on commercial farming and associated trading activities (see later discussion). This institutional dispersal encompasses the executive, judiciary, legislature, nongovernmentals organizations (NGOs), universities, and local governments. ABC covers their operational costs, while employing agencies cede their time and pay their salaries. The lack of specific legislation “clearly defining the objectives, scope, mechanisms, competencies and processes of development cooperation” hampers the construction of a clear set of rules and standards (Leite et al. 2014, 40–​41). There has thus been an ever-​greater reliance on the use of consultants from multilateral bodies such as UNDP and FAO to provide transitory arrangements. In addition to the major “social development” players mentioned in the preceding, one could also highlight the Ministry of Health, as well as the Ministry of Science, Technology and Innovation (Ministério de Ciência, Technólogia e Inovação; MCTI). The Oswaldo Cruz Foundation (Fiocruz) acts as the focal point for health-​based devel­ opment projects in South America and Africa with 18+ projects in capacity-​building and strengthening of health systems, including those in Nigeria, Burkina Faso, Mali, and Tanzania. The Brazilian Agricultural Research Corporation (Empresa Brasileira de Pesquisa Agropecuária; Embrapa) has experimented with soybean expansion in Mozambique, while technical aid has been provided for other endeavors such as the production of cotton (Benin, Burkina Faso, Chad, and Mali) and rice (Senegal). Another high-​ profile project has been supported by the National Service for Industrial Training (SENAI). In 2013, SENAI ran 13 vocational education and training initiatives while numerous others were planned (Leite et al. 2014). These were reputedly very successful, but also expensive in terms of required investments in capital equip­ ment and training costs. Apparently, Brazilian domestic vocational training needs, rather than African demands, are now being prioritized. Peacekeeping operations have seen Brazilian troops deployed in various regions of the world, the vast bulk of resources being allocated to Haiti (under MINUSTAH) since 2004 as a result of the earthquake, in line with national defense policy. Other interventions include in Mozambique, Angola, and East Timor. Cooperation ac­ tivities in Haiti involved over 50 organizations, including the Ministry of Health and well-​known CSOs, and most humanitarian aid under the Coordenação Geral de Ações Internacionais de Combate à Fome (General Coordination for International Action Against Hunger; CGFome), attached to the Ministry of External Relations, was also directed there. In the educational sphere, SSC also supports scientific and technolog­ ical cooperation at a large scale, through regional and bilateral ties and via partnerships. Scholarships for foreign students have favored those from Portuguese-​speaking coun­ tries through the Brazilian Higher Education Coordination Agency (Coordenaçāo de Aperfeiçoamento de Pessoa de Nível Superior; CAPES), with special bilateral projects benefiting East Timor, Cuba, Argentina, Mozambique, Cape Verde ,and Guinea-​Bissau (Leite et al. 2014).

542   Anthony Hall

25.4.  Pursuing Social Development in Africa: Social Protection and Nutritional Security Brazil’s social development agenda in Africa can be broadly divided into two major, overlapping phases. Phase I focused initially on cash transfers, as represented by the em­ blematic Bolsa Família, while Phase II was better balanced in terms of addressing food security more generally. Taken together, combining cash transfers with nutrition meas­ ures was effectively a dual strategy that envisaged (a) stabilizing incomes through cash transfers and other social protection policies on the demand side, and (b) guaranteeing basic food supplies and sustaining small farmer livelihoods. Setting the scene for Brazil’s “social” vision of aid, established in 2004, was Itamaraty’s CGFome, controlled di­ rectly by Itamaraty. Its budget rose from US$1.25 million in 2007 to US$17.5 million by 2010. Reconstruction in Haiti represented almost 70% of all humanitarian cooperation (Leticia et al. 2014). More broadly, it is thought by some observers that those domestic policies that fared well in Brazil itself have had a positive impact in Africa in the achievement of soft power goals (Mares and Trinkunas 2016). For example, in the fields of poverty alleviation and inequality reduction, many of these assumptions are perceived as having been success­ fully replicated across the Atlantic. Unfortunately, there is very little direct evidence to back up these assertions, given the total lack of evaluation studies, and wishful thinking seems to predominate regarding outcomes. As Chapter  24 by Barrientos in this volume shows, Brazil has been relatively successful in preparing a favorable policy context for effective anti-​poverty measures regarding cash transfers. Brazil’s Constitution of 1988 established the basic right to non-​ contributory benefits targeted at the poorest sectors of the population as a vehicle for attacking poverty in a risk-​based approach. The Single Registry system (CadÚnico) was created in 2001 to identify, select, and monitor the progress of beneficiaries. Some pro­ gram interventions have been partially transferred to Africa (as in the cases of Ghana and Kenya). However, given the different contexts and lack of impact evidence, as noted earlier, few comparative conclusions can be drawn. The umbrella program Zero Hunger (Fome Zero) was set up to connect public anti-​ poverty initiatives, including Bolsa Família and others such as the Food Acquisition Program (PAA)—​forerunner of PAA-​Africa—​in 2003. The National School Nutrition Program (Programa Nacional de Alimentação Escolar; PNAE) of 2003 was also part of the package and was eventually exported to Africa. The formation in 2004 of the aptly named Ministry of Social Development and Fighting Hunger aimed to merge these policies. Reactivation of the National Council on Food Security (Conselho Nacional de Segurança Alimentar e Nutricional; CONSEA) also introduced a valuable monitoring role in food supply and nutrition. The ambitious program Brazil Without Extreme

South-South Cooperation for Social Development    543 Poverty (Brazil 2011) was intended to eventually bring these elements together under a more inclusive and production-​oriented perspective (Hall 2017). Yet this “social invest­ ment” approach is arguably still in its infancy and needs to be properly developed. Under the label of social protection, COBRADI (Relatório sobre Cooperaçāo Brasileira para o Desenvolvimento International; Report on Brazilian Cooperation for International Development) identified several types of technical assistance provided by the ministries of Health, Social Development, Agrarian Development, and Social Security and Employment, 23% of which was directed to Africa (Caixeta and Suyama 2015). Most ABC funds were used to cost travel and maintenance, agency staff time, and some equipment. Just 14 of the 215 individual projects were classed as “triangular,” implemented by third parties such as the US Agency for International Development (USAID), the United Nations Population Fund (UNFPA), Oxfam, WHO, and UNDP. The first SSC project with MDS was in partnership with the Department for International Development (DFID) under the Africa-​ Brazil Program for Social Development in 2006 (IPC-​UNDP 2008). Under the scheme, Africans would receive technical assistance to develop social protection programs or enhance existing plans. As a consequence of this initiative, support was provided to the Ghanaian government to design and implement the CCT program Livelihood Empowerment Against Poverty (LEAP). This initiative was a high-​intensity endeavor in which technical assistance from MDS and IPC was mobilized to provide the Ghanaians with intensive support on key aspects such as the setting up of the single registry and anti–​child labor policies. The Brazil-​Ghana pilot program was considered a very successful cooperation expe­ rience by MDS, due to the high level of technical support provided by DFID and IPC. Ambitious targets were set to reach 160,000 families living in extreme poverty within five years, expanding beneficiary groups within LEAP and other social protection programs. On the back of a sound program, budget plans were made to refine bene­ ficiary selection procedures and monitoring and evaluation measures, as well as for follow-​up visits to Brazil by technicians from Ghana. At that time, African countries were anxious to learn about Brazilian social pro­ tection. The intergovernmental regional conference “A Transformative Agenda for the 21st Century: Examining the Case for Basic Social Protection in Africa” was held in Livingstone, Zambia. The final declaration from the conference, co-​hosted by the Zambian government and the African Union, in which Brazil was the only non-​African participant, called for greater exchange of experiences and information on social pro­ tection and cash transfers (Cirillo et al. 2016). A second phase of the Africa-​ Brazil Program for Social Development was implemented in 2008, involving DFID, as well as Brazilian government representatives from the MDS, IPEA, and the IPC-​IG. Three regional expert meetings were held in prep­ aration for the first Ministerial Meeting on Social Protection in Africa, held in Namibia in 2008. Brazil’s then-​minister of social development, Patrus Ananias, also took part, confirming his country’s interest, while the African Union recommended that “Member States should take advantage of [ . . . ] South-​South cooperation and regional and inter­ national best practices” (Cirillo et al. 2016, 6). At around the same time, in 2009, Brazil

544   Anthony Hall and Kenya expressed interest in developing Kenya’s Single Registry or unified database and a proper monitoring and evaluation system, as well as cross-​sector collaboration. Mozambique expressed similar wishes and also requested support in developing Social Assistance Reference Centers (CRAS), based on the Brazilian model. Many of the study tours undertaken at the time focused on such key social protection issues. In 2011, 80 official delegations representing 50 countries visited Brazil to learn about Brazilian social development, followed by 136 delegations from 65 countries the following year. By 2013, there was a substantial decline in these visits (Brazil 2016a). Another major initiative, aided by the Brazilian Ministry of Health/​Fiocruz, was the establishment of a factory in Mozambique for producing anti-​retroviral drugs used in the treatment of HIV/​AIDS. An office was set up in the capital Maputo to monitor this and other similar ventures on the continent (Caixeta and Suyama 2015; Hall 2013). In spite of these individual social policy measures, however, there is little evidence that this formed part of a concerted reorientation of formal social protection measures in Africa, as originally mapped out in the early phase from 2006. Despite the laudable intentions, in most cases, tours and initial expressions of in­ terest did not result in a formal agreement or a structured plan of action. In other words, “[o]‌nly in a few cases did the exchange of experiences during the study tours create the conditions to formalize South-​South cooperation between the countries involved” (Cirillo et al. 2016, 6). This policy transfer from Brazil to Africa, in spite of the technical assistance offered by a range of multilateral organizations such as UNDP, FAO, World Bank, the United Nations Children’s Fund (UNICEF), and IPC-​IG, as well as bilateral support, could not be sustained at existing levels. In order to meet growing demand from African countries for knowledge exchange, it was therefore decided to convene a series of international seminars on “Social Policies for Development.” The last sem­ inar, held in May 2016 in Brasília, received delegations from 43 countries, with 34% from Africa and 48% from Latin America and the Caribbean (MDS 2016). The response from African policymakers to this “economies of scale” attempt at diffusion of ideas has been positive, according to analysts of Brazil-​Africa knowledge sharing. This learning exchange is seen as having produced a “positive influence and has inspired several programs and instruments in Africa,” rather than a long-​term, structured plan of action (Cirillo et al. 2016, 17). In fact, a long list of social protection instruments “inspired by the Brazilian experience” is given in the report just cited. Yet only a small number of these are described as having been partially implemented, and it is readily admitted that “only a few countries received formal technical support from the Brazilian government via South-​South cooperation projects” (Cirillo et al. 2016, 17). These included the LEAP program in Ghana, the National Social Protection Policy in Kenya, and the Cash Transfer Program in Cape Verde. The initial focus on cash transfers via MDS based on the Bolsa Família, as already noted, later shifted toward an emphasis on food and nutritional security. This was only partially due to the understandable resource constraints, in terms of budg­ etary restrictions and limited technical capacity, on the Brazilian government’s ability to respond to growing African demand (the generous assistance provided through

South-South Cooperation for Social Development    545 UN agencies such UNDP notwithstanding). In addition to such strictly technical considerations, it was also a political and an ideological issue. There emerged in Brazil a growing debate within academia and civil society organi­ zations (CSOs) around the need to adopt a more integrated and holistic policy approach toward social protection and poverty reduction. This meant stimulating effective de­ mand for food among poorer sectors of society, while also addressing the supply side, stimulating the family farm sector, and boosting the livelihoods of rural producers. Raising household incomes through cash transfers was now seen as only one part of the social protection approach. These ideas are well conceptualized by Amartya Sen (1999). Furthermore, CCTs came to be perceived as relatively costly, with a complex structure and (despite the early rhetoric of SSC) propelled to some extent at least by a donor-​ driven agenda. Areas of future cooperation were defined at the “Brazil-​Africa Dialogue on Food Security, Fight Against Hunger and Rural Development,” held in Brasília in 2010. Some 40 African ministers defined 10 pilot projects on public food purchase, extending cov­ erage of school meal programs, supporting family farm modernization. FAO, WFP, CGFome, WFP, and DFID are implementing PAA-​Africa via projects in Ghana, Malawi, Senegal, Mozambique, Niger, Kenya, Côte d’Ivoire, Rwanda, and Ethiopia. Pilots in Ethiopia, Malawi, Mozambique, Niger, and Senegal started in mid-​2012 as a three-​year initiative (Leite et al. 2013). This dual strategy is based on government purchase of family produce and its sub­ sequent distribution in school feeding programs, in an attempt to build up local food supply chains. The Brazilian government funds these initiatives through its contributions to the FAO and WFP. A similar policy switch was chosen for 13 countries in Latin America and the Caribbean under the project labeled Support for National and Sub-​regional Strategies for Food Security and Poverty Reduction in Latin America and the Caribbean. One key factor determining the success of PAA-​Africa has been the role of CONSEA. This enabled the broad left in Brazil to demand accountability from the government, and compared and encouraged civil society institutions in Africa to become political actors. It was a structure considered viable in the African context. Unlike the bureaucratic gov­ ernment structures for monitoring CCT programs, CSOs in Africa were encouraged to seek a new lease of life. An expected outcome of PAA-​Africa was to strengthen links be­ tween Brazilian partner countries, civil society, and family farming (Leite et al. 2014). At the same time, Brazilian CSOs were encouraged by Brazil’s Ministry of Foreign Relations to become more involved in developing humanitarian assistance and expanding their role as policymakers. The More Food Africa program was created in 2010 by the MMDA and was intended to boost food production in the small family farm sector in Africa as well as to stimulate the mechanization of small-​scale family farming, in keeping with Ministry of Agrarian Development priorities at that time. It aims to stimulate commercial exchanges by encour­ aging sales of appropriate equipment for use by African famers, and to boost cross sector collaboration between agriculture and industry in support of sustainable national food

546   Anthony Hall security strategies. By 2012, plans were underway for eight countries to join up: Ghana, Zimbabwe, Mozambique, Senegal, Kenya, Cameroon, Namibia, and Tanzania. Yet critics have expressed doubts over the transferability of Brazilian notions of ‘family farming’ to the African context. The MDA targeted Africa’s small-​and medium-​ sized farm sector based on comparisons with Brazil, through the relabeled More Food International (MFI) project (Cabral et al. 2016). This drew upon Brazil’s own More Food Program, placing MFI as the MDA’s main cooperation instrument, and was popular with the Lula administration. It was considered that subsidized credit such as the pop­ ular Programa Nacional de Fortalecimento da Agricultura Familiar (National Program for the Strengthening of Family Farming; PRONAF) for the purchase of Brazilian farm equipment would support the African demand for modernization and increased productivity. This approach, based on a more polarized conception of Brazilian agrarian structures as divided into large-​scale commercial farming at one end of the scale and family farmers at the other, was seen by civil society as representing a reactionary impulse that prioritized the interests of industry and commercial producers. Thus, together with attacks on Embrapa’s support for ProSavana (Regional Development Program in Mozambique to stimulate large-​scale agribusiness in the Nampula corridor), Brazilian academics and CSOs were able to use Africa as a “discursive battlefield” (Cabral et  al. 2006, 56). As such, the suggestion is that the coincidence of interests between African political and bureaucratic elites has been exposed and the small farm sector in Africa is possibly being ignored, at least in the cases of three example studies: Ghana, Mozambique, and Zimbabwe. In all three instances, there was a visible “coincidence of interests between African political and bureaucratic elites and the Brazilian hardware exporters” (Cabral et  al. 2016, 56). MDA support for the More Food International program partially supported the “modernization” take in this diversified framework and thus gained the support of the industrial lobby as well as government budgetary assistance. Until it was effec­ tively abolished in May 2016, the MDA was “eager to reinstate its credentials of family farming, and distance itself from ProSavana and the criticism that the former also serves a business agenda” (Cabral et al. 2016, 55). Brazil’s new emphasis on food security in the wider debate on social protection has been reflected in the statistics on Brazilian humanitarian assistance (CGFome 2016). Brazilian aid channeled through CGFome, which is provided through the Ministry of Foreign Relations, started modestly in 2006 at less than US$0.5 million, peaking at over $93.5 million by 2010. By 2015, this figure had dropped to less than US$8 million. Comparable trends were evident for that period in supplies of food and medicines. One of the major challenges faced by Brazil’s SSC machine is how to maintain the mo­ mentum of engagement with African initiatives in social protection and food security in rapidly changing budgetary and political circumstances. However, the indications are that budgets are shrinking, as exemplified by Brazil’s humanitarian cooperation programs that, as just noted, have seen major cuts since 2013, along with SSC aid opera­ tions in general (CGFome 2016; Brazil 2016a).

South-South Cooperation for Social Development    547

25.5.  Towards Virtual Cooperation? One possible way to help bridge the gap is by providing virtual solutions. For example, the World Without Poverty site was set up by IPEA, the Ministry of Planning, Ministry of Social Development and Fighting Hunger, IPC-​IG, and the World Bank. The Brazil Learning Initiative for a World Without Poverty was launched in 2013 to “share Brazil’s experience of development with social inclusion and promote the exchange of best practices with international partners” (WWP 2013). The WWP produces data and publications in several languages on social protection programs to inform managers, technical staff, and policy practitioners. The site is meant to be user-​friendly, focusing on key aspects such as the single registry, Bolsa Família, social assistance, food security, and Brazil Without Extreme Poverty (wwp.org.br). A second knowledge-​sharing platform is provided by the World Bank, IPC, IPEA, and the UN Inter-​Agency Cooperation Board, or SPIAC-​B (socialprotection.org). This is a far more comprehensive site both thematically and demographically. It covers a wide selection of social protection policies and links many different communities of practitioners across the globe, providing access to a variety of knowledge-​sharing possibilities and acting to increase potential for policy innovation via capacity-​building webinars. Communities of practice for Africa and Latin America are run by a team of assistants and global helpers, catering for a membership of some 1,100 participants. There is little direct evidence that such experiences have been effective, in terms of sharing information that is then used to put specific policies into practice. This is es­ pecially true without the kinds of direct training and face-​to-​face contacts that used to be supported through technical assistance. When 48 African policymakers from 24 countries were recently consulted about how they perceived such knowledge-​sharing experiences, including the international seminars, they were fairly unequivocal. They largely concluded that they did not see this as a structured plan of action, but more as a form of “horizontal cooperation” that might inspire change. Understandably, perhaps, the partners looked forward to signing more formal bilateral agreements as SSC projects (Cirillo 2016). It remains to be seen whether virtual technical assistance is capable of maintaining the rhythm of change in SSC, and in what direction.

25.6.  Conclusion: The Future of South-​S outh Cooperation Without doubt, under Lula there was something of a golden age in which SSC boosted its soft power influence. This was especially true for triangular cooperation, which used to account for one-​fifth of Brazil’s technical assistance projects, and is considered by some in aid circles to have significant potential. It is suggested, for example, that “participation

548   Anthony Hall in triangular cooperation is not only a goal unto itself but also serves the broader pur­ pose of furthering international relations” (Abdenur 2006, 7). Some observers see the main challenge as finding the “equilibrium between policy independence and assertive autonomy, on the one hand, and greater proximity to traditional donors on the other, as it attempts to scale-​up its assistance and consolidate its position as a provider” (Cabral and Weinstock 2010b, 33). There are specific reasons for this in terms of pragmatism, altruism, strengthening cultural ties, and providing regional leadership. At least in the past, Brazil has carved out for itself a distinctive identity based on these alleged characteristics, especially with regard to its triangular role in aid networks:  (a) pragmatism—​Brazil is better equipped to pass on relevant expertise that fits developing country circumstances; (b) altruism—​Brazilian SSC is supposedly motivated by nonprofit goals and has tech­ nical capacity-​building as its main aims rather than profit; (c) cultural ties—​Brazil relies on strengthening cultural ties with Portuguese-​speaking and Spanish-​speaking nations, as well as developed countries with which it has long-​standing ties, such as Japan; and (d) Brazil’s quest for regional leadership in Latin America, as illustrated by its role in the deployment of peacekeepers in Haiti. It looks as if Brazil is unlikely to be able to build upon this SSC legacy in any sustain­ able fashion. A stagnating economy and the need to cut the MRE budget by half has seri­ ously undermined its “soft power.” This includes a failure to keep up with contributions to the very UN agencies on which the country claims to depend for supplying continued technical assistance. Furthermore, the country’s political credibility has been severely damaged by numerous political scandals, including the Petrobrás /​Lava Jato corruption affair and related troubles, such as the impeachment of President Dilma Rousseff. It has long been recognized that the Brazilian Cooperation Agency (ABC) needs to be reorganized to better coordinate its technical cooperation and that of international agencies with which it collaborates. Recent interviews with ABC-​MRE suggested that the present system is being examined to determine whether a more systematic and structured approach could be defined. However, in the present state of political turmoil in Brazil, a timely resolution to this problem seems unlikely. At the same time, how­ ever, the new minister of foreign affairs, José Serra, has indicated that Brazil will seek more conventional, commercial relations with its aid partners in future, while (then) President Dilma also indicated that government-​business relations might regain a foot­ hold in the near future. In the cold light of day, the “war against hunger” pursued under Lula’s presidential leadership seems like an act of enlightened self-​interest designed to help Brazil forge a more distinctive role as an emerging middle power, acting as a bridge between de­ veloped and developing countries (Nina 2006). Policies pursued for domestic purposes have been effectively applied to achieve wider, international aims. This connection appears to have been substantially weakened under the post-​Dilma presidency. Social protection and food security policies, once prioritized at home and overseas, no longer occupy the same pride of place as business and commercial interests seem to be replacing the values of SSC in Brazil’s changing aid strategy.

South-South Cooperation for Social Development    549 Some observers cast doubt on whether Brazilian aid is (or indeed has ever been) quite so altruistic as it is sometimes portrayed. Burgess (2014, 370), for example, suggests that Brazil “buys” support “through the expression of solidarity and the transfer of applied expertise rather than the provision or promise of substantive economic rents.” According to him, this also reduces the risk of aid dependency and corruption. Yet how far this pathway represents a viable strategy for the future must be called into question, given that the budget for technical cooperation has been significantly reduced and di­ rect assistance is being replaced by “virtual” support from a distance. Another line of inquiry concerns the emerging relationships between China, Brazil, and Africa with regard to agricultural development, possibly involving trilateral coop­ eration (Scoones et al. 2016). Strong states aligned with business interests might seek to encourage various models of rural production, such as ProSavana or small ventures under the aegis of a developmental state. However, in Brazil’s case, major doubts would have to be raised as to whether it has the technical capacity, the budget, or the political freedom to make such choices. Silva (Chapter 29 in this volume) seems to suggest that Brazil should develop its ties with the BRICS (Brazil, Russia, India, China, and South Africa) economies, such as China, to strengthen its level of global economic integration. It is worth noting that Africa has been the target not just of Brazil in its SSC endeavors, but also of China. As has been pointed out, China has expanded not just its conven­ tional mining and energy initiatives, but has also developed “hundreds of health, en­ ergy and social infrastructure projects,” whose aim is ostensibly to promote constructive relationships as part of a “charm offensive.” Part of this strategy is to help secure African votes in the UN General Assembly on a range of issues (The Economist, April 16, 2016; August 13, 2016). The “Eight Principles of Chinese Aid” bear a striking resemblance to those that in theory also underpin Brazilian foreign assistance; for instance, equality and mutual benefit, respect for national sovereignty, non-​conditionality, low-​interest loans, and transfer of technical skills (Provost and Harris 2013). Given the sheer volume of resources available to China to pursue its own SSC program, Brazil will have to watch its back. A further possibility could be for Brazil to seek a stronger role on the global stage by extending its international peacekeeping operations through concentrating on capacities that are in short supply, such as intelligence, logistics, and command and con­ trol. It has also been suggested that Brazil should increase its humanitarian and devel­ opment assistance, building upon its ABC experience, expanding the role of BNDES (Banco Nacional de Desenvolvimento Econômico e Social; National Bank for Economic and Social Development) and extending its coverage to other countries such as the BRICS (Mares and Trinkunas 2016; Trinkunas 2015;). The New Development Bank, set up by the BRICS, and the Contingent Reserve Arrangement (CRA) to serve the Global South, may help in this respect. Many stakeholders thus await a reformulation of Brazil’s aid program and its priorities in the near future. Out of the current crisis, it might be that a phoenix will arise from the ashes, with a new development aid policy that combines the best of the new and the old. However, whether any reformulation will be able to galvanize public opinion around

550   Anthony Hall continuing support for South-​South cooperation, given the current political turmoil in Brazil, remains to be seen.

Note 1. The Cadastro Único, “Single Registry,” often referred to as CadÚnico, is an electronic data­ base of the population for use in social policy implementation.

References Abdenur, Adriana. 2006. “The Strategic Triad:  Form and Content in Brazil’s Triangular Cooperation Practices.” International Affairs Working Paper, The New School. Aguilar, Sérgio Luiz Cruz. 2013. “The South Atlantic: Brazil-​Africa Relations in the Field of Security and Defense.” AUSTRAL: Revista Brasileira de Estratégia e Relações Internacionais 10: 47–​67. Brazil. 2011. Brazil Without Extreme Poverty. Brasília: MDS. Brazil. 2013. Cooperação brasileira para o desenvolvimento internacional: 2010. Brasília: IPEA/​ABC. Brazil. 2016a. Cooperação brasileira para o desenvolvimento internacional:  2011–​ 2013. Brasília: IPEA/​ABC. Brazil. 2016b. “Speech of Minister José Serra on the Occasion of His Inauguration to the Position of Minister of Foreign Affairs.” Brasília, May 18. http://​www.itamaraty.gov.br/​en/​ speeches-​articles-​and-​interviews/​minister-​of-​foreign-​affairs-​speeches/​1 4044-​speech-​ by-​minister-​jose-​serra-​on-​the-​occasion-​of-​the-​ceremony-​in-​which-​he-​took-​office-​as-​ minister-​of-​foreign-​affairs-​brasilia-​may-​18-​2016 Burges, Sean. 2014. “Brazil’s International Development Cooperation:  Old and New Motivations.” Development Policy Review 32 (3): 355–​374. Cabral, Lidia, and Julia Weinstock. 2010a. “Brazil: An Emerging Aid Player.” Briefing Paper No. 64, Overseas Development Institute, London. Cabral, Lidia, and Julia Weinstock. 2010b. Brazilian Technical Cooperation for Development: Drivers, Mechanics and Future Prospects. London: ODI. Cabral, Lídia, Arilson Favareto, Langton Mukwereza, and Kojo Amanor. 2015. “Brazil’s Agricultural Politics in Africa:  More Food International and the Disputed Meanings of ‘Family Farming.’” World Development 81: 47–​60. Caixeta, Marina Bolfarine, and Bianca Suyama. 2015. “A cooperação Sul-​Sul em proteção so­ cial.” Observatório Brasil e o Sul (1): 1–​14. CGFome. 2016. Resumo geral de informações financeiras. Coordenação-​ Geral de Ações Internacionais de Combate à Fome. Brasília: MRE. Cirillo, Cristina, Lívia Maria da Costa Nogueira, and Fábio Veras Soares. 2016. “Brazil-​Africa Knowledge-​Sharing: What Do African Policymakers Say?” International Policy Centre for Inclusive Growth Working Paper No. 143. Costa Leite, Iara, Bianca Suyama, and Melissa Pomeroy. 2013. “Africa-​Brazil Co-​operation in Social Protection: Drivers, Lessons and Shifts in the Engagement of the Brazilian Ministry of Social Development.” WIDER Working Paper No. 2013/​022. Costa Leite, Iara, Bianca Suyama, Laura Trajber Waisbich, and Melissa Pomeroy, with Jennifer Constantine, Lizbeth Navas-​Alemán, Alex Shankland, and Musab Younis. 2014. “Brazil’s

South-South Cooperation for Social Development    551 Engagement in International Development Cooperation: The State of the Debate.” Evidence Report 59, Institute of Development Studies, Brighton. Hall, A. 2013. “Political Dimensions of Social Protection in Brazil.” In Social Protection, Economic Growth and Social Change, edited by James Midgley and David Piachaud, 166–​ 184. Cheltenham, UK: Edward Elgar. Hall, Anthony. 2017 “Cash Transfers as Social Investments:  The Brazilian Case.” In Social Investment and Social Welfare:  International and Critical Perspectives, edited by James Midgley, Espen Dahl, and Amy Conley Wright, 141–​159. Cheltenham, UK: Edward Elgar. Lisboa, Carla. 2013 “Brasil ajuda o mundo a reduzir a miséria.” Desafios de Desenvolvimento 10 (77): 22–​29. Mares, David, and Harold Trinkunas. 2016. Aspirational Power: How Brazil Tries to Influence the International Order and Why It So Often Fails. Washington, DC: Brookings Institution Press. MDS. 2016. Seminário internacional políticas sociais para o desenvolvimento. Resumo de delegações confirmadas. Brasília: MDS. Nina, Alexandre. 2007. “Action against Hunger and Poverty: Brazilian Foreign Policy in Lula’s First Term (2003–​2006).” Working Paper Number No. CBS-​83-​07. Centre For Brazilian Studies, Oxford. OECD. 2016. Triangular Co-​operation:  Findings from a 2015 survey. Paris:  DAC Global Relations. OECD Aid Tables. Various years. Available at http://​www.oecd.org/​dac/​dac-​global-​relations/​ brazil-​development-​co-​operation.htm, accessed September 8, 2016). Pino, Bruno Ayllón. 2014. “Evolução histórica da cooperação Sul-​Sul.” In Repensando a cooperação internacional para o desenvolvimento, edited by André de Mello Souza, 54–​86. Brasília: IPEA. Provost, Claire, and Rich Harris. 2013. “China Commits Billions in Aid to Africa as Part of Charm Offensive.” The Guardian, April 29, 2013. Scoones, Ian, Kojo Amanor, Arilson Favareto, and Gubo Qi. 2016. “A New Politics of Development Cooperation? Chinese and Brazilian Engagements in African Agriculture.” World Development 81: 1–​12. Sen, Amartya. 1999. Development as Freedom. Oxford: Oxford University Press. Trinkunas, Harold. 2015. “Brazil’s Global Ambitions.” Americas Quarterly (Winter). http://​ www.americasquarterly.org/​content/​brazils-​global-​ambitions. The Economist. 2016. “A Despot’s Guide to Foreign Aid.” April 16, 2016. The Economist. 2016. “Asia’s Scramble for Africa.” August 13, 2016. UN. 2010. “Narobi Outcome Document of the High-​level United Nations Conference on South-​South Cooperation.” United Nations: February 23, 2010. Wiesebron, Marianne L. 2013. “Blue Amazon:  Thinking the Defense of Brazilian Maritime Territory.” Austral: Brazilian Journal of Strategy and International Relations 2 (3): 1–​124.

Chapter 26

L ab or M a rk et Devel opment i n Bra z i l Formalization at Last? Celia Lessa Kerstenetzky and Danielle Carusi Machado

26.1. Introduction In contrast to the “lost decade” of the 1980s and the sluggish 1990s, the Brazilian economy achieved sizable gross domestic product (GDP) growth rates during the first decade of the twenty-​first century.1 Although far from Chinese or Indian performances, Brazilian growth, at a yearly 4% rate, was nonetheless meaningful: the recent growth path was inclusive or “redistributive” (Kerstenetzky 2014). In the wake of new labor market and social policies, economic growth boosted by favorable external conditions was finally reconciled with poverty reduction and a decline in economic inequality. Poverty dropped to 13% of the population in 2014, down from 36% in 1995; inequality, from a Gini index of .60 in 1995, to .52 in 2014.2 The new growth path was accompanied by a sustained expansion of employment, es­ pecially formal employment. In fact, the elasticity of formal employment in relation to growth exceeded one, while the reduction of the share of informality in employment from 2002 to 2014 amounted to 16.5 percentage points.3 In assessing the significance of this achievement, it is useful to bear in mind that the Brazilian labor market has traditionally featured high levels of informality, and, thus, sizable shares of the workforce in insecure, low-​paying, and socially unprotected jobs, in rural and urban areas alike (Baltar 2015; Barbosa Filho, Ulyssea and Veloso 2016; Cacciamali 2000). Moreover, the debt crises and (on the whole) the trade liberalization eras pushed informality rates to very high levels, reaching over 50% of total employ­ ment in the 1990s. Additionally, while the 1970s—​the zenith of the so-​called import-​ substitution industrialization (ISI) model—​might be hailed for displaying lower levels of informality thanks to impressive growth rates, this period’s poor record in other

Labor Market Development in Brazil    553 related aspects of job quality, such as minimum pay and wage inequality, means it has little else to recommend itself. In fact, in the early twenty-​first century, associated with the significant creation of formal jobs, there was an increase in labor incomes, particularly of earnings at the bottom, and a consistent decline in wage inequality. The contrast with other members of the BRICS (Brazil, Russia, India, China, and South Africa) group suggests different growth models at work, given those countries’ experiences such as slower employment growth (India’s “jobless growth”; Ghate 2014), high incidence of low pay (India and South Africa; OECD 2015), and/​or inequality hikes (China and India; OECD 2015). On the face of it, the Brazilian “jobs–​rich growth” appears significant. A critical element of this equality-​enhancing formalization process was the largely adhered-​to policy of valori­ zation of the minimum wage, which raised the wage floor and shrank wage inequality while defying predictions of negative employment effects (Brito, Foguel, and Kerstenetzky 2017).4 It should be pointed out that much of the mentioned favorable changes happened concomitantly in other countries in Latin America. It is, however, the unusual intensity of the reduction of informality in Brazil—​reaching a level of 36%, while the regional av­ erage still hovers around 55% (Amarante and Arim 2015)—​that arouses interest. In less developed economies, informality hardly represents an autonomous choice of individual workers facing an array of attractive options before them (OECD 2015; Tokman 2011). Informality is often associated with working poverty and inequality (e.g., the formality premium) and is more common in labor-​intensive sectors and among poorly educated and trained workers, vulnerable groups such as women and young workers, ethnic minorities, and in poorer and less developed areas (Alejo et al. 2015). Brazil is no different, with higher prevalence among women, non-​whites, less educated and less skilled workers, those living in rural areas or the poorer regions of the North and the Northeast of the country, and in agriculture, construction, and service sectors. Although the recent phase of formalization lessened this, informality remains more common among these groups. This chapter, after presenting an overview of the evolution of the labor market in Brazil, first documents the formalization process and its main consequences in terms of enhancing job quality and well-​being, and then systematizes what is known about the main driving forces behind it. The chapter also explores the connection with struc­ tural change toward a service-​driven economy in Brazil, examining the growth of service sector employment, the locus of most job creation. As this has been a global phenomenon, it is important to distinguish specific traits of the “Brazilian way,” such as an alleged bias toward the distributive services subsector. Here lies a natural con­ trast to India, where job creation is mostly being driven by the technologically supe­ rior productive services subsector (Eichengreen and Gupta 2011). Ultimately, the chapter sets out to assess the sustainability of the positive changes that have occurred, through the identification of challenges and opportunities opened up by this process. The concluding section notes that the socially progressive scenario is about to change substantially in the wake of the deepest recession the country has suffered in a cen­ tury, starting in 2015. Sometimes it can seem that Brazil is doomed to have “nostalgic longing for the future.”5

554    Celia Lessa Kerstenetzky and Danielle Carusi Machado

26.2.  Recent Changes in Brazil’s Labor Market In recent times the Brazilian labor market has undergone a favorable transformation. The role of labor market and social policies in this process—​in tandem with the com­ modity boom and subsequent export-​led growth—​cannot be overestimated. Those policies helped socially spread the new bonanza and generate demand-​driven growth impetuses of their own. From the ostensive minimum wage valorization policy (which, having started in the mid-​1990s, underwent a substantial acceleration in the 2000s), to incentives to and increasing oversight of formalization, to expansion of social security provisions—​all of these initiatives helped to boost labor market performance. To begin with, after a decade of soaring unemployment rates came a period of res­ pite, particularly in metropolitan areas, to the point that many analysts began sensing a situation of full employment and fearing work scarcity (Barbosa Filho 2015). In fact, rural and urban unemployment rates climbed from 6.4% to 9.7% during the 1990s and up to 2003, only to undergo the inverse trajectory back to 6.2% by 2012 (see Table 26.1).6 Although markedly larger for young workers, for that group it dropped to 14.6% in 2012, after having peaked at 19.3% in 2005. Additionally, employment levels rose 24% from 2002 to 2014, 20% among men and 30% among women. Overall, the participation rate climbed during some of the initial years of the first decade of the new century (2004 and 2005) before starting to reduce in 2006, ultimately declining from 65.7% to 64.2% between 2002 and 2014. But the decline was notably due to men withdrawing from the labor market (from 78.3% to 75.0%), as women were slightly increasing their participation (from 53.9% to 54.3%). The combination of a slow pace of growth of the working-​age population, following demographic transition, with reduction of economic participation has provoked some anxiety among analysts in the country, who have conjectured that the latter might be re­ flecting the former (Barbosa Filho 2015). However, while diminution of unemployment was mainly associated with employment growth (Baltar 2015) and not so much with a shrinking labor force, the diminished economic participation may alternatively be asso­ ciated with enhanced family income. In fact, while in the 1995–​2002 period, real labor income had shrunk more than 10%, it grew 60% in the following decade, thus allowing young family members to leave the labor market and turn to education. Participation rates of youth decreased 7% (from 44% to 37%, deriving both from less employment and unemployment). As with the other part of the story, the expansion of governmental transfers—​pensions and an array of social benefits—​increased household disposable income, thus helping avoid prema­ ture entrance, especially of teenagers, into the labor market. Looking beyond average earnings, low earnings as well as wage compression accompanied the positive changes. Over the recent period, low earnings approached the stat­ utory minimum wage, which in turn underwent substantial valorization, in excess of 100%, reaching the value of 70% of median earnings. (In a group of 10 key emerging economies,7 only Colombia surpassed Brazil on that count (OECD 2015).) As the valorization of the

Labor Market Development in Brazil    555 minimum wage outpaced that of average earnings, there followed a process of considerable wage compression. Thus, wage inequality declined yearly up to 2014: from 2002 to 2014, the Gini index of labor earnings shrank from .55 to .48 (see Table 26.1). Table 26.1 Selected Labor Market Indicators, Brazil (1995–​2014) Indicators 1995, 2002, 2012 and 2014

1995

2002

2012

2014

Participation rate (%)**

68.28

65.71

63.65

64.21

Women

53.61

53.93

53.12

54.26

Men

84.05

78.35

74.97

74.97

Black people

69.47

68.77

66.32

67.11

Non-​black people

67.36

67.04

65.46

65.81

Youngsters (15–​24)

65.19

63.01

59.08

59.03

Unemployment rate (%) *

5.96

9.12

6.18

6.87

Women

7.20

11.56

8.22

8.84

Men

5.11

7.30

4.61

5.33

11.45

18.04

14.62

17.03

6.54

10.33

7.08

7.76

Youngsters (15–​24) Black people Non-​black people

5.51

8.11

5.17

5.86

Employment level ***

66.20

75.31

90.60

93.40

Women

26.50

31.34

38.84

40.75

Men

39.50

43.88

51.82

52.62

White people

36.90

41.39

44.05

43.91

Non-​white People

29.30

33.92

46.72

49.45

Average labor income (R$ 2014)

1,135.69

1,056.80

1,432.58

1,684.00

Women ****

1,033.19

1,034.71

1,355.80

1,430.00

1,654.26

1,470.27

1,865.92

1,923.00

769.39

736.49

1,117.37

1,189.65

1,637.31

1,495.55

1,979.72

2,064.94

0.60

0.59

0.52

0.52

a

Men

****

Black people Non-​black people Gini index (household income) Gini index (labor income)

—​

0.55

0.48

0.48

Poverty (P0) b

35.07

34.38

15.93

13.29

Extreme povertyb

15.19

13.98

5.29

4.20

*  Unemployment rate—​Population aged 15+ (including rural North). Source: IETS. **  Participation rate—​Population aged 15+ (including Rural North). Source: IETS (1995); IBGE (2002; 2012; 2014). IETS (youngsters) ***  Employment level in millions—​Population aged 15+ (including Rural North). Source: Author’s elaboration from PNAD/​IBGE. ****  Average real monthly income in Reais, deflated by INPC/​IBGE—​principal job for employed aged 10+. Source: IBGE. a 

Average real income in reais—​principal job—​R$ October 2014—​IBGE. Note: all data regarding black and non-​black people were compiled from IETS. b 

IPEADATA. Poverty and extreme poverty lines based on caloric needs.

556    Celia Lessa Kerstenetzky and Danielle Carusi Machado Together with the minimum wage valorization policy, the contraction of education premia at all levels assisted this process (Lustig, Ortiz-​Juarez, and Lopez-​Calva 2013), as the schooling of the labor force had been increasing over time (52% had reached 11 years or more of schooling in 2014, compared to only 32.4% in 2002). In fact, despite fears, there is no robust evidence of a scarcity of high-​skilled workers, except in specific sectors (Ulyssea 2015), though the slow advancement of labor productivity suggests that years of schooling may be a misleading indicator in this respect. The relative importance of both factors, minimum wage and education, for the de­ cline of wage inequality in Brazil has been well documented (Barros et al. 2007; Brito 2015; Cunha and Vasconcelos 2012; Ferreira et al. 2006; Saboia 2007). However, some­ what counterintuitively, the preeminence of the minimum wage in explaining the de­ cline in household income inequality in Brazil over the last two decades has been a theme of recent research (Brito, Foguel, and Kerstenetzky 2017).8 The reason for this is that, as in other countries in the region, such as Colombia and Mexico, the minimum wage, in addition to being the official wage floor in the labor market, also indexes basic pensions and constitutionally defined social benefits. It is mainly through these non-​ labor-​market channels that the minimum wage affects the incomes of around 40% of the Brazilian population, most of whom live in households with per capita income below the median income (Brito, Foguel and Kerstenetzky 2017; Foguel, Ulyssea, and Corseuil 2014). An additional inequality-​reducing factor that has received some attention lately is the formalization process and institutional innovation associated with it, as discussed in the following section. Labor market socio-​demography has also improved in recent times, with shrinking gender and racial gaps, in terms of rates of participation and/​or earnings. That women are participating a bit more, and men less, may be associated with a substantially smaller gender pay gap, whose magnitude (38% in 1995, now down to 28%) has traditionally been a factor underscoring (especially, though not exclusively) poor women’s decision to stay at home as housekeepers and/​or caregivers. As for non-​whites, the expansion of schooling and the resulting contraction of educational inequalities may be a factor behind the diminished, though still huge, pay gap (non-​whites earned 49% of whites’ earnings in 2002 and 57% in 2014), as this factor traditionally dominated others to ex­ plain wage inequality between whites and blacks in Brazil (see Table 26.1). Although these wage gaps had already been shrinking in the 1990s, there is some evidence that formalization in recent times has accelerated this process (IBGE 2015). A sectoral analysis of this evolution would further show the absolute and relative diminution of agricultural and domestic employment, such that “bad jobs” contracted substantially. These became costly for employers and unattractive for the now better schooled and socially protected workers, who enjoyed better options in and out of the labor market. In addition, even in those jobs, general conditions improved, especially in terms of labor insertion in the formal economy and pay. It is true that industrial employ­ ment underwent a relative diminution, but not in absolute numbers, and formalization increased therein. The sectoral analysis is dealt with in detail in section 26.4.

Labor Market Development in Brazil    557 In any case, a factor cross-​cutting all of the positive developments is the formalization process that took place in Brazil during the last decade. This process conferred stability on the observed positive changes in the labor market mentioned thus far, in that work relations have been mediated by increasingly effective regulation.

26.3.  Formalization: Driving Forces and Beyond As previously mentioned, in less developed countries informality flourishes as a re­ sult of a lack of better options for making a living. Informality thus overlaps with poverty and underemployment, and disproportionately affects vulnerable groups. Economic growth per se has not been able to eliminate or drastically reduce it (Tokman 2011). In Brazil during the 1970s, rapid industrialization and urbanization did not generate enough regular jobs in urban areas, such that in the absence of unemployment insur­ ance (created only in 1986) excess workforce would turn to low-​paying informal activities, mainly self-​employment and work without a contract (Baltar 2015; Cacciamali 2000; Krein 2010). Famously, this was a period of elevated employment, average earnings, and wage in­ equality. During the debt crises of the 1980s, recession and structural adjustment delivered open unemployment and fueled informality. But, somewhat surprisingly, the really signifi­ cant hike in the level of informality happened in the 1990s, which saw an unprecedented in­ crease in informality, even with the positive economic growth rates between 1993 and 1997. This increase outweighed that of the recession years at the beginning of the 1980s (Ulyssea 2006), challenging cyclical explanations (Amadeo et al. 1994, cited in Ulyssea 2006). Similarly, in Latin America, during the ISI period and with burgeoning growth rates, an informal sector would cohabit with a formal one and absorb excess workforce—​so much so that open unemployment used to be relatively low. But infor­ mality expanded considerably over the 1980s and 1990s (Tokman 2014). With trade liberalization, a peculiar dynamic emerged: in the face of international competition, employment would become an adjustment variable for large employers, implying em­ ployment instability and, thus, substantial open unemployment and the spread of in­ formal work relations. These mushroomed in the wake of labor market reforms or outright unlawful practices. In the trade liberalization period, Brazilian industry was severely affected, and the informality-​intensive tertiary sector flourished (Baltar 2015; Cacciamali 2000; Ramos cited in Ulyssea 2006, 598), leading the way in employment generation. The latter’s participation in employment climbed to 59% in 2000 from 53% in 1989 (it would take 14 years for this proportion to increase another 6%; WDI 2016). But informality also increased in the manufacturing sector itself, challenging the structural change explana­ tion (Ulyssea 2006).

558    Celia Lessa Kerstenetzky and Danielle Carusi Machado Demand for informal workers increased across the whole spectrum of sectors and firms, from large to small. The interaction that was initiated among firms of different sizes illuminates changes in the composition of work demand. To begin with, employ­ ment grew in small firms while decreasing in large companies (Baltar 2015; Krein 2010). Moreover, large employers adapted to trade opening by strongly reducing formal em­ ployment, so that all new employment was either self-​employment or unregistered contracts (Baltar 2015). Overall, large companies partially cut jobs, by directly reducing the number of their lines of operation to face competition, and importing parts and components, but also displaced jobs to an array of smaller providers of industrial serv­ ices who entertained unregistered work relations (Baltar 2015). Informality and the services sector grew in tandem. In a way, the period 2003–​2014 was the mirror image of the 1990s as far as informality goes. Formalization colonized informal work relations at the firm level and presided over the creation of new job posts in the aggregate (Corseuil and Foguel 2016).9 This was assisted by an array of institutional initiatives in a time of sustained, although moderate, economic growth (Alejo et al. 2015; Baltar 2015; Sojo 2015). The background conditions of an open economy, structural change toward a service-​driven economy, and absence of major labor legislation reform were much the same. A mark of the new period is that formalization cut across economic sectors, businesses, and many different groups, so that the “level effect” generally was more im­ portant than the composition one. Thus, informality was more present among women (unregistered domestic work, mostly), blacks, youth, older workers, low-​educated workers (up to 10 years of schooling), those living in non-​metropolitan areas, those in the poorer Northeast and North regions of the country, those in lower-​productivity ec­ onomic activities (agriculture, construction, and some service sectors) (Barbosa Filho and Veloso 2016), and private-​sector workers (Alejo et al. 2015)—​but, within all of these different groups, informality (considered as workers without a contract plus the self-​ employed) substantially declined. However, a high school education gave an advantage to some workers, adding a composition effect to the level one (Ulyssea et al. 2016). Henceforth in this chapter, unless otherwise noted, informality refers to workers without social security coverage. Under this definition, informality dropped from 52.7% to 36.3% between 2002 and 2014 (see Figure 26.1). Although in absolute terms every sector except construction underwent a reduction in the number of informally em­ ployed workers, with industry taking the lead, in relative terms it declined in all of them. In agriculture, from 88.5% to 76.9%; in construction, from 70.7% to 55.5%; in the services sector, from 43.4% to 28.9%; and in industry, from 35.6% to 21.9% (Figure 26.2). Still, it is in the services sector that informality is most strongly concentrated (and within it, in distributive and personal services), to the extent that it is the employment leader (65% in 2014): in 2014, 52% of all informal jobs were in that sector (distributive and personal services accounting for 83% of this share). And since employment has increased in the services sector over the last decade, from 59% to 65%, its share in in­ formal jobs has increased as well, from 49% to 52%. The same happened, though to a smaller extent, with employment in construction, which increased its participation in

Labor Market Development in Brazil    559 0.6 0.53

0.52

0.51

0.5

0.50

0.49

0.47

0.45

0.44 0.39

0.4

0.38

0.37

0.36

2012

2013

2014

0.3

0.2

0.1

0

2002

2003

2004

2005

2006

2007

2008

2009

2011

Figure  26.1. Informal employment participation in total employment (percent):  Brazil, 2002–​2014. Source: Microdata PNAD 2002 to 2014, IBGE.

Agriculture

Industry

Construction

Services

2014

2013

2012

2011

2009

2008

2007

2006

2005

2004

2003

2002

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Figure 26.2.  Informality in each sector. Source: Microdata PNAD 2002 to 2014, IBGE.

overall employment (from 7.4% to 9.4%) and informal employment (from 10% to 14%). Industry and agriculture witnessed the inverse trend: lower participation in overall em­ ployment and informal employment. Between 2002 and 2014, formalization was characterized by a hike in instances of self-​ employment (a variation of 138%) and workers without a contract (a variation of 135%),

560    Celia Lessa Kerstenetzky and Danielle Carusi Machado both with social security, but also in jobs with a contract (a variation of 66%). Domestic workers, a category traditionally deprived of legal contracts, saw a 28% increase in the numbers of those now with contracts and social protection. It turns out that, over the last decade, social protection coverage has been dissociated from legal contracts thanks to legislative innovation that facilitated the contribution of individual workers to social insurance, even in informal units of production or inside households in the case of domestic workers. Thus, these workers were guaranteed access to usual items of social protection (e.g., old age pensions, maternity leave, and work ac­ cident and health insurance).10 The lack of a legal contract, however, still deprives such workers from access to full-​blown labor rights under Brazilian labor law (including legal work hours, paid vacations, a “13th month wage” [akin to a Christmas bonus], over­ time, access to the resolutions of collective agreements and the unemployment insur­ ance system). Thus, to better grasp the quality of the recent formalization process, it is perhaps useful to decompose the reduction of informality into two different dynamics: the ex­ pansion of legal contracts, which include access to social insurance and labor rights (or “full-​blown formalization”), and the expansion of access to social security, albeit without a legal contract (or “second-​best formalization”). In fact, both dynamics have been important, but the larger variation was in the direction of full-​blown formaliza­ tion. This was true of every economic sector, except agriculture, whose reduction of in­ formality relied more on second-​best formalization. In the services sector, whose level of informality declined from 43% to 29%, full-​blown formalization increased from 47% to 56%, and second-​best from 10% to 15%. The systematic investigation of the causes of recent formalization in Brazil records the positive impact of economic activity (Corseuil and Foguel 2016), the average size of firms (Corseuil et al. 2015), the expansion of export firms (Moura et al. 2016), but, above all, an array of institutional measures. These include the following: fiscal simplification (the SuperSimples and the micro-​entrepreneur program, MEI (Alejo et al. 2015; Corseuil et al. 2016; Sojo 2016); access to credit from public banks (Machado and Parreira 2016); expansion of credit for workers and small enterprises (Alejo et al. 2015); the law of do­ mestic servants (Alejo et al. 2015); stricter regulation of tertiarization (Alejo et al. 2015); greater social spending (Saboia 2015); increasingly effective oversight and labor justice (Alejo et al. 2015; Corseuil et al. 2016); and new public-​sector jobs to discharge the new state functions (Alejo et al. 2015). Additionally, increases in the bargaining power of workers and changes in work supply are mentioned, such as the reduction of economic participation and expansion of school enrollment of youngsters (some governmental programs such as the Bolsa Família may have helped as well), and the decrease in the population growth rate (Alejo et al. 2015). The impact of formalization on other social and economic achievements is note­ worthy. The most outstanding is on wage inequality. A  decomposition exercise has estimated this effect to be greater than those of other cleavages, such as localization, sector segmentation, or schooling, between 1999 and 2012 (Alejo et al. 2015). The total effect of −3.2 pp of Gini was decomposed into the “formalization” (change in the share of

Labor Market Development in Brazil    561 the employed who contribute to social insurance) and “formalization premium” effects (change in wage differentials between informal and formal workers), both contributing to deconcentration of wage income, but with the formalization effect dominating (−3.0 pp). In the case of education, reduction of education premia was not of sufficient magni­ tude to compensate for the concentration effect of increased years of schooling. Also in the wake of formalization, working poverty—​here defined as those earning less than the statutory minimum wage—​reduced substantially, from 26% to 20% be­ tween 2002 and 2014, even with the increase in the minimum wage. Finally, formal­ ization has been associated with greater productivity (Barbosa Filho and Veloso 2016b): 87% of the increase in aggregate work productivity between 2000 and 2009 was estimated to be due to workforce displacement from low productivity informal sectors to formal, more productive ones.

26.4.  The Connection with Structural Change: Service Sector Employment Globally, the service sector has been the main generator of new jobs (OECD 2015). While manufacturing has been losing positions in the shares of both value added and, more so, of employment, the shift to services that had begun in the 1960s in the ad­ vanced economies started to spread to less developed ones in the 1980s. This, however, happened at a lower manufacturing share of the GDP and per capita GDP than in the advanced economies. Nowadays, the sector accounts for three-​quarters—​and rising—​ of all employment in the developed world and already two-​thirds in the emerging economies of the G20 group. The service economy has soared, partially as a consequence of businesses and families deciding to not self-​serve anymore (Esping-​Andersen 1999). Thus, externalization of both industrial services, in the wake of decentralization of firms, increasing speciali­ zation, and international competition in the 1970s and 1980s, and household services, following increasing female economic participation starting in the 1970s (Elfring 1989; Esping-​Andersen 1999), stimulated an array of heterogeneous activities that would ini­ tially employ low-​skilled female and youth workers, and later prime-​age workers as well. However, additionally, the sector developed rapidly as a result of new needs generated by increasing per capita incomes and income elasticities of the demand (e.g., for arts, entertainment, recreation). Mass consumption, mass education, mass transportation in the 1960s and 1970s, globalization in the 1980s, but also an aging population with new demands in terms of health and social care, were all additional forces shaping developments of the sector in directions such as distribution of commodities, people and information, and social services. On the supply side, differential productivity levels between the service and the industrial sectors would complement the demand-​side ex­ planation for the increasing share of services employment, implying that at least some

562    Celia Lessa Kerstenetzky and Danielle Carusi Machado services seem not amenable to standardization and “industrialization.” There is even a question as to whether they should be amenable, since, as Jean Fourastié remarked (cited in Elfring 1989), low productivity in the service sector, plus the shift in demand toward services, may be one of the greatest hopes for employment in the contemporary era. This highly heterogeneous sector is usually subdivided into four subsectors—​ productive, distributive, social, and personal (and an array of activities within each of these)—​where the social concentrates up to 40% of the sectoral employment, followed mainly by the distributive and the productive. The productive is the fastest growing in terms of both value added and employment generation. Together with the social, it absorbs highly skilled workers, so that the shift to services has increased the premium to formal education (OECD 2001). The sectoral mix varies among developed countries, suggesting also that other factors are at play, such as institutional settings (the size of the welfare state, employment pro­ tection legislation, union density, collective bargaining coverage, statutory minimum wages, etc.), in addition to the distribution of workers’ characteristics in each country, such as age, gender, and skills (OECD 2001). In advanced economies, three clusters are easily identifiable:  Anglo-​Saxon coun­ tries, with a greater emphasis on the productive but larger share of the personal; Nordic countries, with a greater share of social and smaller of the personal; and Continental European countries, which follow close behind the Nordic, but with less emphasis on social services. These reflect cluster-​specific welfare systems in terms of the mix of pro­ vision by states, markets and families; the Anglo-​Saxons are characterized by stress on market services, whereas the Nordic, by leaning on large welfare states (see Table 26.2). The sectoral mix is different for emerging economies and countries in the LAC (Latin America and the Caribbean) area, where distributive services jobs are predominant, and personal services are much more important than productive ones in terms of em­ ployment opportunities. The outstanding exception seems to be urban China, where productive services already absorb 16% of services employment while social services account for the bulk of it, with over 55%. A question haunting this jobs engine refers to quality. However, according to an OECD report (OECD 2001), in comparison with the goods-​producing sector, the services sector pays higher average earnings, does not have worse working conditions or levels of job satisfaction (in fact, job satisfaction seems to be low across the board; OECD 2001), displays a skill mix similar to the manufacturing sector (Eichengreen and Gupta 2011), and, perhaps more surprisingly, has generated more high-​end than low-​ end jobs as employment has risen faster in the well-​paying productive and social serv­ ices segments. Thus, the remaining problem seems to be dualism: the services sector concentrates the “lion’s share of low-​paid and high-​paid jobs,” leaving the middle rather empty (OECD 2001). When we turn to Brazil, the improvement of labor market indicators, including for­ malization, was concomitant with the growing economic importance of services. In line with the international trend, the services sector has consistently been the Brazilian economy’s jobs engine, and is second only to construction in terms of job variation (36%

Table 26.2 Employment by Service Subsectors and Country Clusters (Percentage) Country

Productive

Distributive

Social

Personal

Anglo-​Saxons Australia

19.3%

31.9%

35.7%

13.0%

Ireland

19.9%

31.5%

35.4%

13.2%

New Zealand

21.2%

33.5%

34.3%

10.9%

United Kingdom

21.7%

28.7%

39.0%

10.6%

United States

23.2%

28.3%

36.8%

11.8%

Mean

22.7%

28.6%

37.0%

11.7%

Nordic Denmark

16.6%

30.2%

44.7%

8.5%

Finland

19.7%

30.9%

40.2%

9.3%

Norway

17.1%

30.1%

46.0%

6.8%

Sweden

21.4%

27.7%

43.2%

7.7%

19.1%

29.3%

43.6%

8.0%

Mean

Continental Europe Austria

19.7%

33.7%

35.2%

11.4%

France

19.5%

28.8%

42.8%

8.8%

Germany

20.5%

32.3%

38.8%

8.4%

Netherlands

22.0%

30.6%

39.2%

8.1%

Switzerland

25.4%

28.7%

37.1%

8.8%

20.6%

30.8%

39.9%

8.7%

Mean

LAC Argentina

13.8%

36.2%

31.9%

18.1%

Costa Rica

17.9%

38.7%

23.8%

19.5%

5.1%

24.9%

55.6%

14.4%

Mexico

11.1%

45.3%

22.5%

21.1%

Peru

10.1%

49.1%

21.8%

18.9%

Uruguay

13.5%

36.7%

28.6%

21.3%

Mean

11.2%

43.0%

25.9%

19.8%

15.5%

22.5%

55.5%

6.5%

5.9%

37.0%

33.6%

23.5%

Brazil

13.8%

37.3%

25.4%

23.5%

Mean

14.2%

29.9%

40.8%

15.1%

Cuba

Emerging Economies China South Africa

Source: Author’s own elaboration from ILO data 2015 and microdata from PNAD/​IBGE, for Brazil. Data regarding 2013 for Australia, New Zealand, Costa Rica, Mexico, Peru, Uruguay, and Korea. Data regarding 2011 for China. Data regarding 2010 for Cuba and the United States. Data regarding 2014 for the other countries.

Table 26.3 Several Indicators of Labor Market by Sectors: 2002 and 2014  

 2002

 

 

Total Employment

  N Total Distributive

 

 

 

 

 

 

 

 

 

 

 

Average Labor Income*

P90 Labor P10 Labor Income Income Gini

% Women in College % Youth Employment Degree Employment

37% 1,319.49 2,739.27 210.71

0.51 31.65

4.47

29.91

Retail and wholesale 13.10 trade

78%

1,232.00

1,200.00

80.00

0.52

37.37

4.45

32.05

Transport

3.17

19%

1,585.40

1,500.00 200.00

0.45

7.83

3.14

20.15

Communications

0.46

3%

1,985.20

1,900.00 280.00

0.48

32.88

14.44

36.33

Total Productive

5.19

12% 2,238.82 5,267.82 421.43

0.55 36.45

22.27

26.16

Business 3.41 professional services

66%

2,115.30

2,500.00 200.00

0.55

35.49

22.49

27.90

Financial services

0.75

14%

3,479.00

3,500.00 300.00

0.49

46.65

36.54

26.65

Insurance services

0.23

4%

2,457.80

2,500.00 250.00

0.47

48.37

24.02

23.29

Real estate services

0.80

15%

1,557.20

1,500.00 200.00

0.52

27.50

7.4

19.10

11.25

25%

706.52 1,264.27 126.43

0.48 74.69

1.88

27.0

Hotels, restaurants, and bars

2.80

25%

909.20

800.00

30.00

0.46

49.81

2.2

26.78

Domestic services, repair, barber, beauty, cleaning

7.34

65%

513.70

400.00

60.00

0.39

89.63

0.38

25.93

Recreation amusement

1.01

9%

1,532.10

1,800.00 100.00

0.55

37.87

11.78

34.45

Miscellaneous personal services

0.11

1%

1,084.90

1,200.00

0.52

45.66

2.9

36.13

Total Personal

Total Social

16.73

%

 

11.47

80.00

26% 1,776.31 3,792.84 421.43

0.52 62.54

25.20

19.08

2,200.00 200.00

0.52

35.61

19.92

16.63

Government proper: civil and military

3.86

34%

2,167.90

Health services

2.29

20%

2,111.30

2,100.00 200.00

0.55

74.94

28.41

20.45

Education services

4.30

37%

1,458.60

1,428.00 200.00

0.47

78.77

31.47

18.98

Miscellaneous social 1.02 services

9%

894.9

0.52

68.10

9.74

25.64

900.00

20.00

Total Services

44.65 100% 1,387.55 3,160.70 210.71

0.54 50.99

11.22

25.95

Agriculture

13.91

18%

398.67 1,053.56 126.43

0.54 33.33

0.53

27.22

Industry

11.03

15% 1,353.62 2,739.27 252.86

0.52 35.18

5.74

27.97

5.54

7% 1,018.38 1,685.70 284.46

0.44

2.37

21.65

Construction

2.65

*  R$ 2014, deflated by INPC/IBGE. ‘N’ in millions. Source: Microdata PNADs. **  Other activities are not present in the table, but in 2002 they represent less than 0.3% in 2002 and 0.1% in 2014.

 

 2014

   

 

Total Employment

 

 

 

 

 

 

 

 

 

 

 

 

 

Union Density

N

%

Average Labor P90 Labor P10 Labor % Women in College % Youth Union Income* Income Income Gini Employment Degree Employment Density

13.75

22.63

37%

1,598.9 3,000.00 500.00

0.43 35.35

7.72 24.14

13.17

10.56

17.34

77%

1,514.8

3,000.00

400.00

0.43 41.91

7.81

27.25

10.98

25.04

4.68

21%

1,831.7

3,000.00

724.00

0.38 11.21

5.32

12.51

20.01

26.58

0.61

3%

2,206.0

5,000.00

724.00

0.44 34.01

23.43

24.84

23.13

23.75

8.40

14%

2,562.8 5,000.00 724.00

0.50 43.51

31.41 22.26

19.64

19.55

6.03

72%

2,432.0

5,000.00

724.00

0.50 43.27

30.75

23.35

17.28

44.81

1.06

13%

3,757.3

8,000.00

815.00

0.48 51.33

50.32

24.52

35.75

27.18

0.23

3%

3,155.1

8,000.00

800.00

0.48 57.62

32.14

19.75

23.17

20.91

1.08

13%

1,997.4

4,000.00

724.00

0.44 34.17

16.40

14.50

16.25

4.96

14.27

23%

1,058.4 2,000.00 250.00

0.42 75.05

4.93 17.54

5.80

8.33

4.44

31%

1,186.6

2,000.00

300.00

0.39 57.70

4.29

23.98

8.95

2.34

8.47

59%

830.5

1,500.00

200.00

0.36 89.56

2.16

12.80

3.32

14.52

1.17

8%

2,190.8

4,300.00

500.00

0.51 40.69

27.21

26.07

11.82

6.02

0.20

1%

1,400.9

3,000.00

400.00

0.41 45.58

5.42

26.08

5.32

27.09

15.43

25%

2,373.0 5,000.00 724.00

0.48 64.72

39.85 13.69

24.88

26.17

5.04

33%

2,923.2

6,300.00

724.00

0.50 42.34

31.95

13.48

23.26

29.54

3.85

25%

2,494.4

5,000.00

724.00

0.50 76.34

36.58

14.90

23.08

29.70

5.73

37%

1,967.8

4,000.00

724.00

0.41 76.28

51.50

12.91

28.99

14.15

0.82

5%

1,251.2

2,500.00

200.00

0.40 67.04

22.23

14.69

14.51

16.12

60.73

100% 1,800.1 3,500.00 450.00

0.48 53.27

18.51 19.67

15.31

20.92

11.28

12%

646.7 2,000.00 160.00

0.51 31.81

1.85 16.70

27.62

20.70

12.48

13%

1,735.0 3,000.00 600.00

0.44 36.12

9.75

21.91

20.35

6.70

8.82

9%

1,478.0 2,500.00 500.00

0.37

3.47 18.75

7.60

3.28

566    Celia Lessa Kerstenetzky and Danielle Carusi Machado versus 59%). Overall, the sector accounts for 65% of all jobs, up from 59% in 2002. Also following international trends, agricultural and industrial jobs have steadily been de­ clining, the former from 18% to 12% and the latter from 15% to 13%, so that services are not only the gateway for newcomers to work, but also a destination for those who have left less job-​dynamic areas (see Table 26.3). Of course, the question is to what extent the new jobs offered quality employment. Although construction was the other significant source of new jobs, the absolute num­ bers of those that are informal increased by 25% (even if formalization increased in the sector as a whole, as the growth in the numbers of formal workers outweighed that of the informal ones) (see Figure 26.2). In any case, services’ contribution was much more significant, creating 16 million jobs, four times the number generated by the growth in construction. It turns out that the level of formalization of services sector employment was im­ pressive, above that of all the other sectors in absolute numbers; inside the sector, “bad” occupations such as domestic work underwent substantial declines (from 13% to 9.3% of services sector jobs), and social and distributive services underwent substantial reductions of informality. Of the four subsectors, it was the more recent high-​value added but smaller produc­ tive subsector, usually identified as high-​end services, that witnessed the biggest in­ crease in employment (62% and 77% specifically in business and professional services), a feat that replicates international records. But in absolute terms, the greatest contribu­ tion came from the huge distributive sector (and from wholesale and retail activities mostly), which in 2014 had 6 million more job posts than in 2002, though social serv­ ices, coming second as the services sector’s jobs generator, in 2014 had 4 million more (see Table 26.3). When it comes to income, the average earnings of the services sector rose above that of all others, except industry, whose average level was similar. Its variation over the decade (30%), however, was somewhat below the average (38%), which was influenced by the improved performances of both agriculture (62%) and construction (45%) (see Table 26.3). Personal services witnessed the largest positive variation, 50%, probably as a consequence of being the activity most sensitive to the variation of minimum wage (see Table 26.3). Generally speaking, services jobs are more affected by the minimum wage since the incidence of workers earning it is the highest in that sector (11.7% versus 9% in manufacturing, 7.5% in construction, and 7.4% in agriculture). However, after manufacturing, services has the lowest level of working poverty (i.e., those earning less than the statutory minimum wage), around 15% (against 65% in agriculture, 16% in con­ struction, and 12% in industry). Overall, more than 73% of the workers in the services sector earn above the minimum wage (79% in industry, 76.5% in construction, but 27.6% in agriculture)—​ranging from 86% in the productive subsector to 63.6% in the personal one. It is nonetheless remarkable that though working poverty is concentrated in agri­ culture, the largest amount of the working poor is still in the services sector: 9 million people, half of whom work in personal services.

Labor Market Development in Brazil    567 The personal services subsector, despite displaying the largest degree of informality (52% in 2014)  and the lowest average income, has witnessed substantial favorable changes on both counts. Productive and social services, on the other hand, had milder earnings improvements but still boast much higher average wages (twice the personal services average) and low informality levels (respectively, 17.6% and 9.8%). An additional income measure of job quality is labor income dispersion. Traditional services sector dualism has been mitigated somewhat. As in every other sector, the Gini coefficient and the 90/​10 ratio have declined consistently over the last decade, while the income of those in the first decile of the labor income distribution (P10) more than doubled in real terms from 2002 to 2014, especially in the low-​end services of the dis­ tributive and personal subsectors. Although services has the second highest Gini (.48), after agriculture (.52), it underwent the second deepest decline (0.07 points), after in­ dustry (0.08 points), where the distributive subsector was responsible for most of the reduction, with personal services following. Industry and distributive services led the decline in the 90/​10 ratio, whereas personal still displayed the highest level (see Table 26.3). Inside the services sector, distributive and personal, besides having lowest average earnings, are the less unequal in Gini indices, but also have the lowest value of low (P10) and high (P90) pays, so that less inequality reflects leveling at low levels. Productive and social, on the other hand, are the more unequal but have the highest average earnings, P10, and P90 remunerations. In comparison with manufacturing, social and productive services pay higher remuneration to workers at the low end. Table 26.4 summarizes relative positions in terms of low pay and wage inequality of all the sectors, taking into account the noted strong heterogeneity inside the services sector. The first value in parentheses refers to the income level at the first decile (i.e., of the monthly income in Brazilian reais of the person occupying this position in the dis­ tribution), and the second value refers to the Gini index of the sector/​subsector, both in 2014. Distribution, personal, and construction form a cluster of low inequality of low P10 income levels, while industry and social make a group of low inequality with rather high levels of P10 income. This suggests that income dualism in the services sector is somehow mitigated by the presence of social services. In terms of workers’ characteristics, the presence of women draws attention; they predominate in the services sector as a whole, but with greater emphasis inside the

Table 26.4 Low Pay (P10) and Gini Indices of Economic Sectors: Brazil, 2014 High inequality (Gini)

Low inequality (Gini)

High low-​pay (R$)

Productive (724, .507)

Industry (600, .441), Social (724, .482)

Low low-​pay (R$)

Agriculture (160, .517)

Distributive (500, .431), Personal (250, .423), Construction (500, .378)

Source: Microdata from PNAD (2014).

568    Celia Lessa Kerstenetzky and Danielle Carusi Machado social and personal subsectors (where they represent three-​quarters of the work­ force). Youth are not especially prevalent, and their participation has been declining substantially, not only in the services sectors, but in every other sector as well. Labor force schooling has been on the rise in Brazil, but the service sector contains the lowest—​and declining—​representation of low-​educated workers (0–​10  years of schooling), even considering its internal variation, while it shares with industry the largest proportion of 11–​14 years of schooling, but exceeds industry in terms of workers with some college education or a college degree (its share is two times that of industry). In the latter case, the outstanding performance comes from the so­ cial (especially) and the productive subsectors. Overall, it is the services sector that concentrates the workforce with the highest years of schooling on average. As for participation of non-​whites, while the sector corresponds on average to their overall participation among the population, non-​whites are overrepresented in the personal and distributive services subsectors, and whites in the social and, especially, the pro­ ductive subsectors. It is also notable that employment of non-​whites has increased in all of the sectors. Other aspects that might reflect the relative strength of employment institutions are union density, average hours worked per week (AHWPW), and the incidence of long working hours (over 48 hours a week). Union density is low in Brazil and has been declining, except in agriculture (where it has risen from 21% to 27.6%) and, somewhat, in construction—​but in any case not enough to compensate for the near stagnation in industry and in the services sector. In services, already the least union­ ized sector (in line with the global norm), the productive and social subsectors have suffered important declines, whereas personal has seen a surge, but from a meager baseline (from 4.9% to 5.8%) (see Table 26.3). But social services workers still make up, after agriculture, the most unionized group, at 25%. The figures seem to suggest that union density is not a credible explanatory factor with regard to the employment improvements of late. This notwithstanding, AHWPW have been declining in Brazil, varying from 34.2 hours in agriculture to 41.5 in industry, while services display the second highest av­ erage, 39.5 hours. In addition, long work hours have abruptly dropped across the all sectors similarly, from 18.2% in industry to 22.2% in services. But the distributive serv­ ices subsector still records the longest AHWPW (43 hours) and the greatest propor­ tion of workers with long work hours (31.7%), followed by personal (though at some distance: 25%). The productive services subsector has a greater share of workers in these conditions (16%) than social services (9%). It may well be that this set of employment quality dimensions, however mediated by specific features of sectors and subsectors, has responded to the advance of registered work. All in all, it seems that the surge in employment in the services sector in Brazil in the wake of the general improvement of the labor market has not detracted from job quality. The jobs generated were mostly formal ones, they were paid the highest average earnings (which had been increasing in the last decade), and they displayed, following a substantial

Labor Market Development in Brazil    569 reduction, the second lowest degree of working poverty. Most of these achievements happened with greater intensity at the lower end of the sector, namely in distributive and personal services. Moreover, although the second most unequal in terms of wage dispersion, the sector has seen inequality reduce throughout the period considered. In terms of indicators of presence and effectiveness of employment institutions, the sector as a whole, while having the lowest level of union density,11 does display AWHPW near the economy’s average and a share of workers in long work hours not much larger than one might have expected, with only the distributive subsector being over the average (21.4%). Bearing in mind the international experience summarized earlier in this section, part of the investigation of job quality in the services sector should turn to the country’s rela­ tive position in terms of the sectoral mix and trends within it. Thus, as in some emerging economies, the distributive subsector predominates as the biggest employer within services (24.4%). This clearly contrasts with the predom­ inance of social services within the services sector in every cluster and country in the advanced economies, but also in some emerging economies, such as Argentina and Uruguay in the Latin American region, and China (see Table 26.4). Moreover, this subsector’s share climbed somewhat in the recent period. It contains the largest share of young workers considering all sectors. Second, personal services, also in line with other emerging economies, predominate over productive in a proportion of two jobs to one, even if its share shrank a little by 15.3% in 2014. Also, while in advanced economies food and lodging predominate within the personal services subsector, in Brazil it is domestic services (59.3% of its employment in 2014). Thus, the worst-​performing subsectors are still in force. When it comes to productive services, though this has been the subsector with the fastest pace of employment growth in Brazil, not only has it re­ corded a meager share in employment (9% in 2014; for comparison, the same level of advanced economies’ shares back in 1985), but also its growth speed is well below that in advanced economies when this subsector began to take off during the 1960s, 1970s, and 1980s (data in Elfring 1989, for 1960–​1973 and 1973–​1985). As for social services, the proportion of employment (16.5% in 2014) dropped somewhat, though not the absolute numbers. However, again, this runs counter to the development trajectory of advanced economies in their postwar takeoff. To give a rough idea of missed strategic opportunities, it is useful to compare Brazil and advanced economies in terms of shares of employment in productive and so­ cial services, for the year 2014 (the “share gap”), as well as the respective employment growth rates in these subsectors in two different periods (the “variation gap”). With the latter gap, the average growth speed between 1973 and 1985 in advanced economies is compared with that in Brazil between 2002 and 2014, as in 1985 advanced economies had a similar share of service employment as Brazil in 2014.12 It turns out that the share gap hovers around 12 percentage points in the productive and over 21 percentage points in the social subsectors, while the variation gap is 0.7 percentage points in the produc­ tive and 4.7 percentage points in the social.

570    Celia Lessa Kerstenetzky and Danielle Carusi Machado Again, the lack of emphasis on social and productive in favor of distributive and per­ sonal subsectors signals lack of strategy: these are employment intensive sectors with prime quality jobs in terms of insertion, earnings, income dispersion, and mitigated du­ alism (the social services in particular). In the productive dwells the core of the knowl­ edge economy and tradable services; in the social, much of the human resources inputs for the latter to happen, including education and health services. Inside the social sector, particularly, the biggest employers are education and health, and they expanded employment in the period considered here, especially the health sector. The segmentation between public and private provision further discloses other strategic opportunities (data not shown here13). Hence, although the biggest employer in social sectors was and still is public education, education (both public and private) lost relative position, while health gained substantially, and the private branch was and remains a more important employer within it. Generally speaking, all of these activities’ jobs, as in every other sector, saw improvements in terms of formalization, average earnings, Gini indices (except for public health), schooling, female and non-​white participation, low pay, and working poverty. But some differences are noticeable. To begin with, employment in the public sector is substantially more formalized, both in education and health, average earnings improvements were more intensely felt in the public sector (the lowest level is in private education), inequality is some 5 Gini percentage points smaller in the public sector, and working poverty is low (half of the proportion displayed in the pri­ vate sector). Last but not least, union density is much more significant in the public than in the private sector. As for workers’ characteristics, the share of those with at least some college education is higher in public education than in private, and higher in private health care than in public. The prevalence of female workers is more a mark of public education (almost 80% of the workers in this activity) than of private, whereas no noticeable difference seems to exist between public and private health care on that count (both around 76%). Also, non-​white workers are more common in the public sector of education and health, and their participation has been on the increase. In spite of overall advantages of the public sector when it comes to job quality in educa­ tional and health-​care activities, the last decade does not exhibit a decisive bias toward the growth of these employment opportunities. Making up over half of social services subsector jobs, they have stayed put (in the case of public education) or enjoyed modest expansion (in the case of public health, considering that the private sector employs 26% more workers than the public) in view of the constitutional mandate of public health care provision.

26.5.  Concluding Remarks The Brazilian labor market had already begun to decelerate by 2014, only to contract substantially in the following two years when a huge recession claimed 7.7% of GDP.

Labor Market Development in Brazil    571 Although the causes of recession may be multiple—​a combination of international crises and the end of the commodity boom, economic cyclicality, domestic policy mismanagement, lack of a clear growth strategy, and political strife, still to be objec­ tively assessed—​its consequences are already affecting many of the previous decade’s achievements. Data from quarterly surveys (the PNAD Contínua from IBGE; see Table 26.5) show employment contracting in the last two quarters of 2015 and first quarter of 2016, amounting to 2% less employment than in early 2015. The participation rate climbed during most of 2015 and the first quarter of 2016, reflecting loss of family income and engagement of new (and young) family members in the labor force. In the end, the con­ sistent trajectory of diminishing unemployment rates was reversed in the beginning of 2015, reaching peak levels of 11% (24% for youth) in 2016. Labor earnings that had already stopped growing in 2014 began to fall in 2015, so much so that the first 2016 quarter registers a 3% decline in relation to the two previous years. Lower earnings re­ flect in part the shift to less secure forms of occupation, as self-​employment rose for the first time in 12 years, even while contributions to social security are still on the rise. Even if services sector employment has shown the greatest resilience, still generating employ­ ment opportunities when all of the other sectors have been shrinking, signs of recovery in the near future are unclear.

Table 26.5 Labor Market Indicators, 2014–​2016

Quarters

Unemployment Rate (%)

Unemployment Rate (%)—​ Average Youngsters * Earnings**

Formality ***

Informality

1st Quarter, 2014

7.2%

15.8%

1,976.00

64.3%

35.7%

2nd Quarter, 2014

6.8%

15.3%

1,939.00

64.6%

35.4%

3rd Quarter, 2014

6.8%

15.3%

1,940.00

64.7%

35.3%

4th Quarter, 2014

6.5%

14.1%

1,959.00

64.7%

35.3%

1st Quarter, 2015

7.9%

17.6%

1,975.00

64.9%

35.1%

2nd Quarter, 2015

8.3%

18.6%

1,963.00

64.9%

35.1%

3rd Quarter, 2015

8.9%

19.7%

1,937.00

64.6%

35.4%

4th Quarter, 2015

9.0%

19.4%

1,909.00

65.7%

34.3%

1st Quarter, 2016

10.9%

24.1%

1,914.00

65.7%

34.3%

Note: Data refer to population aged 14+. Source: Data from SIDRA/​IBGE. Authors’ own elaboration. *  Youth are defined as population aged 18–​24. **  Earnings are defined as the average real monthly labor income in Reais for population aged 14+, R$ for February/​2016 ***  Defined as the population contributing to the pension scheme.

572    Celia Lessa Kerstenetzky and Danielle Carusi Machado In any case, this uncertainty about the future should not prevent us from briefly assessing the preceding golden years of the Brazilian labor market—​in which formal­ ization led by services sector jobs growth was the hallmark—​and from identifying in­ tervenient factors and remaining challenges. The evidence seems to suggest that it was not structural change per se, but rather the way it was mediated by labor market policies that did the trick, as the 1990s had shown this sector’s employment increasing in tandem with growing informality. But this cannot be the whole story. Generally speaking, pre-​ distribution interventions are very important for generating economic returns with greater social equilibrium and thus, to some extent at least, precluding the need for extensive redistribution. In this sense, the Brazilian option has leaned toward regulatory policies (and also social transfers to the working poor, not discussed here) with detriment to service provision to upgrade the labor market. These policies were no doubt important to raise the floor both in terms of better pay and working conditions, but also to price out of the market “bad jobs” such as domestic services. But more appears necessary for the flourishing of the potential of the Brazilian labor market to provide good jobs and well-​being. Thus far, the devel­ opment of the services sector labor market indicates the need for a better balance that prioritizes employment in the high-​quality areas of productive and social services. The latter, in addition to smoothing dualism, is the potential source of inputs in the areas of education and health care (and also social care, not discussed here, but important for women’s economic participation), for the general development of labor market and social welfare. Moreover, it seems in any case important to hold onto the gains in terms of better in­ sertion and social protection for those workers who for one reason or another end up being employed in the personal or distributive subsectors (for many, especially youth, still an important gateway to the urban labor market). In addition, active labor market policies may enable such workers to move on to more attractive occupations. These, though critical, are undeveloped in Brazil.

Notes 1. The authors thank José Antonio Ocampo for detailed comments on a previous version of this chapter, as well as Rodolfo Hoffmann and other participants of the USP workshop for helpful suggestions. We are also grateful to Graciele Guedes for excellent assistance with data. The usual disclaimer applies. 2. Data from IPEADATA. 3. Informality is here calculated as the proportion of workers who do not contribute to social security. But also when measured as the proportion of workers without a formal contract and self-​employed—​the other common definition—​it shrank substantially in this period. 4. Over the 1995–​2014 period, the minimum wage more than doubled in real terms following a policy of recuperation of its past value; most of the increase happened between 2004 and 2014. As we explain in section 26.1, valorization impacts not only labor income but also non-​labor income such as basic pensions.

Labor Market Development in Brazil    573 5. The expression in Portuguese (saudades do futuro) is from Clarice Lispector’s poem Saudades. It was recollected by Gilberto Hochman in private conversation with one of the authors. 6. Data compiled by the authors from the Pesquisa Nacional por Amostra de Domicílio (PNAD—​National Household Sample Survey), IBGE. 7. This includes Brazil, Chile, China, Colombia, India, Indonesia, Mexico, The Russian Federation, South Africa, and Turkey. 8. Saboia (2007) had already found this result for the period 1995–​2005. 9. This happened in spite of allegedly elevated labor costs in Brazil (Pastore 1996). The issue has sparked methodological controversy in the country (see Pastore 1996 and Pochman 1994). 10. Individual contributions to social security were also allowed for unpaid domestic work, work for self-​consumption, and young students. 11. It should be noted that, at any rate, low union density is a reality throughout all sectors in Brazil. 12. The data for the “share gap” is from Table 26.4, whereas that for the “variation gap” is from Elfring (1989), where the following countries were considered: France, the United Kingdom, the United States, Japan, Sweden, Germany, and the Netherlands. 13. Data available from the authors.

References Alejo, Javier, Guillermo Cruces, and Soledad Parada. 2015. “Desigualdad e informalidad en América Latina: El caso del Brasil.” In Desigualdad e informalidad: Un análisis de cinco experiencias latino-​americanas, edited by Verónica Amarante and Rodrigo Arim, 129–​162. Libros de la CEPAL, No. 133 (LC/​G.2637-​P). Santiago de Chile: Comisión Económica para América Latina y el Caribe (CEPAL). Baltar, Paulo. 2015. “Crescimento da economia e mercado de trabalho no Brasil.” Texto Discussão No. 2036. Brasília: IPEA, February. Barbosa Filho, Fernando de Holanda. 2015. “Mercado de trabalho no Brasil: Passado, presente e futuro.” In Panorama do mercado de trabalho no Brasil, edited by Regis Bonelli, and Fernando Veloso, Chapter 2, 35–​54. Rio de Janeiro: Fundação Getúlio Vargas. Barbosa Filho, Fernando de Holanda, and Fernando Veloso. 2016a. “Fatos estilizados da informalidade do trabalho no Brasil.” In Causas e consequências da informalidade no Brasil, edited by Fernando de Holanda Barbosa Filho, Gabriel Ulyssea, and Fernando Veloso, 3–​34. Rio de Janeiro: Elsevier/​FGV IBRE. Barbosa Filho, Fernando de Holanda, and Fernando Veloso. 2016b. “A contribuição da formalização para a elevação da produtividade do trabalho no Brasil nos anos 2000: Uma análise exploratória.” In Causas e consequências da informalidade no Brasil, edited by Fernando de Holanda Barbosa Filho, Gabriel Ulyssea, and Fernando Veloso, 303–​325. Rio de Janeiro: Elsevier/​FGV IBRE. Barros, Ricardo Paes de, Samuel Franco, and Rosane Mendonça. 2007. “A recente queda na desigualdade de renda e o acelerado progresso educacional brasileiro da última década.” In Desigualdade de renda no Brasil: Uma análise da queda recente, edited by Ricardo Paes de Barros, Miguel Nathan Foguel, and Gabriel Ulyssea, Chapter 26, 305–​342. Brasília: IPEA. Brito, Alessandra Scaloni. 2015. “O papel do salário mínimo na redução da desigualdade na distribuição de renda no Brasil entre 1995 e 2013.” PhD dissertation. Economics, Universidade Federal Fluminense, Niterói.

574    Celia Lessa Kerstenetzky and Danielle Carusi Machado Brito, Alessandra, Foguel, Miguel, and Kerstenetzky, Celia. 2017. “The contribution of min­ imum wage valorization policy to the decline in household income inequality in Brazil: A decomposition approach.” Journal of Post Keynesian Economics, 40 (4): 540–​575. Cacciamali, M.C. 2000. “Globalização e processo de informalidade.” Economia e Sociedade 9 (1): 153–​174. Camarano, Ana Amélia, and Solange Kanso. 2012. “O que estão fazendo os jovens que não estudam, não trabalham e não procuram trabalho?” In Boletim de mercado de trabalho: Conjuntura e análise (53), 37–​44. Rio de Janeiro: IPEA. Camargo, José M., and Edward Amadeo. 1996. “Instituições e o mercado de trabalho no Brasil.” In Flexibilidade no mercado de trabalho no Brasil, edited by José M. Camargo, 47–​94. Rio de Janeiro: Fundação Getúlio Vargas Editora. Costa, Joana Simões de Melo, and Gabriel Ulyssea. 2014. “O fenômeno dos jovens Nem-​Nem.” In Desafios à trajetória profissional dos jovens brasileiros, edited by Carlos Henrique Corseuil and Rosana Ulhôa Botelho, 115–​137. Rio de Janeiro: IPEA. Corseuil, Carlos Henrique, and Miguel Foguel. 2016. “Expansão econômica e aumento da formalização das relações de trabalho: Uma abordagem através das contratações.” In Causas e consequências da informalidade no Brasil, edited by Fernando de Holanda Barbosa Filho, Gabriel Ulyssea, and Fernando Veloso, 131–​152. Rio de Janeiro: Elsevier/​FGV IBRE. Corseuil, Carlos Henrique, Rodrigo Leandro de Moura, and Lauro Ramos. 2015. “Expansão do emprego formal e tamanho das empresas:  1995–​2011.” In Panorama do mercado de trabalho no Brasil, edited by Regis Bonelli, and Fernando Veloso, Chapter 7, 149–​168. Rio de Janeiro: Fundação Getúlio Vargas. Corseuil, Carlos Henrique, Marcelo Neri, and Gabriel Ulyssea. 2016. “Uma análise exploratória dos efeitos da política de formalização dos microempreendedores individuais.” In Causas e consequências da informalidade no Brasil, edited by Fernando de Holanda Barbosa Filho, Gabriel Ulyssea, and Fernando Veloso, 209–​235. Rio de Janeiro: Elsevier/​FGV IBRE. Cunha, Marina Silva da, and Marcos Roberto Vasconcelos. 2012. “Evolução da desigualdade na distribuição dos salários no Brasil.” Economia Aplicada 16 (1): 105–​136. Eichengreen, B., and P. Gupta. 2011. “The Service Sector as India’s Road to Economic Growth?” India Policy Forum, National Council of Applied Economic Research 7 (1): 1–​42. Eichengreen, Barry, and Poonam Gupta. 2009. “The Two Waves of Service Sector Growth.” NBER Working Paper Series. Working Paper No. 14968. Cambridge. Elfring, Tom. 1989. “New Evidence on the Expansion of Service Employment in Advanced Economies.” Review of Income and Wealth, 35 (December): 409–​440. Esping-​Andersen, Gøsta. 1999. Social Foundations of Postindustrial Economies. Oxford: Oxford University Press. Ferreira, Francisco H. G., Sergio Firpo, and Julian Messina. 2014. “A More Level Playing Field? Explaining the Decline in Earnings Inequality in Brazil, 1995–​2012.” Working Paper No. 12. International Research Initiative on Brazil and Africa (IRIBA). Foguel, Miguel, Gabriel Ulyssea, and Carlos Henrique Corseuil. 2014. “Salário mínimo e mer­ cado de trabalho.” In Brasil em Desenvolvimento: Estado, Planejamento e Políticas Sociais, Vol. 1, edited by Leonardo Monteiro Monasterio, Marcelo Côrtes Neri, and Sergei Suarez Dillon Soares, 295–​324. Rio de Janeiro: IPEA. Ghate, Chetan, ed. 2012. The Oxford Handbook of the Indian Economy. Oxford:  Oxford University Press. IETS. Parâmetros e resultados da PNAD 2014. Available at http://​iets.org.br.

Labor Market Development in Brazil    575 International Labor Organization. Available at http://​www.ilo.org/​global/​statistics-​and-​ databases/​lang-​-​en/​index.htm. Instituto Brasileiro de Geografia e Estatística. PNAD. Available at http://​www.ibge.gov.br/​ home. Instituto Brasileiro de Geografia e Estatística, ed. 2015. Síntese de indicadores sociais 2015: Uma análise das condições de vida da população brasileira. Rio de Janeiro: IBGE. Instituto de Estudos do Trabalho e Sociedade (IETS). Available at http://​iets.org.br/​. Instituto de Pesquisa Econômica Aplicada. Available at http://​ipeadata.gov.br/​. Kerstenetzky, Celia Lessa. 2014. “The Brazilian Social Developmental State:  A Progressive Agenda in a (Still) Conservative Political Society.” In The End of the Developmental State?, edited by Michelle Williams, 172–​196. New York: Routledge. Krein, José Dari, and Marcelo Weishaupt Proni. 2010. “Economia informal: Aspectos conceituais e teóricos.” Série Trabalho decente no Brasil: Escritório da OIT no Brasil. Available at: http://​ www.oitbrasil.org.br/​sites/​default/​files/​topic/​employment/​pub/​economia_​informal_​241.pdf. Lustig, Nora, Luis F. Lopez-​Calva, and Eduardo Ortiz-​Juarez. 2013. “Declining Inequality in Latin America in the 2000s: The Cases of Argentina, Brazil, and Mexico.” World Development 44: 1558–​1569. Machado, Luciano, Maria Araujo Parreiras. 2016. “Efeito do Cartão BNDES no empregos formal de empresas de menor porte.” In Causas e consequências da informalidade no Brasil, edited by Fernando de Holanda Barbosa Filho, Gabriel Ulyssea, and Fernando Veloso, 237–​ 252. Rio de Janeiro: Elsevier/​FGV IBRE. Menezes Filho, Naercio A., Pedro Henrique Fonseca Cabanas, and Bruno Kawaoba Komatsu. 2015. “A condição ‘nem-​nem’ entre os jovens é permanente?” In Panorama do mercado de trabalho no Brasil, edited by Regis Bonelli, and Fernando Veloso, Chapter 4, 97–​122. Rio de Janeiro: Fundação Getúlio Vargas. Moura, Rodrigo Leandro de, and Fernando Holanda de Barbosa Filho. 2015. “A queda da informalidade no Brasil entre 2002 e 2012.” In Panorama do mercado de trabalho no Brasil, edited by Regis Bonelli, and Fernando Veloso, Chapter 6, 125–​148. Rio de Janeiro: Fundação Getúlio Vargas. Moura, Rodrigo Leandro de, Carlos Henrique Corseuil, and Lauro Ramos. 2016. “Evolução do tamanho médio das empresas formais no período 1999–​2011: Resultados incorporando a sua dimensão exportadora.” In Causas e consequências da informalidade no Brasil, edited by Fernando de Holanda Barbosa Filho, Gabriel Ulyssea, and Fernando Veloso, 153–​168. Rio de Janeiro: Elsevier/​FGV IBRE. OECD. 2001. OECD Employment Outlook 2001. Paris: OECD Publishing. OECD. 2015. OECD Employment Outlook 2015. Paris: OECD Publishing. Available at http://​ dx.doi.org/​10.1787/​empl_​outlook-​2015-​en. OECD, ILO, and the World Bank. 2014. “G20 labor Markets: Outlook, Key Challenges and Policy Responses.” G20 Labor and Employment Ministerial Meeting, Melbourne, Australia, September 10–​11, 2014. Pastore, José. 1996. “A batalha dos encargos sociais.” Folha de São Paulo, February 28, 1996. Pochmann, Márcio.1994. “Cálculo dos componentes dos custos do trabalho nos 10 setores mais importantes da economia dos Países do Mercosul.” Campinas: UNICAMP/​CESIT. 57 p (Working Paper). Saboia, João. 2007. “Efeitos do salário mínimo sobre a distribuição de renda no Brasil no período 1995/​2005: Resultados de simulações.” Revista Econômica (UFF) (2): 270–​295.

576    Celia Lessa Kerstenetzky and Danielle Carusi Machado Saboia, João. 2015. “Evolução recente do mercado de trabalho metropolitano e perspectivas.” In Panorama do mercado de trabalho no Brasil, edited by Regis Bonelli, and Fernando Veloso, Chapter 3, 55–​74. Rio de Janeiro: Fundação Getúlio Vargas. Sojo, Ana. 2015. “Including Informal Economy Workers in Contributory Social Protection: Current Challenges in Latin America.” International Social Security Review 68 (4): 69–​92. Souza, André Portela, Vladimir Ponczek, and Eduardo Zylberstajn. 2015. “Transições no Mercado de trabalho.” In Panorama do mercado de trabalho no Brasil, edited by Regis Bonelli and Fernando Veloso, Chapter 8, 169–​180. Rio de Janeiro: Fundação Getúlio Vargas. Tokman, Victor. E. 2011. “Employment:  The Dominance of the Informal Economy.” In The Oxford Handbook of Latin American Economics, edited by José Antonio Ocampo and Jaime Ros, 767–​789. Oxford: Oxford University Press. Ulyssea, Gabriel. 2006. “Informalidade no mercado de trabalho brasileiro: Uma resenha da literatura.” Revista de Economia Política 26 (4): 596–​618. Ulyssea, Gabriel. 2015. “Panorama do mercado de trabalho brasileiro: Uma análise de temas recentes.” In Panorama do mercado de trabalho no Brasil, edited by Regis Bonelli and Fernando Veloso, Chapter 4, 75–​96. Rio de Janeiro: Fundação Getúlio Vargas. Ulyssea, Gabriel. 2016. “Firmas, informalidade e os impactos de políticas para a formalização.” In Causas e consequências da informalidade no Brasil, edited by Fernando de Holanda Barbosa Filho, Gabriel Ulyssea, and Fernando Veloso, 279–​302. Rio de Janeiro: Elsevier/​ FGV IBRE.

Chapter 27

Environmenta l  I s su e s Ariaster B. Chimeli

27.1. Introduction Brazil is the ninth-​largest economy in the world, with a GDP of US$1.7 trillion and gross domestic product (GDP) per capita of US$8.500, roughly 85% of the world av­ erage.1 This places Brazil as an upper-​middle-​income country, according to the World Bank. Despite declining income inequality in recent years, income remains rather con­ centrated, with the richest 20% receiving 56% of national income and the poorest 20% receiving only 3.6% (World Bank poverty data). To put income inequality in perspective, Brazil ranks as the 16th most unequal country in the world with a Gini coefficient of 51.9, whereas the United States ranks 43rd (Gini coefficient of 45), and the United Kingdom 107th (Gini coefficient of 32.4). These figures can serve as a starting point for our charac­ terization of the environment and development scenario in Brazil. Given the upper-​middle-​income country status that Brazil enjoys, one might argue that improvements in environmental indicators are to be expected. This line of reasoning implicitly or explicitly invokes the “environmental Kuznets curve” (EKC) hypoth­ esis, an income-​pollution relationship postulating that as income increases, pollution increases, reaches a peak, and then eventually begins to decrease. Furthermore, income inequality might suggest that the poorest might be the most vulnerable to the negative consequences of environmental degradation, since wealthier citizens might have the means to avoid pollution damages. These conjectures are, however, far from obvious in the Brazilian context. Wealthy regions of the country still experience serious environmental problems. The Tietê and Pinheiros rivers in the city of São Paulo, the richest city in the country, are heavily polluted and the water unsuited for consumption. This remains the case de­ spite the water shortage crisis the city experienced in 2014 and 2015, the recurrence of which remains a threat. The Guanabara Bay in the Rio de Janeiro metropolitan area is another pollution problem that several governments have pledged to tackle, but failed to solve. Water pollution in Rio de Janeiro gained international visibility during the 2016

578   Ariaster B. Chimeli Summer Olympic Games, with claims that some athletes became sick due to exposure to pollution. Finally, in November 2015, perhaps one of the world’s largest environmental disasters took place in the state of Minas Gerais as a waste retention dam for a mining enterprise broke down, destroying an entire village, killing 19 people, and damaging over 650 kilometers. Still, with respect to the wide income inequality in the country, it is not obvious that the richest are insulated from the consequences of environmental degradation. The wealthiest regions of the country are also the ones that experience high levels of air pollution (Dasgupta et al. 2002), and might be subject to poor water quality during shortage periods. Nevertheless, lower income families are likely to be disproportion­ ately affected even in these wealthier regions, especially due to limited access to sanita­ tion and health services. Although environmental degradation is still an important and often neglected and misunderstood problem in the country, significant advances have taken place in recent years. Most notably, the rate of deforestation in the Amazon region has fallen sharply since 2004. The state of São Paulo, the largest producer of sugarcane in the country, has successfully implemented measures to reduce air pollution during harvest season. For decades now, a number of states have implemented a tax rebate system that rewards municipalities that protect the local environment. Water basin associations around the country have produced positive results with respect to the management of water resources. Environmental data collection efforts still have a long way to go, but some progress has occurred, especially with respect to the monitoring of deforestation. In the case of urban air quality, however, data are sparse, geographically concentrated, and not consistently collected. At the heart of the environmental reality in Brazil is the building up of institutions to protect the environment. Since the inclusion of a chapter on environment in the Brazilian constitution in 1988, a number of institutions have gained strength and im­ portance. The Ministry of the Environment has been a consolidated part of recent governments and has been a player in international environmental negotiations. The Brazilian Institute for the Environment and Renewable Natural Resources (IBAMA) implements national environmental policies, works as a monitoring agent, applies penalties (although enforcement is oftentimes weak), and has the power to authorize or embargo activities depending on their environmental impact. Civil society has grown increasingly engaged in a number of environmental protection actions, ranging from animal rights movements to the preservation and recovery of degraded rainforests.

27.2.  Environment and Development Insights from the environment and development literature provide us with a useful benchmark for the Brazilian case. The state of the environment has well-​documented impacts on economic development through its influence on individuals’ health, their

Environmental Issues   579 ability to generate income, and the direct impact on their well-​being. The environment is also intimately connected to the profitability of firms that use natural resources and services in their production processes. This argument applies both to developed and, perhaps even more so, to developing countries.2 Despite the importance of the environment to economic development and human well-​being, and the potentially large gains yielded by environmental improvements, environmental quality is often very poor in developing countries. This observation led Greenstone and Jack (2015) to identify what they believe is the central question for the environment and development literature: Why is environmental quality so poor and in­ dividual willingness to pay for improvements in environmental quality so low in devel­ oping countries? They offer four main reasons why this is the case. I use these reasons as a departure point in this chapter. The first two reasons, high marginal utility of consumption and high marginal cost of environmental improvements, refer to an optimal path for the environment-​ development nexus. The other two reasons, political economy and market failures, appeal to the development literature in stressing the market failures that are common and important in developing countries. The balance between the marginal utility of consumption and marginal abatement cost as a driver for the environment-​development path is present in a number of the­ oretical works on the environmental Kuznets curve (e.g., Chimeli and Braden 2009; Selden and Song 1995; Stokey 1998). Meanwhile, political economy factors and market failures such as missing markets, incomplete information, and corruption may affect total factor productivity, pollution intensities, and the effectiveness of environmental spending. These economy-​specific parameters can then determine the path for environ­ mental quality as economies grow.3 The first reason for low environmental quality and low willingness to pay for en­ vironmental improvements in developing countries postulated by Greenstone and Jack is the high marginal utility of consumption when incomes are low. That is, at low levels of income, the return from consumption dominates the return from environ­ mental protection spending. This factor is likely to be at play in lower-​income parts of Brazil, such as the Amazon region, where deforestation and ecosystem simplification may be associated with key activities for the local economy. In several instances where the federal government attempted to crack down on illegal logging in the Amazon region, unemployment increased in certain towns and an immediate local political reaction took place.4 However, the high marginal utility of consumption argument is less likely to be a dominating force in the wealthier polluted cities of the Southeast region of Brazil. Second, if the marginal cost of environmental improvement is high, we can expect small investments in environmental quality improvements. A high marginal cost of en­ vironmental improvement is due, most likely, to the lack of the capacity and infrastruc­ ture needed to design, implement, and enforce environmental policies. Thus, protecting the environment might be costly, even if the cleanup and recovery of degraded areas would be relatively inexpensive. In fact, we can expect a number of “low-​hanging fruit”

580   Ariaster B. Chimeli opportunities to be present in developing countries with respect to environmental improvements in highly degraded areas. Plans for the cleanup of the iconic Guanabara Bay and Lagoa Rodrigo de Freitas in the city of Rio de Janeiro, and the Tietê and Pinheiros rivers in the city of São Paulo, have been discussed for decades now. Implementation of these plans has begun in some cases in partnerships with experts from other countries, and the availability of finan­ cial resources does not seem to be the main obstacle. However, years after the stated conclusion date, the plans have failed to produce tangible results.5 The inability of local, state, and federal governments to implement, monitor, and enforce these environmental policies might have raised the marginal cost of environmental protection in the country. In contrast, a successful set of measures have reduced deforestation in the Amazon re­ gion, as discussed later in the chapter. The third candidate explanation for low environmental quality in developing coun­ tries, according to Greenstone and Jack, is the disconnection between social aspirations and the actions of politicians. In many developing countries, political economy forces provide the incentives and the power for government officials and policymakers to act for the benefit of particular interest groups and not necessarily the collectivity. Feler and Henderson (2011) find evidence that governments strategically excluded the provision of services such as water collection in Brazilian city suburbs to deter immigration. This, in turn, contributed to low environmental quality in these excluded localities, whose inhabitants were mainly lower-​income families. Their data reflect decisions taken during the 1964–​1985 dictatorship in Brazil. It is generally difficult to measure political motivations and policy distortions since it is difficult to measure the protection of interest groups. Given recent large-​scale corruption scandals in the country, it is not difficult to imagine that political economy forces might be playing a role in aggravating environmental degradation in Brazil. On the other hand, an active media, a civil society that is increasingly more organized and vocal, as well as progressively stronger federal and local environmental protection agencies in Brazil, all contribute to greatly mitigating the influence of interest groups on environmental issues.6 In fact, environmental protection agencies are often denounced as an impediment to economic growth by some interest groups in Brazil, a sign of their relative independence of political forces. In 2016, the then-​president Dilma Rousseff inaugurated the Belo Monte hydropower plant, the third largest in the world. Environmental groups had been highly vocal in denouncing the impact of the enterprise, and the original project was modified to have the plant operate as a run-​of-​the-​river dam, avoiding the buildup of a large lake that would flood forests and indigenous lands. Giving up on the lake has significant economic costs to the country, since it would work as an energy reservoir and make the national power system more resilient to climatic and demand shocks. It is unclear whether the run-​of-​the-​river dam would pass a cost-​benefit test, after accounting for environmental gains, relative to the alternative. Nevertheless, the Belo Monte case serves as an example of the fact that environmental concerns were not totally dismissed in favor of catering to specific interest groups. Furthermore, and in contrast to the original observation by

Environmental Issues   581 Greenstone and Jack, the Belo Monte case seems to indicate that the Brazilian public was effectively willing to make a substantial sacrifice for the protection of a relatively small portion of the Amazon forest. To put things in perspective, the Belo Monte dam generated a reservoir of 516 square kilometers, which is about 60% smaller than the res­ ervoir in the original project (1,225 square kilometers). In contrast, the Amazon forest spreads over an area of approximately 3.3 million square kilometers in the Brazilian ter­ ritory and, according to the National Institute for Space Research (INPE), deforestation in the region totaled 4,848 square kilometers in 2014 alone. Finally, Greenstone and Jack (2015) identify the market failures that are recurrent in developing countries as the fourth candidate reason for low environmental quality and low willingness to pay for environmental improvements. These market failures include limited information, as well as missing land, capital, labor, credit, and risk markets. These add to the classical public goods and externalities market failures that are funda­ mental to environmental problems everywhere in the world. Governments may fail to provide accurate information about the state of the envi­ ronment and its impact on health and income. Surprisingly, only a small number of large cities in Brazil collect information on local levels of air pollutants, despite visible indications of pollution and well-​documented impacts on health. In addition, data for some of these cities are sparse, and communication of the consequences of the different air pollutants to the population is in general inadequate or nonexistent.7 Without accu­ rate and well-​understood information, citizens are less likely to take protective meas­ ures, including pressuring policymakers to act on their behalf. Finally, although the level of illiteracy has fallen in Brazil over recent decades, functional illiteracy, where individuals fail to comprehend information passed on to them, likely remains high. This also contributes to a low willingness to pay for environmental quality and weak propen­ sity to demand environmental improvements. Land markets can be affected by ill-​defined property rights. In these cases, the incentives to invest in environmental protection and recovery of eroded land might be absent, given uncertainty over the return of these investments. Alston et al. (2000) study a conflict between the Brazilian Civil Code and the Constitution that gives rise to ill-​ defined property rights over land in the Amazon region. They show that this, in turn, increases deforestation and violent land conflict. Market failure in credit markets due to the difficulty in writing and enforcing contracts contributes to high interest rates and credit rationing. Liquidity constraints then limit the ability of agents to make environmental quality investments that will only pay off in the future. In the Brazilian case, a different distortion may have taken place as governments have subsidized rural credit, which might have caused an ex­ pansion in deforestation. In fact, Assunção et al. (2015) show that a forest protection policy limited rural credit in critical areas in the Amazon and successfully reduced deforestation. Missing insurance markets may reduce investments in environmental quality improvements if returns are uncertain. These missing markets may thus increase exposure to environmental risk with significant consequences to health and

582   Ariaster B. Chimeli income. In this case, low measured marginal willingness to pay for environmental improvements reflects inability to make investments rather than true marginal will­ ingness to pay. Although Greenstone and Jack’s framework is very useful for our current purposes, we should also use caution in applying it to any specific case. Consider, for example, the stark difference between China and Brazil with respect to environmental quality, with air quality reaching much worse levels in the former country. This may be a con­ sequence of the different paths of economic development that these different countries travel, even after we consider differences in endowments and types of market failure. Chimeli (2003) argues that this is the case, for example, with the economies of the former Soviet bloc and the market-​oriented economies. Although at one point these different groups of countries might have converged with respect to the perceived role of markets, they pursue development paths conditional on various very distinctive legacies from diverse political regimes. In these heterogeneous contexts, optimal development paths may have similar fundamental drivers, but economic agents will respond to mark­ edly different incentives, given their different starting points and inherited institutions. Therefore, a single framework of analysis may be misleading, and policy prescriptions that are adequate in one case are not necessarily useful in another case. In the next section, I discuss some recent economic studies applied to Brazil that focus on the environment. In doing so, I  relate them to the Greenstone and Jack framework whenever it makes sense to do so, keeping in mind the caveat from the previous paragraph. Also important is the institutional context for the papers discussed in the following, as it determines the timing of environmental quality improvements in different countries (Chimeli 2007; Chimeli and Braden 2005). The institutional context reflects the country’s institutions, missing markets, and political economy forces at play.

27.3.  Recent Work on Environmental Economics Applied to Brazil I do not attempt to provide a comprehensive review of the recent literature on environ­ mental economics applied to Brazil. Instead, I will focus on a subset of recent empirical work applied to selected topics. Although this is a small subset of a large and growing literature on environmental economics in Brazil, I believe it is representative of the main trends in recent studies. The reduced number of topics covered here is partly due simply to the scope of this chapter—​but it also reflects the limited availability of information on several environmental indicators of potential interest. For example, as mentioned earlier, information on air quality is still disappointingly meager. By contrast, INPE has made remarkable progress in generating information on deforestation for the Amazon. This has spurred a number of studies applied to that region.

Environmental Issues   583

27.3.1.  Deforestation in the Amazon Perhaps one of the most salient topics in the study of the Amazon rainforest is the drastic decline in deforestation beginning in the middle of the first decade of the twenty-​first century. Between 2004 and 2016, the annual rate of deforestation fell from around 27, 000 square kilometers to just under 10,000 square kilometers, according to data pro­ vided by the INPE. Assunção et al. (2015) investigate the causes for this drastic decline and focus on two main candidates: the decline in commodity prices, and the introduction of novel forest protection policies for the Amazon. In 2004, the federal government implemented the action plan for combating deforestation in the Amazon (PPCDAm). This plan inte­ grated federal, state, and municipal governments, as well as satellite imagery pro­ vided by INPE, which allowed real-​time monitoring of illegal deforestation and the subsequent dispatching of law enforcement. The plan focused on territorial manage­ ment and land use, command and control policies, and the promotion of sustainable practices. In late 2007 and 2008, Presidential Decree 6,321 and National Monetary Council Resolution 3,535 established a priority list of municipalities with critical levels of deforestation, required farmers in these areas to prove land title and compliance with environmental laws as a condition for access to credit, and defined harsh penalties for illegal logging, including the destruction of products, materials, and vehicles used by loggers. Assunção et al. (2015) estimate that both declining commodity prices and govern­ ment policies contributed to the reduction in deforestation. Deforestation between 2005 and 2009 totaled 57,100 square kilometers and their counterfactual simulations suggest that policies prevented an additional 73,000 square kilometers of deforestation in the same period.8 The results in Assunção et al. (2015) suggest an improvement in the up-​to-​then weak monitoring and enforcement capacity. As noted by Greenstone and Jack (2015), weak institutions contribute to a higher cost of environmental improvements and lower envi­ ronmental quality. In fact, Chimeli and Boyd (2010) explore a policy introduced before the 2004–​2008 shift in monitoring and enforcement and find evidence that it caused even larger environmental damages. In 2001, the federal government prohibited the market for big leaf mahogany, a tree species from the Amazon biome, but this policy was not backed up by effective monitoring and enforcement capacity. Chimeli and Boyd collect evidence that a large illegal market flourished after prohibition, causing illegal extraction of the species to actually increase. Chimeli and Soares (2017) further explore this policy and estimate a large increase in homicides associated with the ensuing illegal market for mahogany. Macpherson et  al. (2010) use a simulation model to evaluate the advantages and disadvantages of different policy instruments (royalties, performance bonds) that regu­ late logging. They calibrate their model to an area in the Eastern Amazon (Paragominas) and reinforce the importance of institutional weakness/​strength in determining the ef­ fectiveness and impact of these policy instruments.

584   Ariaster B. Chimeli Other articles have focused on more recent evidence on the drivers of deforest­ ation. Lopez and Galinato (2005) study the impact of opening for international trade and deforestation. Their starting point is the three major direct causes of deforestation identified by previous studies: access to roads or other infrastructure,9 poverty, and ag­ ricultural expansion. They then estimate the impact of trade openness and economic growth on these three drivers. They apply their model to Brazil, Indonesia, Malaysia, and the Philippines. They estimate that openness to trade increases forest cover in Brazil (and the Philippines). This happens because the net effect of trade openness is to shift resources to the production of export goods that are not dependent on deforestation. Lopez and Galinato (2005) also estimate a negative impact of economic growth on the forest cover. On the one hand, growth reduces poverty and the pressure on forests, but on the other hand, growth promotes agricultural expansion, which is a driver of deforestation. The net effect in the Brazilian case is an increase in deforestation due to economic growth. Their results are in general not the same for the different countries in their sample, suggesting that specificities of the local economy matter in linking trade and growth to deforestation. Bowman (2016) also investigates international markets and deforestation. She explores exogenous variation in the status of municipalities with respect to the presence of foot and mouth disease to link cattle ranching and deforestation in the Amazon. Bowman uses a panel data set and accounts for spatial correlation to perform a causal test of the linkage between cattle ranching and deforestation. The results highlight the linkage between global beef markets and regional deforestation. Hansen and Naughton (2013) estimate a spatial econometric model with ecological, fertility, and socioeconomic indicators associated with land clearing in the Amazon. Based on their results, they postulate that accessibility to land and ease of clearing is a better predictor of land clearing than soil fertility. Alternatively, they speculate that more fertile land is cleared first, and that this type of land is more productive for more years, which makes additional clearing in the area less necessary. They also emphasize the role of spatial modeling in studying land clearing and designing appropriate policies. Other studies have focused on survey data to explore the interaction between local communities and the Amazon forest. This type of study can provide useful insight for the design of policies that aim at reducing deforestation, such as several REDD (Reducing Emissions from Deforestation and forest Degradation) projects. Pattanayak and Sills (2001) explore survey data and find evidence that locals in the Tapajós region use the collection of forest products as insurance against agricultural subsistence risk; that is, they find that forest product collection contributes to con­ sumption and income smoothing. Further, they estimate that families with more non-​ agricultural income sources (livestock and family members in urban areas) are less likely to resort to forest collection activities, but these practices are not limited to rela­ tively poor households. Klemick (2011) analyzes survey data at the farm level in the Bragantina region in the eastern Amazon. Through a spatial econometric model she estimates a positive impact of fallowing on only long-​run farmers’ income through increased productivity of their

Environmental Issues   585 properties. In addition, Klemick reports a positive externality to downstream farmers, whose productivity increases through impacts on the local hydrological system. She estimates that farmers efficiently allocate land between cultivation and fallowing, and associates this result with the private property regime, as opposed to common property. However, she does not claim that farmers intentionally internalize the external effects of land fallowing in their decision-​making processes. Weber et al. (2011) use an econometric approach to study the impact of the integrated conservation and development project (IDCP) “ProManejo” in the Tapajós region. This program was meant to provide alternative sources of income and economic activ­ ities to improve local population welfare and promote forest conservation. The idea is to diversify income sources and compensate participants for engaging in conservation strategies or avoiding activities that contribute to deforestation. They find evidence of a positive impact of the program on family incomes, but no clear effect on conservation. Finally, Sant’anna (2017) investigates the relationship between land concentration and deforestation in the Brazilian Amazon. He identifies different alternatives that rural workers and farmers face and estimates that land inequality affects deforesta­ tion positively. He estimates that this effect was stronger before the PPCDAm policies implemented in 2004. In concluding this section on Amazon deforestation, it is important to look at Figure 27.1. There is an upward trend in deforestation that begins in 2012 and intensifies after 2014. Existing studies point to a number of determinants of deforestation, including pov­ erty, corruption among government officials, political considerations, and commodity prices. The upsurge in deforestation in the Amazon coincides with the beginning

35,000 30,000 25,000

Km2

20,000 15,000 10,000 5,000 0 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 Year

Figure 27.1.  Annual rate of deforestation in the Brazilian Amazon. Source: National Institute for Space Research (INPE/​PRODES).

586   Ariaster B. Chimeli and deepening of the economic and political crises that Brazil has experienced in re­ cent years. It is still too early to say what is causing this upsurge, after almost 10 years of reductions in deforestation; an important and fertile avenue for research is therefore opening up.

27.3.2. Ethanol Ethanol production in Brazil has spurred national and international interest. Ethanol is derived from processing sugarcane, a renewable and efficient source of fuel produc­ tion with small net carbon emissions in its production and consumption processes. However, recent studies have improved our understanding of the impact of sugarcane production on land use and deforestation, as well as human health. We are also starting to have a better understanding of the demand for ethanol and what drives it. Along these lines, Andrade de Sá et al. (2013) use a dynamic panel to link cattle herds in the state of São Paulo and the Amazon region, sugarcane production in São Paulo, and deforestation in the Amazon. The state of São Paulo is the largest pro­ ducer of sugarcane and is located in the Southeast region of the country, miles from the Amazon region. Nevertheless, Andrade de Sá and colleagues find evidence of an indirect impact of sugarcane expansion on deforestation in the Amazon region that occurs over a 10-​to 15-​year period. Their results invite us to ponder the actual net impacts of ethanol on carbon emissions and other environmental impacts associated with deforestation.10 Nicolella and Belluzzo (2015) explore the passing of a law in the state of São Paulo in 2002, which required the phasing out of the burning of sugarcane before harvesting. Although the compliance period was relatively large, adoption of mechanical harvesting not relying on burning was relatively quick. Using respiratory illness hospital data for 2000, before the passing of the law, and for 2007, they estimate a large reduction in inpa­ tient admissions with respiratory diseases (but not an impact on outpatient admissions or length of stay in the hospital). Chagas et al. (2016) study the same issue using a spatial econometrics approach to account for spatial effects. They estimate stronger effects of the policy and significant impacts on neighboring municipalities that do not produce sugarcane. Salvo and Huse (2013) focus on the characteristics of the demand for ethanol and gasoline in Brazil, two substitute fuels for owners of flex-​fuel vehicles that can operate with any combination of the two fuels. They use data from interviews of fuel consumers at gas stations in six large cities in Brazil and verify that consumers are heteroge­ neous and respond to several factors, in addition to price, when choosing between ethanol and gasoline. They conjecture that some of these factors are environmental preferences, whether ethanol consumers are from the state where it is produced, en­ gine performance concerns, age, and education. Salvo and Huse stress the need for attention to demand and not only supply when it comes to stimulating clean fuel con­ sumption. For example, they show that tax cuts to lower the price of alternative fuels

Environmental Issues   587 might have a limited impact on consumption. Furthermore, consumer education and information might be important aspects to be targeted when policymakers attempt to promote green fuels.

27.3.3. Climate Change Studies on the impact of climate change are gaining momentum in Brazil. Several of these studies use input-​output or computable general equilibrium (CGE) models (e.g., Haddad et al. 2013). A promising line of research in this area is an in-​depth explora­ tion of the double-​dividend hypothesis for climate policy. Early versions of the double-​ dividend hypothesis conjectured that Pigouvian taxes could generate two benefits:  a Pigouvian effect where a pollution tax corrects externalities and increases welfare, and a revenue recycling effect where tax revenues from the Pigouvian tax can be used to reduce other distortionary taxes. However, the double-​dividend hypothesis must be placed in a context where other distortionary taxes already exist and will not be eliminated. In this case, interactions between the Pigouvian taxes and other distortionary taxes could exac­ erbate market distortions and thus reduce or even eliminate the gains that were initially hoped for. The net effect of carbon taxes on welfare is an open empirical question that deserves attention. Assunção and Chein (2016) take a different approach and study the regional impacts of climate change on agricultural productivity in Brazil. They first estimate the impact of climate variables on agricultural productivity at the municipal level for Brazil. Then, they use the Intergovernmental Panel on Climate Change (IPCC) forecasts for temper­ ature and precipitation for the country to simulate the impact of climate change on na­ tional agricultural productivity. They estimate an average productivity loss of 18% for the whole country, with larger losses occurring in the North, Center-​South, and Northeast regions, and the smallest losses occurring in the South and Southeast regions. These spatially heterogeneous impacts as simulated by Assunção and Chein would deepen re­ gional inequalities in the country, given the importance of the agricultural sector to the regional economies. Timmins (2007) notes that global warming can introduce non-​marginal changes that endorse several valuation techniques such as the hedonic method. He then uses a model of endogenous sorting where heterogeneous individuals choose where to live according to several attributes of different regions. These choices are influenced by mi­ gration costs and affect local markets for labor and locally traded commodities. Within this framework, Timmins estimates that climate change produces annual costs ranging from US$1.6 billion to US$8.1 billion for a moderate climate change scenario, depending on migration costs. Finally, Rocha and Soares (2015) investigate how water scarcity in the semiarid region of Northeast Brazil affects early life health. They find that negative rainfall shocks cor­ relate with higher infant mortality, lower birth weight, and shorter gestation periods. Their results suggest severe impacts of climate change in the region in the absence of

588   Ariaster B. Chimeli strategies for adaptation and policies to provide basic health services and sanitation infrastructure.

27.3.4.  Other Studies The relatively large number of studies on land use and deforestation of the Amazon reflects the importance of that biome to both Brazil and the entire world, and also the richer data sets that are available for such studies. There are other pressing environ­ mental problems in Brazil that do not receive the same attention, in great part due to limited data availability. Nevertheless, some promising avenues of research are opening up regarding the economic analysis of these other important problems, as seen in some recent work. Lipscomb and Mobarak (2016) concentrate on the effects of local government decen­ tralization on water pollution in Brazil. Decentralization can be a tool to promote better public service to local populations, but it can also open the possibility of the generation of additional externalities, as pollution is allowed to take place close to jurisdictional borders. This is what Lipscomb and Mobarak find as they explore a large number of splits that have changed the borders of municipalities in Brazil, and the resulting cross-​ border water pollution in rivers. Water pollution increases as informal residential settlements such as favelas are allowed to grow close to downstream borders, and poor enforcement of water emissions occurs in these areas. They confirm this effect through an increase of light intensity close to downstream borders. Also, the more borders a river crosses, the more polluted it is. The additional pollution effect is mitigated when there is room for Coasian bargaining in water-​basin committees that successfully imple­ ment negotiations among stakeholders. Pollution externalities are also mitigated when upstream and downstream municipalities have mayors from the same political parties. Lipscomb and Mobarak’s contributions are important in different dimensions. First, they contribute to the dialogue on decentralization as a policy tool. Second, they high­ light weak institutional settings with selective enforcement of pollution control, which, as discussed in Greenstone and Jack (2015), contributes to increased pollution in devel­ oping countries in general and in Brazil in particular. Third, they provide a testament to an effective institutional setting, the water-​basin committees, which can lower pollution and improve governance over important natural resources. Féres and Reynaud (2005) focus on industrial water use. Their study contributes to the estimation of industrial water demand and water-​related cost structure using a sample of large firms from the state of São Paulo. They estimate an elastic demand and a small share of water expenditures on total cost. Based on this, they propose that water charges would have a low impact on cost and would be effective in promoting water conservation. Additionally, in a simulation exercise, Féres and Reynaud conclude that legislation pushing for reductions in pollution discharge also increase water demand. This happens because stringent water pollution legislation leads to a cost increase and a reduction in the use of materials (sources of pollution). Since materials and water are

Environmental Issues   589 estimated to be substitutes, water use increases. However, their exercise does not take into account the possibility of technological adaptation. Dasgupta et al. (2002) take on the problem of industrial air pollution in Brazil and Mexico. They use data on industrial air pollution for different plant sizes in Mexico and estimate that smaller plants are more pollution intensive. Then they explore data on manufacturing plants of different sizes in Brazil. They use their results from emissions intensities found in Mexico and simulate emissions in Brazil. Dasgutpa and colleagues find that smaller and dirtier plants are located in lower income areas in Brazil. However, harm to humans from industrial air pollution occurs in higher income areas as a conse­ quence of larger volumes of emissions from larger plants. They claim that this result is not a reflection of the concentration of larger plants in larger cities, because they control for pollution density in their analysis. Dasgutpa et al. (2002) revisit the impact of international trade on the environment and document an increase in the use of pesticides during the 1990s, when trade liber­ alization played an important role in the expansion of cash crops for exports in Brazil. These findings raise concerns with respect to the environmental and health effects of the increase in the use of pesticides in the country, an issue that might have played a role in the recent commodity boom that promoted substantial growth of the agricultural sector in Brazil.

27.4. Conclusion This chapter has reviewed the recent Brazilian experience with respect to the environ­ ment from an economic perspective. The framework proposed by Greenstone and Jack (2015) can be used as a starting point to explain the state of the environment in devel­ oping countries. High marginal utility of consumption, high marginal cost of abate­ ment, imperfect representation of citizens by politicians, and market failures are put forth as candidate explanations for poor environmental quality and low willingness to pay for environmental improvements in developing countries, even when large gains to health and income could result. As we have seen, these arguments are applicable in several contexts in Brazil, but not universally. A review of recent empirical work in envi­ ronmental economics applied to Brazil illustrates important lines of research and trends that have prevailed in recent years, as well as several important possible avenues for fu­ ture research. I therefore hope that this chapter can serve as a starting point for those interested in learning more about environmental issues in Brazil.

Notes 1. I thank Filipe Vasconcelos for invaluable research assistance. 2. Recent papers investigating the connection between the environment, health, and welfare in the Unites States include Deschenes et al. (2009), Chay and Greenstone (2005), and

590   Ariaster B. Chimeli Graff Zivin and Neidell (2013). Greenstone and Hanna (2014) study air pollution and in­ fant mortality in India, and Arceo-​Gomez et al. (2016) study the impact of pollution on in­ fant mortality in Mexico City. The US dust bowl of the 1930s, the Samarco dam breakdown in Brazil, and the expected impact of climate change on global markets are all examples of systemic threats of environmental degradation to production and welfare. 3. Chimeli and Braden (2005) and Chimeli (2007) provide a theoretical treatment of the impact of total factor productivity, pollution intensity of capital, and the effectiveness of spending on environmental protection on the optimal dynamic path for environmental quality. 4. For example, in 2008 the Brazilian media reported that the mayor of the town of Tailândia in the state of Pará, a logger himself, distributed food provisions for 6,000 workers in the illegal logging industry who had lost their jobs due to the effective federal government policies to curb deforestation. (For example, in G1 (Globo), February 28, 2008: “Prefeito de Tailândia-​PA dá cestas básicas a demitidos”; in Folha de São Paulo, February 29, 2008: “Em cidade tomada por fiscais, prefeito Macarrão distribui pão às massas”; in EcoDebate, March 11, 2008: “Quadrilha da madeira lucra R$ 90 mi. Este é o valor que grupo criminoso arrecadou com fraudes no Pará”.) 5. As covered in the media, e.g. UOL, June 12, 2016: “RJ gasta R$ 40 mi para limpar Baía de Guanabara, mas projeto está parado”; Folha de São Paulo Especial, March 22, 2016: “Rio Tietê é a memória e a vergonha de SP.” 6. Féres and Reynaud (2012) report the effect of formal (inspections and sanctions) and informal (social pressure) regulation on the environmental performance of Brazilian manufacturing firms. They state that formal regulation is largely influenced by community pressures. 7. The Brazilian Institute for Geography and Statistics (IBGE) reported emissions of total sus­ pended particulates (TSP), PM10, SO2, O3, NO2, and CO for the cities of Belo Horizonte, Curitiba, Distrito Federal, Porto Alegre, Rio de Janeiro, Salvador, São Paulo, and Vitória between 1995 and 2003. The series are not complete for all cities; in fact, they reduce to a few observations in several cases. The data are compiled by IBGE in the Indicadores de desenvolvimento sustentável—​Brasil 2004. Dimensão ambiental—​Atmosfera. 8. Hargrave and Kis-​Katos (2013) also report the role of commodity prices and the actions of IBAMA in determining deforestation in the Amazon region. 9. See, for example, Pfaff (1999) for a study applied to the Brazilian Amazon. 10. Seyffarth (2015) explores a different dimension of the impact of expansion of ethanol pro­ duction. She estimates an increase in the production of sugarcane and decrease in the pro­ duction of rice, beans, and soybeans.

References Alston, Lee J., Gary D. Libecap, and Bernardo Mueller. 2000. “Land Reform Policies, the Sources of Violent Conflict, and Implications for Deforestation in the Brazilian Amazon.” Journal of Environmental Economics and Management 39: 162–​188. Andrade de Sá, Saraly, Charles Palmer, and Salvatore di Falco. 2013. “Dynamics of Indirect Land-​Use Change: Empirical Evidence from Brazil.” Journal of Environmental Economics and Management 65: 377–​393. Arceo-​Gomez, Eva O., Rema Hanna, and Paulina Oliva. 2016. “Does the Effect of Pollution on Infant Mortality Differ between Developing and Developed Countries? Evidence from Mexico City.” The Economic Journal 126 (591): 257–​280.

Environmental Issues   591 Assunção, Juliano, and Flávia Chein. 2016. “Climate Change and Agricultural Productivity in Brazil: Future Perspectives.” Environment and Development Economics 21 (5): 581–​602. Assunção, Juliano, Clarissa Gandour, and Rudi Rocha. 2015. “Deforestation Slowdown in the Brazilian Amazon:  Prices or Policies?” Environment and Development Economics 20 (6): 697–​722. Bowman, Maria S. 2016. “Impact of Foot-​ and-​ Mouth Disease Status on Deforestation in Brazilian Amazon and Cerrado Municipalities between 2000 and 2010.” Journal of Environmental Economics and Management 75: 25–​40. Chagas, André L. S., Carlos R. Azzoni, and Alexandre N. Almeida. 2016. “A Spatial Difference-​ in-​Differences Analysis of the Impact of Sugarcane Production on Respiratory Diseases.” Regional Science and Urban Economics 59: 24–​36. Chay, Kenneth Y., and Michael Greenstone. 2005. “Does Air Quality Matter? Evidence from the Housing Market.” Journal of Political Economy 113 (2): 376–​424. Chimeli, Ariaster B. 2003. “Optimal Dynamics of Environmental Quality in Economies in Transition.” The Economics of Transition 11 (1): 123–​152. Chimeli, Ariaster B. 2007. “Growth and the Environment: Are We Looking at the Right Data?” Economics Letters 96 (1): 89–​96. Chimeli, Ariaster B., and Roy G. Boyd. 2010. “Prohibition and the Supply of Brazilian Mahogany.” Land Economics 86 (1): 191–​208. Chimeli, Ariaster B., and John B. Braden. 2005. “Total Factor Productivity and the Environmental Kuznets Curve.” Journal of Environmental Economics and Management 49 (2): 366–​380. Chimeli, Ariaster B., and John B. Braden. 2009. “A Capital Scarcity Theory of the Environmental Kuznets Curve.” Environment and Development Economics 14 (5): 541–​564. Chimeli, A. B, and R. R. Soares (2017). “The Use of violence in Illegal Markets: Evidence from Mahogany Trade in the Brazilian Amazon.” American Economic Journal: Applied Economics 9 (4): 30–​57. Dasgupta, Susmita, Robert E.  B. Lucas, and David Wheeler. 2002. “Plant Size, Industrial Air Pollution, and Local Incomes:  Evidence from Mexico and Brazil.” Environment and Development Economics 7 (2): 365–​381. Deschenes, Olivier, Michael Greenstone, and Jonathan Guryan. 2009. “Climate Change and Birth Weight.” American Economic Review 99 (2): 211–​217. Feler, Leo, and J. Vernon Henderson. 2011. “Exclusionary Policies in Urban Development: Under-​Servicing Migrant Households in Brazilian Cities.” Journal of Urban Economics 69 (3): 253–​272. Féres, José, and Arnaud Reynaud. 2005. “Assessing the Impact of Environmental Regulation on Industrial Water Use: Evidence from Brazil.” Land Economics 81 (3): 396–​411. Féres, José, and Arnaud Reynaud. 2012. “Assessing the Impact of Formal and Informal Regulations on Environmental and Economic Performance of Brazilian Manufacturing Firms.” Environmental and Resource Economics 52: 65–​85. Graff Zivin, J., and M. Neidell. 2013. “Environment, Health, and Human Capital.” Journal of Economic Literature 51: 689–​730. Greenstone, Michael, and K. Jack. 2015. “Envirodevonomics:  A Research Agenda for an Emerging Field.” Journal of Economic Literature 53 (1): 5–​42. Haddad, Eduardo A., Alexandre A. Porsse, and Paula C. Pereda. 2013. “Regional Economic Impacts of Climate Anomalies in Brazil.” Revista Brasileira de Estudos Regionais e Urbanos 7 (2): 19–​33.

592   Ariaster B. Chimeli Hansen, Winslow D., and Helen T. Naughton. 2013. “Social and Ecological Determinants of Land Clearing in the Brazilian Amazon: A Spatial Analysis.” Land Economics 89 (4): 699–​721. Hargrave, Jorge, and Krisztina Kis-​Katos. 2013. “Economic Causes of Deforestation in the Brazilian Amazon:  A Panel Data Analysis for the 2000s.” Environmental and Resource Economics 54: 471–​494. Klemick, Heather. 2011. “Shifting Cultivation, Forest Fallow, and Externalities in Ecosystem Services:  Evidence from the Eastern Amazon.” Journal of Environmental Economics and Management 61: 95–​106. Lipscomb, Molly, and Ahmed M. Mobarak. 2016. “Decentralization and Pollution Spillovers: Evidence from the Re-​drawing of County Borders in Brazil.” Review of Economic Studies 84 (1): 464–​502. López, Ramón, and Gregmar I. Galinato. 2005. “Trade Policies, Economic Growth, and the Direct Causes of Deforestation.” Land Economics 81 (2): 145–​169. MacPherson, Alexander J., Douglas R. Carter, Marco W. Lentini, and Mark D. Schulze. 2010. “Following the Rules:  Brazilian Logging Concessions under Imperfect Enforcement and Royalties.” Land Economics 86 (3): 493–​513. Nicolella, Alexandre C., and Walter Belluzzo. 2015. “The Effect of Reducing the Pre-​harvest Burning of Sugar Cane on Respiratory Health in Brazil.” Environment and Development Economics 20 (1); 127–​140. Pattanayak, Subhrendu K., and Erin O. Sills. 2001. “Do Tropical Forests Provide Natural Insurance? The Microeconomics of Non-​Timber Forest Product Collection in the Brazilian Amazon.” Land Economics 77 (4): 595–​612. Pfaff, A. 1999. “What Drives Deforestation in the Brazilian Amazon? Evidence from Satellite and Socioeconomic Data.” Journal of Environmental Economics and Management 37: 26–​43. Rocha, Rudi, and Rodrigo R. Soares. 2015. “Water Scarcity and Birth Outcomes in the Brazilian Semiarid.” Journal of Development Economics 112: 72–​91. Salvo, Alberto, and Cristian Huse. 2013. “Build It but Will They Come? Evidence from Consumer Choice between Gasoline and Sugarcane Ethanol.” Journal of Environmental Economics and Management 66: 251–​279. Sant’anna, André A. 2017. “Land Inequality and Deforestation in the Brazilian Amazon.” Environment and Development Economics 22 (1): 1–​25. Selden, Thomas M., and Daqing Song. 1995. “Neoclassical Growth, the J Curve for Abatement and the Inverted U Curve for Pollution.” Journal of Environmental Economics and Management 29: 162–​168. Seyffarth, Anelise Rahmeier. 2016. “The Impact of Rising Ethanol Production on the Brazilian Markets for Basic Food Commodities: An Econometric Assessment.” Environmental and Resource Economics 64 (3): 511–​536. Stokey, Nancy L. 1998. “Are There Limits to Growth?” International Economic Review 39 (1): 1–​31. Timmins, Christopher. 2007. “If You Cannot Take the Heat, Get Out of the Cerrado . . . Recovering the Equilibrium Amenity Cost of Nonmarginal Climate Change in Brazil.” Journal of Regional Science 47 (1): 1–​25. Weber, Jeremy G., Erin O. Sills, Simone Bauch, and Subhrendu K. Pattanayak. 2011. “Do ICDPs Work? An Empirical Evaluation of Forest-​Based Microenterprises in the Brazilian Amazon.” Land Economics 87 (4): 645–​681.

Chapter 28

The Ec onomi c s of Health i n Bra z i l Antonio Carlos Coelho Campino, Maria Dolores Montoya Diaz, and Flavia Mori Sarti

28.1. Introduction The inception of the modern study of the economics of health is frequently attributed to an article published by Kenneth Arrow (1963) approximately 50 years ago: “Uncertainty and the Welfare Economics of Medical Care.” In Brazil, an early forerunner in the field of the economics of health was the Panamerican Health Organization’s pioneering publi­ cation, Economia de La Salud—​Perspectivas para América Latina (1989), which achieved a significant impact in the area. Meanwhile, a particularly important contribution came in the shape of the first book in Brazil specifically dedicated to the topic, entitled Economia da Saúde and published by IPEA (Instituto de Pesquisa Econômica Aplicada) in 1995 (Piola and Vianna, eds. 1995). Nevertheless, the literature on the economics of health in Brazil, and in Portuguese, is still comparatively scarce today (E. I. G. Andrade et al. 2007; Ministério da Saúde 2013; Vianna 1998), although the relatively small cohort of professionals specializing in this area includes a number of noteworthy authors.1 On the agenda of health policy discussion around the world, a prominent place is occupied today by the growth in health expenditure and health services financing. In developed countries, the concern centers on the growing costs of medical hospital as­ sistance. However, in developing countries, the size of the population with no regular access to a formal health assistance system must also be considered, in addition to the costs of health services. As such, in these countries—​Brazil included—​we must add the concern over equal access to the questions related to economic efficiency. In this vein, a growing number of authors have addressed health inequalities in Brazil.2 This chapter examines the historical background to the Brazilian health system and analyzes its characteristics from an economic perspective, considering the magnitude

594    Antonio Carlos Coelho Campino et al. and evolution of inequities in health outcomes and the utilization of health services provided through universal health coverage policies in Brazil during recent decades. It also looks at questions regarding financial protection provided to the population by the Brazilian health system, and challenges for the future.

28.2.  Country Context 28.2.1.  Geography and Socio-​Demography The Brazilian population has been growing at slower rates over the most recent decades (Table 28.1). Although the economically active still represent a significant share of the Table 28.1 Social and Demographic Characteristics of the Brazilian Population, 1995–​2015 Sociodemographic Indicators

1995

Population

161,848,162 174,425,387 185,986,964 194,946,470 207,847,528

Population ages 0–​14 (%)

32.40%

29.52%

27.51%

25.45%

23.03%

Population ages 65+ (%)

5.02%

5.55%

6.27%

7.00%

7.84%

Population ages 80+ (%)*

0.90%

1.05%

1.22%

1.51%

1.90%

Population growth (avg annual rate)

1.53%

1.44%

1.14%

0.88%

0.86%

Population density (indiv per km²)*

2000

2005

2010

2015

19.01

20.48

21.84

22.89

24.87

2.50

2.36

2.07

1.83

1.79

Age dependency ratio (pop 0–​14 y & 65+ y: pop 15–​64 y)

58.85%

53.21%

50.10%

47.18%

44.66%

Age dependency ratio (pop 65+ y: pop 15–​64 y)

7.12%

7.74%

8.79%

9.83%

11.35%

22.20%

18.80%

15.80%

13.50%

14.31%

Fertility rate (births per woman)**

Rural population (%)

Source: World Bank. World dataBank. World Development Indicators (WDI) & Global Development Finance (GDF). Washington, DC: World Bank, 2012. Available at http://​databank.worldbank.org/​ddp/​ home.do, accessed February 1, 2017. Obs.: (*) Source: Instituto Brasileiro de Geografia e Estatística (IBGE). Sistema IBGE de Recuperação Automática (SIDRA). Available at www.sidra.ibge.gov.br accessed: February 1, 2017. Information regarding individuals 80 years old and above In 1995 and 2005 refers to Brazilian population estimates from IBGE, and in 2015 refers to PNAD 2014 data. (**) The fertility rate in 2015 refers to 2014.

The Economics of Health in Brazil    595 population, an accelerated decrease in fertility rates and a steady increase in the propor­ tion of elderly people show that the Brazilian population is rapidly undergoing dem­ ographic transition (Monteiro 2000; Nishijima and Biasoto Jr. 2006). That is, the age dependency ratio is decreasing, mainly due to the decrease in the proportion of the pop­ ulation under 14 years old. The Brazilian elderly population, comprising individuals older than 60  years according to Law 10.741/​03, has increased considerably, reaching 10% of the popula­ tion in 2010, and is projected to reach 18.7% by 2030, 29.4% by 2050, and 33.7% by 2060, according to Instituto Brasileiro de Geografia e Estatística (IBGE). Thus, the share of elderly individuals in the Brazilian population will increase by around 239% between 2010 and 2050, and approximately 275% between 2010 and 2060, while the overall Brazilian population in the same periods is projected to increase by 15% and 11.6%, respectively. The population aging process in Brazil has been accelerating rapidly due to varia­ tions in the fertility rate, infant mortality rate, and overall mortality rate associated with increases in longevity. The notion of “demographic dividend” refers to the impact of changes in age structure and the longevity of the population on economic growth. The first demographic dividend results from a reduction in the fertility rate, which causes faster growth of the labor force in comparison to the dependent population, leading to a reduction in the dependency rate. Studies performed in Brazil indicate that this divi­ dend should extend until 2025.3 Such socio-​demographic evolution patterns may lead to significant increases in health expenditures during the coming decades. Although there is evidence of little cor­ relation between health expenditures and the evolution of health indicators in Brazil during 1980s and 1990s (Nishijima and Biasoto Jr. 2006), several studies indicate a different scenario: advances observed in health technology, associated with significant changes in the demographic and epidemiologic characteristics of populations, result in major impacts on national health expenditures (Dormont et al. 2006; Gandjour et al. 2005; Heffler et al. 2002).

28.2.2. Health status Brazilian health indicators have improved during recent decades (Table 28.2). Life expectancy among Brazilian men remains significantly lower than among women, as a consequence of high mortality rates associated with external causes (especially violence), which primarily affect young males. Nevertheless, life expectancy among men presented a significant increase from 1995 to 2015. The rise in female life expect­ ancy is due mostly to a reduction in mortality rates related to maternal and perinatal problems. Public policies to reduce infant mortality during recent decades have yielded signif­ icant results due to government actions focusing on sanitation infrastructure for clean water supply and sewage treatment in Brazilian cities, as well as on health programs such as the Family Health Program (Macinko et al. 2006; Rocha and Soares 2010).

596    Antonio Carlos Coelho Campino et al. Table 28.2 Health Characteristics of Brazilian Population, 1995–​2010 Health Indicators

1995

2000

2005

2010

2015

Life expectancy at birth, Female†

71.6

74.0

75.8

77.2

78.3

Life expectancy at birth, male†

63.8

66.3

68.1

69.6

70.7

Life expectancy at birth, Total†

67.6

70.0

71.8

73.3

74.4

Mortality rate, infant (per 1,000 live births)

39.7

28.1

19.5

14.8

14.6

Mortality rate, under 5 years of age (per 1,000 live births)

46.3

32.0

21.9

16.6

16.4

Mortality rate, maternal (per 100,000 live births)*§

96.0

81.0

67.0

56.0

58.2

Burden of communicable diseases, maternal, perinatal, and nutritional conditions (% of total DALYs per 100,000 pop)**

—​

13,736

10,268

8,809

9,017

Burden of noncommunicable diseases (% of total DALYs per 100,000 pop)**

—​

35,726

37,163

39,745

42,056

Burden of external causes of death (% of total DALYs per 100,000 pop)**

—​

8,148

8,247

8,646

9,317

Source: World Bank. World dataBank. World Development Indicators (WDI) & Global Development Finance (GDF). Washington, DC: World Bank, 2017. Available at http://​databank.worldbank.org/​data/​ reports.aspx?source=world-​development-​indicators ,accessed February 1, 2017]. Obs.: (†) 2015: Data referring to 2014. (§) 2015: Data referring to 2012. (*) Source of data 1995–​2010: World Health Organization. Maternal mortality in 1990–​2010. Available at http://​www.who.int/​gho/​maternal_​health/​countries/​bra.xls, accessed February 1, 2017. (**) Source of data 2000–​2015: World Health Organization. Disease burden: Estimates for 2000–​2015. Geneva, WHO, 2017. Available at http://​www.who.int/​healthinfo/​global_​burden_​disease/​estimates/​en/​ index2.html, accessed February 1, 2017.

Brazilian health indicators reveal a partial epidemiologic transition during recent decades. The burden of communicable diseases is decreasing and life expectancy is rising. Nevertheless, there is still a significant burden of communicable diseases in cer­ tain regions of the country. The coexistence of high rates of both communicable and non-​communicable diseases is a reflection of persistent inequalities as well as a mixed pattern of improvement in health conditions of the Brazilian population.

28.3.  The Brazilian Health System 28.3.1.  Community Services In a fascinating study evaluating diverse characteristics of various health systems around the world in order to consider the possible components of a notional “ideal

The Economics of Health in Brazil    597 healthcare system,” Britnell (2015) highlights the quality of community services in Brazil. The SUS (Sistema Único de Saúde, or Unified Health System) was created by the 1988 Constitution, and according to Britnell is one of the largest public health systems globally. It aims to provide universal health care, largely through PSF (Programa Saúde da Familia, or Family Health Program), which is a radical form of community empow­ erment. It is seen as an excellent example for low-​and middle-​income countries that wish to develop universal health coverage. However, developed countries may also ben­ efit from greater community activism to respond to pressures originating from the rise of chronic diseases and population aging at an accelerated pace. PSF provides a model on a national scale for community services for geographi­ cally defined population groups, usually around 4,000 inhabitants. The teams of health professionals within the program usually include a physician, a nurse, a nursing assis­ tant, and six health community agents, recruited locally. They serve 100–​150 households with monthly visits, regardless of need or demand. The PSF provides support to households for immunization programs, chronic disease management, health promo­ tion and screening, especially the monitoring of hypertension and diabetes for elderly individuals. Data from the 2013 National Health Survey (Pesquisa Nacional de Saúde [PNS] 2013)  shows that 53.37% of Brazilian households are registered in PSF (34  million households); 86.46% of these households having been registered for more than one year. During the 12 months prior to the survey, these households received at least one visit from a community agent or other PSF team member. Community health agents have been responsible for dramatic reductions in infant mortality and hospitalizations for chronic diseases and mental health problems (Rocha and Soares 2010). According to Britnell (2015, 5), “[w]‌hile developed countries continue to struggle with fragmented community care and disengaged communities, Brazil offers an innovative solution.”

28.3.2. Resources There were 2.35 beds per 1,000 inhabitants in Brazil in 2012:  1.49 beds per 1,000 inhabitants in the private sector and 0.86 in the public sector. Approximately 51% of hos­ pital beds within SUS were private, referring to private organizations with agreements to provide health care to SUS. SUS made available 343,640 hospital beds in 2010, 1.68 hospital beds per 1,000 habitants. In August 2016, SUS had 309,095 beds, representing a 10% decrease in com­ parison to 2010. Almost 84% of hospital beds are located in three regions: the Brazilian Southeast (37%), Northeast (30%), and South (16.8%). The reduction in hospital beds could be attributed to the economic crisis experienced by Santas Casas de Misericordia, nonprofit private health organizations that received less financial resources transferred from SUS in relation to costs of health services provided.

598    Antonio Carlos Coelho Campino et al. Considering the purpose of hospital beds in Brazil, out of 336,997 beds serving SUS, over 22% are surgery and general clinic beds, and over 12% are pediatric and obstetric beds (Table 28.3). The same pattern occurs among non-​SUS beds. The variation in percentages is higher (26% and 20%; 8.6% and 7.2%), and the order of weighting also differs in the case of non-​SUS beds, surgery being more emphasized than general practice and obstetrics more emphasized than pediatrics. Psychiatry also has practically the same importance as pediatrics. In September 2016, there were 6,690 hospitals in Brazil, 70% private, 21% municipal, 8% state, and 1% federal; that is, 30% were public and 70% private (Table 28.4). There were 57 hospital admissions per 1,000 inhabitants and 16.9 outpa­ tient procedures per capita carried out by the SUS in 2010. Approximately 51% of

Table 28.3 Hospital Beds in Brazil, According to Type of Utilization and Sector, 2016 SUS Type

Non-​SUS

Total

Quantity

%

Quantity

%

Quantity

%

Surgery

74,633

22.1

40,609

26.3

115,242

23.4

General clinic

84,150

25.0

31,756

20.5

115,906

23.6

Psychiatry

25,960

7.7

10,950

7.1

36,910

7.5

Paediatry

42,115

12.5

11,201

7.2

53,316

10.8

Obstetrics

40,822

12.1

13,244

8.6

54,066

11.0

Others

69,317

20.6

46,830

30.3

116,147

23.6

TOTAL

336,997

100.0

154,590

100.0

491,587

100.0

Source: CNES, September 2016.

Table 28.4 Hospitals in Brazil, According to Administrative Sphere, 2016 Sphere

Number

%

Municipal

1,405

21.0

535

8.0

State Federal

67

1.0

Private

4,683

70.0

TOTAL

6,690

100.0

Source: CNES, September 2016.

The Economics of Health in Brazil    599 hospitalizations were performed within the public sector in Brazil, up to July 2016: 49.8% within organizations of the public administration and 1.25% by government controlled companies. State hospitals (8% of Brazilian hospitals) and municipal hospitals (21% of Brazilian hospitals) were responsible for the major part of the hospitalizations within the public sector. Private organizations were responsible for 5.7% of hospitalizations, and nonprofit entities were responsible for 43% of hospitalizations in the private health sector. Regarding expenditures on hospitalizations, internment in municipal hospitals presents the lowest value (R$771.05),4 followed by state hospitals (R$1,318.04), non­ profit entities (R$1,489.30), private organizations (R$1,517.49), government companies (R$1,672.96), and, finally, hospitals of the federal administration (R$1,735.73). However, these figures must be viewed in light of other considerations, especially weighing the average value of internment against the complexity of health services provided by the hospital. During 2016, up to August, 268,955 outpatient attendances were performed; 28.6% by SUS (77,007) and 71.4% by the private sector (191,948). The Brazilian Northeast re­ gion was responsible for 35.5% of outpatient attendances in SUS, followed closely by the Southeast (30.5%) and South regions (18.2%). Since 27.6% of the Brazilian population lives in the Northeast, the poorest region in Brazil, 41.9% in the Southeast, and 14.3% in the South, these three regions comprise 83.8% of the population and 84.2% of outpatient attendances, thus corresponding to what we would expect. In the private sector, the prevalence of outpatient attendances is concentrated in the Southeast region (52%), due to typically higher incomes in this region allowing for the payment of health plans. Following this, the South region presents 22.6% of outpatient attendances, these two regions accounting for three-​quarters of the outpa­ tient attention provided by the private sector in Brazil. The Northeast region accounts for 13.9% of private sector attendances, the Center-​West region 8.3%, and the North region 3.3%. Approximately 67% of the Ministry of Health’s total expenditure in 2012 was on health attention, 23.5% being spent on basic attention and 43.6% on medium-​and high-​ complexity procedures. Public sector health-​ care expenditure per capita amounted to R$300.00 in the same period, R$104.99 per capita on basic attention and R$195.01 on medium-​and high-​complexity procedures (in 2012 R$). Public-​ sector health expenditure was around 4% of gross domestic product (GDP). Federal expenditures in health amounted to 1.64% of GDP in 2010,5 while state expenditures were 0.99% of GDP and municipal expenditures 1.04% of GDP (summing to a total of 3.67% of GDP in 2010).

28.3.3. Private Health Sector The private sector is permitted to participate in the SUS structure in complementarity to public health services (Article 199 of the Constitution of 1988). Private-​sector health

600    Antonio Carlos Coelho Campino et al. organizations can be categorized into two groups: the for-​profit private sector and the nonprofit private sector. Institutions that participate in the latter are either philan­ thropic or religious, whereas the former is comprised of services offering health plans, hospitals, and private doctors. There are four types of private health plans in Brazil:  group medicine with pre-​ payment, medical cooperatives, employers’ health plans, and traditional insurance plans. The regulation of the supplementary health sector in Brazil was established by Provisional Measure 1.801-​10,6 which institutes regulations for the functioning of the supplementary health market. The new rules alter previous regulation, emphasizing that private health plans must cover all types of diseases. The private nonprofit sector is “instituted generally on the initiative of the communities (civil or religious) or of social groups, most having a source of financing dependent on State resources, whether by fee for services or by utilization of fiscal mechanisms (exemptions from several federal, State, and municipal taxes), tax, and contributive waiver” (Elias et al 1996, 75–​76). Additionally, due to long-​standing problems with health services supplied by the public sector (queues, shortages of doctors in basic health services, etc.), there was a growth in the use of services provided by the for-​profit private sector in the period 2008–​2014 (Table 28.5), especially regarding medical or dental plan beneficiaries. Medical-​hospital health plan beneficiaries amounted to 24.5% of the total popula­ tion in the first half of 2015 and fell to 23.5% in the first half of 2016 (Table 28.6). This fall represents approximately 1.6 million individuals leaving health plans, an effect very likely caused by the increase in unemployment during this period due to the crisis in the Brazilian economy, given that a significant proportion of the 12 million individuals who lost jobs had corporate health plans.

Table 28.5 Evolution of Health Plan Beneficiaries in Brazil, 2008–​2014 Year

Beneficiaries (millions)

% Variation

2008

52.5

—​

2009

55.8

6.3%

2010

59.5

6.5%

2011

62.9

5.8%

2012

66.7

6.0%

2013

69.6

4.3%

2014

71.5

2.8%

2015

71.1

–0.6%

Source: Data about beneficiaries IESS (February 1, 2017).

The Economics of Health in Brazil    601 Table 28.6 Health Plan Beneficiaries, According to Type of Plan in Brazil, 2015–​2016 Type

2015 1st Half

2016 1st Half

Medical-​hospital plan beneficiaries

50,129,325

48,487,129

Dental plan beneficiaries

21,549,811

21,960,006

Population

204,450,649

206,081,432

Medical-​hospital plan beneficiaries/​Population (%)

24.52

23.53

Dental plan beneficiaries/​Population (%)

10.54

10.66

Sources: Data on beneficiaries: IESS (July 26, 2016); population: IBGE.

Table 28.7 Tickets, Expenses, and Margins of Health Plans, According to Type, Brazil, 2010–​2015 Average Ticket R$

Average Expense R$

Gross Margin R$

2010

134.59

109.30

25.29

2011

149.07

122.88

26.19

2012

162.22

137.85

24.37

2013

179.78

151.49

28.29

2014

204.60

173.96

30.64

2015

236.65

200.24

36.41

2010

9.63

4.42

5.21

Year Medical-​Hospital

Exclusively Dental 2011

10.10

4.91

5.19

2012

10.11

4.93

5.18

2013

10.42

4.77

5.65

2014

10.96

5.01

5.95

2015

11.40

5.33

6.07

Source: ANS/​TABNET, information collected by IESS Data (September 5, 2016).

The pattern was different among exclusively dental plans, possibly due to their lower costs. The number of beneficiaries grew from 10.5% of the Brazilian population in the first half of 2015 to 10.7% in the first half of 2016. Health plans submit to ANS (Agência Nacional de Saúde Suplementar –​National Agency of Supplementary Health) information on revenues and the average expense of attendances (Table 28.7). The standard expression for average revenue is “average ticket,”

602    Antonio Carlos Coelho Campino et al. defined as “revenue of monthly average services per beneficiary” by IESS (Instituto de Estudos de Saúde Suplementar, or Supplementary Health Study Institute); while average expense is “monthly average expense with attendance per beneficiary.” Medical-​hospital health plans reported a gross margin of R$36.41 per beneficiary in 2016, corresponding to 15.4% of revenues. Exclusively dental plans obtained a gross margin of R$6.07, corresponding to 53.2% of revenues. However, according to average ticket and average expense definitions presented, it is important to note that this does not represent the companies’ profit, since they have to deduct administrative and other expenses from this gross margin.

28.4.  Health Inequalities 28.4.1. Concepts Having sketched the main features of Brazil’s health-​care system we now turn to a quan­ titative analysis of the evolution and magnitude of inequities in health outcomes and the utilization of health services in the Brazilian health system over recent decades. Models commonly used to evaluate inequities at each period were based on estimations of concentration curves and concentration indices. Concentration curves display the share of health outcome or utilization accounted for by the cumu­ lative proportion of individuals in the population, ranked from poorest to richest. The curves show variation in health outcomes or utilization according to socioec­ onomic status; if everyone, irrespective of living standards, has exactly the same value of the health variable, the concentration curve will be a 45-​degree line (line of equality). The further a curve is above (under) the line of equality, the more the cor­ responding health variable is concentrated among the poorest (richest) households. For example, Figure 28.1 shows the concentration curves of two variables constructed with data from the 2013 National Health Survey (PNS): having chronic health problem (Yes = 1 and No = 0) and self-​assessed health (“Regular,” “Bad,” or “Very Bad” = 1, “Excellent” or “Good” = 0). Both variables present more concentra­ tion among the poorest, as the curves are above the line of equality. As the variables represent “not good health,” this indicates that health problems are more concen­ trated among the poor. The concentration index quantifies the degree of socioeconomic-​related inequality in a health variable. It is directly related to the concentration curves: twice the area be­ tween the concentration curve and the 45-​degree line equals the concentration index. The sign of the index is negative (positive) when the line is above (under) the line of equality (O’Donnell et al. 2008). The standardization procedure7 applied to the concentration index intends to reduce or “eliminate the inequality arising from variables that are beyond the control of policy

The Economics of Health in Brazil    603

Cumulative % of outcome variable

1.00

0.80

0.60

0.40

0.20

0.00 0.00

0.20

0.40

0.60

0.80

1.00

Cumulative % of population, ranked from poorest to richest Line of equality Less than good self-assessed health status Chronic

Figure 28.1.  Concentration curves of inequalities in health outcomes of Brazilian population, PNS 2013. Source: Elaborated by the authors based on PNAD data

makers, notably demographic factors such as age and gender. [ . . . ] A positive value (of concentration index) indicates that the health variable is more concentrated among richer individuals. In the case of a variable for ill health, such as impairment in activi­ ties of daily living, a positive value of the concentration index means that the rich are in worse health than the poor. A negative value indicates the opposite, whereas a concen­ tration index not very different from zero reflects no relationship between income and health status” (Wagstaff et al. 2011, 45). It is important to highlight that because of the standardization procedures, it is pos­ sible to find apparently “incoherent” results between the concentration indices that should prevail and the percentages found among the quintiles. Thus, for example, we can have an index with a negative value, while the percentages of the socioeconomic variable among the quintiles are higher among the richest. Models based on those developed by Wagstaff et al. (2011) for analysis of national health systems regarding equity in health-​care access and financial protection are found, for example, in Sarti et al. (2015), Almeida et al. (2013), and Dmytraczenko and Almeida (2015). This investigation of inequalities in health outcomes and health-​care utilization was achieved by the estimation of standardized concentration indices using Brazilian databases, detailed in the next section.

604    Antonio Carlos Coelho Campino et al.

28.4.2. Databases and Variables Data from three National Household Sample Surveys (Pesquisa Nacional por Amostra de Domicílios, or PNAD) and the 2013 National Health Survey (Pesquisa Nacional de Saúde, or PNS 2013) from IBGE were analyzed. PNAD databases are household-​and individual-​ based surveys that have been conducted annually by IBGE since the 1970s. Household samples for the survey are na­ tionally representative, since sample size is based on population projections from the Brazilian Demographic Census. Thus, PNAD samples may be expanded to represent the Brazilian population. PNAD samples are based on three stages of a probabilistic sam­ pling process: (1) city, (2) census sector, and (3) household. PNAD also includes supple­ mentary data in chosen themes; health status and health system utilization themes from PNAD were investigated in 1998, 2003, and 2008. The National Health Survey, PNS 2013, is a nationally representative survey on popu­ lation health, partially released until September 2016 and encompassing various health dimensions of the Brazilian population, including lifestyle, access to and use of health services, preventive actions, continuity of care, and type of health care (public, private, or private facilities financed by SUS). The PNS questionnaire contains three parts, the first part referring to household characteristics, the second part including information on socioeconomic status and health of household residents, and the third part regarding health data from one adult resident (of 18 years or older), selected with equal likelihood among adult residents. The first two parts of the questionnaire are answered by a house­ hold representative; and the third part is answered by one resident and includes mor­ bidity and lifestyle, anthropometric measures, blood pressure, and laboratory exams (Szwarcwald et al. 2014). Independent variables selected from PNAD and PNS were the following:  age (according to birth date in relation to interview date), gender (self-​ reported gender: male = 1), race (self-​reported race: white/​yellow = 1), educational attainment (higher education  =  1), activity status (individuals with occupation or working  =  1), health insurance coverage (presence of private health-​ care plan, individual or collective = 1), number of household residents (sum of household members in adult equivalents), and region (geographical location of household). Regarding income, the variable selected in PNAD databases was household income (income from household residents, except household employees), while the variable selected in PNS was wages earned in the labor market by household residents (except household employees).8 Outcome variables selected from PNAD and PNS were as follows: probability of outpa­ tient preventive health care (any preventive consultation in the last two weeks), and type of health care (private facilities, private facilities financed by SUS, and public facilities). Additional data from PNAD databases in 2003 and 2008 included smoking habits and accidents and injuries (road traffic and other). PNS presents information on smoking habits and obesity (waist circumference).

The Economics of Health in Brazil    605 It is also important to note that the declaration of chronic diseases in PNAD 1998 in­ cluded undiagnosed chronic diseases perceived and reported by individuals, while data from PNAD 2003, PNAD 2008, and PNS 2013 registered only chronic diseases reported by individuals as diagnosed by health professionals. Data on household income (PNAD) or household labor market income (PNS) was converted into household income per capita in adult equivalents, using a 0.75 weight for individuals younger than 14 years old in order to generate an “equivalence scale” (eh):

eh = ( Ah + φK h )θ (1)

where Ah represents the number of adults in the household h and K h represents the number of children in the household h. The parameters φ and θ should be equal to 0.75, weight previously defined according to Deaton (1997). Samples of PNAD and PNS were selected to include adult individuals (18 years or older), resulting in information from 217,742 individuals in PNAD 1998; 254,942 individuals in PNAD 2003; 260,449 individuals in PNAD 2008; and 145,580 individuals in PNS 2013.

28.4.3.  Inequalities in Health Outcomes During the period 1998–​2013, instances of “less than good self-​assessed health status” were concentrated among the poorest individuals. As found by Almeida et al. (2013), “individuals from lower income quintiles in general reported worse health status and severe physical difficulties more frequently than higher income.” The results indicated that in the surveys carried out in 1998, 2003, and 2008 the proportion of individuals in the lowest income quintile who self-​assessed as having “Regular, Bad or Very Bad” health status was 36%. This proportion was around 17% among individuals belonging to the highest income quintile. Important chronic health problems like hypertension and diabetes and non-​road-​ traffic accidents show mixed patterns during the period. Diabetes and non-​road-​traffic accidents presented positive concentration indices, despite not being statistically signif­ icant in two of the four analyzed periods. It is important to recall that negative concen­ tration indices indicate higher concentration among the poorest individuals. Therefore, there is a concentration of diabetes and non-​road-​traffic accidents among the richest individuals, and hypertension among the poorest, for all years but 2013.9

28.4.4.  Inequalities in Risky Behavior There is lack of information in population-​based surveys regarding risky behav­ ior among the Brazilian population. The only variable included in PNAD was

606    Antonio Carlos Coelho Campino et al. Table 28.8 Evolution of Inequalities in Health Outcomes of Brazilian Population, 1998–​2013 Adult Health

Year

Q1

Q2

Q3

Q4

Q5

Total

CI

Sig

Hypertension

1998 0.17644 0.18110 0.17320 0.16307 0.13342 0.16545 –​0.05240 *** 2003 0.19007 0.19260 0.19212 0.17582 0.15043 0.18021 –​0.04313 *** 2008 0.20660 0.20635 0.20612 0.19712 0.17496 0.19823 –​0.02996 *** 2013 0.22858 0.24176 0.22950 0.22790 0.19867 0.22528

0.01071 *

1998 0.02445 0.02964 0.03436 0.03539 0.02955 0.03068

0.04283 ***

2003 0.03578 0.03772 0.04140 0.04016 0.03679 0.03837

0.00934 —​

2008 0.04731 0.05068 0.05146 0.05260 0.04903 0.05022

0.00953 *

2013 0.07092 0.06340 0.06355 0.05786 0.04908 0.06141

0.00055 —​

Non-​road-​ 2008 0.02339 0.02205 0.02292 0.02182 0.02484 0.02300 traffic accident 2013 0.02973 0.03113 0.02785 0.03167 0.02281 0.02865

0.01395 —​

Diabetes

0.00033 —​

Source: Elaborated by the authors, based on data from PNAD (IBGE 1999, 2004b, 2009) and PNS (IBGE 2014). * Significant at 10%, *** Significant at 1%.

Table 28.9 Evolution of Inequalities in Risky Behavior among Brazilian Population, 2008–​2013 Risk Factors

Year

Q1

Smoking

2008 0.23834 0.20459 0.18367 0.15668 0.12925 0.18250 –​0.12091 *** 2013 0.15710

Q2

Q3

Q4

Total

CI

Sig

0.17626 0.14954 0.12471 0.10646 0.14333 –​0.04906 ***

Smoking 2008 0.17259 0.14426 0.12715 0.11708 among women 18–​49 years old Obesity

Q5

0.09539 0.13386 –​0.10671 ***

2013 0.40745 0.44310 0.44890 0.44546 0.44391 0.43622

0.03655 ***

Source: Elaborated by the authors, based on data from PNAD (IBGE 1999, 2004b, 2009) and PNS (IBGE 2014). *** Significant at 1%.

smoking habits (in 2008), whereas PNS contains data on smoking and obesity (waist circumference). It was possible to identify a higher concentration of risky behavior among the poorest quintiles of income, and women (Table 28.9). Concentration indices results indicate that smoking has a higher prevalence among the poorest, while obesity has a higher prevalence among the richest.

The Economics of Health in Brazil    607

28.4.5.  Inequalities in Health-​Care Utilization Concentration indices for curative health-​care variables (probability of physician visits and of dentist visits, and the absolute value of physician visits)10 measured by Almeida et al. (2013) are positive and statistically significant for 1998, 2003, and 2008, indicating a higher use among richer individuals. The results for the variable representing any preventive health assistance (in the two weeks prior to the interview) were positive and statistically significant during the period analyzed for private health care, showing that inequalities in the utiliza­ tion of preventive health care are pro-​rich, even when controlling for need variables (Table 28.10). There are still substantial differences in health-​care utilization patterns of public and private facilities among the richest and poorest quintiles of the Brazilian popula­ tion. A negative concentration index for public health care indicates more intensive use among the poorest; the opposite is the case for private health care. That notwithstanding, there is a consistent reduction in the concentration index in preventive private health care. Voluntary health insurance assures the possibility of health assistance in the private health sector, and data confirm a growth in the number of beneficiaries of health plans that could be related to the reduction in inequalities in private health care (Table 28.5).

Table 28.10 Evolution of Inequalities in Preventive Health Care among Brazilian Population, 1998–​2013 Adult Preventive Care

Year

Q1

Q2

Q3

Q4

Q5

Total

CI

Sig

Preventive outpatient public health care

1998

0.008

0.010

0.011

0.009

0.006

0.009

–​0.043

**

2003

0.030

0.029

0.030

0.024

0.014

0.026

–​0.122

***

2008

0.023

0.021

0.022

0.020

0.011

0.019

–​0.109

***

2013

0.077

0.059

0.087

0.055

0.015

0.060

–​0.114

***

1998

0.001

0.002

0.005

0.010

0.020

0.008

0.503

***

2003

0.005

0.007

0.013

0.026

0.058

0.022

0.488

***

2008

0.005

0.006

0.011

0.019

0.043

0.017

0.460

***

2013

0.041

0.019

0.023

0.051

0.116

0.050

0.310

***

1998

0.000

0.001

0.001

0.001

0.000

0.000

–​0.050

—​

2003

0.000

0.001

0.001

0.001

0.001

0.001

0.015

—​

2008

0.000

0.000

0.000

0.000

0.000

0.000

–​0.076

—​

2013

0.001

0.000

0.001

0.001

0.001

0.001

0.106

—​

Preventive outpatient private health care

Preventive outpatient private health care financed by SUS

** Significant at 5%, *** Significant at 1%.

608    Antonio Carlos Coelho Campino et al.

28.5. Challenges The Brazilian health system has been evolving positively during recent decades, showing signs of democratization within the initial proposition of universal health coverage. Changes in population characteristics, especially demographic transition, have been posing challenges in terms of public policy, directing the country toward the epidemi­ ological transition—​that is, toward a higher prevalence of chronic non-​communicable diseases (including obesity, diabetes, and hypertension). However, the Brazilian population still faces unresolved issues regarding infectious diseases, including traditional diseases whose incidence is increasing in the country (es­ pecially yellow fever, leprosy, measles, and diarrhea), and emerging infectious diseases (particularly dengue, zika, and chikungunya)11 (Gautret and Simon 2016; Portes et al 2016; Silva Nali et al. 2016). Additionally, the occurrence of infectious diseases like AIDS remains unchanged, potentially influencing public exposure to risk (Luz et  al. 2016; Silva Lizzi et al. 2016; Sousa and Pinto Jr. 2016).

28.5.1. Universal Health Coverage in the Brazilian Health System A considerable number of studies analyze in depth the main challenges in uni­ versal healthcare coverage in post-​1990s Brazil (Almeida et al. 2013; Lucchese 1996; Macinko and Lima-​Costa 2012; Marques 1999; Suárez-​Berenguela 2001). Most studies concerning equity and health care universalization in the Brazilian health system are based on conceptual discussions, qualitative approaches, and/​or opinionated arguments published in scientific literature (Bahia 2009; Bravo 2006; Cordeiro 2001; Escorel 2001; Kerstenetzky 2006; Marsiglia et al. 2005; Noronha et al. 2004; Santos and Gerschman 2004). Regarding the main challenges in universalization of health-​care coverage in Brazil since the inception of SUS, a literature has emerged considering aspects of the financing and management of the public health system, as well as the regulation of the private health sector. Nevertheless, there is still a lack of comprehensive research analyzing these different aspects of the impacts of universal health coverage policies in the reduc­ tion of inequities during recent decades in Brazil.

28.5.2. Financing and Management of the Public Health System Universalization of health care has been the main focus of the public sector within the Brazilian health system since the 1990s. Government expenditure on social policies are

The Economics of Health in Brazil    609 substantial, when compared to other emerging economies, yet most public disburse­ ment is focused on direct cash transfers and other social benefits (Afonso 2006). Although overall Brazilian health system expenditure (% of GDP) has been increasing during recent decades, government expenditures in health were considerably lower than in other American countries. Government health expenditures were higher in Argentina, Uruguay, and Colombia during the 1990s, yet, at the same time, public health care in Brazil offers universal public health-​care coverage (Afonso 2006; Gerschman and Santos 2006; Medici 2005). A combination of factors stressing public health expenditures (universalization of public health care and demographic transition, among others) and problems in govern­ ment financing during the 1990s have threatened operationalization of SUS in its early stages (Afonso and Junqueira 2009). Furthermore, there is a perception of differences in the quality of health care provided in the public and private sectors, discrediting claims of successful universalization in health assistance (Santos and Gerschman 2004). Simultaneously, the Brazilian public health system features a complex structure of fi­ nancing and distribution of resources. A set of funding sources, including numerous taxes and contributions, are dedicated to financing social security policies in Brazil, including health care. The universalization of social policies was the justification for creating and incorporating taxes and contributions in order to ensure the maintenance of public policies through diversification of funding sources (Dain 2007). The management of the Brazilian public health system also presents challenges in establishing universal coverage. An absence of incentives to promote efficiency through control of costs at certain quality levels, allied to open access to health care and au­ tonomy ceded to decentralized government levels, leads to problems in determining ideal subsidies for public healthcare (Lavras 2011; Pinheiro Filho and Sarti 2012). The decentralization process in the Brazilian health system, associated with the uni­ versalization of health care, has faced barriers in relation to regional inequalities and the absence of infrastructure for health-​care supply in various localities (Arretche 2000), although studies show that decentralization has not influenced health inequities (Arretche et al. 2007; Bahia 2005; Costa 2001). Strategic actions necessary for the consolidation of SUS ought to be taken in the near future, including financing schemes, redefinition of priorities in contracting and waging, and the establishment of tools for monitoring and evaluation (Lucchese 1996).

28.5.3.  Regulation of the Private Health System Global health-​care expenditures per capita in Brazil appear to be significantly larger due to the substantial share of health expenditures from the private sector (Afonso 2006), and an increasing share of the Brazilian population has been adopting health insurance in recent decades (Fleury 2011). Although there are some studies connecting the increase in private health insur­ ance to internal macroeconomic policies and the international context (Cordeiro 2001;

610    Antonio Carlos Coelho Campino et al. Marques 1999), the recent expansion of private health care in Brazil seems to be linked to the dynamics of the national health system, especially adjustments in the public health-​ care structure, independent of other factors (Afonso 2006; Menicucci 2003). Private sector operations in health care, including health insurance, health plans, and health service providers, have been regulated by the National Agency of Supplementary Health (Agência Nacional de Saúde Suplementar, or ANS) since 2000 (Macera and Saintive 2004). Nevertheless, during the last decade, several problems have been re­ ported in relation to the extent and quality of health assistance provided by private institutions. Additionally, a number of private health insurance institutions have declared bank­ ruptcy or difficulties in covering costs, interrupting health assistance to patients, re­ ducing health insurance coverage, or imposing barriers to the supply of specific types of health service (Carvalho and Cecílio 2007; Macera and Saintive 2004).

28.5.4. The Judicialization of Health Article 196 of the 1988 Brazilian Constitution states that Health is everyone’s right and duty of the State, guaranteed through social and ec­ onomic policies aimed at reducing the risk of disease and other ailments and equal access to actions and services for its promotion, protection and recovery.

Thus, Brazil is the only country to have a health system intended to be public, free, and universal for over 200  million individuals. Almost 30  years after its inception, the Brazilian health system presents complex challenges to be handled by authorities, managers, experts, researchers, and general society. One of the most critical aspects is the dilemma that has arisen due to the interpretation of constitutional provisions by managers against citizens’ right to full health assistance. There is an increasing adoption of lawsuits by individuals in need of various types of health care or medication, using arguments based on the aforementioned constitutional article. This process has been called the judicialization of health, known to be a recent phe­ nomenon in which significant issues from a political, social, or moral point of view are decided by courts (Barroso 2010). According to a report by the Brazilian attorney gen­ eral about judicial intervention in public health, there have been a growing number of lawsuits against the federal government aiming to receive medications or various forms of health care, especially directed toward the federal court and some states’ courts (Advocacia-​Geral Da União 2013) (Table 28.11). From October 2011 until September 2012, only 30% of 7,773 decisions were favor­ able to the federal government. Although not every lawsuit results in additional health expenditures, the costs of the Ministry of Health have been increasing for some time given the purchase of medication, equipment, and supplies necessitated by court decisions (Figure 28.2).12

The Economics of Health in Brazil    611 Table 28.11 Lawsuits Regarding Health Care against the Brazilian Federal Government, 2009–​2012 Year

Number

2009

10,486

2010

11,203

2011

12,811

2012

13,051

350.00 300.00

R$ million

250.00 200.00 150.00 100.00 50.00 -

2005

2006

2007

2008

2009

2010

2011

2012

Figure 28.2.  Costs of the Brazilian Ministry of Health referring to court decisions, 2005–​2012. Source: Elaborated by the authors based on PNAD data.

According to recent media reports,13 federal government spending on lawsuits re­ garding health care increased by 797% in five years (from R$122.6 million in 2010 to R$1.1 billion in 2015). The Brazilian attorney general’s report also suggests that high rates of success in lawsuits referring to health care may be related to an increase in judicial demands, and highlights unreasonable high costs imposed on SUS, resulting in a loss of ability to handle budgetary constraints in view of unanticipated purchases, thereby generating inefficiencies in the public health system. As a rule, the anticipation of a protective effect is irreversibly granted by Brazilian courts.

612    Antonio Carlos Coelho Campino et al. However, the most perverse consequence has been the loss of the intended logical op­ eration in the health system, resulting from the creation of special treatments directed to a limited number of individuals, contrary to the conception of a policy of universal health care. The Brazilian Supreme Court will soon judge two cases involving the supply of medicines unavailable within SUS or without registration in Brazil, reopening the dis­ cussion on the right of access to health (care) and the impact of lawsuits on govern­ ment accounts. These will have significant implications in guiding future decisions over lawsuits, and favorable decisions for the government must be hoped for in order to safe­ guard the economic sustainability of the Brazilian health system.

28.6.  Final Remarks This chapter has surveyed the economics of health in the Brazilian context, encompassing information on the current state of research, population characteristics, attributes of the health system, evidence on health disparities, and challenges relating to the management of health care in the country. It is important to emphasize that demographic transition in Brazil has compelled the country toward an incomplete epidemiological transition, marked by the coexistence of infectious diseases (traditional unresolved illnesses and emerging maladies) and chronic non-​communicable diseases derived from technological advances and income growth (obesity, diabetes, hypertension, among others), without benefits from sustain­ able economic development. In mid-​December 2016, a Proposed Amendment to the Constitution (PEC) was approved, imposing a ceiling on public spending for the next 20 years. In particular, the health budget received special treatment. The legislation prior to this PEC allocated 13.2% of the Union’s Net Current Revenue to the area of health, and this would have grown to reach 15% in 2020. The PEC anticipates this schedule, moving this percentage to 2017. From 2018, theoretically until 2036, the floor of health spending will be adjusted annually for inflation, guaranteeing in real terms the volume of resources of 2017. This scenario reveals challenges for the ambitious proposal of universal health coverage pro­ vided through public funding, defining priorities for health assistance in association with accountability in government health expenditures.

Notes 1. These include André Medici (World Bank), Mônica Viegas Andrade (Universidade Federal de Minas Gerais), Maria Alicia Ugá (Fundação Oswaldo Cruz/​Escola Nacional de Saúde Pública, FIOCRUZ/​ENSP), Denise Cyrillo (School of Economics, Business Administration and Accounting of the University of São Paulo), and the authors of this chapter. Werner Baer included a chapter on the role of health in the Brazilian development

The Economics of Health in Brazil    613 process, co-​authored by Antonio Campino and Tiago Cavalcanti, in the fifth edition of The Brazilian Economy: Growth and Development (Baer, ed. 2001), published in Brazil as A Economia Brasileira, 2nd edition. 2. Articles discussing and measuring health inequalities in Brazil include Almeida et al. 2013; Almeida and Sarti 2013; Andrade et al. 2013; Andrade et al. 2015; Campino 2011; Campino et al. 1999; Maria Dolores Montoya Diaz 2002, 2003; Macinko and Lima-​Costa 2012; Neri and Soares 2002; Noronha and Andrade 2005; Szwarcwald et al. 2005; Victora et al. 2000; among others. 3. The second demographic dividend corresponds to an increase in the savings ratio in re­ lation to capital in the economy. Thus, it results from a reduction in mortality and con­ sequent increase in life expectancy, given that individuals living longer need to increase savings to cover consumption at older ages. However, it is necessary that wealth should be accumulated in the form of savings or assets for this second dividend to occur. Nevertheless, the magnitude of the second dividend depends on how wealth is created. For example, if the elderly receive private transfers (family members) or public transfers (pensions), the second dividend is reduced; this is the case with Brazil. 4. This value is slightly higher than the “others” category, of R$716.93, but it is unclear what falls into this category. 5. 2010 is the most recent year for which data on expenditure per public entity are available. 6. Published in Diário Oficial da União no. 38 (Friday, June 2, 1999). 7. In the literature, there are direct and indirect methods of standardization: “Direct stand­ ardization provides the distribution of health or utilization across living standard groups that would be observed if all groups had the same age structure, for example, but had group-​specific intercepts and age effects. Indirect standardization, however, ‘corrects’ the actual distribution by comparing it with the distribution that would be observed if all individuals had their own age, but the same mean age effect as the entire population” (Wagstaff et al. 2011, 79). More details about the indices can be found in Kakwani (1977, 1980), Kakwani et al. (1997), Wagstaff et al. (1991). 8. Unfortunately, until September 2016, data on household income and laboratory exams were still unavailable and thus were not included in the analysis. Information on labor market participation and wages was released in June 2016 and included in this analysis. 9. This result for 2013 may be the consequence of the difference in the database used for that year (PNS as opposed to PNAD). 10. Two other variables were analyzed (probability and days of hospitalization) but the indices were not statistically significant. 11. Dengue, chikungunya, and zika viruses are disseminated by the same vector (Aedes aegypti mosquito); although presenting similar clinical signs, they have significant differences in terms of health outcomes for infected individuals. Recently, the substantial rise in the in­ cidence of infectious diseases has been the subject of investigation in Latin America and especially in Brazil due to its health impacts. 12. Indeed, expenses pertaining to purchasing procedures and medication delivery (e.g., car­ rier payment for delivery of medication to patients, insurance payment for medication transportation, and, if applicable, costs of imports) are not included in the values presented. 13. E.g., Folha de São Paulo, September 24, 2016. “Decisão do STF sobre fornecimento de remédios guiará ações pelo país.” Available at http://​www1.folha.uol.com.br/​cotidiano/​ 2016/​09/​1816469-​decisao-​do-​stf-​sobre-​fornecimento-​de-​remedios-​guiara-​acoes-​pelo-​ pais.shtml.

614    Antonio Carlos Coelho Campino et al.

References Advocacia-​ Geral da União. 2013. Intervenção judicial na saúde pública. Report. Brasília: Consultoria Jurídica/​Ministério da Saúde. Available at http://​portalarquivos.saude. gov.br/​images/​pdf/​2014/​maio/​29/​Panorama-​da-​judicializa-​-​-​-​o-​-​-​2012-​-​-​modificado-​em-​ junho-​de-​2013.pdf. Afonso, J. R. 2006. “Universalização do gasto e diversificação fontes de financiamento–​o caso da seguridade social no Brasil.” Seminario Internacional Cohesion Social en America Latina y el Caribe. Santiago, Chile: CEPAL. Afonso, J. R., and G. G. Junqueira. 2009. “Reflexões a respeito da interface entre seguridade so­ cial e fiscalidade no Brasil.” In Seguridade Social, Cidadania e Saúde, edited by L. Lobato and S. Fleury. Rio de Janeiro: CEBES. Almeida, G., and F. M. Sarti. 2013. “Measuring Evolution of Income-​Related Inequalities in Health and Health Care Utilization in Selected Latin American and Caribbean Countries.” Revista Panamericana de Salud Pública 33 (2): 83–​89. Almeida, G., F. M. Sarti, F. F. Ferreira, M. D. M. Diaz, and A. C. C. Campino. 2013. “Analysis of the Evolution and Determinants of Income-​Related Inequalities in the Brazilian Health System, 1998—​2008.” Revista Panamericana de Salud Pública 33:  90–​97. doi:  10.1590/​ S1020-​49892013000200003. Andrade, E. I. G., F. de A. Acúrcio, M. L. Cherchiglia, S. A. Belisário, A. A. Guerra Júnior, D. A. C. Szuster, D. R. Faleiros, et al. 2007. “Pesquisa e produção científica em economia da saúde no Brasil.” Revista de Administração Pública 41 (2): 211–​235. Andrade, M. V., K. Noronha, A. C. Q. Barbosa, T. A. H. Rocha, N. C. da Silva, J. A, Calazans, M. N. Souza, et al. 2015. “A equidade na cobertura da Estratégia Saúde da Família em Minas Gerais, Brasil.” Cadernos de Saúde Pública 31: 1175–​1187. Andrade, M. V., K. V. M. de S. Noronha, R. de M. Menezes, M. N. Souza, C. de B. Reis, D. R. Martins, and L. Gomes. 2013. “Desigualdade socioeconômica no acesso aos serviços de saude no Brasil: um estudo comparativo entre as regiões brasileiras em 1998 e 2008.” Economia Aplicada 17 (4): 623–​645. Arretche, M. T. S. 2000. Estado federativo e políticas sociais: Determinantes da descentralização. Editora Revan. Arretche, M., E. Marques, G. Hochman, M. Arretche, and E. Marques. 2007. “Condicionantes locais da descentralização das políticas de saúde.” In Políticas públicas no Brasil, edited by G. Hochman, M. Arretche, and E. Marques, 173–​204. Rio de Janeiro: Editora Fiocruz. Arrow, K. 1963. “Uncertainty and the Welfare Economics of Medical Care.” American Economic Review 53 (5): 941–​973. Baer, W., A. C. C. Campino, and T. Cavalcanti. 2001. “Health in the Process of Development of Brazil.” In The Brazilian Economy: Growth and Development, edited by W. Baer, 405–​425. Westport, CT: Greenwood. Bahia, L. 2009. “O sistema de saúde brasileiro entre normas e fatos: universalização mitigada e estratificação subsidiada.” Cien Saude Colet 14 (3): 753–​762. Barroso, L. R. 2010. “Constituição, democracia e supremacia judicial:  direito e política no Brasil contemporâneo.” Revista Jurídica da Presidência 12 (96): 5–​43. Brasil. Ministério da Saúde. 2013. A produção de conhecimento em economia da saúde : Uma perspectiva bibliográfica (2004–​ 2012). Brasília:  Ministério da Saúde/​ Organização Pan-​ Americana da Saúde/​Associação Brasileira de Economia da Saúde.

The Economics of Health in Brazil    615 Bravo, M. I. S. 2006. “Política de saúde no Brasil.” In Serviço social e saúde: Formação e trabalho profissional, edited by A. Mota, M. Bravo, R. Uchoa, V. Nogueira, R. Marsiglia, L. Gomes, and M. Teixeira, 88–​110. São Paulo: Associação Brasileira de Ensino e Pesquisa em Serviço Social/​Organização Pan-​Americana da Saúde. Britnell, M. 2015. In Search of the Perfect Health System. London: Palgrave Macmillan. Campino, A. C.  C. 2011. “Gastos catastróficos, iniquidade e proposta de reformulação do sistema de saúde.” In Brasil: A nova agenda social, edited by E. L. Bacha and S. Schwartzman, 104–​108. Rio de Janeiro: LTC. Campino, A. C. C., M. D. M. Diaz, L. M. Paulani, R. G. Oliveira, S. F. Piola, and A. Nunes. 1999. Poverty and Equity in Health in Latin America and Caribbean: Results of Country-​Case Studies from Brazil, Equator, Guatemala, Jamaica, Mexico and Peru, 1–​82. Washington, DC: Pan American Health Organization. Carvalho, E. B., and L. C. de O. Cecílio. 2007. “A regulamentação do setor de saúde suplementar no Brasil: A reconstrução de uma história de disputas.” Cad. Saúde Pública 23 (9): 2167–​2177. Cordeiro, H. 2001. “Descentralização, universalidade e eqüidade nas reformas da saúde.” Ciência and Saúde Coletiva 6 (2): 319–​328. Dain, S. 2007. “Os vários mundos do financiamento da Saúde no Brasil:  uma tentativa de integração.” Cienc. saude coletiva 12 (Supl): 1851–​1864. Diaz, M. D. M. 2002. “Socio-​economic Health Inequalities in Brazil: Gender and Age Effects.” Health Economics 11 (2): 141–​154. doi: 10.1002/​hec.649. Diaz, M. D.  M. 2003. “Desigualdades socioeconômicas na saúde.” Revista Brasileira de Economia 57 (1): 7–​25. Dmytraczenko, T., and G. Almeida, G. 2015. Toward Universal Health Coverage and Equity in Latin America and the Caribbean: Evidence from Selected Countries. Washington, DC: World Bank Publications. Dormont, B., M. Grignon, and H. Huber. 2006. “Health Expenditure Growth: Reassessing the Threat of Ageing.” Health Economics 15 (9): 947–​963. Elias, P. E., A. Cohn, A., and P. E. M. Elias. 1996. “Estrutura e organização da atenção à saúde no Brasil.” In Saúde no Brasil: políticas e organização de serviços, edited by A. Cohn and P. E. Elias, 57–​117. São Paulo: Cortez/​CEDEC. Escorel, S. 2001. Os dilemas da equidade em saúde: Aspectos conceituais. Brasília: OPAS. Fleury, S. 2011. “Direitos sociais e restrições financeiras:  Escolhas trágicas sobre universalização.” Ciência and Saúde Coletiva 16 (6): 2686–​2688. Gandjour, A., J. Greb, E. Bomsdorf, and K. W. Lauterbach. 2005. “Impact of Demographic Changes on Healthcare Expenditures and Funding in the EU.” Applied Health Economics and Health Policy 4 (1): 1–​4. Gautret, P., and F. Simon. 2016. “Dengue, Chikungunya and Zika And Mass Gatherings: What Happened in Brazil, 2014.” Travel Medicine and Infectious Disease 14 (1): 7–​8. Gerschman, S., and M. A. B. Santos. 2006. “O Sistema Único de Saúde como desdobramento das políticas de saúde do século XX.” Revista Brasileira de Ciências Sociais 21 (61): 177–​227. Heffler, S., S. Smith, G. Won, M. K. Clemens, S. Keehan, and M. Zezza. 2002. “Health Spending Projections for 2001–​2011: The Latest Outlook.” Health Affairs 21 (2): 207–​218. Kakwani, N. 1977. “Measurement of Tax Progressivity:  An International Comparison.” The Economic Journal 87 (345): 71–​80. Kakwani, N. 1980. Income Inequality and Poverty. Oxford: Oxford University Press. Kakwani, N., A. Wagstaff, and E. Van Doorslaer. 1997. “Socioeconomic Inequalities in Health: Measurement, Computation, and Statistical Inference.” Journal of Econometrics 77 (1): 87–​103.

616    Antonio Carlos Coelho Campino et al. Kerstenetzky, C. L. 2006. “Políticas sociais:  Focalização ou universalização.” Revista de Economia Política 26 (4): 104. Lavras, C. 2011. “Descentralização, regionalização e estruturação de redes regionais de atenção à saúde no SUS.” In Política e gestão em saúde, edited by N. Ibañez, P. E. M. Elias, and P. H. D. Seixas, 285–​316. São Paulo: Hucitec. Lucchese, P. T. R. 1996. “Descentralização do financiamento e gestão da assistência à saúde no Brasil:  a implementação do Sistema Único de Saúde—​ retrospectiva 1990/​ 1995.” Planejamento e Políticas Públicas 14: 75–​158. Luz, P. M., M. P. Girouard, B. Grinsztejn, K. A. Freedberg, V. G. Veloso, E. Losina, C. J. Struchiner, et  al. 2016. “Survival Benefits of Antiretroviral Therapy in Brazil:  A Model-​ Based Analysis.” Journal of the International Aids Society 19 (20623): 1–​11. PMID 27029828 doi: 10.7448/​IAS.19.1.20623. Macera, A. P., and M. B. Saintive. 2004. O mercado de saúde suplementar no Brasil. Brasília: Secretaria de Acompanhamento Econômico, Ministério da Fazenda. Macinko, J., F. C. Guanais, and M. de F. M. de Souza. 2006. “Evaluation of the Impact of the Family Health Program on Infant Mortality in Brazil, 1990–​2002.” Journal of Epidemiology and Community Health 60 (1): 13–​19. Macinko, J., and M. F. Lima-​Costa. 2012. “Horizontal Equity in Health Care Utilization in Brazil, 1998–​2008.” International Journal for Equity in Health 11 (1): 1. Marques, R. M. 1999. O financiamento do sistema público de saúde brasileiro. Santiago, Chile: Naciones Unidas Comisión Económica para América Latina y el Caribe (CEPAL). Marsiglia, R. M.  G., C. Silveira, and N. C. Junior. 2005. “Políticas sociais:  Desigualdade, universalidade e focalização na saúde no Brasil.” Saúde e Sociedade 14 (2): 69–​76. Medici, A. 2005. Financiamento público e privado em saúde na América latina e Caribe: Uma breve análise dos anos noventa. Washington, DC: Inter-​American Development Bank. Menicucci, T. M.  G. 2003. “Público e privado na política de assistência à saúde no Brasil: atores, processos e trajetória.” Ph. D Dissertation, UFMG. http://​hdl.handle.net/​1843/​ VCSA-​7NGKNX. Monteiro, C. A. 2000. Velhos e novos males da saúde no Brasil: A evoluçäo do país e de suas doenças. São Paulo, Hucitec: Nupens/​USP, 435 p. (Saúde em Debate 91). Musgrove, P. 1989. Economía de la salud:  Perspectivas para América Latina. Washington, DC: Organización Panamericana de la Salud. Neri, M., and Soares, W. 2002. “Desigualdade social e saúde no Brasil.” Cadernos de Saúde Pública 18: 77–​87. Nishijima, M., and G. Biasoto, Jr. 2006. “O padrão de financiamento da saúde nos países da América.” In Regulação do setor saúde nas Américas: As relações entre o público e o privado numa abordagem sistêmica, edited by G. BiasotoJr., P. L. D. B. Silva, and S. Dain, 107–​140. Washington, DC: OPAS. Noronha, J. C. de, L. D. de Lima, and C. V. Machado, C. V. 2004. “A gestão do Sistema Único de Saúde: Características e tendências.” In Saúde no Brasil—​Contribuições para a agenda de prioridades de pesquisa. Série B. Textos Básicos de Saúde. Brasília: Ministério da Saúde. Noronha, K. V.  M.  de., and M. V. Andrade. 2005. “Desigualdades sociais em saude e na utilização dos serviços de saúde entre os idosos na América Latina.” Revista Panamericana de Salud Pública 17: 410–​418. Pinheiro Filho, F. P., and F. M. Sarti. 2012. “Falhas de mercado e redes em políticas públicas: Desafios e possibilidades ao Sistema Único de Saúde.” Ciência e Saúde Coletiva 17 (11): 2981–​2990.

The Economics of Health in Brazil    617 Piola, S. F., and S. M. Vianna, eds. 1995. Economia da saúde: Conceitos e contribuição para a gestão da saúde. Brasília: Instituto de Pesquisa Econômica Aplicada (IPEA). Portes, S. A. R., N. M. Lanzarini, C. B. Vieira, J. Fernandes, A. dos Santos Silva, R. C. de Oliveira, E. V. da Costa, E. R. S. de Lemos, and T. M. Fumian. 2016. “Viruses in Mass Gatherings: Viral Waterborne and Foodborne Diseases.” Virus Reviews and Research 21 (2): 32–​40. Rocha, R., and R. R. Soares. 2010. “Evaluating the Impact of Community Based Health Interventions:  Evidence from Brazil’s Family Health Program.” Health Economics 19 (1): 126–​158. Santos, M. A.  B.  dos, and S. Gerschman. 2004. “As segmentações da oferta de serviços de saúde no Brasil: Arranjos institucionais, credores, pagadores e provedores.” Ciência e Saúde Coletiva 9 (3): 795–​806. Sarti, F. M., T. M. Ivanauskas, TM. D. M. Diaz, and A. C. C. Campino. 2015. “Towards Universal Health Coverage in Latin America and the Caribbean: A Case Study Measuring Inequalities in Health in Brazil.” Background paper for the project Toward Universal Health Coverage and Equity in Latin America and the Caribbean:  Evidence from Selected Countries, ed­ ited by Tania Dmytraczenko and Gisele Almeida. Washington, DC: World Bank. https://​ openknowledge.worldbank.org/ ​ b itstream/ ​ h andle/ ​ 1 0986/ ​ 2 2026/ ​ 9 781464809200. pdf?sequence=6&isAllowed=y. Silva Lizzi, E. A. da, A .A. Nunes, and E. Z. Martinez. 2016. “Spatiotemporal Analysis of AIDS Incidence among Adults in Brazil.” Current HIV Research 14 (6): 466–​475. Silva Nali, L. H. de, D. M. Fujita, F. S. Salvador, M. C. D. da Silva Fink, H. F. de Andrade, C. S. Pannuti, and E. J. de Albuquerque Luna. 2016. “Potential Measles Transmission Risk in Mass Gatherings: Are We Safe for the Olympic Games—​Rio 2016?” Journal of Travel Medicine 23 (4). https://​doi.org/​10.1093/​jtm/​taw026. Published 12 May 2016. Sousa, A. I. A. de, and V. L. Pinto Jr. 2016. “Análise espacial e temporal dos casos de aids no Brasil em 1996-​2011: Áreas de risco aumentado ao longo do tempo.” Epidemiologia e Serviços de Saúde 25 (3): 467–​476. Suárez-​Berenguela, R. M. 2001. “Health System Inequalities and Inequities in Latin America and the Caribbean: Findings and Policy Implications.” In Investment in Health: Social and Economic Returns, edited by Pan American Health Organization, 119–​142. Washington, DC: Pan American Health Organization. Szwarcwald, C. L., M. do C. Leal, G. C. Gouveia, and W. V. de Souza. 2005. “Desigualdades socioeconômicas em saúde no Brasil:  Resultados da Pesquisa Mundial de Saúde, 2003.” Revista Brasileira de Saúde Materno Infantil 5: 11–​22. Szwarcwald, C. L., D. C. Malta, C. A. Pereira, M. L. F. P. Vieira, W. L. Conde, P. R. B. D. Souza Jr., G. N. Damacena, L. O. Azevedo, G. Silva, M. M. Theme Filha, and C. D. S. Lopes. 2014. “Pesquisa Nacional de Saúde no Brasil: Concepção e metodologia de aplicação.” Ciência e Saúde Coletiva 19 (2): 333–​342. Ugá, M. A.  D., and I. S. Santos. 2007. “An Analysis of Equity in Brazilian Health System Financing.” Health Affairs 26 (4): 1017–​1028. Vianna, S. M. 1998. Evolução e perspectivas da pesquisa em economia da saúde no Brasil. Brasília: IPEA. Victora, C. G., J. P. Vaughan, F. C. Barros, A. C. Silva, and E. Tomasi. 2000. “Explaining Rrends in Inequities: Evidence from Brazilian Child Health Studies.” The Lancet 356 (9235): 1093–​1098. doi: http://​dx.doi.org/​10.1016/​S0140-​6736(00)02741-​0

618    Antonio Carlos Coelho Campino et al. Wagstaff, A., M. Bilger, Z. Sajaia, and M. Lokshin. 2011. Health Equity and Financial Protection:  Streamlined Analysis with ADePT Software. World Bank Publications. Washington, DC: The World Bank. Wagstaff, A., O. O’Donnell, E. Van Doorslaer, and M. Lindelow. 2007. Analyzing Health Equity Using Household Survey Data: A Guide to Techniques and Their Implementation. Washington, DC: World Bank. Wagstaff, A., P. Paci, and E. Van Doorslaer. 1991. “On the Measurement of Inequalities in Health.” Social Science and Medicine 33 (5): 545–​557.

Pa rt  V I

BRAZIL AND T H E  WOR L D E C ON OM Y

Chapter 29

B r azil, the BRI C S , a nd the Changing L a nd s c a pe of Gl obal E c onomi c Governa nc e Peri Silva

29.1. Introduction In 2001, the economist Jim O’Neill of the investment bank Goldman Sachs coined the acronym BRIC to represent the economies of Brazil, Russia, India, and China (O’Neill 2001). Apparently, his idea was to use this acronym to better encapsulate his basic conclusion that the global economy would be increasingly dependent on the performance of these key developing economies in future decades. Two main factors are important in understanding O’Neill’s main argument. The first is that the combination of the burst of the high-​technology (“dot-​com”) bubble in 2000 and the terrorist attacks of September 11, 2001, had severely limited developed economies’ expectations of economic growth for the remainder of the decade. The second is that the BRIC economies were significantly large by the end of 2000, and, in contrast to the situation faced by developed economies, their short-​term economic forecast was bright. By 2001, all BRIC economies were among the 20 largest economies in the world,1 and, given their economic performance in the years 2000 and 2001, O’Neill felt that that “over the next 10 years, the weight of the BRICS and especially China in world GDP will grow” (2001, 2).2 By the end of the 2010s, the acronym BRIC had become one of the most popular in discussions involving the world economy.3 The BRIC economies even agreed to initiate a sequence of annual summits with the participation of their heads of state, with the inaugural summit taking place in 2009 in the city of Yekaterinburg, Russia. In 2011, South Africa officially joined the group, extending the original acronym to BRICS. Currently, the governments of the BRICS state that they have two main objectives: (1) coordination

622   Peri Silva in meetings and international organizations; and (2) the development of an agenda for multisectoral cooperation among its members.4 From a Brazilian perspective, some influential policymakers view the BRICS as a very ambitious project. According to former Brazilian president Luiz Inácio Lula da Silva, “Brazil, Russia, India, and China have a fundamental role to play in building a fairer, more representative and safer new world order.”5 As is clear from their stated objectives, cooperation among the BRICS in international forums is a central objective for this group of countries. In this chapter, we focus on the current and potential degree of cooperation among the members of the BRICS in shaping global economic governance. In particular, attention is paid to economic cooperation in multilateral institutions such as the International Monetary Fund (IMF) and the World Trade Organization (WTO), as well as to economic cooperation in bilateral terms, such as in the formation of preferential trade agreements (PTAs). We do not consider arguments related to the effects of membership of the BRICS on the international prestige of a country’s “brand name.”6 It is also important to recognize that while our interest centers on the Brazilian economy, cooperation among the BRICS economies depends on the characteristics of all members and on how they relate to each other. Interestingly, cooperation among BRIC countries was neither expected nor encouraged in O’Neill’s discussion. His argument was simply that a greater share of the global economy would be represented by the BRIC economies over time, and, consequently, “world policymaking forums should be re-​organized and, in particular, the G7 should be adjusted to incorporate BRIC representatives” (2001, 2). The reason, according to O’Neill, is that these economies would become more important in determining the effectiveness of monetary and fiscal policies in the global economy. However, he recognizes that “clearly, the four countries under consideration are very different economically, socially and politically, and incorporating all four of them into a G7 style club might not be straightforward” (2001, 10), an argument that was raised later with different nuances by other analysts, as we discuss later in the chapter. Thus, the ability and willingness of the BRICS countries to cooperate is far from clear; it is this issue that we investigate here. The fact that the economies of the BRICS have been recently linked to potential changes in global governance should be taken neither lightly nor for granted. From a Brazilian perspective, its mere participation in this group of nations makes it evident that a substantial improvement in the country’s international credibility has taken place. The Brazilian economy originally reached a certain degree of prominence in world forums in the 1970s when the country’s GDP gained the position of 8th worldwide. This large economy was then regarded as “the country of the future” given its sizable and young population, continental dimensions, and its high economic growth rates during the 1970s.7 But several serious macro-​economic imbalances in the Brazilian economy were becoming obvious in the late 1970s. As discussed by Baer (2008), the Brazilian economy showed clear stress signs in the late 1970s in the form of resurgent inflationary pressures and substantial increases in the country’s foreign indebtedness (Baer 2008, Chapters 5

Brazil and the BRICs    623 and 6).8 This situation deteriorated further in the first half of the 1980s as the tightening in US monetary policy made it more difficult for Brazil and other Latin American countries to service their sovereign debts.9 Moreover, the social pressure for increasing government spending dramatically increased in Brazil in the second half of the 1980s as the first civilian policymakers replaced the military-​supported leadership that had reigned since 1964.10 As is well known, the Brazilian economy faced serious economic challenges during the second half of the 1980s, as well as during the early years of the 1990s. The most obvious evidence of these challenges was the successive failures of several monetary stabilization plans to control the inflation rate, which culminated in hyperinflation rates during the years of 1990 and 1993 in Brazil.11 Macroeconomic conditions clearly began to improve in Brazil with the launching of the Real Plan in 1994, which successfully brought down inflation rates through a combination of sound monetary policy and economic reforms to support sound fiscal policy over time.12 Yet, for several years after the implementation of the Real Plan, Brazil could not aspire to playing a major role in shaping global governance.13 This perception only changed in the first decade of the 2000s, when low inflation rates as well as more widely accepted fiscal and monetary practices have become the norm in Brazil. In spite of the problems that developing countries such as Brazil, China, and India still face, it is undeniable that the increasing popularity of the BRICS acronym during the last decade is somewhat justifiable, and for the reasons originally outlined by the creator of that acronym. As a group, the BRICS economies have substantially increased their participation in the world economy. Table 29.1 compares the performance of the BRICS group against the United States, the European Union, and against all non-​BRICS economies (including the US and EU), in terms of economic growth and in terms of participation in world trade. The information is provided by splitting the years of 2000–​ 2014 into three periods. Columns 1–​3 suggest that, on average, the BRICS economies have grown at a faster pace than the rest of the world in every period, in particular when compared to the United States and to the European Union. This was particularly true during the 2006–​2010 period, which coincides with the financial crisis initiated with the meltdown of the US real-​estate sector and with the bulk of the most recent commodity boom.14 As would be expected, the superior economic growth performance of the BRICS group has translated into greater participation in world markets. Columns 4–​6 show that the participation of the BRICS economies in world trade has advanced significantly, increasing from 8.49% in the years 2000–​2005 to almost 16% in the period 2011–​2014. On the contrary, US participation in world trade has been stable, while EU participation has significantly decreased. In a nutshell, the BRICS group has indeed become a more significant player in the global economy. Next, we consider the degree of economic interdependency among BRICS members, focusing on possible cooperative measures that may explain such connections. This analysis considers the agenda agreed upon by these countries in their annual summits, plus other sources of information.

624   Peri Silva Table 29.1 BRICS: General Trends Share World Trade 2006–​2010 (5)

Share World Trade 2011–​2014 (6)

8.49%

13.10%

15.98%

2.13%

11.14%

10.67%

10.83%

0.91%

0.74%

26.11%

24.75%

23.53%

3.88%

3.37%

91.51%

86.90%

84.02%

GDP Growth 2000–​2005 (1)

GDP Growth 2006–​2010 (2)

GDP Growth 2011–​2014 (3)

BRICS

5.92%

6.19%

4.29%

United States

2.79%

0.78%

European 2.23% Union Non-​ BRICS

4.35%

Share World Trade 2000–​2005 (4)

Calculations made by the author using information from the World Development indicators (World Bank). Columns 1–​3 use the simple average of GDP growth across countries and years. Columns 4–​6 define the share of world trade by taking the simple average between the share of world imports and the share of world exports across years. The share of world trade for the European Union uses the membership information for this preferential agreement from 2008. Moreover, it takes into account intra-​and inter-​EU trade flows in calculating world trade. Calculations made by the author using information from the World Development indicators (World Bank).

29.2.  Economic Interdependency and Annual Meetings 29.2.1.  Economic Interdependency As the previous section shows, the BRICS economies have increased their economic significance in the world economy. The data in Table 29.1 show that these countries have grown faster than developed economies and the rest of the world. By the end of the first decade of the 2000s, and to the surprise of some analysts,15 the members of this group have agreed to hold annual summits with the presence of their heads of states. The first annual summit was held in the Russian city of Yekaterinburg in 2009, followed by summits in Brasília in 2010 and in the Chinese city of Sanya in 2011.16 The decision to hold the initial annual summit coincides with the deepening of economic ties among the BRICS economies as measured by the trade flows of goods and services, as well as flows of foreign direct investment (FDI) among these economies. This information is discussed in Table 29.2, where column 1 describes the share of intra-​BRICS trade flows over the years, while column 2 shows the share of intra-​BRICS FDI flows. It is clear from column 1 that the share of intra-​BRICS trade flows increases

Brazil and the BRICs    625 Table 29.2 Trade Flows and FDI for the BRICS group

Year

Share Intra-​BRICS Trade (1)

Share Intra-​BRICS FDI Inflows (2)

Inter-​BRICS Share of Intra-​BRICS Share of Preferential Trade Preferential Trade (3) (4)

2001

4.22%

0.05%

2.27%

0.16%

2002

4.59%

0.03%

2.28%

1.34%

2003

4.86%

0.04%

1.25%

0.02%

2004

5.14%

0.30%

1.38%

0.01%

2005

5.81%

0.50%

1.58%

0.00%

2006

6.08%

1.25%

2.39%

0.04%

2007

7.02%

2.01%

5.24%

0.99%

2008

7.57%

1.27%

5.84%

2.88%

2009

7.85%

5.81%

6.54%

3.20%

2010

8.43%

4.87%

1.81%

0.00%

2011

8.88%

4.54%

7.13%

5.43%

2012

8.63%

7.16%

8.51%

0.82%

2013

8.51%

N.A.

5.35%

1.78%

2014

8.50%

N.A.

9.39%

1.46%

Calculations made by the author. Columns 1, 3, and 4 use data from the World Integrated Trade Solution, and they were originally aggregated at the 6-​digit of the Harmonized System. Column 2 uses data from the United Nation Conference on Trade and Development (UNCTAD). Preferential access is determined by using the difference between the applied MFN and the preferential tariffs. Bilateral trade for a good is deemed preferential if this tariff difference is equal or greater than 2 percentage points.

significantly and continuously from 2001 to 2011 by about 110%, and has, since then, slightly decreased. A similar situation is seen in column 2 but, in this case, the initial share of intra-​BRICS FDI was excessively low and seems more volatile than that measured using the share of trade flows. Notice that the peak years of the intra-​BRICS trade share coincide with the initial annual meetings involving the heads of states of these countries. The important thing is to assess whether these annual summits, as well as official declarations of cooperation among the BRICS countries, are actually responsible for the greater interdependence among these countries. In this regard, we offer three arguments that the greater interdependence among the BRICS economies described in Table 29.2 is not the direct result of policies implemented in consultation (or in cooperation) among them. First, note that Table 29.1 indicates that the BRICS economies have on average grown faster than the non-​BRICS economies during that time frame. This is particularly true in the period 2006–​2010, when the BRICS initiated their annual summits. The well-​known gravity model of international

626   Peri Silva trade17 suggests that the larger the trade partners, the more they tend to trade, ceteris paribus. Thus, since the BRICS economies have grown faster than the rest of the world, it is not surprising that they have traded more with each other over time relative to trade with non-​BRICS members.18 Second, the BRICS could be promoting trade among themselves by forming PTAs. This could be important since, as Baier and Bergstrand (2007) show, PTAs can significantly promote trade among members over time.19 This issue is examined in columns 3 and 4 of Table 29.2, where the former column displays the share of trade taking place on a preferential basis by BRICS members with non-​members, while the latter column displays the same share of trade among BRICS members. It is clear from comparing columns 3 and 4 that BRICS members undertake more preferential-​basis trade with non-​BRICS countries than with their fellow BRICS. It is easy to understand why this is the case. The World Trade Organization (WTO) indicates20 that 423 PTAs that include goods and services have been notified and are in place. However, the same source indicates that only two partial scope agreements—​ one agreement between India and the Common Market of the Southern Cone (MERCOSUR) and the Asia Pacific Trade Agreement—​officially involve more than one member of the BRICS.21 The former partial scope agreement links Brazil and India in a preferential manner, while the other links China and India in a preferential manner.22 At the same time, China has joined several more comprehensive PTAs in the Pacific region, including free trade areas (FTAs) with the Association of Southern Asian economies (ASEAN) in 2005, with Chile in 2007, and with New Zealand in 2008, among others. Likewise, Brazil has been a member of MERCOSUR since 1991, and this agreement has been enlarged by granting membership to Venezuela, as well as by forming separate FTAs with Bolivia and with Chile.23 Moreover, Brazil has exchanged preferential access with all other Latin American countries in a partial scope manner through the Latin American Integration Association (LAIA) since 1982. A  similar situation applies to India, Russia, and South Africa.24 In sum, the members of the BRICS have shown no inclination to promote closer ties through cooperation by forming PTAs, but have displayed considerable appetite for forming PTAs with other regional members.

29.2.2. Annual Meetings As indicated in the preceding, the third argument used to demonstrate that the greater economic interdependency among the BRICS from 2001 to 2011 is not a direct result of policy measures adopted by the members of this group is based on their joint statements issued after the initial annual summits. These joint statements allow us to evaluate other possible measures of cooperation suggested during the most recent annual summits. This is important when considering both the current degree of cooperation and the prospects for future cooperation. In this section, we consider both issues. The first BRICS annual summit with heads of state present was only in 2009. Thus, it is unlikely that any sort of policy measures jointly agreed upon during these summits can explain the increase in the trade and FDI flows among these economies between 2001 and 2011. Still, it is important to look into the content of the joint declarations issued

Brazil and the BRICs    627 by BRICS members during the annual summits. The former Brazilian chancellor Luiz Felipe Lampreia argues (2009, n.p.) that the inaugural “summit in Yekaterinburg did not generate any results, and there should not have been any expectation for a different result.” Similarly, Lampreia declares (2010a, n.p.) that the second BRICS annual summit in the Brazilian capital in 2010 was successful, but that “it would not have been possible to expect concrete results from it, at least in the short run, nor to refer to a ‘bloc.’ . . . ” This sense that the initial annual summits of the BRICS did not generate concrete proposals can also be detected in its members’ joint declarations.25 The joint declarations issued after the first, second, and third annual summits stress very diverse topics ranging from statements of support for coordination of economic policies to overcome the global financial crisis and the recognition of the important role played by international trade and FDI in this, to vows of support for promoting dialogue on energy efficiency, climate change, humanitarian relief, and on the fight against terrorism, among several other topics. Of particular importance is that these joint statements mention support for a multilateral global order, where reforms of multilateral institutions such as the United Nations, the IMF, and the World Bank would be key. For example, the declaration from the second annual summit held in 2010 calls for “a substantial shift in voting power in favor of emerging market economies and developing countries to bring their participation in decision making in line with their relative weight in the world economy.” In the case of the WTO, the declaration issued after the second annual summit states, “We (BRICS) concur in the need for a comprehensive and balanced outcome of the Doha Round of multilateral trade talks, in a manner that fulfills its mandate as a ‘development round,’ based on the progress already made, including with regard to modalities.” As for the United Nations, the declaration issued after the second annual summit states, “we (BRICS) reiterate the importance we attach to the status of India and Brazil in international affairs, and understand and support their aspirations to play a greater role in the United Nations.” This statement seems to acknowledge the aspiration from the governments of Brazil and India to become permanent members of the United Nations Security Council.26 Note that these initial joint statements neither mention any desire to pursue preferential trade agreements among these economies nor mention any specific policies that the BRICS wish to jointly pursue in well-​established multilateral forums. In a nutshell, the three initial summits did not produce concrete policy results that could explain the increase in economic interdependence among these economies from 2001 to 2011. We can now consider the joint statements issued after the more recent annual summits that took place between 2012 and 2015. The joint declaration issued after the fourth annual summit held in New Delhi (India) in 2012 also contains many diverse vows of support for vaguely defined policies in many areas, while including additional concerns related to the deterioration of the economic situation in certain countries of the Eurozone, to the civil war in Syria, and to the Arab-​Israeli conflict. This joint statement also expresses concern with delays in implementing the Governance and Quota IMF reform approved in 2010. This reform aims at increasing the resources available to the IMF, while also modifying each member’s IMF Quota shares, thereby changing their voting power within that institution. Most important, the fourth annual joint meeting declaration states that the BRICS countries were considering the possibility of setting up a New

628   Peri Silva Development Bank (NDB). This seems to be the first specific new policy advocated by the BRICS since the inaugural annual summit held in 2009. The fifth BRICS annual summit was held in 2013 in the city of Durban, South Africa. Previous joint statements’ vows of support for vaguely defined policies, as well as similar statements of concern about (lack of) reforms in multilateral institutions, can also be found in the joint statement issued after the fifth summit. In this case, the important point seems to be the confirmation that the BRICS would indeed be establishing the NDB with the objective of “mobilizing resources for infrastructure and sustainable development projects in BRICS and other emerging economies and developing countries, to supplement the existing efforts of multilateral and regional financial institutions for global growth and development.”27 Moreover, this joint statement announces that these nations would be establishing a Contingent Reserve Agreement (CRA) requiring the amount of US$100 billion to “help BRICS countries forestall short-​term liquidity pressures, provide mutual support, and further strengthen financial stability.” The sixth annual BRICS summit was held in 2014 in the city of Fortaleza, Brazil. The main outcome of the summit was the definition of the basic parameters for the NDB. According to the joint declaration of the sixth annual summit, the new development bank would have initial authorized capital of US$100 billion, with initial subscribed capital of US$50 billion to be equally shared among members. Moreover, the members also announced that the headquarters of the NDB would be located in the city of Shanghai in China, and also announced the citizenship of the top staff positions of this bank. Likewise, the BRICS announced the signing of the treaty for the establishment of the CRA that had originally been announced in the fifth summit in Durban. The joint statement also stresses the disappointment of the group with the continued delays in implementing the approved 2010 IMF reforms. The 2015 BRICS summit was held in Ufa, Russia. The main addition of this joint summit declaration relative to previous ones was the announcement of the entry into force of the NDB and of the CRA; both initiatives had been discussed and announced in previous summits. Importantly, the joint statement expressed deep disappointment over the delay in implementing the 2010 IMF reforms highlighted earlier, with a special emphasis on the delay by the US Congress in ratifying this reform. As in previous summits, there were no other specific policies announced during the summit that would be jointly pursued by these nations in multilateral forums.

29.3.  Limited Cooperation and Asymmetries The previous two sections detailed the increase in the economic importance of the BRICS group in the global economy, as well as the steady increase in their economic

Brazil and the BRICs    629 interdependence from 2001 to 2011, while noting that it appears that this interdependence is not the result of explicit joint policies adopted by the group. Furthermore, the joint statements issued after each annual summit seem to suggest a very narrow body of specific policies that these countries have agreed to cooperatively pursue over time. This section proceeds in two steps. First, we outline the most visible specific policies that the BRICS have agreed to pursue in a cooperative way. Second, we argue that this limited degree of cooperation in policy terms reflects strong asymmetries within the group. We group these asymmetries into macro asymmetries and micro asymmetries. The former focuses on the importance of each member of the BRICS in the global economy. The latter focuses on asymmetries in the economic structure of the economies and on asymmetries related to the way in which each member of the group tends to interact with multilateral institutions such as the WTO.

29.3.1.  Current Cooperative Measures As per the previous section, the main specific items of cooperation agreed between the BRICS countries appear to be the following: • The formation of a New Development Bank (NDB) and the establishment of a Contingent Reserve Agreement (CRA) • Support for the implementation of the Governance and Quota IMF reform approved in 2010. By any standard, this list is rather short and modest, particularly in light of the prominence of the BRICS acronym in many policy discussions across various fields. Among other things, the list is short since it leaves out cooperation regarding specific policies related to important multilateral institutions such as the WTO and the United Nations. It is important to note just how modest the items in this list are. Consider the formation of the NDB, an initiative that Rodrik (2013) labels as the most ambitious plan of this group thus far. First, note that each member of the BRICS already has financial institutions whose stated objective is to either promote economic development or related targets such as exports. These existing financial institutions are of considerable magnitude, some of them dwarfing in size the proposed NDB. In 2013, when the BRICS confirmed that the NDB would be established, the Brazilian development bank made loans totaling $88 billion, its Chinese counterpart $240 billion. The details announced during the BRICS sixth annual summit suggests that the NDB would be significantly smaller than the existing banks and, according to The Economist (2014), the NDB “joins an alphabet soup of regional and national development banks the lending of which already dwarfs the $52.6 billion the World Bank disbursed last year.” As posited earlier, another way of considering the relative importance of the list of joint policy objectives is to consider its omissions. This article focuses on economic cooperation but, even if we were to expand our interests to other areas, we would not

630   Peri Silva be able to add much to the aforementioned list. Let us consider the issue of cooperation among the BRICS in the United Nations. There is no doubt in policy circles that the governments of Brazil and India aspire to a permanent seat on the United Nations Security Council (UNSC). Patriota (2013) elaborates on the reasons to expand the number of permanent members of the Security Council.28 However, Pant (2013) concludes that “China and Russia have little incentive to seek a change in the global political institutional fabric. They have a stake in preserving the status quo. . . .” Stuenkel (2013) explains that Russia is reluctant to support modifications due to “its fear that an expansion of permanent UNSC voices might diminish its influence. This sentiment is particularly strong since permanent membership in the UNSC is, aside from its nuclear capacity, Russia’s only basis for global power status.” In conclusion, the main opposition to the expansion of the role played by BRICS countries in the United Nations may actually come from certain BRICS members themselves.29 Of course, these are not the only omissions that confirm the limitations in the cooperative policy efforts undertaken by the members of the BRICS. Another obvious example involves the lack of a clear coordinated effort involving WTO negotiations, an issue that we return to later. Next, we elaborate on the asymmetries in the characteristics of the BRICS members that we believe are hampering this group’s ability to create and pursue objectives in a cooperative fashion.

29.3.2. Asymmetries among the BRICS As discussed in the introduction, policymakers in the BRICS group had considerable expectations that these countries could jointly exert significant influence in determining important world affairs. Luiz Felipe Lampreia argues that the BRICS group had the non-​explicit objective of offering a counterbalance to the United States’ weight in delineating international affairs (Lampreia 2012).30 To pursue this objective, it would be imperative that this group become more influential in the world economy over time. This necessary condition seems to be satisfied in part since the members of this group represent the largest economies among non-​OECD countries,31 but also because their aggregated economic performance over time has been satisfactory, as suggested by the numbers in Table 29.1. Still, the members of the BRICS are independent political entities, and cooperation among them depends on their mutual interest. As such, any collaboration between the group will be based on it being the case that each member can play an important role in pursuing international affairs; this could imply either members occasionally supporting the pursuit of some objectives not especially prioritized by themselves in exchange for support on other objectives that form their true priorities, or, more simply, a situation where the BRICS countries happen to share the same preferences and goals in terms of international affairs outcomes. The former condition requires that there exists a balanced situation within the group in terms of the importance of each member in the world economy. The extent to which this is lacking can be labeled the degree of “macro

Brazil and the BRICs    631 asymmetry” among the BRICS. The latter condition requires that the BRICS countries share similar preferences in international affairs. The extent to which this is lacking can be labeled the degree of “micro asymmetry.” Next we discuss evidence and examples suggesting that both macro and micro asymmetries are very pronounced among BRICS members. These asymmetries have made it highly challenging for this group to effectively coordinate their behaviors in international forums.

29.3.3. Macro Asymmetries Table 29.1 demonstrates why the acronym BRICS has become so popular in the international media; on average, the group grew faster than the rest of the world in the period 2000–​2014. Accordingly, its proportional participation in world trade has also increased. However, this evidence hides substantial macro asymmetries within the group. Table 29.3 provides evidence that the macro asymmetries among the BRICS members have always been significant and actually have increased over time. Columns 1 and 2 show the GDP of the members of the BRICS in the years 2000 and 2014, respectively. A comparison between these columns reveals that China was by far the largest economy of this group in the early years of the first decade of the 2000s, when the BRICS acronym was created, and its size relative to the other BRICS economies had significantly increased by 2014. Column 2 shows that the Chinese economy is now larger

Table 29.3 BRICS: Macro Asymmetries

GDP 2000 (1)

Growth GDP 2000–​2005 (4)

Growth GDP 2006–​2010 (5)

Growth GDP 2011–​2014 (6)

Share World Exports 2000–​2014 Polity IV (7) (8)

1,212 3.36%

3.15%

4.51%

2.24%

0.98%

8

GDP 2014 (2)

Growth GDP 2000–​2014 (3)

Brazil

771

China

1,424

5,270 9.72%

9.54%

11.27%

8.05%

7.74%

0

India

603

1,598 7.01%

6.23%

8.34%

6.48%

1.41%

9

Russia

567

999 4.59%

6.78%

3.72%

2.45%

2.03%

5

South Africa

214

329 3.21%

3.89%

3.13%

2.30%

0.59%

9

40,980 58,180 3.99%

4.39%

3.94%

3.40%

—​

3

World

Columns 1–​7 use information from the World Development indicators (World Bank). Columns 1 and 2 use GDP measures based on 2005 US$ billion. Columns 3–​7 provide the average by country of the respective variables for the years 2000 to 2014. Column 8 uses the indicator of democracy (DEMOC) compiled by the Polity IV Project. This indicator varies from 0 to 10 with higher values indicating greater popular participation in the political process.

632   Peri Silva than the other four BRICS economies combined, and is more than 15 times larger than the economy of South Africa. Moreover, a comparison between columns 1 and 2 also shows that the Indian economy has also performed better than the other BRICS, since in 2000 it was the third largest economy, after the Chinese and the Brazilian economies, but by 2014 had become the second largest economy of the group. Column 3 helps to grasp these dynamics. This column shows that the Chinese and Indian economies have grown at a much higher pace than both the rest of the BRICS and the world economy.32 At the same time, column 3 shows that Russia has grown at a slightly higher pace than the world economy, while Brazil and South Africa have actually grown at an average pace slightly below that of the world economy.33 Columns 4–​6 show the average growth rates of each economy during various yearly intervals. As discussed in the preceding, the Chinese economy grew at a faster pace than the other BRICS across all years, followed by India. The economies of Brazil, China, and India grew at their fastest pace from 2006 and 2010 when the decision was made by this group of economies to hold annual summits.34 This time period coincides with the most recent commodity boom, where astonishing economic growth in China helped to propel the price of certain commodities (for example, iron ore) to record levels, which benefited the exports of well-​endowed natural resource economies such as Brazil’s (e.g., Foroohar 2016). In column 6, we can see that all BRICS economies slowed down from 2011 to 2014. This was most pronounced in the case of the Brazilian economy, which has faced the greatest drop in the rate of growth among these economies and had essentially stagnated by 2014. The Brazilian economy has suffered from a multiplicity of factors, including the end of the commodity boom, the mismanagement of its economic policy, and a series of corruption scandals related to its state-​owned oil company (Petrobrás).35 The spectacular performance of the Chinese economy is also related to the fact that it has become a major player in world markets. Column 7 shows that over these years the Chinese share of world exports has surpassed the share of exports from the other BRICS combined. Needless to say, the greater presence of China in world markets has been noted both by developed countries and by the other BRICS members.36 Growing exposure to Chinese products tends to represent competition to exports from the other BRICS in third-​party markets as well as in their own domestic markets. In the case of Brazil, Bown (2015) states that about 40% of the Brazilian anti-​dumping tariffs imposed from 2008 to 2013 were directed against Chinese products. This time frame coincides with the initial five annual summits held by the BRICS, revealing important sources of conflicts among these countries due to their macro asymmetries. Moreover, Facchini et al. (2010) show that the greater the share of imports from China in Brazilian imports, the higher the tariffs and nontariff barriers imposed by the Brazilian government tend to be. In sum, China has been the main target of protectionist devices applied by the Brazilian government. Column 8 reveals another major asymmetry among the BRICS members: the group comprises countries with very diverse political systems. In that column, we use an index with information on the degree of democracy across countries from the Polity IV project.37 The information shown in column 8 reveals that Brazil, India, and South

Brazil and the BRICs    633 Africa are well-​established democracies, while China’s political system is highly undemocratic. Russia has a more hybrid political system with elements typical of both democratic and dictatorial systems. This political asymmetry may also affect the economic side of the BRICS group. For instance, that the Chinese political regime is clearly not democratic makes it difficult to construct a preferential agreement between China, Brazil, and Brazil’s MERCOSUR partners, since one of MERCOSUR’s pillars is its democratic clause. As pointed out by Pant (2013), these asymmetries have made it more difficult for the BRICS to devise a unified response to important world affairs such as the UNSC’s decision on the international intervention in Libya in 2012. On that occasion, China, Russia, Brazil, and India abstained from voting, while South Africa voted in favor. Moreover, the BRICS group also has been very explicit in its criticism of the lack of reforms in multilateral institutions such the IMF and the World Bank, as is clear from the joint statements issued after their summits. However, they have been unable to offer a unified alternative suggestion for how to allocate the leadership position of these institutions.38 In sum, the identified macro asymmetries make it more challenging for the BRICS to cooperate, since the Chinese economy is far more influential in world affairs than the rest of the BRICS. This fact makes it more difficult for the BRICS to pursue objectives not supported by the Chinese government. Second, Chinese economic success tends to generate frictions between other BRICS countries and China. Finally, differences in the political regimes of the BRICS countries tend to make their relationship even less cooperative.

29.3.4. Micro Asymmetries As noted in the preceding, macro asymmetries arise out of the fact that each of the BRICS countries has different types and amounts of influence on world affairs. In this section, we focus on asymmetries in the micro-​economic structure of the economies. We begin by considering the degree to which each member of the BRICS depends on intra-​BRICS trade. Table 29.2 shows that the degree of economic interdependence has increased, in particular between 2001 and 2011. However, this fact hides import differences between countries, as seen in Table 29.4. Column 1 shows the share of total trade (imports plus exports) dedicated to trade with other members of the BRICS group for the year 2014. This measure informs us about the trade dependency of each country on the other members. According to column 1, China is the least dependent on trade with other BRICS countries, followed (in a distant second) by India. The most dependent country on trade with other BRICS is Brazil, followed by South Africa. The relatively high degree of dependency of the Brazilian economy on other BRICS is clearly related to its pattern of trade, as discussed in the following. Columns 2 and 3 show the share of intra-​BRICS trade that each country undertakes specifically with Brazil and China. In column 2, it is clear that the only BRICS country that significantly relies on trade with Brazil is China, while, conversely,

634   Peri Silva Table 29.4 BRICS: Micro Asymmetries Share of Intra-​ BRICS Trade with China 2014 (3)

Trade Imbalance with China /​ Trade with BRICS (4)

Correlation of RCA with Brazil 2014 (5)

Average Applied MFN (Imp + Exp)/​ Tariff GDP 2014 (6) (7)

Share of Tariff Lines with Water 2014 (8)

—​

25.61%

14.01%

90.3%

–​0.43

60.05%

9.35%

3.46%

Share of Trade with BRICS 2014 (1)

Share of Intra-​BRICS Trade with Brazil 2014 (2)

Brazil

16.22%

—​

78.62%

China

6.18%

30.14%

—​

—​

India

10.96%

13.91%

68.41%

–​24.26%

0.15

42.52%

12.07%

94.91%

Russia

11.09%

6.12%

85.51%

3.82%

0.30

55.20%

7.59%

11.90%

South Africa

14.92%

4.98%

68.58%

–​19.64%

0.49

58.63%

7.44%

63.19%

World

9.94%

92.05%

Calculations made by the author. Columns 1–​5 and 7–​8 use information from the World Integrated Trade Solution (WITS). Columns 1–​5 rely on data originally aggregated at the 3-​digit of the ISIC. Columns 7 and 8 use data aggregated at the 6-​digit of the Harmonized system. Column 4 uses an average of the trade imbalance (exports-​imports) by country over the years 2000–​2014. Column 6 uses information from the World Development Indicators over the years 2000–​2014. In column 7, average tariff is defined as the simple average by country for the year of 2014, while in column 8, tariff water is determined by the presence of a bound tariff five percentage points higher than the applied MFN tariff.

all BRICS members have most of their intra-​BRICS trade flows with China. Again, this is particularly important for the Brazilian economy, which displays the greatest intra-​ BRICS share of trade, and, at the same time, is considerably dependent on China. Thus, trade among the BRICS is centered on China, especially in the case of the Brazilian economy. Columns 4 and 5 focus on other, albeit related, sources of friction among the BRICS. In column 4 we consider the average degree of trade imbalance between China and the other BRICS from 2000 to 2014, while in column 5 we consider the degree of correlation between measures of revealed comparative advantage for Brazilian industries and for the other BRICS. In column 4, we can see that India and South Africa have accumulated severe trade deficits with China over the years, a fact that has been a source of concern for all governments involved (BRICS Business Council 2016). This is particularly important given that China pursues an exchange rate policy that can be blamed for keeping the value of its currency relative to the US dollar, a policy not followed by the other BRICS economies.39 In column 5, we compare the revealed comparative advantage of Brazil with the other BRICS. It is clear that Brazil tends to share similar (positive) revealed comparative advantages with other natural resource–​ abundant economies (South Africa and Russia), while the opposite is true for China. As it is well known, China is

Brazil and the BRICs    635 very competitive in manufacturing products, and its expansion in the world economy is primarily driven by exports of manufactured products. Table 29.5 confirms the information given in column 5 of Table 29.4, by providing information on the most important exporting industries for Brazil, China, and Russia in the year 2014. Brazil’s main exports are agricultural products and mineral resources, while China primarily exports manufactured products, in particular electrical equipment and machines. Russia has the most concentrated export mix, since its exports of mineral resources account for more than 65% of total exports. This information also explains why Brazil has become so dependent on trade with China. As China grew at impressive rates while simultaneously building basic infrastructure, the Brazilian economy was able to sell agricultural products (e.g., soybeans) and minerals (e.g., iron ore) to the resource-​scarce Chinese economy. As the Chinese economy slowed down, and the price of commodities decreased from 2012 to 2014, the Brazilian economy was one of the most affected. Thus, these countries’ trade policies face asymmetric challenges generated by different export and import mixes, different exchange rate regimes, and different paces of economic growth. There are also significant asymmetries pertaining to how each member of the BRICS has interacted with the WTO over time. These asymmetries have existed since long Table 29.5 Main Exports from BRICS Share of Exports—​ Russia 2014 (6)

Main Exports—​Brazil 2014 (1)

Share of Exports—​Brazil 2014 (2)

Main Exports—​China 2014 (3)

Share of Exports—​China 2014 (4)

Main Exports—​ Russia 2014 (5)

Agriculture and livestock products (ISIC 111)

20.62%

Electrical equipment (ISIC 383)

22.22%

Crude petroleum and natural gas (ISIC 220)

45.34%

Food manufacturing (ISIC 311)

17.07%

Machines except electrical (ISIC 382)

18.12%

Refined oil products (ISIC 353)

22.90%

Metal ore (ISIC 230)

15.60%

Apparel (ISIC 322)

6.42%

Chemical products (ISIC 351)

5.12%

8.35%

Metal products (ISIC 381)

6.10%

Non-​ferrous metal products (ISIC 372)

3.86%

Crude petroleum and natural gas (ISIC 220)

Calculations made by the author using information from the World Integrated Trade Solution. The data were originally aggregated at the 3-​digit of the International Standard Industrial Classification (ISIC, revision 2). Share of exports corresponds to exports at the industry level over the country’s total exports for 2014.

636   Peri Silva before the decision to hold annual BRICS summits. Brazil, India, and South Africa have been WTO members for several decades, while China joined the WTO in 2001 after a 15-​year negotiation process, and Russia became a member only in 2012. These are very important differences since the “old WTO members” represented by the democracies of the group have had significant leeway in binding tariffs at very high levels and also in applying tariffs at higher than average levels over time due to the Enabling Clause.40 The same does not apply to the “new WTO members,” who have had to substantially liberalize their economies compared to the old members.41 Columns 6 and 7 in Table 29.4 show two measures of the degree of openness of these economies, and column 8 shows the share of tariff lines with significant amounts of “tariff water.”42 Columns 6 and 7 show that Brazil and India are the least open economies of the group, while China is the most open considering the share of trade over GDP; South Africa is the most open using average tariff levels. Likewise, column 8 shows that the “old WTO members” have tariff water available in at least 90% of their tariff lines, while China has tariff water only in 3.46% of its tariff lines. This implies that the old WTO members of the BRICS group have much more leeway in changing trade policy under the WTO than the new WTO members of the group. Tariff levels in Brazil are particularly higher on manufacturing products, where the average protection level applied is 17.79%, as against only 5.88% applied to non-​manufacturing products. On the contrary, India’s average tariff level on non-​ manufacturing products is 16.77%, which is by far the highest among the BRICS members.43 As for China, its protection level is well-​distributed, with an average tariff of 9.53% in manufacturing products and an average of 7.28% in non-​manufacturing products. These micro asymmetries among the BRICS have certainly made it more difficult for these countries to cooperate within the WTO. After eight rounds of successful multilateral negotiations sponsored by the WTO (as well as by its predecessor, the General Agreement on Tariffs and Trade [GATT]), the Doha round of negotiations collapsed in 2008. Hopewell (2014) explains that the key issue was discussions involving the “security safeguard measures (SSM),” an important tool for food-​importing countries like India to protect domestic producers from surges in imports. In this case, India championed the SSM, a position that was not supported by Brazil, the United States, and the European Union. Still, an agreement could have been reached if China had agreed to concessions on market access to manufacturing and agricultural products of interest to the United States. Instead, it sided with India on the SSM while refusing to offer any assurance sought by the United States. This clearly pitted Brazil, which has a comparative advantage on agricultural products, against India and China, which are major importers of agricultural products but still also have a significant peasant population (for a similar point, see Hoekman and Kostecki 2009, 286). While this section has focused on trade-​related micro asymmetries, extending the analysis to FDI-​related asymmetries reveals even more pronounced differences across BRICS members. Table 29.6 focuses on intra-​BRICS FDI inflows for the year 2012.44 According to column 1, China received no FDI inflows from other BRICS members in the year 2012. On the other hand, South Africa received 23.67% of its FDI inflows from

Brazil and the BRICs    637 Table 29.6 BRICS: Asymmetries in FDI Share of FDI Inflows from BRICS 2012 (1)

Share of FDI Inflows—​ Brazil 2012 (2)

Share of FDI Inflows—​ China 2012 (3)

Share of FDI Inflows—​India 2012 (4)

Share of FDI Inflows—​ Russia 2012 (5)

Share of FDI Inflows—​ South Africa 2012 (6)

Brazil

0.49%

—​

0.00%

0.03%

0.00%

0.64%

China

0.00%

0.49%

—​

0.81%

0.89%

18.15%

India

0.89%

0.00%

0.00%

—​

0.03%

1.81%

Russia

0.93%

0.00%

0.00%

0.00%

—​

3.07%

South Africa

23.67%

0.00%

0.00%

0.05%

0.00%

—​

Calculations made by the author. Columns 1–​6 use information from the United Nations Trade and Development agency.

other BRICS members, with more than 80% of these flows originating in China. Brazil, India, and Russia received less than 1% of their FDI inflows from other BRICS members in the year 2012, and these FDI flows originated mostly in China.45 Thus, intra-​BRICS FDI inflows also primarily originate from China, and South Africa is the BRICS country most dependent on them.46

29.4. Looking Ahead The BRICS group has become a popular topic of discussion in international affairs. This group of countries represents some of the largest economies outside of the OECD group and, taken together, they have grown relatively faster than the rest of the world from 2000 to 2014. It is commendable that the leaders of large developing countries are regularly meeting to discuss a wide range of international affairs, and understandable that the citizenry of these nations take pride in being part of this select group of nations. However, it is surprising that the most visible step taken by the members of this group has been to establish another development bank that simply joins the ranks of many others already in existence. It is not even clear exactly how this development bank would operate differently from well-​established institutions such as the World Bank. As mentioned earlier, behind these efforts seems to lie a desire to create a counterbalance to the weight of the United States and the European Union in international affairs. It is certainly the case that the BRICS can contribute toward desirable changes in the balance of power. However, that requires a certain cohesion around ideas and values that the group can defend in unison. The asymmetries that exist at the macro and micro economic

638   Peri Silva levels among the members of this group tend to prevent them from reaching their main objective, which is to use the power of the group to steer more effectively the course of events in world affairs. This conclusion does not imply that China will not become increasingly important in international affairs; all indications suggest that China is becoming an essential partner in multilateral forums. One example is that China’s quota at the IMF is now the third highest, almost identical to that of the Japanese, and just below the US quota share, with the implementation of the 2010 IMF reforms in 2015.47 This appears simply to reflect the sheer size of the Chinese economy. Moreover, China also displayed its importance by being a member of the select group of countries that decided the fate of the Doha round. It is also a permanent member of the UNSC, which gives it veto power over policies that are consequential in world affairs. Its continued superior economic performance poises it to increase its central role in international affairs. However, this remains true whether or not the BRICS continue to hold annual summits, as discussed in this chapter. Brazil’s participation in the BRICS has not yet generated many concrete outputs for the country. Before the initial annual summits were held, the expectations of Brazilian authorities were very high regarding the BRICS. Former chancellor Celso Amorim commented in 2008 that “the UN represents the only political space that incorporates all different value systems. There exists a consensus among the BRICS members that is fundamental to implementing a widespread reform of the UN, in such a way as to keep it as the center of the international order to which we aspire. Indefinitely delaying its reform, including the UNSC reform, will increase the risk of diminishing its authority. At the WTO, the articulation that already exists among Brazil, China, and India in the G20 shows the potential of these cooperative efforts.”48 Unfortunately, Brazilian expectations regarding the negotiations for reform of the UNSC did not materialize, and it is certainly the case that China and Russia have incentives not to promote them, as discussed earlier; likewise, conflicts of interest among the BRICS members were fundamental in derailing the Doha round. This was a clear setback for Brazilian diplomatic efforts to promote Brazilian agribusiness interests using a multilateral platform. Brazilian diplomacy currently faces multiple challenges. First is a challenging domestic economic situation, where GDP growth slowed from 2012 to 2014, before a severe recession hit in 2015 and 2016. This certainly calls into question any influence that Brazil might exert on international partners, diminishing Brazilian diplomatic bargaining power. Second, it is clear that Brazilian diplomatic efforts have focused on South-​South relationships, as evidenced by Brazil’s trade alliance with Southern Cone economies (MERCOSUR), its dialogue with BRICS economies, and its efforts in becoming a leader of the so-​called G20 group in WTO negotiations.49 This chapter has not considered MERCOSUR, but it is well known that Brazilian economic ties with its Southern Cone partners have become less important over time. More important, the expansion of MERCOSUR has been limited to accepting Venezuela as the fifth member of the trade bloc in 2012, while no other preferential agreement has been finalized with other major economies or with other major trade blocs (EU,

Brazil and the BRICs    639 NAFTA, ASEAN, etc.). This situation may become problematic for Brazil given the rapid pace at which other countries, as well as existing PTAs, have negotiated new preferential agreements, including the advanced negotiations to conclude the Transpacific Partnership and the Transatlantic Partnership, linking NAFTA, the European Union, Japan, and other countries in Asia through preferential market access. There is a good chance that Brazil may become increasingly economically isolated. The failure of the BRICS group to deliver on Brazil’s high expectations of it certainly does not help on that front.

Notes 1. More precisely, the economies of Brasil, Russia, India, and China represented the 11th, 16th, 14th, and 7th largest economies in the world in 2001 using GDP measures based on 2005 US$. 2. Notice that O’Neill’s argument reflects a straight-​line projection of economic growth for the BRICS based on their economic performance for the years 2000 and 2001. Sharma (2012) criticizes this approach by arguing that “as with previous straight-​line projections of economic trends, however, such as forecasts in the 1980s that Japan would soon be number one economically, later returns are throwing cold water on the extravagant predictions”, p. 1. 3. For instance, the Wall Street Journal mentions Western fears of a new world order with growth in the BRICS nations. “BRICS’ New World Order Is Now on Hold.” Wall Street Journal, January 19, 2016. http://​www.wsj.com/​articles/​BRICS-​new-​world-​order-​is-​now-​on-​hold-​1453240108. 4. See the Brazilian government’s website for the BRICS cooperation, at http://​BRICS.itamaraty. gov.br/​. 5. Information available at http://​blog.planalto.gov.br/​bric-​tem-​papel-​fundamental-​na-​ construcao-​de-​uma-​nova-​ordem-​mundial/​. 6. Luiz Felipe Lampreia, who was the Brazilian chancellor from 1995 to 2001, has publicly stated that it is very positive for Brazil to belong to the BRICS group since he believes that it is the first time in Brazilian history that the country has been part of such a prestigious group (Lampreia 2009). 7. Brazil is the fifth largest country in the world according to the World Bank (http://​data. worldbank.org/​indicator). By the end of the 1970s, its total population reached more than 122 million, which placed Brazil as the sixth most populous country in the world. 8. One could expand the list of serious macroeconomic challenges by adding the increasing level of income inequality in Brazil, which led the country to be listed as one of the most unequal countries in the world. 9. Mexico announced a default on its sovereign debt in 1982, while Brazil officially announced its decision to default in 1987. 10. Franco (1999, 237)  describes this social pressure as an attempt to rescue what some policymakers interpreted as the social debt in Brazil. As he explains, this social pressure led politically organized groups to benefit from the increasing government expending levels at the expense of the least organized members of the Brazilian society. Baer (2008, 121) discusses the particular role played by the approval of the 1988 constitution in Brazil in increasing the demand for fiscal resources. 11. See Chapter  6 by Fernando de Holanda Barbosa in this volume for details of these stabilization plans.

640   Peri Silva 12. The reforms included the deepening of the privatization process initiated in 1991 by auctioning additional state-​owned enterprises in the manufacturing and service sectors, the continuation of liberalization reforms in the energy sector of the economy, and important modifications to the Brazilian social security system, among others. Details about the privatization process in Brazil can be found in Pinheiro (2011). 13. Lampreia (2010b, 242)  indicates that the personal interactions between then-​Brazilian president Fernando Henrique Cardoso and his US counterpart Bill Clinton in 1997 did not include any ideas to use the relationship between Brazil and the United States to improve global governance. Lampreia argues that what made it possible to include Brazil in discussions involving global governance later in the first decade of the 2000s was the gen­ eral improvement of the economic conditions of Brazil over time, conditions that were not yet fully present in the 1990s. 14. Commodity prices peaked in February 2011 and then strongly decreased until 2015. http://​ www.indexmundi.com/​commodities/​. 15. Rodrik (2013) states, “Yet, in a strange case of life imitating fantasy, BRICS—​the original four countries, now joined by South Africa—​have formed a grouping of their own with regular meetings and policy initiatives”, p. 1. 16. Joint declarations for each of the BRICS annual meetings can be found on the Brazilian government website at http://​BRICS.itamaraty.gov.br/​declarations-​action-​plans-​and-​ communiques/​listadecplan. 17. Feenstra (2016) elaborates on the theoretical derivation, as well as on key empirical applications, of the gravity model of international trade. 18. An exception would be if the group performance were actually explained by only one member of the BRICS growing much faster than the rest of the world while the others all grew more slowly than the rest of the world. But this was certainly not the case for the periods between 2001–​2005 and 2006–​2010. 19. They find that PTAs can double trade between members relative to non-​members considering a 10-​year time frame to account for phase-​in periods of bilateral tariffs reduction. 20. See details at http://​rtais.wto.org/​UI/​publicsummarytable.aspx. 21. Sperlich (2015) states that India and the Southern African customs union (SACU) are considering the formation of a PTA. 22. The WTO interprets a PTA as being partial in scope if it involves preferential liberalization in a portion of sectors only. On the contrary, Article XXIV of the WTO states that customs unions (CUs) and free trade areas need to promote substantial duty free trade among members. Moreover, the distinction between a CU and an FTA is that the former requires a common external tariff while the latter allows members to set external tariffs independently. 23. Similar information is obtained by considering the comprehensive dataset by Scott Bair and Jeffrey Bergstrand that is available at https://​kellogg.nd.edu/​faculty/​fellows/​ bergstrand.shtml. 24. For instance, India formed an FTA with the Association of Southern Asian Economies (ASEAN) in 2009, while Russia formed an FTA (Common Economic Zone) with Belarus, Kazakhstan, and Ukraine in 2004, and then agreed to expand this FTA in 2012 to additionally include Armenia, Kyrgyz Republic, Moldova, and Tajikistan. The original Common Economic Zone involving Russia and other members became a CU (The Eurasian Economic Union) in 2015. Likewise, South Africa has been a member of the Southern African Customs Union since 2004 and joined an FTA with members of the European free trade agreement (EFTA) in 2008.

Brazil and the BRICs    641 25. Text of the declarations from the first and second annual meetings of the BRICS can be found at, respectively, http://​www.BRICS.utoronto.ca/​docs/​090616-​leaders.html and http://​www.BRICS.utoronto.ca/​docs/​100415-​leaders.html. 26. Patriota (2013) offers arguments in favor of expanding the number of permanent and non-​ permanent members in the United Nations Security Council. 27. Text of joint statement available at http://​www.BRICS.utoronto.ca/​docs/​130327-​statement.html. 28. Lampreia (2010b, 39, 183–​184) explains that some Brazilian policymakers have held aspirations toward securing a permanent seat on the UNSC since the end of World War II, and some of these aspirations peaked during the celebration of the 50th anniversary of the UN in 1995. 29. In the same vein, Pant (2013) argues that “[n]‌ot surprisingly, it’s US President Barack Obama—​not Hu Jintao or Vladimir Putin—​who promised India that he would help in this goal during his visit to New Delhi in November 2010.” 30. This argument seems to be confirmed by other important Brazilian political figures. For example, the former Brazilian chancellor Celso Amorim claims that the BRICS contribute in a very important way by leaving the world less dependent on a sole source of power in the G7 (França 2015). 31. With the exception of South Africa. 32. The fact that (only) China and India had consistently strong economic performances among the BRICS economies from 2000 to 2014 has not gone unnoticed. Recently Jim O’Neill, the creator of the BRIC acronym, suggested, “I might be tempted to call it just ‘IC’ or if the next three years are the same as the last for Brazil and Russia I might in 2019!” Bloomberg News, January 8, 2015. http://​www.bloomberg.com/​news/​articles/​2015-​01-​08/​ bric-​in-​danger-​of-​becoming-​ic-​says-​acronym-​coiner-​jim-​o-​neill 33. IPEA (2011, 4) suggests that the fact that China has not only become the locomotive of global growth, but also the locomotive for growth of the other members of the BRICS, has important implications for the stability of the group. 34. IPEA (2011) reveals that even during the peak years of economic growth, the heterogeneity among the BRICS was remarkable since China was responsible for 40% of the worldwide growth of demand between the years 2008 and 2009, while Brazil and Russia were responsible for 6.3% and 3.5%, respectively. 35. According to the IMF (2016), Brazilian GDP actually shrank by 3.8% in 2015 and another estimated decrease in GDP of 3.5% is predicted for 2016. If this information is confirmed, it would be the first time since the financial crisis of 1929 that the Brazilian economy shrank during two consecutive years. 36. Autor, Dorn, and Hanson (2013) find negative effects of the US increasing exposure to Chinese exports on employment and wages across its local labor markets. 37. The democracy index varies from 0 to 10, with higher scores indicating a greater degree of democratic values in a political system. 38. Pant (2013) points also to vague statements made by the BRICS on the case of the crisis in Syria’s civil war and on the nuclear dispute between Iran and the West as evidence of the divergent views that are present in the group. 39. Pant (2013) details the disenchantment of the other BRICS with China’s manipulation of the currency. 40. The Enabling Clause was approved by the signatories of the General Agreement on Tariffs and Trade in 1979. It grants differential and preferential treatment to developing countries in multilateral negotiations, and also allows developed nations to extend preferential

642   Peri Silva access to their markets to developing nations on a unilateral basis. Details can be found on the WTO website at https://​www.wto.org/​english/​docs_​e/​legal_​e/​enabling1979_​e.htm. 41. See Subramanian and Wei (2007) for a detailed discussion of why this distinction matters in determining the effects of WTO membership on bilateral trade flows. 42. Defined as the difference between the bound and applied MFN tariffs greater or equal to 5%. Notice that changes in the cutoff difference between bound tariffs and applied MFN tariffs in defining the presence of “water” does not change the relative importance of tariff water across BRICS. 43. Calculations made by the author using information at the 6-​digit level of the harmonized system from the World Integrated Trade Solution (WITS). 44. The latest year of data available. 45. The low importance of intra-​BRICS FDI inflows for the larger four members of the groups contrasts with the increase in importance of these countries as recipients of FDI during the first years of the 2000s (Kappor and Tewari 2010). 46. Pant (2013, 6) points to the fact that “it was never obvious why South Africa, with a population of 40 million and an economy one-​twentieth the size of China, was added to the BRICS group in the first place, apart from the fact that China wanted South Africa in, as China views South Africa as an access point for the whole African continent.” The information on Table 29.6 seems to provide support for the idea that it was in the interest of China to bring South Africa into this group of nations. 47. That does not mean that cooperation among the BRICS members has led to this quota allocation. China already had the 6th highest quota share at the IMF, given the quota reform implemented in 2006 prior to any official BRICS annual summits. Brazil has also had its quota increased from the 14th highest in 2006 to the 10th highest with the 2010 IMF reform. Other large developing countries that are not part of the BRICS group have also had their IMF quota increased. This is so in the cases of Mexico, Indonesia, and Malaysia, for instance. On the other hand, South Africa has had its IMF quota share reduced despite being part of the BRICS group. See IMF (2011). 48. Newspaper article by Celso Amorim, “Os Brics e a reorganização do mundo.” In Folha de São Paulo, June 8, 2008. Translation by author of this chapter. 49. The G20 in the WTO corresponds to a negotiating group formed by 20 developing countries, where Brazil and India play leadership roles. See Hoekman and Kostechi (2009, 285).

References Autor, David H., David Dorn, and Gordon H. Hanson. 2013. “The China Syndrome: Local Labor Market Effects of Import Competition in the United States.” American Economic Review 103: 1553–​1597. Baer, Werner. 2008. The Brazilian Economy: Growth and Development, 6th edition. Boulder, CO: Lynne Rienner. Baier, Scott, and Jeffrey Bergstrand. 2007. “Do Free Trade Agreements Actually Increase Members’ International Trade?” Journal of International Economics 71: 72–​95. Bown, Chad. 2015. “Temporary Trade Barriers Database.” The World Bank. Available at http://​ econ.worldbank.org/​ttbd/​ BRICS Business Council.2016.“India-​ChinaTradeDeficit at $44.7 Billion in April-​January.” Available at http://​www.brics-​info.org/​india-​china-​trade-​deficit-​at-​44-​7-​billion-​in-​april-​january/​.

Brazil and the BRICs    643 The Economist. 2014. “An Acronym with Capital.” Available at http://​www.economist.com/​ news/​finance-​and-​economics/​21607851-​setting-​up-​rivals-​imf-​and-​world-​bank-​easier-​ running-​them-​acronym. Facchini, Giovanni, Marcelo Olarreaga, Peri Silva, and Gerald Willmann. 2010. “Substitution and Protectionism: Latin America’s Trade Policy and Imports from China and India.” World Bank Economic Review 24: 446–​473. Feenstra, Robert. 2016. Advanced International Trade:  Theory and Evidence, 2nd edition. Princeton, NJ: Princeton University Press. Foroohar, Rana. 2015. “Why the Mighty BRIC Nations Have Finally Broken.” Time, November. http://​time.com/​4106094/​goldman-​sachs-​brics/​ França, Francis. 2015. “Brics deixam o mundo menos dependente de uma única fonte de poder.” Interview with Celso Amorim for DW.com. Available at http://​www.dw.com/​pt-​br/​brics-​ deixam-​o-​mundo-​menos-​dependente-​de-​uma-​%C3%BAnica-​fonte-​de-​poder/​a-​18567022. Franco, Gustavo H. B. 1999. O desafio brasileiro: Ensaios sobre desenvolvimento, globalização e moeda, Vol. 34. São Paulo: Editora. Hoekman, Bernard, and Michel Kostechi. 2009. The Political Economy of the World Trading System. Oxford: Oxford University Press. Hopewell, Kristen. 2015. “Different Paths to Power: The Rise of Brazil, India and China at the World Trade Organization.” Review of International Political Economy 22: 311–​338. International Monetary Fund. 2011. “Quota Formula Review: Data Update and Issues (Statistical appendix).” International Monetary Fund. 2016. “World Economic Outlook.” January 2016 update. IPEA. 2011. “Relações comerciais e de investimentos do Brasil com os demais países do BRICS.” Comunicados do IPEA 86. http://​repositorio.ipea.gov.br/​bitstream/​11058/​4597/​1/​ Comunicados_​n86_​Rela%c3%a7%c3%b5es_​comerciais.pdf Kapoor, Radhika, and Ritika Tewari. 2010. “FDI in the BRICS:  Changing the Investment Landscape.” The Perspective of the World Review 2: 147–​170. Lampreia, Luiz Felipe. 2009. “BRICS: O assunto da semana.” O Globo blog, 18/​06/​2009. http://​ blogs.oglobo.globo.com/​lamp,reia/​post/​BRICS-​assunto-​da-​semana-​196698.html Lampreia, Luiz Felipe. 2010a. “Brics.” O Globo blog, April 16, 2010. http://​blogs.oglobo.globo. com/​lampreia/​post/​brics-​284366.html Lampreia, Luiz Felipe. 2010b. “Depoimento.” CPDOC, Fundação Getúlio Vargas, Rio de Janeiro. Lampreia, Luiz Felipe. 2012. “Os BRICS em Sua Dimensão Real.” O Globo blog, March 29, 2012. http://​blogs.oglobo.globo.com/​lampreia/​post/​os-​brics-​em-​sua-​dimensao-​real-​438207.html O’Neill, Jim. 2001. “Building Better Global Economic BRICS.” Global Economic Paper No. 66, Goldman Sachs. Patriota, Antonio. 2013. “Globalizing the Security Council.” Project Syndicate. Embassador Anotio de Aguiar Patriota, June 03, 2013. https://​www.project-​syndicate.org/​commentary/​t he-​urgent-​need-​to-​reform-​t he-​un-​s ecurity-​council-​by-​antonio-​de-​aguiar-​ patriota?barrier=accessreg. Pinheiro, Armando Castelar. 2011. “Two Decades of Privatization in Brazil.” In The Economies of Argentina and Brazil:  A Comparative Perspective, edited by Werner Baer and David Fleischer, 252–​281. Cheltenham, UK: Edward Elgar. Pant, Harsh V. 2013. “The BRICS Fallacy.” The Washington Quarterly 36 (3): 91–​105. Rodrik, Dani. 2013. “What the World Needs from the BRICS.” Project Syndicate. Dani Rodrik, October 04, 2013. https://​www.project-​syndicate.org/​commentary/​the-​brics-​ and-​global-​economic-​leadership-​by-​dani-​rodrik?barrier=accessreg

644   Peri Silva Sharma, Ruchir. 2012. “Broken BRICS: Why the Rest Stopped Rising.” Foreign Affairs 91 (6): 1–​7. Sperlich, Yvonne. 2015. “From Power Transition to Economic Integration Theory: A Review of the BRICS Literature.” Working Paper No. 15-​10-​1, University of Geneva. Stuenkel, Oliver. 2013. “Brazil and UN Security Council Reform: Is It Time for Another Big Push?” Post-​Western World. Available at http://​www.postwesternworld.com/​2013/​06/​30/​ brazil-​and-​un-​security-​council-​reform-​is-​it-​time-​for-​another-​big-​push/​ Subramanian, Arvind, and Shang-​Jin Wei. 2007. “The WTO Promotes Trade, Strongly but Unevenly.” Journal of International Economics 72: 151–​175.

Chapter 30

B r azilian Tr a de a nd International E c onomi c Prospects i n a n Anti-​G l obaliz at i on E ra Donald V. Coes

Werner Baer’s many contributions to our understanding of Brazilian development include a career-​long interest in that nation’s trade and external economic relations. Those international connections often played a central role in the development process. This chapter attempts to continue in that tradition, examining Brazil’s sometimes reluctant embrace of world markets in recent decades. Although both economists and the Brazilian population as a whole appear to have convinced themselves that fuller participation in international trade, capital, and labor markets is generally a good thing, this comes at the same time that some of Brazil’s major international economic partners are moving in the opposite direction, increasingly questioning the advantages of “globalization.” We begin with a brief review of some of the major trends in Brazilian trade and international economic policy, including its reaction to international commodity market and capital market shocks in recent decades and the politically driven emphasis on preferential trade. In section 30.2 we examine the question of how “open” the Brazilian economy is, even after some moves toward greater linkages to world markets. Despite the timid liberalization that developed over several decades, it is an economy that could benefit from much more from participation in global markets. In this section we also consider Brazil’s links to other economies through labor markets. Section 30.3 examines some of the major anti-​globalization trends in Brazil’s principal economic partners, and attempts to identify some of their causes. We argue that Brazil’s links to other economies through capital and labor markets are at least as important as are its commodity trade links. Trends in these markets may help explain some of the anti-​globalization attitudes it faces in the future. This section concludes the chapter with some basic observations

646   Donald V. Coes on how an economy like Brazil might best react in a contemporary anti-​globalization world. With the half-​century consensus in support of internationally open trade, capital, and labor markets seemingly under siege, the way ahead for Brazil is far from clear.

30.1.  Major Trends in Brazilian External Economic Policy: Reluctant Globalization As one of the world’s important economies, Brazil is something of a latecomer to globalization in recent times. This was not always the case, and much of Brazil’s five centuries of economic experience may be understood well through its relationship to international markets. Indeed, a popular interpretation of Brazil’s economic history since 1500 is the view that it is a series of commodity-​driven export cycles, beginning with sugar in the sixteenth and seventeenth centuries in the Northeast. Subsequent gold and diamond cycles in the eighteenth century helped move much of the population and economic action further south. The rise of coffee production, at first dependent on slave labor in the Rio de Janeiro area and then on a tide of immigrants in the late nineteenth century in São Paulo, occupies the center stage of Brazilian economic history until 1930. One of the key themes in Werner Baer’s numerous and widely cited interpretations of twentieth-​century Brazilian experience, building in part on seminal earlier work by Celso Furtado and others, was import-​substitution industrialization (ISI) (see, for instance, Baer 1972, 2013; Furtado 2007). This profound change in Brazil’s economic history was inadvertent in its early stages before World War II, as the collapse of coffee prices after 1929 laid the foundations for nascent industrialization centered on São Paulo. By the 1950s, however, the economic distancing from world markets induced by ISI was consciously driven by successive policymakers. Brazil used a number of trade and exchange rate policy instruments in its drive to industrialize. In the 1950s multiple exchange rates were used for different transactions. (This and a number of other important post–​World War II events in Brazilian international economic policy are summarized in Table 30.1.) The system reflected the clear hierarchy through which policymakers viewed different economic activities, with manufacturing heavily favored over traditional agriculture. The multiple exchange rate system did not survive the decade, but the implicit hierarchy continued in the structure of tariff protection. This policy bias was confirmed in studies of “effective rates of protection” later made in the 1960s and 1970s. They suggested that the level of value added in some industrial activities was a multiple of levels in more traditional activities.1 If the benefits of ISI were initially more apparent than were its efficiency costs, its limits had come into view by the end of the decade. After the authoritarian military government took power in Brazil in 1964, domestic stabilization understandably occupied

Brazilian Trade and International Economic Prospects    647 Table 30.1 Selected Events in Brazilian External Trade and Exchange Rate Policies 1953–​1957

Multiple exchange rate system, with different fixed nominal exchange rates for different transactions

1957

Replacement of many fixed duties, NTBs, and other trade restrictions with an ad valorem tariff system. Lei de Similaridade Nacional (Law of Similars), existing since 1911, effectively implemented

1967

Generalized reduction in wide range of ad valorem tariffs

1967–​1973

Extensive system of export incentives created, among them “drawbacks” for domestic taxes paid on production of exports, and exemption from domestic value-​added taxes and taxes on manufactured inputs

1968

Crawling peg (minidesvalorações) policy linked nominal exchange rate to domestic inflation in Brazil and its major trading partners

1974

First “petroleum shock,” with government attempt to prolong the boom (milagre econômico) of the 1968–​1973 period

1979

Second oil shock and further balance of payments deterioration, with 30% nominal devaluation in December 1979

1982

Mexican external debt moratorium and virtual closing of external debt markets to most other Latin American borrowers, including Brazil

1988

Initiation of a number of smaller trade policy changes, among them the elimination of redundant tariffs and simplifications of the tariff schedule

1990–​1992

Series of reductions in ad valorem tariffs, significantly lowering effective levels of protection

1991

Treaty of Asunción among Brazil, Argentina, Uruguay, and Paraguay, with the eventual objective of forming a common market (MERCOSUL)

1994

Plano Real implemented, partly supported with fixed real exchange rate “anchor”

1999

Abandonment of exchange rate anchor, after sharp deterioration in currency reserves, with replacement by domestic macroeconomic targets

2012

Suspension of Paraguay from MERCOSUL, followed by admission of Venezuela

more of the government’s attention than did trade and international economic policy. But there were a number of important changes in external economic policy, which together moved Brazil cautiously in the direction of greater economic openness. In 1967 the complex and messy system of preceding decades was overhauled, with a generalized reduction in nominal tariffs. With the combination of fixed nominal exchange rates and high levels of domestic inflation, however, Brazil’s price competitiveness in international markets could not be maintained. This problem was addressed in 1968, when Brazil adopted a successful system of administered changes in the nominal exchange rate, or a “crawling peg.” This policy helped maintain some of Brazil’s price competitiveness internationally, despite the military government’s difficulties in fully stabilizing the domestic price level. Trade policy in this period continued to be highly interventionist. In contrast to earlier years, however, there was more emphasis on export expansion, rather than simply on import restriction. The instruments used included special

648   Donald V. Coes tax treatment for manufactured exports and other market interventions. Few of these measures extended to other tradable activities, so that the “hierarchy” favoring industrial activity implicit of earlier ISI policies continued on the export side. Despite these reforms, Brazil’s external balance was threatened by the first “petroleum shock” in 1973–​1974. The response of policymakers was partly facilitated by two major changes in international capital markets in the immediately preceding years. First, short-​term loans from international banks to “sovereign borrowers” like Brazil developed rapidly in the late 1960s. They were made on a short-​term (usually six months) basis at the US prime or the London Interbank Offered Rate (LIBOR) interest rate plus a “risk premium” that varied by country. This effectively protected lenders from real losses resulting from rises in world inflation when loans were renewed. Second, major banks shared their participation in given loans, thus lowering the risk for a single bank. This combination, allied with a perception that Brazil would be a major market in ensuing decades, permitted Brazil to finance most of the current account deficit that resulted from the first petroleum shock with only a small impact on domestic economic growth. External adjustment proved to be much harder—​and ultimately, impossible—​after the second oil shock of 1979. A large nominal devaluation late that year effectively ended a decade of real exchange rate stability that been established by the 1968 crawling peg. It did not produce either domestic or external balance, ushering in a “lost decade” that in reality lasted much longer. Brazil’s external payments position deteriorated further, raising sharply the costs of servicing the external debt. This turbulent period was marked by another large devaluation in early 1983. The policy consequences of this loss of control were a series of failed stabilization plans, with trade and international economic policy questions receiving little attention. Some modest changes in trade policy, however, were implemented beginning in the government of José Sarney (1985–​1990), consisting primarily of a “cleaning up” of a number of redundant tariffs (i.e., those nominal tariffs at well above nearly prohibitive levels). During the Collor de Melo government (1990–​1992), several trade policy changes were implemented that were more than just cosmetic. Most of the list of suspended imports was eliminated, and administrative controls were reduced. Despite domestic macroeconomic chaos and the forced resignation of Collor, the reforms were carried out and maintained. Between 1989 and 1993, the simple average of tariffs fell from 32.1% to 13.5%, while the tariff average weighted by value added in each sector fell from 29.4% to 12.5%. The other major trade policy trend of this period was the development and extension of the Southern Cone Common Market (MERCOSUL), whose origins dated to agreements in the late 1980s between the Argentine and Brazilian governments of Raul Alfonsin and José Sarney. This initiative, too, survived the macroeconomic and political turmoil of the early 1990s. Whether or not it had a significant impact on Brazilian trade and economic growth is still a matter of debate, and raises the larger question of whether or not Brazil’s interests are served by geographically preferential trading arrangements like MERCOSUL. During the governments of Lula (2003–​2011) and Rousseff (2011–​2016), preferential trade arrangements became a major focus of trade policy in a political context, even though trade with non-​MERCOSUL partners continued to be far more important. Some of this was due to the domestic political orientation of both governments, which

Brazilian Trade and International Economic Prospects    649 viewed the expansion of trade with some of Brazil’s Latin American neighbors and with Africa as a welcome alternative to alleged “dependence” on high-​income trade partners like the United States and Europe. This period of subordination of trade policy based on more traditional growth and efficiency-​based objectives to more ideological ones was well illustrated in 2012 by Brazil’s support for Venezuela’s admission to MERCOSUL. Despite the attempts of successive Partido dos Trabalhadores (PT) governments to use trade policy as a link to geographical and ideological neighbors, the real trade story of the last decade was driven by commodity price change and the expansion of markets outside the MERCOSUL area, especially that of China. In reality, these two trends in the last decade are not really separable. Their major effects are shown in Table 30.2 and Figure 30.1. The “China effect” dates back at least to the preceding decade, as that nation’s

Table 30.2 Brazilian Export Revenue, 2006–​2015 Rate of Growth from Preceding Year

Billions US Dollars 2006

137.8

16.3

2007

160.7

16.6

2008

197.9

23.2

2009

153.0

–​22.7

2010

201.9

32.0

2011

256.0

26.8

2012

242.6

–​5.3

2013

242.0

–​0.2

2014

225.1

–​7.0

2015

191.1

–​15.1

200 180

140 120

Index Points

160

100

2006

2008

2010

2012

2014

2016

80

Figure 30.1.  Price index for Brazilian exports, 2006–​2016. Source: Fundação Centro de Estudos do Comércio Exterior (FUNCEX).

650   Donald V. Coes high rate of gross domestic product (GDP) growth allowed it to gradually surpass the United States as Brazil’s major trade partner. The growth in the value of Brazilian exports was briefly interrupted by the world financial crisis in 2008–​2009. Its strong and rapid resumption, however, provided the euphoric “boom” foundations for domestic macroeconomic policies that would be unsustainable were commodity prices to fall. In such a buoyant external market, efficiency in trade policy received little attention from policymakers. By 2012, however, the party was ending, as the export price index in Figure 30.1 shows so clearly. This time the interruption was not temporary, and bust replaced boom. In retrospect, the boom was a missed opportunity to move Brazil toward the kind of openness that already had been achieved by most of its trade partners. In the following section, we look in greater detail at the question of how open an economy Brazil really is.

30.2.  Is Contemporary Brazil an Open Economy? The major trends in Brazil’s trade and external economic policies reviewed in the preceding section suggest that it has become a more open economy since the heyday of ISI. Although correct in the context of Brazil’s own history over more than half a century, a conclusion that Brazil is an open economy does not stand up as well in the context of the contemporary world economy. Defining “openness” in an external economic sense raises a number of questions, in part because Brazil’s connections to the rest of the world run through much more than merchandise and commodity trade. A commonly used starting point, which is easy to use with available data, is the ratio of trade flows—​usually exports—​to the size of the economy. The dependence between country size and the relative importance of trade, however, is a serious issue, complicating international comparisons of trade/​GDP ratios like those considered in the following. Despite this difficulty, a comparison of Brazil’s export/​GDP ratio to that of most other large economies shows that it is not a very open economy when compared with most of its trade partners. Table 30.3 shows the export coefficients and purchasing power parity (PPP) GDP (in PPP estimates) of 22 countries. Although data are available for more than 200 countries, this group accounts for most of the world’s trade and GDP, and includes Brazil’s most important trading partners. The countries in this group are ordered in decreasing order of the share of exports in GDP. As might be expected, large economies like the United States or China have lower export coefficients than do smaller open economies like the Netherlands or Belgium, a trend made even more obvious when the data is plotted as in Figure 30.2. But the size of the economy is only part of the story. The GDP of the United States or of China, measured in PPP terms, is over five times that of Brazil. Yet the export coefficients of both of these much larger economies are higher than those of Brazil,

Table 30.3 Exports as a Share of GDP and PPP GDP Export Share (%)

PPP GDP ($ Trillions)

Netherlands

86.7

0.73

Russia

29.4

3.46

Belgium

85.7

0.45

Venezuela

28.1

0.55

Korea, Rep.

55.3

1.66

China

27.4

16.16

Germany

51.0

3.49

France

27.2

2.44

Chile

34.9

0.39

Turkey

25.3

1.42

Spain

32.5

1.50

India

24.2

6.77

Mexico

31.9

2.01

Australia

20.7

1.01

United Kingdom

31.7

2.32

Argentina

16.0

0.78

South Africa

30.6

0.66

Japan

14.9

4.62

Canada

30.3

1.52

United States 13.5

16.80

Italy

29.8

2.05

BRAZIL

Export Share (%)

12.3

PPP GDP ($ Trillions)

3.01

Sources: World Bank development indicators. Export coefficients are averages of 2011–​2013 data; GDP is PPP estimate from 2013 data.

100 90

Netherlands Belgium

Export share of GDP (%)

80 70 60

S. Korea

50 40

Germany

Chile Russia

30

India

France

20

Australia

China

Japan

U.S.

Argentina

10

BRAZIL

0 0

2

4

6

8

10

12

14

16

18

PPP GDP ($ trillion)

Figure 30.2.  Export share and GDP. Source: Table 30.3. For clarity, country labels for Canada, Italy, Mexico, South Africa, Spain, Turkey, the United Kingdom, and Venezuela have been omitted. Export coefficients and GDP for these countries are between 25% and 33% and between $0.39 trillion and $2.3 trillion, respectively.

652   Donald V. Coes which is in last place among this group of 22 countries. Although the countries listed in Table 30.3 account for most of the world’s trade, our conclusions would be little changed by expanding the sample to all 208 of those countries for which World Bank export coefficient data are available. In this case, the Brazilian economy would be more open than only eight countries, none of them important participants in the world economy.2 It is difficult—​perhaps impossible—​to escape the conclusion that even today, and even after significant trade policy reforms in the 1960s and 1990s and active diplomatic participation in discussions of multilateral trade liberalization, Brazil is a less active participant in the world economy than it could be. Some of this may be the legacy of early decades of inward-​looking industrial and commercial policy. Other factors may include the sharp swings in the real exchange rate in the past several decades. An alternative way of evaluating the Brazilian economy’s openness rests not on the size of trade in relation to GDP, but the degree to which its markets are connected to the rest of the world. High degrees of connection do not necessarily imply large movements of goods or services for which there is an integrated market. Gold markets in the heyday of the gold standard, for example, were almost perfectly connected among the major economies on the standard, yet physical movements of the commodity were often relatively small. The mere possibility of price arbitrage may be enough to connect prices within countries. When transport costs are small relative to the value of the good, markets may have a high degree of “openness.” As the costs of obtaining price information have fallen steeply in the last several decades, partly through the development of the Internet, this kind of connectedness or openness has grown significantly. At both a wholesale and a retail level, Brazilian buyers are much more aware of price differences than they were in past decades. This change, which might be characterized as the “Miami shopping mall” or the “Google it” effect, is largely due to the fall in price acquisition costs. It is a product both of the Internet and of the much greater numbers of Brazilians visiting the United States and other foreign markets. Viewed in this way, government trade and commercial policy are degrees of interference in the connection, imposed in pursuit of specific political or economic objectives. The vast literature on the efficiency effects of such forms of interference, and the ranking of different degrees of intervention in terms of their welfare cost to the economy as a whole are clearly relevant to the Brazilian case. Our more limited objective here, however, is simply to gauge this degree of interference in contemporary Brazil compared to its major trade partners and a number of other middle-​income economies. Estimating internationally comparable degrees of trade policy restrictiveness that take into account both tariffs and nontariff barriers (NTBs) is an empirically daunting task. An important contribution has been made, however, by Lee et al. (2009, 2012), based on earlier theoretical work by Anderson and Neary (1994). Their estimates convert NTBs into an overall measure of trade restrictiveness.3 Table 30.4 shows some of these estimates for 2009, and is a subset of their estimates for over 100 countries. Among those shown are the BRICS (Brazil, Russia, India, China, and South Africa), as well as several comparable middle-​income countries and Latin American trade partners. Since

Brazilian Trade and International Economic Prospects    653 Table 30.4 Indices of Overall Trade Restrictiveness All

Agricultural

Manufactured

Goods

Goods

Goods

Brazil

21.7

24.7

21.5

Mexico

15.2

28.3

13.8

Russian Federation

15.2

22.5

13.7

India

14.9

69.5

13.1

Venezuela

14.4

43.9

9.9

Australia

11.3

29.5

9.8

China

9.7

14.4

9.3

Argentina

9.2

9.6

9.1

Korea, Rep.

9.0

49.8

4.5

Japan

8.9

38.3

4.7

Turkey

7.3

32.0

6.0

Chile

7.2

22.3

5.9

United States

5.7

17.0

4.5

European Union

5.6

33.6

3.4

South Africa

5.3

12.8

4.7

Canada

4.8

18.5

3.3

Unweighted average

9.6

29.5

7.7

Source: Lee, Neagu, and Nicita (2012). The estimates are percentage measures of the overall level of protection, with NTBs converted to tariff-​equivalent percentages. The unweighted averages for the countries of the table exclude Brazil.

the members of the European Union have a common external tariff and NTBs, there is one single estimate for this group. Only a few smaller African economies had higher indices of trade restrictiveness than did those of the countries shown in Table 30.4.4 But among major economies, Brazil stands out for its comparatively high level of overall import restrictiveness, estimated for all categories of imports to be over 21%. Unlike many of its major trade partners, including the United States and Europe, trade restrictions in Brazil are less concentrated in the agricultural sector. Since this sector has a lower weight in Brazil’s imports than do other sectors, especially manufacturing, the estimate of the overall level of trade restrictiveness is heavily weighted by non-​agricultural imports. The continued high level of restriction on agricultural trade, often a multiple of the level in manufacturing, suggests that Brazil’s focus on protection in this sector is justified. The comparatively greater trade restrictiveness in Brazil, even after several periods of trade liberalization in the 1970s and 1990s, has produced an economy in which not only imports, but also exports, are a smaller part of national product. This is a consequence of

654   Donald V. Coes the fact that, over the longer run, approximate current account balance would result in changes in export values moving in the same direction as import values, even if the two trade flows diverge significantly in short time periods. As argued previously despite significant trade policy reforms in the 1960s and 1990s Brazil is a less active participant in the world economy than it could be. There appears to be a clear link between recent levels of trade restrictiveness, as noted in Table 30.4, and Brazil’s current position among the most closed of the world’s major economies. Brazil has come a long way from the heyday of ISI of the 1950s and 1960s, as our survey of trade and exchange rate trends in section 30.1 suggests. But the rest of the world has not stood still, as the preceding data show. When put in the context of the contemporary world economy, rather than in the context of its own past, Brazil could probably participate considerably more in world trade than it does. Connection to the rest of the world through trade, however, is only part of Brazil’s of the economy’s international linkage. We conclude this section with a brief consideration of labor market linkages. In the context of the forces behind anti-​globalization now appearing in some of Brazil’s major economic partners, linkages through labor markets may be at least as important as trade in goods and services. During the nineteenth and early twentieth centuries, Brazil was one of the world’s important destinations for international migration, with migrants from the Mediterranean supplying much of the labor necessary for the expansion of the coffee sector. With the fall in activity in this sector after 1930, international migration to Brazil slowed down significantly, and did not return to earlier levels even after World War II. The result of this change was a transformation of Brazil from a population with a relatively high proportion of foreign-​born residents in the early twentieth century to one with relatively few immigrants a century later. In 1900, 6.2% of the Brazilian population was foreign born, with 21% of the population of São Paulo and 24% of the Distrito Federal (now the city of Rio) foreign born. By the 1920 Federal Census, these proportions had fallen slightly, to 4.9% nationally and to about 18% and 20%, respectively, in São Paulo and in the Distrito Federal.5 Although some immigration, primarily from Portugal, continued in the decades after World War II, Brazil’s role in the world economy as a destination for emigrants from Europe was behind it. In recent times, the much smaller foreign-​born population of Brazil has been dominated by “neighborhood” immigration, with significant numbers of Bolivians, Paraguayans, Argentines, and Uruguayans. With the erosion of macroeconomic stability in Brazil after the two oil shocks, Brazilians began to emigrate in significant numbers in the 1980s for the first time in its history. Parallel to the case in goods markets, some of this reflected the greater familiarity of Brazilians with conditions in foreign markets, including labor markets. By the 1970s, many more Brazilians had traveled to other countries than in earlier decades, and by the end of the twentieth century many had personal contact with family members or friends residing in higher-​income foreign countries, notably the United States and Western Europe. In contrast to earlier decades, this familiarity was not limited to higher-​income elites.

Brazilian Trade and International Economic Prospects    655 Table 30.5 Migrant Population and Share of Population Population (Thousands)

Percent Share

United States

46,627

14.5

Germany

12,006

14.9

Russian Federation

11,643

8.1

Saudi Arabia

10,186

32.3

United Kingdom

8,543

13.2

United Arab Emirates

8,095

88.4

Canada

7,836

21.8

France

7,784

12.1

Australia

6,764

28.2

Spain

5,853

12.7

Italy

5,789

9.7

India

5,241

0.4

Ukraine

4,835

10.8

Thailand

3,913

5.8

Pakistan

3,629

1.9

BRAZIL

1,847

0.9

Source: Migration Policy Institute, tabulated from UN, Dept. of Economic and Social Affairs (2015).

Cross-​country comparison of the share of immigrants in total population can be ambiguous. This is partly due to the fact that in some areas, particularly in the Middle East, many immigrants are temporary residents, rather than people who are likely to spend the rest of their lives in the host country, as did so many Italians, Portuguese, and other Europeans in Brazil a century ago. Another factor is the reallocation of populations in the former Soviet Union, with the Russian Federation both sending and receiving population to and from other members. Even with these two caveats, however, Brazil is no longer an important immigrant destination. Compared to the other countries ordered by the size of their immigrant population in Table 30.5, only India has a lower relative share of immigrants in its population. Table 30.5 also indicates that the United States has been by far the major destination for international migrants. Brazilian emigration patterns are no exception to this global pattern, with the United States the principal receiving country for Brazilians who move abroad. Table 30.6 presents some recent estimates of the numbers of Brazilians resident outside Brazil. Despite the rough agreement among the three non-​independent estimates for most countries, the effect of undocumented emigration produces some wide variation in estimates of Brazilian emigrants in the United States. Since the

656   Donald V. Coes Table 30.6  Estimates of Brazilian Population by Country of Residence (figures are in 1000s)

United States Japan

Itamaraty (Late 2010)

Migration Policy Institute (Late 2015)

Wikipedia (Undated; Accessed 2016)

1,067

348

451

210

182

210

Paraguay

202

74

202

Portugal

140

130

140

Spain

128

100

128

United Kingdom

118

56

118

Germany

96

64

95

Italy

67

104

85

France

45

57

80

Switzerland

44

47

44

Belgium

43

8

43

Argentina

41

49

27

Bolivia

32

26

31

Netherlands

27

16

27

Uruguay

26

10

26

Canada

25

29

23

236

NA

NA

Other countries

Sources: Ministério das Relações Exteriores, SGEB/​DCB/​DBR (June 2011); Migration Policy Institute, based on UN, Dept. Economic and Social Affairs tabulation; Wikipedia (unsourced).

Itamaraty (Foreign Ministry) estimates include extensive embassy and consular estimates of the undocumented US resident population, this estimate may be the most informative one in an economic context.

30.3.  Anti-​Globalization: What Is Behind It and What Might It Mean for Brazil? Brazil today is unquestionably more open to world markets than it was a half-​century ago, at the height of the ISI period. But to become as open as are most of the world’s major economies, it still has a significant way to go, as some of the evidence presented in the preceding section shows. Disillusionment with both populist protectionism and with the focus on second-​level world markets like MERCOSUL, moreover, has opened

Brazilian Trade and International Economic Prospects    657 the domestic political possibility in Brazil of a refocus on its largest trading partners. Yet those partners, especially the United States and Western Europe, may now be moving in the opposite direction. Both the “Brexit” vote in the United Kingdom and the explicitly protectionist message successfully put across by Donald Trump in the United States suggest this. Anti-​globalism attitudes are apparent in the domestic politics of a number of other important Brazilian trade partners as well. Much of the reaction to “globalization” in the high-​income countries has been expressed as opposition to trade agreements. Although the North American Free Trade Agreement (NAFTA) is the most prominent target of anti-​global forces, other actual and proposed trade agreements like the Trans-​Pacific Partnership (TPP) are also either in the crosshairs or already dead. If Brazil is to avoid the most serious consequences of a rising political tide of anti-​globalization, it is essential that its origins be better understood. We look briefly here at some of the main features of the anti-​globalization trend, asking whether it is really trade in goods and services that is the problem. Despite Brazil’s sometimes reluctant participation, the growth of world trade since the end of World War II has produced a significant rise in living standards in all countries that were part of it. Countries that took the most advantage of more open markets, most notably China after 1980, saw per capita income rise to unprecedented levels. Some of the increase can be explained by the diffusion and acquisition of new technologies. But there is little doubt that a substantial share the world’s postwar real income growth accords well with the centuries-​old pro-​trade nostrums of economists ever since Adam Smith and David Ricardo. Why then are “neoliberal” trade policies the villains for significant parts of the electorate in high-​income countries, both on the left and on the right? The answer may lie in part in the failure of economists to explain more clearly to the public what more liberal trade policies—​whether in the United States, the United Kingdom, or in Brazil—​can be expected to do, and what they probably will not do. Another answer may be that opposition to open trade has political and cultural roots that may not be based at all on actual trade patterns. In this latter interpretation, anti-​globalization is not explained by any basic economic argument, an issue to which we return after considering some purely economic arguments for contemporary opposition to more open trade. The basic “gains from trade” argument is that if there are differences in prices between separate national markets, then there is a potential welfare gain for both economies. This is a deductive argument, not an inductive one, and no amount of case studies, new data sets or econometric techniques, or new sets of politicians can overturn this conclusion. As a guide to either making policy or understanding politics, however, the indisputable validity of the gains from trade argument is not very helpful. There are several important limitations implicit in the trade liberalization argument. Together they help explain some of the opposition to more open markets, both in “emerging markets” like Brazil or in high-​income ones like the United States. The analysis of both winners and losers from any market rearrangement has been a veritable sub-​industry of the economics profession for at least a century. Major

658   Donald V. Coes contributions to our understanding of the income distribution effects of liberalizing trade came from the “factor proportions” theory, which links expansion of exportable production to increases in the incomes of factors used relatively more intensively in those activities, with inverse effects on the importable products side. Subsequent refinements of this argument emphasized the importance of factor inputs of limited mobility, specific to particular activities. Despite the insights this class of models provides, the fact that there are both winners and losers has often been submerged in simplistic debates over free trade versus no trade. A related limitation arises from asymmetries in the identification of winners and losers. The activities that will suffer from greater international competitive pressure are already in existence and are often vocal and well organized. Activities that will expand exportable production in a more open trade environment are less obvious, especially in the initial stages of trade expansion. And even when they do become established, trade liberalization is unlikely to get much of the credit for their success. A third limitation that helps to explain popular mistrust of more open trade arises from the distinction between short-​run and long-​run effects. Although the context is not often stated explicitly, the orthodox “gains from trade” argument rests on a comparison of two economies in which price flexibility has had enough time to allow all productive factors to be fully employed. This assumption is in fact necessary to build the basic model. But it is obviously not a good description of Brazil, the United States, or any other real-​world economy in the short run. This point emerges more clearly in the basic “2 x 2” general equilibrium model that has been the workhorse of international trade theory for many decades. Although usually stated in term of two factors—​labor and capital—​producing two tradable outputs, it can better be reinterpreted here as two broad collections of inputs used relatively more intensively to produce either “exportables” or “importables,” even though such inputs can eventually be used productively in the other sector. Greater opening to trade, like that supported by the United States and Western European policymakers since World War II, pushed by the World Bank and other multilateral institutions, adopted dramatically by China after 1980, or reluctantly and partially by Brazil after the exhaustion of the ISI phase, can be modeled as a shift in production from importable goods toward exportables. The long-​run view of such a shift compares two equilibria in which all factor inputs are fully employed. In this view, trade agreements are not related to changes in net employment of either labor or capital. If NAFTA is a good agreement for the United States or Mexico, it is because it replaces less efficient activities with more efficient ones in a larger common market. There should be no expectation that it has either positive or negative effects on employment. In the long run, the net employment effect of trade liberalization is zero. But this is small comfort in the short-​run real world in which we all must live, rather than in the long run in which we are all dead. Any real-​world path between the relatively closed economy long-​run equilibrium and that of the relatively open one leave a variety of different inputs without work for some time during the economy’s transition.

Brazilian Trade and International Economic Prospects    659 During this transition period, employment of a number of inputs falls in the first stages, resulting in a contraction of production in “importable” sectors. Economic expansion of the “exportable” sectors requires labor training and hiring, capital investment, and repositioning of other economically mobile inputs. If mobility is hampered, moreover, by the inability of workers to sell a house and move to an area of expanding production, by hesitation in financial markets, or by long lead times in investment in human capital, the transition to the new, theoretically more efficient open-​economy equilibrium may take even longer. If this temporary period of less-​than-​full employment were the only consequence of moving toward greater openness, however, we would not have a good explanation of the political opposition to more open trade. Changes in trade policy would be just one more temporary shock to the macro-​economy, similar to changes in fiscal or monetary policy or shocks from financial markets. Domestic demand shocks in countries like the United States, however, tend to generate much larger effects on employment than do changes in trade policy, yet they are far less likely to generate the kind of controversy that surrounds NAFTA or Brexit. The much deeper problem, especially when we recognize specialized factors in different sectors, is that the losers are a very different set of people from the winners. The losers, moreover, may have to bear the burden of their losses for a long time—​possibly permanently—​long after the economy as whole has returned to “full employment.” These sectoral differences, especially in the United States and higher-​income Western European countries, have major demographic, educational, and geographic features. The model sketched in the preceding suggests that factors that can move more easily out of contracting sectors into expanding ones are more likely to be winners than losers. Younger workers with only a few years of experience employed in a declining sector may be better positioned to escape from it than are workers whose career is mostly behind them. Educational attainment, which also better positions people for new activities, has a parallel effect. Voter behavior in the UK plebiscite on Brexit, or in the recent US presidential elections, is consistent with these two hypotheses. Continued membership in the European Union would have passed among younger voters in the United Kingdom; it failed among older ones. In the United States, Trump’s “tear up NAFTA” and “make America great” rhetoric resonated more with older voters than it did with younger ones. Similarly, the anti-​globalism message appears to have received relatively more support among voters with only a high-​school education that it did among those with a college degree. The distribution of people who think themselves “losers” from greater international market integration also has noticeable geographic features. In recent decades, increases in US international competitiveness have come in part from high-​technology centers on both coasts, like California’s Silicon Valley or Massachusetts’s Route 128 Corridor. Although protectionist sentiment is difficult to isolate from other political factors in the recent US elections, it appears to have been significantly stronger in the Rust Belt states of the upper Midwest, like Michigan, both in the presidential primaries and in the gen­ eral election.

660   Donald V. Coes The preceding argument attempts to explain protectionist sentiment primarily as an economically rational reaction by those domestic factors that lose with greater opening to international markets. Anti-​globalization, however, cannot be characterized simply as traditional protectionism against import competition. To explain recent protectionism as “economically rational” is conceding too much. An alternative explanation is that trade policies are the wrong target of workers who feel threatened. A much better explanation of job loss in manufacturing in the United States and most other high-​ income economies is not that the jobs went to China or Mexico, but that they were taken over by capital. Although there is much debate about the size of this effect, there is no question that manufacturing output in the United States has increased sharply over recent decades, even as employment in manufacturing has fallen.6 This strong rise in productivity is much more consistent with greater use of robots and computerization, with intense use of specialized capital, than it is with a transfer of the same jobs to other countries. In the context of Brazilian international trade, this confusion among US and European electorates about the source of the threats to their employment in manufacturing has several implications. First, an erroneous diagnosis that “the jobs are all pouring out of the US,” when many of them are simply being replaced by capital, raises the possibility of an equally erroneous remedy on the part of the United States and other countries where this misperception takes root. If relatively lower real wage countries like Mexico or Brazil are perceived as the destinations for the allegedly transferred jobs, they are likely to be the target of new protectionist pressures. Second, if the major explanation of the fall in manufacturing employment in the high-​income economies is really the substitution of lower and moderately skilled labor with capital and with highly trained labor, then countries like Brazil stand to lose in this respect also. Not only do middle-​level emerging economies become the whipping boys for the jobs they have allegedly stolen, but they are badly placed to supply the factors in increasing demand—​primarily specialized capital and highly skilled labor. The failure of Brazil’s educational system to produce the kind of skill sets that are in ever greater demand only aggravates this difficulty. The high cost of capital, due in part to its being “squeezed out” by public-​sector fiscal disequilibrium, further subtracts from Brazilian competitiveness in this new kind of global market. In a US context, NAFTA may simply be a lightning rod for much more extensive and deeper anti-​foreign sentiments that may have relatively little purely economic content. It may in fact be a waste of time to try to explain anti-​NAFTA positions among voters as the result of even rough calculation of the actual costs or benefits of this trade arrangement for wage earners and consumers. A more likely and more nationalistic possibility is that opposition to NAFTA or to free EU trade is a proxy for concern not with the entry of foreign goods, but with the entry of foreign people. Viewed in this way, anti-​globalization is not an economic phenomenon at all, but a political and even cultural one. The data in Table 30.5 show that the United States is by far the destination of the largest share of the world’s international migrants. With the exception of China, most of Brazil’s

Brazilian Trade and International Economic Prospects    661 major trade partners are among the other principal destination countries, with some of them having an even larger share of foreign-​born residents than does the United States. As a sending country, Brazil now is clearly part of the global pattern, with the United States easily the preferred destination, despite some of the greater legal and bureaucratic obstacles for Brazilians to move there. What might all this portend for Brazil in a world of rising anti-​globalization? The arguments sketched in the preceding provide some measure of our ignorance of the path ahead. In both the United States and in the European Union, the proposed dismantling or curtailment of trade agreements has become a popular vehicle for economic nationalism. But as the cumbersome withdrawal procedure from the European Union suggests, actual implementation of Brexit is likely to take years. Unless the process leads to a larger breakdown in European trade, immediate consequences for Brazilian trade are probably small. In the context of anti-​NAFTA sentiment among populists in the US and defensiveness among its past proponents, however, there may be greater concerns for Brazil. In its early years, the three-​country agreement was viewed by its supporters as the beginning of a movement that could eventually include all the countries in the hemisphere. With NAFTA itself under strong pressures, such a hope today seems way out of date. Brazil will be fortunate if barriers to trade in the Western Hemisphere fall as they have in the past, and in a worst-​case scenario a US tariff war with Mexico could escalate beyond North America. If the hypothesis is correct that it is as much movements of people, as it is goods and services trade, that are the underlying cause of anti-​globalization pressures both in Western Europe and in the United States, Brazil faces some additional obstacles. As a country that has moved over a century from being a receiver to a sender of migrants, Brazil is vulnerable to some additional anti-​globalization headwinds. For Brazilians, particularly those who are younger and more highly educated, who hope to participate in the markets for higher-​skilled labor that have opened internationally in recent decades, there are potential disappointments. As an example, European citizenship based on immigrant ancestors, particularly from Italy and the Iberian countries, has permitted some Brazilians to become residents of the high-​ income economies like the United Kingdom. Implementation of Brexit would close off this possibility in the UK case. Although rising anti-​immigrant sentiment in Europe is not directed at Brazilians, they nevertheless will become the targets of it if European restrictions increase. In the United States, limited visa quotas for highly specialized foreign workers face opposition, although employers may lobby for their continuation. It is far too early to know if the current wave of anti-​globalization will have any lasting effects on Brazil’s economic relations with the world outside, either through trade or through labor markets. But it is clear that the political consensus that underlay liberal trade, investment, and immigration policies in many of Brazil’s major international partners cannot be taken for granted.

662   Donald V. Coes

Notes 1. Early examinations of the effective protection rate by sector, which is the net effect of trade policy on the value added generated in a given activity, include those of Bergsman (1970) and Tyler (1976). 2. Afghanistan, Burundi, Central African Republic, East Timor, Kiribati, Nepal, São Tomé Principe, and Sudan. 3. The estimates by Lee et al. also include tariff-​only measures, as well as the corresponding levels of trade restriction faced by exporters from the respective country to its trade partners. 4. Among them were Tanzania, Sudan, and Senegal (53.3%, 44.1%, and 42.3% overall levels of trade restrictiveness, respectively). 5. Censos Demográficos, 1900 and 1920. 6. Some of this debate is summarized in Autor (2015). For a possibly overly large estimate of the vulnerability of many manufacturing jobs to computerization, see Frey and Osborne (2013).

References Anderson, James E., and J. Peter Neary. 1994. “Measuring the Restrictiveness of Trade Policy.” World Bank Economic Review 8 (2): 151–​169. Autor, David H. 2015. “Why Are There Still So Many Jobs? The History and Future of Workplace Automation.” Journal of Economic Perspectives 3: 3–​30. Baer, Werner. 1972. “Import Substitution and Industrialization in Latin America: Experiences and Interpretations.” Latin American Research Review 7 (1): 95–​122. Baer, Werner. 2013. The Brazilian Economy: Growth and Development, 7th edition. Boulder, CO: Lynne Reiner. Bergsman, Joel. 1970. Brazil:  Industrialization and Trade Policies. Development Center, Organisation for Economic Co-​ operation and Development. Oxford:  Oxford University Press. Bonelli, Regis, Pedro da Motta Veiga, and Adriana Fernandes de Brito. 1997. “As políticas industrial e de comércio exterior no Brasil: Rumos e indefinições.” IPEA Texto para Discussão No. 527. Rio de Janeiro:  Instituto de Pesquisa Economica Aplicada. Available at http://​ repositorio.ipea.gov.br. Coes, Donald V. 1995. Macroeconomic Crises, Policies and Growth in Brazil, 1964–​ 90. Washington, DC: World Bank. Corseuil, Carlos, and Honório Kume, eds. 2003. A abertura comercial nos anos 1990. Rio de Janeiro: IPEA. Available at www.repositorio.ipea.gov.br. Frey, Carl B., and Michael A. Osborne. 2013. “The Future of Employment: How Susceptible Are Jobs to Computerization?” Working paper, Oxford Martin School. Available at oxfordmartin. ox.ac.uk. Furtado, Celso. 2007. Formação econômica do Brasil, 1st edition. Rio de Janeiro:  Fundo de Cultura. Kee, Hiau Looi, Christina Neagu, and Alessandro Nicita. 2010. “Is Protectionism on the Rise? Assessing National Trade Policies during the Crisis of 2008.” World Bank Policy Research

Brazilian Trade and International Economic Prospects    663 Working Paper No. WPS 5274. Washington, DC: World Bank. (Updated 2012.) Available at becon.worldbank.org/​WBSITE/​EXTERNAL/​EXTDEC/​EXTRESEARCH/​0. Kume, Honório, Guida Piani, and Carlos Frederico Bráz de Sousa. 2003. “A política brasileira de importação no período de 1987–​1998: Descrição e avaliação.” In A Abertura Comercial nos Anos 1990, edited by Carols Corseuil and Honório Kume, 9–​39. Rio de Janeiro: IPEA. Tyler, William G. 1976. Manufactured Export Expansion and Industrialization in Brazil. Mohr: Kieler Studien. World Bank, International Comparison Program database. Washington, DC: The World Bank. Available at http://​data.worldbank.org/​indicator.

Chapter 31

The Evolu tion of Fore i g n Direct Inv e stme nt in Bra z i l Breno Augusto da Silva e Silva

31.1. Introduction Foreign direct investment (FDI) is widely considered an important contributor to economic growth (Petri 1997). It is different from other types of capital flow, including short-​term flows, in that it is not speculative. FDI, unlike short-​term capital, is related to longer horizons in terms of planning, execution, and the time necessary to reap returns. This type of investment has become more important globally since the early 1990s. The globalization process promoted FDI, providing opportunities for many countries, whether developed or developing. The possibilities of reaching new markets and of cost reductions led transnational companies (TNCs) to invest abroad to achieve these goals. One of the fundamental characteristics of FDI is that it can facilitate the transfer of technologies to the host country of the investment. Thus, the entry of new companies into an existing industry, or even a new industrial activity that did not previously exist in the country, could also increase the international competitiveness of extant companies in that country. According to Cravino et al. (2007), FDI has grown at a remarkable rate over the past 20 years. In the second half of the 1990s, the growth rate reached 40% per year. In 2004, FDI flows reached an amount of US$648 billion. The stock of foreign capital multiplied roughly fivefold from 1990 to 2004, from US$1.77 trillion in 1990 to $9 trillion in 2004. When looking specifically at developing countries (including Brazil), it is observed that FDI stocks in such countries increased from US$364 billion in 1990 to US$2.23 trillion in 2004, or around six times higher. From World War II onward, Brazil stood out as an FDI recipient among the countries of Latin America. The program of industrialization through import substitution,

The Evolution of Foreign Direct Investment in Brazil    665 emphasized by the government of Juscelino Kubitschek (1956–​1961), meant that Brazil was the developing country that received the most investments from developing countries. However, this position was lost in the early 1980s with the debt crisis and the exhaustion of the import-​substitution model. This situation was only reversed in the early 1990s, when foreign direct investment flows rebounded in Brazil and Latin America. In the Latin American context, Brazil is a major recipient of FDI. However, despite the increase of FDI in Brazil, it is important to note which sectors received investment. Not only the amount of investment, but also the sector to which the investment is directed is very important in determining the contribution to the economic growth of the country.

31.2.  Theoretical Context 31.2.1.  Foreign Direct Investment and Economic Growth According to Kaldor (1975), productivity growth in the economy is related to the growth of employment in the industrial sector and the reduction in the employment rate in non-​industrial sectors. Thus, it is understood that the growth of industrial production is very important to economic growth. Furthermore, the relationship between the variables demand and productivity would be a two-​way street; that is, increases in demand would lead to increases in productivity and increases in productivity would lead to increases in demand. Kaldor’s growth theory focuses on the issue of cumulative causation, where growth is based on the interaction between the division of labor and the extent of the market. Thus, demand would serve as an inducer of changes in labor division, which in turn would change in terms of the efficiency with which goods are produced. The Kaldor model also recognizes that technological changes are endogenously determined, but additionally notes the restriction imposed by demand. Kaldor follows Verdoorn’s law, whereby the productivity growth rate depends on the output growth rate. Kaldor’s empirical analysis takes into account the importance of the manufacturing sector to economic growth. For Kaldor, the growth of manufacturing would lead to an increase in labor productivity, both in the manufacturing sector, and between it and other sectors of the economy (Knell 2004). FDI is directly related to economic growth. Some authors analyze the importance of which sectors receive this kind of investment. Depending on the sector, the investment could bring more competitiveness and returns to the economy. Blomstrom and Kokko (1998) argue that FDI provides competitive gains to the industry of a particular country in which it occurs, as national companies benefit from the spillover of knowledge, benefits that are not exclusively internalized within the company that made the investment. When domestic and transnational companies are operating

666    Breno Augusto da Silva e Silva in the same sector, the latter have no incentive to disseminate their knowledge to local firms, and look for protection through intellectual property, through higher wages to keep their workers, and through seeking countries where the imitative capacity of local companies is not strong. This case is related to the horizontal diffusion of knowledge. In the analysis of a production chain, FDI can occur at different points of the chain: it can be made in segments at the beginning of the chain, for instance, in activities that produce inputs or primary sectors; or it can occur in the final section of the production chain, the processing stage. These allocations of foreign investment imply different results in terms of the appropriation of knowledge by local firms. If the investment takes place at the beginning of the production chain, transnational corporations would not have incentives to provide knowledge to the segments that will receive those supplies, since these segments could, after acquiring such knowledge, produce those inputs used in the production process without purchasing them from the TNCs that were initially suppliers. On the other hand, if FDI occurs at the end of the production chain, TNCs could have incentives to provide knowledge to national companies located in the previous segments in the production chain, since these companies are their inputs suppliers. In this case, the better the relationship in the supply chain, the better the quality of inputs to the investing company. Alfaro (2003) attempts to show that FDI alone, in a general sense, does not necessarily imply positive results in terms of gains for the receiving economy. Such gains depend on the response capacity of the sectors in which these investments are made. The author utilized cross-​sectional data for a group of Organisation for Economic Co-​operation and Development (OECD) countries for the period 1981–​1999. Investment in the industrial sector would be the most desirable, since, according to this study, this would bring various positive results; investment in the service sector, by contrast, showed ambiguous results. According to Blalock and Simon (2009), multinational corporations (MNCs) have greater incentives to disseminate their knowledge to input suppliers than do domestic competitors. This is so owing to the fact that the multinational companies have an interest in making their input suppliers more competitive, which would indirectly contribute to increasing their own competitiveness. Javorcik (2004) analyzes the investments made in Lithuania, considered a transitional economy. The analysis was based on annual data from surveys of companies in the country. The companies included in the survey accounted for 85% of the production in each sector. Through data in an unbalanced panel for the period 1996–​2000, the estimates were made. Lithuania was chosen because it is an economy in transition, in which the skilled labor endowment makes it an appropriate site for the study of the diffusion of knowledge and productivity gains. The results are consistent with the existence of diffusion of FDI through backward linkages, but show no connection to forward linkages in the case of vertical diffusion. In the case of horizontal diffusion, the results show that there is no diffusion. Chakraborty and Nunnenkamp (2008) mention that there are differences in terms of causality depending on which sector FDI arrives in. The authors analyzed FDI directed

The Evolution of Foreign Direct Investment in Brazil    667 to specific industries in India, with the aim of finding linkages between investment and economic growth, using a Granger causality methodology and panel data. The results show that the effects of economic growth provided by FDI differ among sectors. In the case of the primary sector of the economy, according to the tests performed, there was no causality between investment and output growth. In the case of the manufacturing sector, the results showed that there is causality between FDI and economic growth. In the service sector only transient effects were found; nevertheless, an interesting result is that it seems that investments in the service sector caused growth in the manufacturing sector through the effects of knowledge diffusion by backward linkages. For Crespo and Fontoura (2007), an obvious channel of externalities would be the demonstration/​imitation effect promoted by transnational corporations. Introducing new technology can be costly and can involve a number of risks to a national company, as there are information costs relating to the technology, as well as risk related to not knowing whether or not the introduction of this new technology will be generally successful. As such, a transnational company’s introduction of a new technology may represent a demonstration of success, thereby encouraging local companies also to adopt the technology. Haskel et  al. (2002) undertake a panel analysis at the firm level for the UK manufacturing sector for the period 1973–​1992, in order to examine whether there is a diffusion of productivity of FDI to local firms, and whether, in this way, it would be feasible for the countries to pay a certain value. The authors found a positive correlation between total factor productivity (TFP) of local firms and the share of industrial activity of foreign affiliates, at the plant level. Correlation between TFP of foreign plants with the share of industrial activity of local firms was not found. The estimates showed that for each increase of 10 percentage points of foreign presence in the United Kingdom, the domestic industry plants increased productivity by around 0.5 percentage points. However, the value of diffusion by workers seems to be lower than the necessary government incentives—​in some cases, much lower. For Hadiwibowo (2010), the benefits of FDI flows to developing countries are certain: economic growth and capital accumulation. However, there are also risks involved in capital flows and financial integration. Among these risks is a drop in capital inflows, which could result in crises in developing countries. (Generally, although not exclusively, “capital inflows” means resources inflows.) In the specific case of Indonesia, domestic investment moves in the same proportion as savings move. This does not support the issue of open capital markets, since when an investment is directly correlated with savings, there is an indication of capital immobility. A balance of payments analysis showed that the amount of capital that entered the country was lower than the net payment of factor income abroad. As such, there was a net outflow of funds concurrently with capital inflows. Kugler (2000) investigates whether the inflow of FDI generates technological externalities, using data from the Manufacturing Census of Colombia. The author argues that, so far, evidence of new technological opportunities for countries receiving foreign investment through transnational operations is scant. One limitation is how to

668    Breno Augusto da Silva e Silva measure such spillovers; this work attempted to measure intra-​industry externalities instead of inter-​industry externalities. Transnational companies, seeking to avoid risks such as profit reduction, will have no incentive to spread their technology to competitors (i.e., so-​called intra-​industry externalities). However, suppliers of inputs for this MNC located in other industries could benefit from these externalities (i.e., inter-​industry externalities). For Waldkirch (2008), FDI has grown in Mexico since the creation of the North America Free Trade Agreement (NAFTA), making it important to reflect on the effects of its creation on the Mexican economy. Waldkirch seeks to study the impact of FDI on industry productivity and wages in Mexico in the first 10 years following the creation of NAFTA, using data from 1994–​2005. The author finds evidence of an increase in productivity, especially in total factor productivity (TFP). The effects on wages were zero or negative. According to Aitken and Harrison (1999), governments would only promote the entry of foreign investment to take advantage of the spillovers that foreign companies would provide to local companies. In a study of Venezuela, with panel data, the authors find a relationship between foreign ownership and productivity, but that this relationship would only be robust for small firms. For joint venture spillovers for local firms, the results were negative. Thus, the net impact appears to be small. Joint ventures seem to capture the spillovers. For Liu (2008), it is necessary to emphasize the differences between short-​term and long-​term spillovers to analyze whether the effects of productivity gains are positive or negative. According to Liu, the diffusion process initially involves costs, which could have a negative outcome in the short term. However, this process would help companies improve their production capacity, which would lead to productivity gains in the long run. Another important distinction to be made is that between the spillovers inter-​ industry and intra-​industry. This distinction between inter-​industry and intra-​industry must be made due to the fact that often one of these two types of spillover might not be significant, while the other can be quite significant. The literature shows that spillovers of the inter-​industry type through backward linkages can have far greater importance than intra-​industry diffusion, also called horizontal diffusion. The diffusion between industries that occurs in the production chain is called vertical diffusion. Bonelli (1998) states that, from the point of view of economies receiving such benefits, FDI, trade, and financial liberalization would have the purpose of increasing economic growth, promoting greater development, and increasing welfare. However, such benefits would not only be linked to the amount of investment received by a country. Features such as initial market conditions and the production efficiency of both existing companies in the country and new entrants would be critical in determining the earnings appropriation level that would be provided by FDI. Imbs and Wacziarg (2003) offer an analysis of industry concentration and levels of per capita income. According to these authors, economies tend to be more diversified in an early stage of development, with economic activity relatively well distributed among

The Evolution of Foreign Direct Investment in Brazil    669 sectors, but at some point, economies tend to specialize, in a process of sectoral concentration in a U-​shape. Fagerberg (2000) argues that countries that specialize in technology-​ intensive industries such as the electronics industry have a higher growth rate compared to countries that do not pursue this type of expertise. The author analyzes the manufacturing sector for a group of 39 countries, which is composed of different industries, in different stages of development; 24 industries are analyzed from 1973 to 1990. However, the results indicate that, on average, structural change did not lead to productivity growth.

31.2.2.  Functions of Foreign Direct Investment According to Dunning (1993), FDI can be classified into four categories: market-​seeking projects, efficiency-​ seeking projects, resource-​ seeking projects, and asset-​ seeking projects. Market-​seeking projects are investments designed to reach new markets (i.e., those of other countries). Efficiency-​seeking projects are investments that are geared toward seeking greater efficiency in the production process (i.e., lower production costs). Here, the direction is made for both the domestic market and the international market. Resource-​seeking projects aim to obtain abundant raw materials and manual labor at low costs. Investor companies tend to use intensively these abundant resources and allocate the products for export. Asset-​seeking projects are related to companies seeking to acquire new assets through mergers, acquisitions, joint ventures, or even through the installation of new plants. For Laplane and De Negri (2004), from the 1990s in Brazil, FDI was mostly of the market-​seeking type, aimed at the local/​regional market. This type of FDI aims to obtain new markets and tends not to be export-​oriented, especially when there are protections so that more sales are made in the domestic market, which means that there is discrimination in relation to exports and the focus is really the domestic market. In the opinion of Laplane and Sarti (1999), this would have meant loss of attractiveness for the Brazilian industrial sector in relation to the service sector, with regard to FDI.

31.2.3. Institutional Environment as Foreign Direct Investment Attraction Factor Hall and Jones (1999) hypothesize that the difference in the accumulation of capital productivity and thus the output per worker would be related to differences in social infrastructure in different countries, encompassing institutions and government policies that determine the economic environment that agents inhabit. It is in this environment that people learn skills and firms carry out investments and production. Adequate social infrastructure is a determining factor for the level of workers’ skills, and the level of investment by enterprises and technology transfer. Somewhat

670    Breno Augusto da Silva e Silva paradoxically, the state can be not only the greatest inducer of development, through promoting adequate social infrastructure, but also the main obstacle to development when it ceases to do so. The authors conclude that long-​term economic performance depends on institutions and government policies that result in an adequate social infrastructure that is able to facilitate the acquisition of new skills by workers and to create an environment conducive to new investments, technological transfer, and better production by companies.

31.3.  Foreign Direct Investment in Brazil and Latin America Compared to Other Regions It is important to assess the situation of both Brazil and Latin America more broadly in a comparative perspective. How does FDI relate to economic growth nationally and in the region? What are the different functions of FDI, and which are emphasized? What relevance does the institutional environment in this part of the world have? Do Brazil and Latin America follow the same path as other countries and regions in terms of FDI and its effects, or is there a different path? What difference does this make in relation to outcomes such as growth? A number of writers argue for the importance of FDI to Brazil, identifying various gains. For instance, Bonelli (1998) argues that FDI increases competition and competitiveness due to the greater market contestability that is generated by this type of investment. This can be seen in Brazil in the case of nontradable goods such as products of banks and insurance companies. FDI in the nontradables sector has generated significant positive implications for that sector. For Laplane and Sarti (1999), investments made in the 1990s had an important impact on the competitiveness of Brazilian industry. This was especially so in the case of durable and nondurable goods, where investments promoted new products, more modern processes, and in some areas, such as the automotive industry, the creation of new businesses. Rocha (2004) assesses the Brazilian case in the 1970–​2000 period, in order to investigate gains associated with structural change in the manufacturing and mining industries. The author concludes that while there is no evidence that Brazilian industry in this period reversed the trend of growth in productivity, there is also no evidence of a structural bonus, defined as the acceleration of growth due to the reallocation of resources from lower productivity sectors to more productive sectors. Moreover, the creation of jobs in Brazil during the period appears to have been concentrated in low-​ productivity sectors. Lacerda and Oliveira (2009) find that, in terms of the sectorial distribution of FDI, it can be observed that industry is not generally the featured sector. FDI arriving

The Evolution of Foreign Direct Investment in Brazil    671 in Brazil in 2000–​2008 was aimed mainly at the tertiary sector, to the detriment of manufacturing. The share of manufacturing in FDI was reduced between 1995 and 2000, while the services sector increased its share from 30.9% in 1995 to 64% in 2000. For these authors, the redirecting of FDI during this period represents a significant restructuring of Brazilian production. This direction of FDI, and more broadly the decision-​making of companies considering opening operations in Brazil, relates to issues beyond narrowly economic aspects. Lacerda and Oliveira (2009) argue that these include the institutional environment as a whole, which includes, for instance, bureaucracy, patent law, and judicial procedures. Such factors can be hard to assess; empirical studies typically only emphasize locational factors, since obtaining data on issues such as bureaucracy requires a joint effort by companies. Thus, the most commonly used variables are market size and gross domestic product (GDP) growth rate. However, to create an environment more conducive to investment, in order to attract foreign investors, Brazil needs to tackle questions regarding excessive bureaucracy and a lack of human capital. These issues are among those leading other authors to a more equivocal assessment of the impact of FDI in Brazil and Latin America, stressing potential positive impacts’ contingency on a number of variables. For the United Nations Conference on Trade and Development (UNCTAD 1999), managing a policy related to FDI in the context of greater international competitiveness is an essential issue. Allowing foreign investors to decide where to invest may not be suitable, especially in a context of market failures and institutional malfunction. The formulation and implementation of effective strategies by governments requires a development vision with coherence and coordination. According to Marino et al. (2002), there is a need to target the Brazilian economy toward the external sector, as well as investing in human capital and research and development, necessitating the political means to encourage such investments. Only in this way will the economy gain greater dynamism in the longer term. For these authors, an increase in industrial output over time is important in determining the dynamism of an economy. An Economic Commission for Latin America and the Caribbean (ECLAC 2005) study finds that the presence of multinational companies in Latin America will only bring the hoped-​for benefits to those countries of the region that improve the policies and national institutions in place to deal with international commitments on investments, that offer incentives to attract investment, and that evaluate the outcomes of FDI policies. Up to the time of the study, much FDI in Latin American countries had not been of a desirable standard. Thus, a lack of a good environment for investment leads to investments based only in the comparative advantages of countries; Brazil tends to receive more investment in the primary sector compared to other countries. This is the finding of Carvalho and Kupfer (2007), who seek to determine the trajectory of the structural change experienced by Brazil in recent years compared to other countries. They analyze the sectoral composition of Brazilian industry and find that, although there are certain

672    Breno Augusto da Silva e Silva structural rigidities, there has been specialization in lower-​technology sectors. For these authors, Brazil has specialized in a lower stage of development, in terms of GDP per capita, compared to developed countries, which is bad for the country’s economy and overall prospects. Despite the importance of FDI in contributing to increased competitiveness, in Brazil the expected gains were not fully promoted. Competitiveness, in global terms, seems not to follow a single model in different economies around the world in the sense of which variable is most important for encouraging it. Nevertheless, FDI has played an important role in increasing the competitiveness of nations. FDI can, in receiving countries, internalize technology and know-​how in the production of goods and services that facilitate access to the most sophisticated markets. In addition, it brings with it experience of the activities of transnational business in international trade. However, in Brazil specifically, foreign investment has not by itself promoted the gains that might have been expected, such as greater growth in exports of higher value-​added products. Investments in Brazil have generally prioritized the supply of the domestic market, rather than contributions to the generation of foreign exchange, such as occurs with investments in emerging economies in Asia. The Brazilian process of productive internationalization in the first half of the 1990s was characterized by a strong asymmetry in equity and commercial plans. At the financial level, domestic companies were transferred to foreign capital, and Brazilian companies have not invested in the same proportion abroad. Thus the participation of foreign capital in Brazilian companies increased, while the service sector was denationalized. Commercially, there was an increase in the share of imported products in domestic production, such as the increasing import of intermediate products, but without a proportional increase of domestic production for export. Although there have been efficiency gains for companies, with consequent increases in productivity, there has been no proportionate increase in exports. These asymmetries represent a marked difference between the internationalization process of Brazil and those of Mexico and Asian countries such as South Korea, where a large portion of production has been exported. The Brazilian process has been based on the domestic market. Brazil’s trajectory also differs from that of China. In China most investment has been for the construction of new plants, called greenfields, rather than the purchase of existing assets. In China, although supply of the domestic market is contemplated, an important part of the production of transnationals is oriented to the foreign market (Sarti and Laplane 2002). One of the most remarkable issues is that in Brazil, FDI has promoted more transfer of national firms to international capital, with the increase in importation of inputs, without the same contribution to exports. In countries such as Mexico and South Korea, the increase in production provided by FDI was directed to exports. Veiga (2004) argues that the Brazilian industrialization of the past two decades had as one of its drivers the FDI that entered the country, being directed to high-​technology industries and capital goods. However, according to data from the Central Bank of

The Evolution of Foreign Direct Investment in Brazil    673 Brazil, the industrial sector has been losing its relative share of FDI flows since at least 1996, which is a matter of concern for the external competitiveness of the Brazilian economy. A further important question is, as touched on earlier, the quality of FDI and the sectors to which it is directed; it is not only the amount of FDI, but also its purposes that are important. Ideally one should be able to invest in more sophisticated technological products. In ECLAC’s (2005) assessment, 2004 was the first year to see growth of FDI flows to Latin America between 1999 and 2004. However, this does not mean that the countries of Latin America and the Caribbean, including Brazil, have solved the issue of the limited benefits they received through the presence of multinational companies in the region. It is noteworthy that FDI flows to the region were qualitatively insufficient, and were targeted at traditional sectors rather than new sectors with higher added value. FDI attracted to Brazil in search of natural resources was of great importance in the country’s consolidation as an important commodities exporter. However, the benefits of this type of investment are limited in relation to other national objectives, such as increasing international competitiveness in sophisticated technological products. Cravino (2007) analyzed the trajectories of FDI stocks in China and Latin America and noted that the accumulation of FDI stocks in China was much broader than the accumulation of stocks in Latin America, when looking from 1990, but not since 1997, when analyzing all sectors. When specifically analyzing the manufacturing sector, the difference between China and Latin America remains striking even in the period between 1997 and 2003, a period in which the growth rates of China’s stocks are higher than those in Latin America and a period in which China experienced a great rate of GDP growth. It is important to consider explanations for such differences between China and Latin America; it is perhaps the case that China’s growth rates are more affected by FDI than are those of Latin American countries. For Nonnenberg (2003) much of the foreign investment received by Brazil from 1996 to 2000 resulted from policy regarding regulations and the privatization process. A portion corresponding to less than 20% of foreign investments was directed to industrial activities. The author notes that one of the effects of this small share of foreign investment in industry would be lower transfer of technology, as the service sector has less potential for this than the industrial sector. Another criticism made by Nonnenberg is that the service sector is not able to generate foreign exchange for the country, either through exports or import substitution. Shafaeddin (2005) argues that the productive and export sectors of most Latin American countries have regressed, rather than upgrading their structures. According to Shafaeddin, in the prior two decades the productive structure of most of the region had been changing toward becoming more focused on food processing, on industries based on the processing of raw materials, or on production of nontradable goods. Shafaeddin calls this process premature specialization.

674    Breno Augusto da Silva e Silva

31.4.  Foreign Direct Investment Data Analysis in Brazil In Brazil, according to the Brazilian Central Bank Bulletin, the amount of FDI to Brazil increased in the period 1995–​2015 (Table 31.1). According to Liu (2008), FDI is often seen only as a way to address the lack of domestic savings. However, although an increase in the amount of FDI received by a country is important, as noted earlier, it is more important to observe which sectors this kind of investment is going to. Gregory and Oliveira (2005) analyze the sectoral destination of FDI in the Brazilian economy and find that the industrial sector received most of this capital in the 1980s and the early 1990s. In 1995, according to the Foreign Capital Census of the Central Bank of Brazil, 67% of such funds were for industry, 31% for the service sector, and around 2% arrived in agriculture and mining. In industrial terms, it appears that the investments were directed to the chemicals, automotive, metallurgy, food, and machinery and equipment industries. In the services sector, there are the financial and wholesale trade activities. However, after 1996, as a result of the privatization process in the Brazilian economy, the service sector is now the largest foreign investment recipient. Thus, as per the Foreign Capital Census of the Central Bank from 2000 to 2005, the sectors of industry and services virtually reversed their positions in terms of

Table 31.1 Foreign Direct Investment (US$ Billion) Year

FDI

Year

FDI

1995

4.7

2006

18.8

1996

16.1

2007

34.6

1997

20.8

2008

45.1

1998

20.6

2009

25.9

1999

30.1

2010

48.4

2000

32.8

2011

66.7

2001

22.6

2012

65.3

2002

16.6

2013

64.0

2003

10.1

2014

96.9

2004 2005

18.1 15.1

2015

75.1

Source: Brazilian Central Bank Bulletin, Annual Report (1995–​2014).

The Evolution of Foreign Direct Investment in Brazil    675 Table 31.2 FDI Stock and Sector Participation in 1995, 2000, 2005, and 2010–​2014 (in Million Dollars) Economic Activity Year

Agriculture, Livestock, Mining

Stock 1995 Partic. (%)

924.99 2.22

Stock 2000 Partic. (%)

Industry

Service

Total

27,907.09 66.93

12,863.54 30.85

41,695.62 100.00

2,401.38 2.33

34,725.62 33.71

65,887.81 63.96

103,014.51 100.00

Stock 2005 Partic. (%)

5,890.67 3.62

53,763.05 33.03

102,820.26 63.15

162,807.27 100.00

Stock 2010 Partic. (%)

92,371 16.0

227,857 39.0

266,981 45.0

587,209 100.00

Stock 2011 Partic. (%)

85,775 15.0

228,882 39.0

274,935 47.0

589,592 100.0

Stock 2012 Partic. (%)

70,983 12.0

253,686 41.0

289,770 47.0

614,439 100.0

Stock 2013 Partic. (%)

68,590 12.0

231,761 41.0

266,938 47.0

567,290 100.0

Stock 2014 Partic. (%)

63,061 12.0

207,594 39.0

260,789 49.0

531,445 100.0

Source: Brazilian Central Bank—​Census of Foreign Capital—​base years: 1995, 2000 2005, 2010, 2011, 2012, 2013, and 2014.

internalization of external investments (Table 31.2). This drop, in relative terms, in FDI to the Brazilian industrial sector can lead to various problems in relation to international competitiveness. Currently, there are few Brazilian companies able to compete in world exports of processed products not intensive in natural resources. A rare successful example is Embraer, the aircraft industry, which in addition to units in Brazil has two joint ventures abroad: the Harbin Embraer in China, and OGMA in Portugal. The data in Table 31.2 show that the stock of FDI in Brazil, in 1995, was concentrated in the sectors of manufacturing and services. Considering the three sectors (primary, industry, and services), it can be seen that the primary sector increased its relative share of FDI from 2.22% in 1995 to 12.0% in 2014; industry’s share reduced from 66.93% to 39%; the service sector’s share increased from 30.85% to 49%. The service sector expanded its relative share in the stock of FDI directed to Brazil from 30.85% in 1995 to 63.96% in 2000 and 63.15% in 2005. The results of the Central Bank’s (BCB) 2011 census, for the base year 2010, show that on December 31, 2010, the stock of FDI reached US$660.5 billion, corresponding to 30.8% of GDP. According to the BCB, that amount is US$188 billion higher than

676    Breno Augusto da Silva e Silva Table 31.3 FDI Inflows to Brazil, 2001–​2010 (Million Dollars) Economic Activity Primary

Manufacturing

Services

Total

2001 (absolute)

1,493.55

7,000.98

12,547.17

21,041.70

2001 (%)

7.10

33.27

59.63

100

2002 (absolute)

637.86

7,555.30

10,585.15

18,778.30

2002 (%)

3.40

40.23

56.37

100

2003 (absolute)

1,487.01

4,506.02

6,909.37

12,902.41

2003 (%)

11.53

34.92

53.33

100

2004 (absolute)

1,072.82

10,707.82

8,484.70

20,265.34

2004 (%)

5.29

52.84

41.87

100

2005 (absolute)

2,194.37

6,402.81

12,924.38

21,521.57

2005 (%)

10.20

29.75

60.05

100

2006(absolute)

1,363.12

8,743.78

12,124.40

22,231.30

2006 (%)

6.13

39.33

54.54

100

2007 (absolute)

4,750.80

13,481.28

16,102.77

34,334.86

2007 (%)

13.84

39.26

46.90

100

2008 (absolute)

12,994.82

14,013.04

17,499.20

44,457.06

2008 (%)

29.23

31.52

39.25

100

2009 (absolute)

4,596.71

13,481.30

13,601.49

31,679.50

2009 (%)

14.51

42.56

42.93

100

2010 (absolute)

18,158.35

19,345.58

15,103.25

52,602.18

2010 (%)

34.52

36.77

28.71

100

2011 (absolute)

10,297

26,837

32,397

69,530

2011 (%)

14.81

38.60

46.59

100

2012 (absolute)

6,532

22,243

31,767

60,543

2012 (%)

10.79

36.74

52.47

100

2013 (absolute)

9,990

15,218

24,134

49,341

2013 (%)

20.25

30.84

48.91

100

2014 (absolute)

5,621

16,922

33,556

56,099

2014 (%)

10.02

30.16

59.82

100

2015 (absolute)

8,310

20,967

28,630

57,907

2015 (%)

14,35

36,21

49,44

100

Sources: Brazilian Central Bank (2010) and Brazilian Central Bank Bulletin—​Annual Report (1995/​2015).

previously estimated in the International Investment Position (IIP) of US$472.6 billion, corresponding to 22% of GDP (this major change in value for the same period is due to methodological changes in the calculation, which began to take into account market value and not only book value).

The Evolution of Foreign Direct Investment in Brazil    677 An important issue in the analyzed period, between 2001 and 2010, is that the primary goods sector accounted for 34.52% of FDI flows in 2010, whereas in 2007 this percentage was only 7.10%. This result therefore displays a trajectory of FDI that supports the thesis that the Brazilian economy is emphasizing the primary sector. The manufacturing sector oscillated a little in this period, but on the whole remained stable, at 33.27% in 2001 and 36.77% in 2010. Its percentage was more than four times higher than that of the primary sector in 2001, but in 2010 was similar. The service sector experienced a pronounced drop in its percentage share, from 59.63% in 2001 to 28.71% in 2010. This result shows that there was a marked decline of FDI directed to the services sector, with no increase in the industry sector. Foreign investment, in 2010, was largely directed to the primary sector, particularly in relation to the three previous years. According to Kruger (2008), the structural change of economies over time tends to follow a path called Hypothesis of the Three Sectors (Figure 31.1). Under this hypothesis, during the early development of an economy the participation of the primary sector would typically account for more than 50% of all production. The industry and services sectors combined have a small share, below 50% percent. Subsequently, the share of the primary sector falls, while that of the industry sector increases almost in proportion to the decline of the primary sector. In turn, the services sector remains stable. Subsequently the industry sector increases its share, overtaking the primary sector to become the majority sector; at this stage, the services sector undergoes a small increase in participation. In a final stage, the services sector would increase its participation to become the majority share of production, with a fall in the share of the industry sector, returning to values ​​close to the initial stage; the primary sector, already at a low share, experiences a further slight decrease before more or less stabilizing. In the case of the Brazilian economy in recent years, has FDI followed a path in line with Kruger’s hypothesis? To examine this question, we use data on FDI provided by the Central Bank of Brazil (BCB 2011).

1 Industry sector

Services sector

0

Primary sector

Figure 31.1.  Hypothesis of the three sectors. Source: Kruger (2008).

678    Breno Augusto da Silva e Silva The data on FDI received by Brazil between the years 2001 and 2010 are shown in Figure 31.2, divided between the three main sectors—​primary, manufacturing, and services—​in percentage terms. Therefore, at each point of the x (horizontal) axis, the sum of the three sectors is 100%. As this shows, the path followed by FDI received by Brazil during these 10  years follows a different direction from that proposed by Kruger. Regardless of whether GDP causes investment or investment causes GDP, if investments in Brazil were in accord with the hypothesis of the three sectors, the primary sector should be reducing its participation and the service sector should be increasing its participation. In fact, though, according to Figure 31.2, the opposite occurs. The industry sector undergoes no important fluctuations in the period, more or less retaining a stable share. It is in observing the primary and services sectors that the striking result arises. The services sector sees a decline in its share in the period, from 59.63% in 2001 to 49.44% in 2015. Conversely, the primary goods sector doubles its share from 7.10% in 2001 to 14.35% in 2015. We can also observe that the primary sector reached around 30% of participation in 2008–​2010, the period colored by the global economic crisis of 2008. These results are worrying, pointing to a possible re-​primarization phenomenon in the Brazilian economy. The increase in FDI in the primary sector between 2001 and 2015 can be linked to the comparative advantage that the country has in this industry. Brazil, in producing more primary goods, would serve merely as a supplier of inputs for companies that will produce higher value-​added goods in their countries of origin.

100.00 90.00 80.00 70.00 60.00 50.00 40.00 30.00 20.00 10.00 0.00 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) Primary

Services

Manufacturing

Figure 31.2.  FDI in sectorial terms in Brazil. Source: Produced by the author based on survey data.

The Evolution of Foreign Direct Investment in Brazil    679

31.5. Conclusions For many countries, FDI has become increasingly important in recent years, as a way to seek new markets (for investors’ countries), and as a development opportunity (for investment-​receiving countries). Thus, this type of investment has become even more important in analyzing the development process of emerging economies such as Brazil. Although the amount of FDI is certainly important to the development prospects of countries, the sector to which investment is directed can be considered more important still. In the 1990s the industrial sector lost participation in FDI in comparison to the services sector in Brazil. In the first decade of the 2000s, flows of FDI to the industrial sector maintained their proportional participation in comparison to primary and services; toward the end of the decade, the services sector’s share of flows decreased and the primary sector’s increased. In the 2010s, the primary goods sector again increased proportional participation in FDI flows. The causes of these changes in FDI flows and stocks can be found in the social infrastructure and in the incentives (or barriers) provided by government. Much of the FDI arriving in Brazil has cleaved to the country’s natural comparative advantage in the primary sector, since the government has not directed investment to sectors with a higher technological level, and has been unable to provide more qualified labor, or an environment apt to attract investment to more dynamic sectors. When governments do not influence the direction of FDI, investment patterns will typically revert to following comparative advantages. Brazil’s recent tendency toward re-​ primarization, reflected in the direction of FDI flows, cannot be a good recipe for the country in structural terms. Brazil may well be forgoing opportunities to provide a good environment for investment in key sectors of the economy in industry or services, and with this the prospect of a positive structural change leading to sustainable long-​term economic growth.

References Alfaro, L. 2003. “Foreign Direct Investment and Growth: Does the Sector Matter?” Mimeo. Harvard Business School, Boston. Aitken, B., and A. Harrison. 1999. “Do Domestic Firms Benefit from Foreign Direct Investment? Evidence from Venezuela.” The American Economic Review 89 (3): 605–​618. BCB (Brazilian Central Bank). Various years (1995–​ 2015). Brazilian Central Bank Bulletin: Annual Report. https://​www.bcb.gov.br/​?BOLETIMANO. BCB (Brazilian Central Bank). Various years (1995–​2014). Census of Foreign Capital. http://​ www.bcb.gov.br/​pt-​br/​#!/​n/​CENSUS. Blalock, G., and D. H. Simon. 2009. “Do All Firms Benefit Equally from Downstream FDI? The Moderating Effect of Local Suppliers’capabilities On Productivity Gains.” Journal of International Business Studies 40: 1095–​1112. Blomstrom, M., and A. Kokko. 1998. “Multinational Corporations and Spillovers.” Journal of Economic Surveys 12 (2): 1–​31.

680    Breno Augusto da Silva e Silva Bonelli, R. 1998. “A Note on Foreign Direct Investment (FDI) and Industrial Competitiveness in Brazil.” Text for Discussion No., 584, Rio de Janeiro. Available at http://​www.ipea.gov.br/​ pub/​td/​td0584.pdf, accessed March 10, 2016. Brewer, T. A. 1993. “FDI in Emerging Market Countries.” In The Global Race for FDI: Prospects for the Future, edited by L. Oxelheim, 11–​49. Berlin: Springer-​Verlag. Carvalho, L., and D. Kupfer. 2007. “The Structural Transition of Brazilian Industry: Diversification to Specialization.” Proceedings of the XXXV National Meeting of Economics 054, ANPEC—​National Association of Postgraduate Centers in Economics. http://​www. anpec.org.br/​encontro2007/​artigos/​A07A054.pdf Chakraborty, C., and C. Nunnenkamp. 2008. “Economic Reforms, FDI, and Economic Growth in India: A Sector Level Analysis.” World Development 36 (7): 1192–​1212. Cravino, J., D. Lederman, and M. Olarreaga. 2007. Foreign Direct Investment in Latin America during the emergence of China and India. Washington, DC: World Bank, Latin America and Caribbean Region. Crespo, N., and M. P. Fontoura. 2007. “30 Years of Investigation surrounding the effects of FDI Externalities on National Companies: What Conclusions?” Economic Studies São Paulo 37 (4): 849–​874. Dunning, J. 1993. Multinational Enterprises and the Global Economy. Harlow: Addison-​Wesley. ECLAC (Economic Commission for Latin America and the Caribbean). 2005. “Foreign Direct Investment in Latin America and Caribbean 2004.” Available at http://​www.cepal.org/​en/​ publications/​1129-​foreign-​investment-​latin-​america-​and-​caribbean-​2004, accessed March 10, 2016. Fagerberg, J. 2000. “Technological Progress, Structural Changes and Productivity Growth: A Comparative Study. Structural Change and Economic Dynamics 11: 393–​411. Gregory, D., and M. F. B. A. Oliveira. 2005. “The Development of Enabling Environment in Brazil for Foreign Direct Investment Attraction.” Executive Resume. CEBRI. Available at http://​www.cebri.com.br/​cebri/​, accessed March 10, 2016. Hadiwibowo, Y. 2010. “Capital Inflows and Investments in Developing Countries: The Case of Indonesia.” The International Journal of Applied Economics and Finance 4 (4): 220–​229 Hall, L, R. E., and C. I. Jones. 1999. “Why Do Some Countries Produce So Much More Output per Worker Than Others?” The Quarterly Journal of Economics 114 (1): 83–​116. Haskel, J., S. Pereira, and M. Slaughter. 2002. “Does Inward Foreign Investment Boost the Productivity of Domestic Firms?” Working Paper No. 8724. National Bureau of Economic Research, Cambridge, MA. Imbs, J., and R. Wacziarg. 2003. “Stages of Diversification.” American Economic Review 93 (1): 63–​86. Javorcik, B. S. 2004. “Does Foreign Direct Investment Increase the Productivity of Domestic Firms? In Search of Spillover Backward Linkages.” The American Economic Review 94 (3): 605–​627. Kaldor, N. 1975. “Economic Growth and the Verdoorn Law: A Comment on Mr. Rowthorn’s Article.” Economic Journal 85 (340): 891–​896. Knell, M. 2004. “Structure Change and the Kaldor-​Verdoorn Law in the 1990s.” Revue d’économie industrielle 105: 71–​83. Kruger, J. J. 2008. “Productivity and Structural Change: A Review of the Literature.” Journal of Economic Surveys 22 (2): 330–​363. Kugler, M. 2000. “The Diffusion of Externalities from Foreign Direct Investment:  Theory Ahead of Measurement.” Discussion Paper in Economics and Econometrics. Department of Economics, University of Southampton.

The Evolution of Foreign Direct Investment in Brazil    681 Lacerda, A. C., and A. Oliveira. 2009. “Inflows of Foreign Direct Investment (FDI) in Brazil: An Analysis of Denationalization of the Productive Structure in the 2000s.” Seventh Debate Series, PUC-​SP. Available at http://​www.pucsp.br/​eitt/​downloads/​vii_​ciclo_​debates_​2009/​ Microsoft_​Word_​-​_​VIICiclo2009_​ArtLacerda-​Alexandre_​23.04.09.pdf, accessed March 10, 2016. Laplane, M. F., and F. De Negri. 2004. “Impact of Foreign Companies on the Brazilian Foreign Trade: Evidence from 90s Decade.” Economy 30 (1): 31–​48. Laplane, M., and F. Sarti. 1999. “Foreign Direct Investment and the Impacts on the Balance of Payments in the 90s.” Text for Discussion, No. 629, 7–​45. Institute of Applied Economic Research (IPEA), Rio de Janeiro. Liu, Z. 2008. “Foreign Direct Investment and Technology Spillovers: Theory and Evidence.” Journal of Development Economics 85: 167–​193. Marinho, E. L. M., C. A. G. Walnut, and A. L. T. Rosa. 2002. “Empirical Evidence of the Kaldor-​ Verdoorn Law for Brazil Industry Transformation.” RBE, Rio de Janeiro 56 (3): 457–​482. Nonnenberg, M. J.  B. 2003. “Determinants of Foreign Direct Investment and the Impacts of Multinationals in Brazil: The Decades of 1970 and 1990.” Text for Discussion No. 969. Institute of Applied Economic Research (IPEA), Rio de Janeiro. Petri, P. A. 1997. “Foreign Direct Investment in a Computable General Equilibrium Framework.” Making APEC Work: Economic Challenges and Policy Alternatives, Keio University, Tokyo, March 13–​14. Rocha, F. 2007. “Labour Productivity and Structural Changes in Mining and Manufacturing Industries, 1970–​2000.” Applied Economics 27 (2) (106): 221–​241 Shafaeddin, S. M. 2005. “Trade Liberalization and Economic Reform in Developing Countries:  Structural Change or De-​Industrialization?” United Nations Conference on Trade and Development, Discussion Paper No. 179. UNCTAD. 1999. World Investment Report 1999: Foreign Direct Investment and the Challenge of Development. Geneva: UNCTAD. Veiga, P. M. 2004. “Foreign Direct Investment in Brazil: Regulation, Flows and Contribution to Development.” International Institute for Sustainable Development. Available at http://​ www.iisd.org/​pdf/​2004/​investment_​country_​report_​brazil.pdf. Waldkirch, A. 2008. “The Effects of Foreign Direct Investment in Mexico since NAFTA.” MPRA Paper No. 7975.

Chapter 32

Multinat i ona l C orp orat i ons from  B ra z i l Edmund Amann

32.1. Introduction One of the most striking features of the global economy over the past quarter-​century has been a relative shift in the origin of direct investment flows, away from the traditional industrialized economies of Western Europe, North America, and Japan and toward the emerging market economies of Latin America and Asia. The scale of these flows—​ which increased by more than a factor of 20 between 1990 and 2015 (UNCTAD 2016)—​ has been such as to remake the global corporate and economic landscape. Not only have emerging market-​based multinational corporations (MNCs) increased their presence in their customary domains (the regions surrounding their home markets); they also have increasingly become important investors in advanced industrial economies. As this process has unfolded, such European and US corporate icons as Jaguar Land Rover and GE Appliances have been the subject of takeovers by fast-​expanding Indian and Chinese MNCs. At the same time, partly as a result of the quest to secure raw-​material supply chains, there has been a surge in interregional South-​South resource-​seeking investment. Perhaps the most notable example of this has been the rise of Chinese mining investment in Sub-​Saharan Africa. Latin America, too, has been the subject of significant Chinese inward investment in the natural resources sector over the past decade (ECLAC 2015). The changing profile of global FDI flows carries broad significance in at least two respects. First, it provides a clear demonstration of the fact that what Hymer termed “ownership advantages” are no longer the exclusive preserve of enterprises based in the “North.” Rather, it is evident that such advantages—​whether in terms of technology, brand identity, or other forms of intellectual property—​are ever more globally

Multinational Corporations from Brazil    683 dispersed. Second, the increasing global role of emerging market-​based multinationals is reshaping the geopolitical landscape, providing the means through which a range of hitherto marginalized countries have been able to exercise soft power. As might be expected, Latin America’s largest economy, Brazil, has very much contributed to this global shift. Since the early 1990s, according to UNCTAD data (UNCTAD 2016), its stock of outward foreign direct investment (OFDI) has multiplied fourfold. This has occurred as such well-​known (and not so well-​known) enterprises as Petrobrás, Embraer, Marco Polo, and JBS have launched and/​or expanded their international operations. Yet, compared with the significant coverage in the academic literature and the financial press accorded to their Chinese and Indian counterparts, for example, relatively little attention (at least in the English-​language literature) has been given to the rise of the Brazilian MNC. This chapter seeks to redress this balance to some extent through a structured survey of the rise of Brazilian OFDI and the enterprises behind it. To accomplish this objective, the chapter is structured as follows. First, following an overview of relevant theoretical considerations centering on Mathews’s view of the potential sources of emerging market MNC advantage (section 32.2), the chapter presents a brief survey of statistical trends surrounding Brazilian OFDI over the past 15 years or so (section 32.3). Accompanying this, the section characterizes the sectoral orientation of Brazilian MNCs, pointing out the significant natural-​resource base (NRB) focus of many of the largest enterprises. It also considers the broad policy-​ related factors that have helped propel the recent surge in OFDI. Further disaggregating the analysis, section 32.4 presents selected case studies of individual Brazilian MNCs, tracing their origins, patterns of internationalization, and sources of ownership advantage. Enterprises surveyed in this section comprise Embraer, Petrobrás, Vale, Tigre, and Sabó. Drawing the chapter toward a conclusion, section 32.5 interprets in theoretical terms the emergence of Brazilian MNCs and considers the policy shifts that have contributed to this process. Finally, in the concluding section (32.6), the arguments are summed up and some consideration are given to the challenges currently facing Brazilian MNCs. Not the least of these is the current wave of corruption scandals surrounding key MNCs in the energy and construction sectors. It will be argued that these partly underlie a process of consolidation and divestment that is taking place in many of Brazil’s largest MNCs.

32.2.  Theoretical Considerations Perhaps surprisingly, the considerable body of theoretical literature that has built up concerning the internationalization of firms is comparatively recent in origin. The seminal contributions of Stephen Hymer in the 1960s provided the springboard for a series of conceptual developments, which ultimately gave rise to the so-​called eclectic or OLI (Ownership-​Location-​Internalization) paradigm of international production (Dunning and Lundan 2008). It is the latter that currently represents the dominant

684   Edmund Amann analytical framework used to study the phenomenon of foreign direct investment (FDI) and the motivations of the MNCs that lie behind it. According to the theoretical precepts surrounding this approach, the internationalization of production by firms may only take place if they possess some form of “ownership advantage.” Such an advantage—​which may stem from the development of product or process technology, a trademark, brand name, or other salient form of intellectual property—​confers the ability to exercise monopolistic power in a market and, hence, to engage in non-​price-​based forms of competition (see Caves 2007). It is this which provides the basis for depriving competitors of market share and, ultimately, the generation of profits through international expansion. Whether or not such expansion is achieved through directly owned subsidiaries, joint ventures, or licensing will depend on a range of factors, including the degree of informational asymmetry, the degree of legal recourse, transportation costs, and the scope of transactions costs in general (Dunning and Lundan 2008). Assuming the decision is made to leverage extant ownership advantages through the OFDI route, the determination of which territory ends up being the recipient will depend on location-​specific factors, which will vary dependent on the character of the investment involved. On this basis Dunning famously identified different forms of motivations for foreign investment:  efficiency seeking, market seeking, natural resource seeking, and strategic asset seeking (Dunning and Lundan 2008). Thus, for example, OFDI aimed at tapping domestic consumers in overseas markets would be characterized as market-​seeking, whereas that targeted at the mining sector would be termed natural resource seeking. In passing, it is worth noting at this juncture that as far as inward FDI is concerned, Brazil has been primarily associated with the latter two motivations (see Silva e Silva, Chapter 31 in this volume). From the early 1990s onward, as previously noted, the nature of FDI flows has changed remarkably, with emerging market-​based firms accounting for an increasing share of the global total. This has posed a challenge for established theories of international production, the conventional assumption having been that emerging-​market-​ economy MNCs possessed no discernible ownership advantages (Williamson et  al. 2013, 1). However, a considerable body of theoretical and empirical work shows that this is palpably not the case. For one thing, as Amann and Cantwell (2012) make clear, there is ample evidence that the innovative capabilities that often underpin ownership advantages are increasingly dispersed globally. That this is the case is partly the result of deliberate public policy intervention in emerging market economies, which has sought to create national and sectoral systems of innovation (Malerba and Mani 2009). Second, as Mathews (2002) famously argued, the character of the core ownership advantage may be intrinsically different in many Emerging Market-​Based Multinational Corporations (EMNCs) from their advanced industrial-​economy-​based counterparts. Specifically, Mathews considers that the capacity of EMNCs to engage in technological and organizational catchup through a process of learning and emulation is itself a form of ownership advantage. The South Korean “model” through which Chaebols such as Samsung and LG achieved preeminence is clearly the source of Mathews’s inspiration,

Multinational Corporations from Brazil    685 but enterprises elsewhere have also engaged in similar learning processes—​and with equal success (see Li and Cantwell 2012; Rasiah 2012). Given all of this, what analytical interpretation can be given to the growth of the Brazilian MNC phenomenon? This question will be examined in due course, but not before the recent trajectory of Brazilian OFDI is considered. It is to this that the discussion now turns.

32.3.  A Brief Statistical Overview of Brazilian Outward Foreign Direct Investment Trends As Figure 32.1 makes clear, Brazil has followed the general pattern established by other emerging market economies in that its OFDI stock has increased significantly over the past two decades. Underlying this trend is not only a tendency toward an acceleration in outflows (though this is not a uniform trend), but also a growth in the valuation of overseas assets. The sixfold increase in value in overall foreign holdings of Brazilian enterprises between 2001 and 2015 represents one of the most dramatic indicators of the degree to which Brazil’s economy—​and some of its leading enterprises—​have seized the opportunities presented by globalization, whether in terms of declining barriers to investment or trade. Still, as Table 32.1 indicates, the outflows contributing to the accumulation of OFDI stock have been highly volatile. In this sense Brazil, again, shares much in common with its emerging market counterparts. As Vodusek (2002) points out, FDI flow series tend to exhibit much greater year-​on-​year variation than stock data since, in any particular year, flows are much more likely to be influenced by lumpy individual 350 300 250 200 150 100 50 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Figure 32.1  Brazilian external foreign direct investment stock (US$ billion), 2001–​2015. Source: Author’s elaboration from Banco Central do Brasil data.

686   Edmund Amann Table 32.1 FDI Outflows, 2001–​2015 US$ Billion 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 –2.26 2.48 0.25 9.8

2.5

28.2 7.1

20.5 –10.1 22.1 11.1

–5.3 –1.2 2.2

3.1

Source: Author’s elaboration from UNCTAD data.

transactions, such as large merger and acquisition deals or, more rarely, significant greenfield investments. Even so, the Brazilian outflow data series in Table 32.1 appears especially volatile, with a notable trend being an uptick in negative flows toward the end of the series. These negative flows represent a strong tendency toward divestment among some enterprises. According to an important recent report (FGV-​Columbia 2017, 19) this divestment activity can be seen as a way of “refocusing [ . . . ] core activities and raising liquidity” at a time of financial stress associated with the domestic economic and political crisis. Perhaps the most notable example of this trend has been Petrobrás, the state-​controlled oil and gas company, and one of Brazil’s largest multinationals. Petrobrás has been forced to engage in a large-​scale (US$35 billion) divestment program, in response not only to a domestic recession but also to more firm-​specific factors, notably an artificial cap on domestic resale prices for fuels, and fallout from the Lava Jato corruption scandal. The latter centered on Petrobrás and involves the skimming of funds from infrastructure contracts and their diversion to political parties (see Prado and Carson, Chapter 35 in this volume, on the impact of corruption scandals on Brazil’s economy). Although the divestment phenomenon is an important one, it should not detract from the earlier observation that, in overall terms, Brazil has become a much more significant source of OFDI in recent years. Later in the chapter, some of the firm-​level drivers of this process will be examined with reference to some representative examples of Brazilian MNCs. For now, however, it is worth briefly considering the meta-​level factors that have propelled the process. At the root of the tendency toward greater OFDI among Brazilian enterprises lie broader forces of internationalization. Prior to the start of the 1990s, Brazil remained a comparatively closed economy, characterized by elevated tariff and nontariff barriers and significant barriers to inward FDI, notably in the transportation and public utilities sectors (Baer 2013). Historically, this policy configuration—​commonly termed import-​ substitution industrialization (ISI)—​had tended to favor inward rather than outward FDI. The FDI inflows realized tended to be characterized by market-​seeking motivations and were crucial in establishing critical industrial sectors such as automobiles and electrical goods (Baer 2013). From the late 1980s onward, as Brazil embarked upon a new course of civilian rule, the implementation of Washington consensus-​style policies radically changed the incentive structures facing firms. Trade and market liberalization reduced trade barriers and encouraged exports across a range of sectors from capital goods (Amann 2000) to

Multinational Corporations from Brazil    687 agricultural exports (Mueller 2016). As this process unfolded, so began growing internationalization of Brazilian firms as external markets were sought, whether as an attempt to compensate for reduced domestic market share, or simply to take advantage of newly realized competitiveness. The commodity price boom of the first decade of the 2000s and the early 2010s lent further force to the export and accompanying investment surge being experienced by a range of Brazilian enterprises. As many authors in the field observe—​among them Dunning and Lundan (2008)—​ exposure to the export market for many firms may prove the first step along the road to establishing an external presence via the location of foreign production facilities. Against this background, it is perhaps not hard to appreciate why accelerated OFDI has become such a marked feature of the Brazilian and other emerging market countries’ corporate landscapes since the global wave of economic liberalization in the final decade of the twentieth century. While such an overall assessment of the meta-​level forces propelling accelerated OFDI may be persuasive, it does not provide a convincing account of the firm-​level dynamics and decision-​making processes that determine individual investment decisions. Nor can such an assessment properly account for the observed differences in sectoral focus among investor countries; practically every emerging economy has faced similar policies of trade and market liberalization over the past 25 years, and yet particular sectors are far more represented in some countries than others. Table 32.2 indicates that in the Brazilian case, the very largest home-​grown MNCs tend to be concentrated overwhelmingly in natural-​resources-​based (NRB) sectors. Thus mining, oil and gas, and food processing are heavily represented, while manufacturing

Table 32.2 Top 10 Brazilian Multinational Corporations, Ranking by Foreign Assets, 2014 Rank (2014)

Name

Sector

Foreign Assets (2014)

1

Vale

Mining

62,704

2

Gerdau

Steel

15,254

3

Petrobrás

Oil and gas

13,676

4

JBS

Food processing

9,632

5

BRF

Food processing

5,450

6

Marfrig

Food processing

3,084

7

Minerva

Food processing

2,307

8

Magnesitsa

Mining

1,789

9

Embraer

Transport equipment

1,504

Tupy

Transport equipment

1,141

10

Source: Author’s adaptation from FDC-​Columbia (2016, 2).

688   Edmund Amann and finance are less so. This pattern remains more or less intact if we examine a larger sample of Brazilian multinationals (FGV-​Columbia 2017). However, if we were to examine a list of the largest Chinese or Indian MNCs (UNCTAD 2015), the sectoral focus would appear very different, with manufacturing, business services, and technology featuring far more commonly. Such a non-​NRB specialization is even more evident when one examines patterns of outward investment from Southeast Asian economies (see, for example, Rasiah 2012). What might account for this very distinctive pattern of specialization, at least as far as the Brazilian case is concerned? The most obvious answer is that patterns of MNC sectoral focus reflect in part underlying national comparative advantages. Given that the literature (e.g., Buckley and Ghauri 1999, x–​xi) argues that export success can be a springboard for subsequent OFDI, it should not be a surprise that such a link exists. Reinforcing this link, Amann and Baer (2010) point out the critical role played by access to international capital markets. The argument here is that enterprises located in the NRB sector typically represent a good credit risk where borrowing in foreign currency—​especially US dollars—​is contemplated. This is because revenue streams tend to be heavily biased toward earnings in dollars given that internationally traded commodities are usually bought and sold in that currency. Hence, from the perspective of the lender, or for that matter, the foreign investor seeking to acquire a stake, currency risk is minimized. Emerging market-​based enterprises outside the NRB sector are unlikely to benefit from such dollar-​denominated revenue streams to the same extent and thus may find tapping international capital markets more problematic (Amann and Baer 2010). Another feature that becomes apparent from even the most cursory scrutiny of any list of Brazil’s largest MNCs is the degree to which current or former state-​owned enterprises (SOEs) are represented. Thus, in Table 32.2, two of the top 10 largest Brazilian MNCs are former SOEs (Vale and Embraer), while Petrobrás is majority state owned. More extensive lists of the largest Brazilian MNCs (e.g., FDC 2015) also feature the presence of other current or former SOEs, such as Banco do Brasil (finance) and CSN (steel). Moreover, the state, through the BNDES (National Development Bank), continues to hold significant stakes in large Brazilian enterprises, many of which have never been considered SOEs in their own right. An important example here is provided by Aracruz, a key MNC in the pulp and paper sector. According to Mussachio and Lazzarini (2014, 29) the BNDES played a critical role in funding the investment program that eventually allowed Aracruz to become a major international player in its sector. The role of the state—​either as a current or former owner or investor—​is of special relevance to the emergence of Brazilian MNCs in an economy typified by a low savings ratio, high costs of financial intermediation, and, as a consequence, high real interest rates. Given these long-​standing structural factors, enterprises typically are not able to fund large investments—​including foreign acquisitions—​through recourse to domestic private capital markets. Instead, such transactions are usually funded through a mixture of retained earnings, funding from the BNDES, and, as we have seen, through accessing international capital markets. In addition, the BNDES, as Torres and Zeidan (2016)

Multinational Corporations from Brazil    689 point out, also provides an invaluable source of funding where the financing of exports is concerned. Through mechanisms such as interest rate equalization, the BNDES assists Brazilian enterprises in accessing foreign markets and in financing export sales at internationally competitive rates. Once again, the state, therefore, has exercised considerable influence on the process of Brazilian corporate internationalization, even though in the majority of cases the enterprises concerned were never SOEs in themselves. The close relationship between the state and major Brazilian MNCs has, of course, come under detailed scrutiny in the wake of the corruption scandals surrounding Petrobrás in the second half of the 2010s. This chapter will consider this important issue in the concluding section.

32.4.  Patterns of Internationalization, Technological Capability Building, and Ownership Advantages: Firm-​L evel Evidence The discussion thus far might lead one to surmise that the growth of the Brazilian MNC phenomenon in recent years is simply one facet of an extended boom in the natural resources sectors—​that, in some sense, the enterprises concerned have essentially been riding a wave of internationalization not of their own making. This assessment is, in fact, very wide of the mark, since it ignores important firm-​level dynamics that center on the creation and leveraging of ownership advantages. As will be seen, such advantages center to a surprising extent on the development and acquisition of technological capabilities. Before we examine some illustrative firm-​level examples, it is worth reflecting on the systematic attempts that have been made over the years via public policy to stimulate technological progress among Brazilian enterprises. A reflection on this topic serves to underscore the point that the emergence of Brazilian MNCs has had a long gestation period in which the state has played a critical role. The development of indigenous technological capabilities at firm level has benefited for several years from an elaborate policy framework designed to boost innovation and technological self-​sufficiency across Brazil’s diversified industrial, agricultural, mining, and services sectors. The current policy architecture dates from the 1960s and the 1970s, when the military government of the time established a series of special agencies that, collectively, continue to form part of Brazil’s National System of Innovation (for further details, see Mani 2002). Among these agencies, FINEP (Agency for Studies and Projects) has played a vital role in providing seed money to enterprises with which to launch innovation projects, while FINAME (Financing Agency for Machinery and Equipment) has provided subsidized lines of credit designed to facilitate the sales of technologically complex equipment (Dahlman and Frischtak 1993). In a similar vein, the BNDES has,

690   Edmund Amann through two further lines of credit, BNDES-​Ex (BNDES Exportações) and FINAMEX (Financiadora de Maquinas e Equipamentos), provided export finance that has proved a vital ingredient in sealing the sales success of Brazil’s advanced manufacturing sector, not least in aerospace (da Silva 2007). The ISI era under the military also saw the foundation of research and development institutes located within major SOEs. Thus it was that Cepel and Cpqd, the nationally famous research laboratories of Eletrobrás and Telebrás, were formed. Despite the privatization of large swaths of Eletrobrás’s assets and the complete privatization of Brazil’s telecommunications system (formally the domain of Telebrás), both research institutes remain operational to this day as stand-​alone foundations. Similar research institutes operate within the petroleum, life sciences, and agricultural industries, while the Brazilian university system remains populated with technology institutes with close ties to industry, for example the University of São Paulo’s renowned IPT (Institute for Technological Research). More recently, in 2013, the administration of President Rousseff formed Embrapii, an agency designed to foster yet closer relationships between private enterprise and universities and research institutions. For Dahlman and Frischtak (1993) and Mani (2002), this institutional eco-​system has played a vital role in the support of domestic technological capability development and, by extension, the forging of developed comparative advantages across a range of sectors, by no means all of which could be said to lie within Brazil’s traditional NRB purview. To gain a more disaggregated sense of the ways such capabilities have been developed and absorbed at firm level and then employed in the course of internationalization, it is worth considering the cases of Petrobrás, Embraer, Braskem, and Gerdau. The case of Embraer has been widely researched (see, for example, Amann and Figueiredo 2012; Oliveira et al. 2013), if for no reason other than it demonstrates the feasibility of developing a home-​grown MNC in a technologically dynamic sector in which Brazil, up until the 1970s, had little developed comparative advantage. Amann and Figueiredo (2012) demonstrate how the process of internationalization was intimately bound up with associated processes of learning and accumulation of technological capability. Ultimately, this has allowed the enterprise to exhibit what Lall in a seminal article (Lall 1992) would have termed world-​leading capabilities in its defined sector (regional jet transports). The means through which Embraer acquired its technological capacity is an interesting case study in its own right since, as a former SOE, it shares key common features with other Brazilian MNCs, notably Petrobrás and Vale. Having been founded as a government-​owned concern in 1969, Embraer benefited from targeted state funding aimed at establishing a national capability to enter the production of simple transport aircraft (Amann and Figueiredo 2012). Such aircraft were designed to cope with the rugged short airstrips typical of the more remote regions of North and West Brazil. The research and development (R&D) effort needed to develop the first model of such aircraft (the Bandeirante) was considerable, but benefited from close ties to a public research institute (the CTA—​Aerospace Technology Centre). Subsequently, in the 1970s, a second generation of higher performance pressurized aircraft (the Brasilía series) saw the development of more advanced capabilities, and by the 1980s and 1990s, in collaboration with foreign technology

Multinational Corporations from Brazil    691 providers and subcontractors such as CASA of Spain, Embraer had successfully penetrated the fast-​growing global market for regional jet transports (Amann and Figueiredo 2012). As might be expected in the light of the internationalization literature, Embraer’s technological success-​driven growing presence in the export market proved a trigger for eventual FDI: in 1994, the year it was privatized, Embraer acquired the Portuguese aircraft maintenance firm, OGMA. The objective here could be considered market-​seeking in Dunning terms, the idea being to provide a European base for the servicing of Embraer aircraft. In a similar market-​seeking venture, Embraer in 2002 established an assembly facility for its EMB 145 RJ model in China. The first Chinese-​ completed aircraft rolled off the assembly line in 2004. Petrobrás, another enterprise originally created by the state, corresponds to a more stereotypical Brazilian MNC in that its activities center on the NRB sector (Amann 2012; Ramsey and Almeida 2010, 222). However, the creation of distinct ownership advantages through the acquisition of technological capabilities has strongly influenced its growth as an international entity. Like many other SOEs, Petrobrás traces its roots back to the 1950s, when it was founded as part of a policy aimed at creating domestically owned capacity in a sector that, due to rapid industrialization, was experiencing fast-​ growing demand. Petrobrás was granted sole domestic rights to upstream oil and gas exploration and production, with the established—​mostly foreign-​owned—​private sector still allowed to function, but only in distribution and retail activities. The surge in oil prices experienced in the early 1970s proved a strong stimulus to internationalization, with Petrobrás deploying its newly acquired exploration and production capabilities in Libya and, with considerable success, in Iraq; there, in 1975, it discovered the extensive Majnoon oil field (Antonio and Lara 2005, 12). Regarding international expansion, Petrobrás experienced a period of retrenchment in the late 1980s as it concentrated on the development of a significant new oil field discovered off the coast of Rio de Janeiro and Espírito Santo states. This offshore field, located in the Campos Basin, proved both technically challenging and a stimulus to accelerated innovation and capability building for the simple reason that it involved drilling exploration wells at depths of up to 2 kilometers. Such depths were far greater than those encountered in previous offshore oil fields such as the Gulf of Mexico or the North Sea. In order to meet these challenges, Petrobrás developed a special technology development program—​PROCAP—​in association with Brazilian universities and research institutes (Leite 2005). However, Petrobrás also drew on the expertise of foreign oil services firms. The acquisition of such outstanding offshore exploration and production capability has indisputably placed Petrobrás on the global technological frontier, and it has set a number of world records for deep-​sea drilling (World Bank 2008). More significantly, in the context of the present discussion, the development of these world-​leading capabilities has created genuine ownership advantages that Petrobrás has been able to exploit as it, following the 1980s, pursued an aggressive strategy of expanded foreign investment (World Bank 2008). Thus, by the end of the last decade, Petrobrás had expanded exploration and production work to Angola, Argentina, Bolivia, Colombia, Nigeria, Trinidad and Tobago, and the United States, and had acquired

692   Edmund Amann offshore exploration blocks in Equatorial Guinea, Libya, Senegal, and Turkey (World Bank 2008)  Beyond the acquisition of technological capability, such a push toward intensified internationalization was stimulated further by more competitive conditions at home: following constitutional changes in the late 1990s, Petrobrás lost its statutory monopoly in domestic offshore and onshore exploration and production and began to face competition in these spheres from the likes of Royal Dutch/​Shell and BG group. Despite the successes, Petrobrás’s model of international expansion has not proved entirely sustainable. As shall be seen in the concluding section, the post-​2011 decline in the oil price, combined with the emergence of the Lava Jato scandal, has forced a period of divestment and retrenchment. Another of Brazil’s leading MNCs, Vale, has in some ways experienced a similar trajectory of foreign expansion to Petrobrás. Vale, one of the world’s largest mining groups, was founded in the late nineteenth century, and was later nationalized during World War II (Ramsey and Almeida 2010, 205). During the subsequent four decades, as an SOE, Vale, like Petrobrás, invested heavily in technology and gained some international exposure thanks to a joint venture in the 1970s with US Steel. However, it was only following privatization in 1997 that Vale began a major wave of overseas expansion, initiating greenfield operations in countries as diverse as Peru, Gabon, and Mozambique. Vale also became a major player in the North American market, acquiring the Canadian enterprises Canico and Inco in 2005 and 2006, respectively. While Vale’s technological competence and organizational agility are explanatory factors regarding its outward push, Ramsey and Almeidia (2010) focus on two especially important driving forces: a dynamic management team post-​privatization, and Vale’s corporate vision centering on becoming a complete materials supplier—​a one-​stop shop—​for the global steel industry. Vale’s success, as far as its internationalization is concerned, is thus only partially connected with Brazil’s undoubted natural comparative advantage in mineral extraction; technology, strategy, and entrepreneurial vision were necessary to leverage core strengths in the interests of overseas expansion. So far the discussion has centered on well-​known, high-​profile MNCs that are, or have at some point been, SOEs. What of smaller, less well-​known enterprises that have always been located in the private sector? Two interesting case studies are provided by Sabó, an automotive components manufacturer, and Tigre, a manufacturer of draining and plumbing products for the construction industry. Sabó, although perhaps not the most widely known in its sector, is nevertheless one of the most established Brazilian MNCs (Forúm de Líderes 2007, 236). Starting with an acquisition in Argentina in 1992, the enterprise subsequently acquired capacity in Germany and Austria and entered into the joint ownership of a Japanese manufacturer. The global expansion of Sabó occurred over a period in which the Brazilian economy itself—​including the automotive sector—​ was being exposed to more intensified global competition thanks to trade and investment liberalization (Baer 2013). One of the most interesting features of the Sabó case is the core role played by technological capability building. The group has invested heavily in R&D and has used its expansion into Germany to tap expertise there, establishing

Multinational Corporations from Brazil    693 an R&D center in that country (Forúm de Líderes 2007, 236). Sabó has invested a considerable portion—​40%—​of its overall R&D budget in external activities, engaging such leading institutions as the University of São Paulo and the Instituto Tecnológica de Aeronáutica (IPT) to participate in its innovative activities (Forúm de Líderes 2007, 236). The capabilities so acquired have conferred the necessary ownership advantages to facilitate Sabó’s overseas expansion while, as noted, overseas expansion itself has unlocked opportunities for engaging in local knowledge networks, facilitating further capabilities building. Compared with Sabó, Tigre has a much longer history of internationalization through FDI. Having been founded during World War II—​though not as an SOE—​ Tigre soon acquired considerable experience manufacturing extruded and injected plastic products (Oliveira et al. 2013, 18). In that respect, it was something of a regional pioneer, entering a new technological field in which the global frontier at the time was moving quite rapidly. Further product innovation into the 1960s and 1970s saw Tigre become an effective specialist in plastic piping and tubing used in the construction industry. According to Oliveira et al. (2013, 18) this sector’s acute market fragmentation, due in part to geographically differentiated soil and climate conditions, “creates a specific microcosm favoring innovative behavior, and that was true for Tigre.” On the basis of this innovative capacity, and facing more competition in its domestic market from 1997 onward, Tigre established production facilities in Chile, Argentina, Bolivia, Peru, Ecuador, Colombia, and the United States. This outward expansion followed initial internationalization experience gained with the establishment of the Paraguayan plant in 1977 (Oliveira et al. 2013, 18).

32.5.  Interpreting the Brazilian Multinational Corporation Phenomenon As noted previously, the emerging-​market-​based MNC phenomenon appeared to present challenges to established theoretical perspectives on the basis that such enterprises were held to possess no discernable ownership advantages. At least as far as our brief review has established, however, such a sweeping contention is clearly not tenable in the case of all Brazilian MNCs. For the enterprises surveyed, whether within the NRB sector or in the nontraditional manufacturing sector, there is clear evidence that the creation of technological capabilities and/​or the development of organizational innovations provided a springboard for internationalization. In this sense, the dynamics involved bear significant resemblance to those associated with MNCs originating in advanced industrial economies. Of course, it is also the case that, as Mathews (2002) would suggest, some enterprises (e.g., Sabó) display learning and “catchup” behavior involving the acquisition of

694   Edmund Amann technological capability through foreign expansion, whether through mergers and acquisitions (M&A) or via the founding of foreign research facilities. Sabó is by no means alone in this external quest for the acquisition of ownership advantages, as Fleury and Fleury (2016) make abundantly clear. However, the literature (e.g., Amann and Figueiredo 2012; Williamson et al. 2013) suggests that domestically acquired capabilities more often than not play a vital role and that these capabilities are very often the outcome of fruitful interactions between enterprises, Brazil’s technology policy framework, and, more broadly, national and sectoral systems of innovation. More generally, Fleury et al. (2013, 270) draw attention to a very important point. They argue that the creation of ownership advantages through innovation in the Brazilian context is not confined to conventional product and process innovations but, critically, also includes the development of new business models, administrative processes, and marketing design. The relative role played by each form of ownership advantage for any given enterprise in any given situation where the expansion of overseas operations is contemplated will, of course, tend to vary. In this sense, as Parente (2013, 462) argues, it can be quite difficult to generalize about the trajectories of Brazilian MNCs. In other words, it would be wide of the mark to talk about a “Brazilian model” of MNC. In fact, Parente himself terms the trajectories of Brazilian EMNCs “quite unstructured” (2013, 462). While it is undoubtedly true that the trajectory and sequencing of internationalization will vary from MNC to MNC and that there is essentially no Brazilian model, there are some distinctive features of the Brazilian experience that are worth highlighting, if only because they pose some significant policy issues. First, as stated toward the beginning of this chapter, it remains the case that Brazil’s very largest MNCs remain rooted in NRB activities. Second, it is notable that the shift away from ISI from the late 1980s onward was associated with a strong impulse to initiate or expand further international operations among Brazilian enterprises (Amann and Figueiredo 2012). The reasons for this are numerous, but are likely to center on the challenge-​response mechanism engendered by greater domestic market competition; greater freedom of operation following privatization; and, more generally, a global regulatory and trade environment more amenable to direct capital flows. Third, it can be argued that despite the supposed rolling back of the state associated with the end of ISI, public policy has remained a vital determinant of the growth of Brazilian MNCs. Musacchio and Lazzarini (2014) persuasively set out the close links that have continued to bind large Brazilian enterprises to the state, whether these comprise actual ownership in whole or in part, the provision of subsidized credit via the BNDES, or significant public-​sector contracts. Thus, the demise of ISI has not necessarily destroyed the edifice of state capitalism; indeed, a close relationship with the state appears to be a strong if not defining characteristic of Brazil’s largest MNCs, regardless of whether they were ever full-​blown SOEs. The sense that proximity to the state has represented a form of prequalification for internationalization is only strengthened when one considers the open advocacy of “national champions” under the Lula and Rousseff governments (Bazuchi et al. 2013). The objective here was to facilitate the expansion of Brazilian MNCs as a means of exercising international soft

Multinational Corporations from Brazil    695 power. This was as often as not achieved through soft loans channeled via the BNDES (Bazuchi et al. 2013). The fact that the largest MNCs more often than not enjoy close links to the state and/​ or are located in the NRB sectors poses at least two important policy issues. First, given the need to diversify Brazil’s economic base in the face of a commodity price crash, the desirability of large MNCs emerging from outside the NRB sector must surely be stronger than ever. Second, at a time when the state’s resources are under severe strain, and business-​government relationships under unprecedented legal scrutiny, there is a strong case for developing a new generation of MNCs that are less reliant on state largesse and unhealthily close contacts.

32.6.  Conclusions: The Challenges Now Facing Brazilian Multinational Corporations By the second half of the 2010s, it had become clear that Brazilian MNCs, which had been riding high until the middle of the decade, were facing a set of unprecedented challenges, some connected with external events in the global economy, some much closer to home. Taken together, these events have forced a period of retrenchment among several Brazilian MNCs, including some of those surveyed here. Thus, as already noted, Petrobrás has announced plans to divest US$19.5 billion in 2017–​2018, while Vale has announced a US$1 billion sale of one of its subsidiaries plus, the sale of 12 of its VLOCs (very large ore carriers) to a Chinese bank. Sabó in 2014 concluded the sale of its German subsidiary, KACO, to another Chinese enterprise (FGV-​Columbia 2017, 20; Sabo 2017). The sharp decline in commodity prices since 2011 has, of course, presented a particular challenge to Brazil’s largest MNCs given their relative focus on NRB extraction and processing. Under the circumstances, it was only to be expected that such enterprises would have to contend with the inevitable challenges posed by capacity-​demand mismatches. The processes of divestment in which such enterprises are involved are, in this sense, typical of what is going on globally. However, given the fallout from the Lava Jato corruption scandal, Brazilian MNCs are having to face a more existential challenge; this is more domestic and enterprise-​specific in nature. Lava Jato and related corruption investigations have shed light on the extent to which close contractual relationships between the government and large MNCs such as Petrobrás, Odebrecht, and Braskem have morphed into channels for improper payments to politicians. As the investigations have continued, senior executives alongside politicians have been arrested and jailed. The ultimate impact on the future evolution of Brazilian MNCs is unknown. However, what does seem certain is that the form of state capitalism in which such enterprises

696   Edmund Amann were nurtured cannot avoid a fundamental overhaul. This is welcome to the extent that it may level the playing field and allow for the emergence of a new generation of large Brazilian MNCs.

References Amann, E. 2000. Economic Liberalization and Industrial Performance in Brazil. Oxford: Oxford University Press. Amann, E., and Baer, W. 2010. “Multinationals from the Periphery: An Analysis of the Latin American Experience.” World Economics 11 (1): 125–​146 Amann, E., and Cantwell, J., eds. 2012. Innovative Firms in Emerging Market Countries. Oxford: Oxford University Press. Baer, W. 2013. The Brazilian Economy: Growth and Development. Boulder, CO: Lynne Rienner. Bazuchi, K., S. Zacharias, L. Broering, M. Areola, and R. Bandeira de Mello. 2013. “The Role of Home Country Resources for Brazilian Multinational Companies.” Brazilian Administrative Review 10 (4): 416–​438. Buckley, P., and P. Ghauri. 1999. The Internationalization of the Firm. London:  Cengage Learning. Caves, R. 2007. Multinational Enterprises and Economic Analysis. New  York:  Cambridge University Press. Dahlman, C., and C. Frischtak. 1993. “National Systems Supporting Technical Advance: The Brazilian Experience.” In National Innovation Systems: A Comparative Analysis, edited by R. Nelson, 414–​450. New York: Oxford University Press. Dunning, J., and S. Lundan. 2008. Multinational Enterprises and the Global Economy. Cheltenham, UK: Edward Elgar. ECLAC. 2015. Latin American Investment Report 2015. Santiago: Economic Commission for Latin American and the Caribbean. FGV-​Columbia. 2016. “Leading Brazilian Multinationals:  Trends in an Era of Significant Uncertainties and Challenges.” Available at http://​ccsi.columbia.edu/​f.iles/​2013/​10/​EMGP-​ Brazil-​Report-​2015-​Jan-​27-​final.pdf FGV-​ Columbia. 2017. “The Top 20 Brazilian Multinationals:  Divestment under Crisis.” Available at http://​ccsi.columbia.edu/​files/​2013/​10/​EMGP-​Brazil-​Report-​March-​21-​2017-​ FINAL.pdf. Fleury, A., and M. Fleury. 2016. “O desenvolvimento das multinacionais Brasileiras no cenário global.” GVExecutivo 15 (1): 35–​37. Fleury, A., M. Fleury, and F. Borini. 2013. “The Brazilian Multinationals’ Approach to Innovation.” Journal of International Management 19: 260–​275. Fórum de Líderes. 2007. Internacionalizaçao das Empresas Brasileiras. São Paulo: Fórum de Líderes Empresariais. Lall, S. 1992 “Technological Capabilities and Industrialization.” World Development 20 (2): 165–​186. Mueller, B. 2016. “The Political Economy of the Brazilian Model of Agricultural Development: Institutions versus Sectoral Policy.” Quarterly Review of Economics and Finance 62 (November): 12–​20. Leite, L. 2005. Inovação: A combustível do futuro. Rio de Janeiro: Qualitymark.

Multinational Corporations from Brazil    697 Li, H., and J. Cantwell. 2012. “China.” In Innovative Firms in Emerging Market Countries, edited by E. Amann and J. Cantwell. Oxford: Oxford University Press. Malerba, F., and S. Mani. 2009. Sectoral Systems of Innovation and Production in Developing Countries. Cheltenham, UK: Edward Elgar. Mani, S. 2002. Government, Innovation and Technology Policy: An International Comparative Analysis. Cheltenham, UK: Edward Elgar. Mathews, J. 2002. Dragon Multinational: A New Model for Global Growth. New York: Oxford University Press. Musacchio, A., and S. Lazzarini. 2014. State-​Owned Enterprises in Brazil: History and Lessons. Paris: OECD. Oliveira M., Jr., F. Borini, and A. Fleury. 2013. “Innovation by Brazilian EMNEs.” In The Competitive Advantage of Emerging Market Multinationals, edited by P. Williamson et al, 11–​28. New York: Cambridge University Press. Parente, R., A. Cyrino, N. Spohr, and F. Carvalho de Vasconcelos. 2013. “Lessons Learned from Brazilian Multinationals’ Internationalization Strategies.” Business Horizons 56: 453–​463. Ramsey, J., and A. Almeida. 2010. A ascensão das multinacionais brasileiras. Rio de Janeiro: Elsevier. Rasiah, R. 2012. “Malaysia.” In Innovative Firms in Emerging Market Countries, edited by E. Amann and J. Cantwell, 191–​222. Oxford: Oxford University Press. Silva, O. 1998. A decolagem de um sonho: A história da criação de Embraer. São Paulo: Lemos. Torres, E., and R. Zeidan. 2016. “The Life Cycle of National Development Banks: The Experience of Brazil’s BNDES.” Quarterly Review of Economics and Finance 62 (November): 97–​104. UNCTAD. 2016. World Investment Report 2015. Geneva: United Nations Conference on Trade and Development. Voduzek, Z., ed. 2002. Foreign Direct Investment in Latin America:  The Role of European Investors. Washington, DC: Inter-​American Development Bank. Williamson, P., R. Ramamurti, A. Fleury, and M. Fleury, eds. 2013. The Competitive Advantage of Emerging Market Multinationals. New York: Cambridge University Press. World Bank. 2008. Global Economic Prospects 2008. Washington, DC: World Bank.

Pa rt  V I I

T H E C HA N G I N G ROL E OF T H E  STAT E

Chapter 33

The Rise an d Fa l l of State Ent e rpri se s Armando Castelar Pinheiro

33.1. Introduction Brazil’s first state-​owned enterprises (SOEs) predate the country’s independence in 1822. Until the early twentieth century, however, there were few SOEs, and private firms dominated nearly all sectors of the economy, save banking. Gradually but consistently, in the following decades SOEs gained importance and dominated key economic areas, from steelmaking to infrastructure. This process peaked in the 1970s when powerful SOEs controlled a large share of corporate savings and expanded into new sectors. In 1971, eight of Brazil’s 25 largest corporations were SOEs, accounting for 31% of the group’s sales. Several factors drove the expansion in the number and size of Brazil’s SOEs. The most important was the state-​led industrialization push that marked the 1930–​1980 period. Policymakers saw industrialization as the only way to develop Brazil economically, and used SOEs to produce manufactured goods and to support private manufacturing companies by acquiring and selling goods and services at favorable prices. In this way, SOEs were one of a basket of industrial policy instruments—​which also included subsidized loans and trade and exchange rate policies—​that channeled resources to foster the domestic production of manufactured goods. However, already by the 1970s, SOEs had begun to be criticized on the grounds that they were competing with private firms for consumers and key resources such as long-​ term financing, skilled workers, and foreign exchange. The business sector, in particular, argued that SOEs should focus more on supporting private firms and less on expanding their activities into sectors that private firms could develop, providing they had access to the same financial and human resources. In the following years, opposition to Brazil’s large SOE sector grew until, timidly in the 1980s and then more resolutely in the 1990s, Brazil embarked upon privatization.

702   Armando Castelar Pinheiro The ascent and eventual retreat of Brazil’s SOEs closely followed the pattern seen in more advanced economies, particularly those of Europe. This parallel highlights the influence of mainstream ideas, particularly those advocated by multilateral organizations, upon domestic policy. Also critical, however, was the shift in domestic priorities. I refer particularly to the priority ascribed to enhanced social spending since Brazil’s return to democracy in 1985. Thus, while until the early 1980s the focus was on channeling resources to expand manufacturing, since then, governments have directed an increasing volume of funds to expand education, health, and social security coverage. SOEs are a poor instrument to carry out that sort of policy. As such, somewhat counter intuitively, the size of Brazil’s state expanded significantly in the last three decades, at the same time that its SOE sector shrank. Privatization has substantially changed the country’s economic landscape, notably in infrastructure. In 2015, only three of Brazil’s 25 largest companies were SOEs. Yet, contrary to what the statistics on SOEs may suggest, the state has retained a substantial influence on Brazil’s corporate sector. This influence stems from the large stock holdings that government banks and SOE pension funds maintain on various corporations (Lazzarini 2011). Moreover, government banks answer for most of the credit in the economy and, together with the pension funds of the remaining SOEs, for almost all of the long-​term financing available in domestic financial markets. Therefore, long-​ term financing reflects not the market’s but the government’s priorities, among which are fostering industrialization and the creation of “national champions.” Thus, while SOEs are no longer the preferred instrument for, nor industrialization the only target of, transferring government subsidies, several of the ideas that led to the rise of Brazil’s SOE sector remain present in policymaking. This chapter examines the rise and fall of SOEs in Brazil, focusing on the drivers of each movement. Section 33.2 chronicles the rise of Brazilian SOEs until their peak in the 1970s and discusses the reasons behind this process. Section 33.3 examines the factors behind the decline of SOEs in the late 1970s and in the 1980s. Section 33.4 analyzes the privatization process. The final section (section 33.5) concludes, arguing that the state remains extremely influential on corporate investment decisions.

33.2.  The Rise of State Enterprises State-​owned enterprises are an instrument of public policy. Whether or not they play an important role in the economy depends on how active public policy is and whether SOEs are seen as the best means to achieve policy goals. The dominant views on both these issues have changed considerably over the last century and a half. Before the Great Depression, the consensus favored little state intervention in the economy. Starting in the post–​World War I period, though, and more forcefully with the Great Depression, the state became a more important economic actor through ownership and regulation

The Rise and Fall of State Enterprises    703 of commercial enterprises. Particularly in Western Europe, the 1930s saw a massive wave of nationalization of private firms (Toninelli 2008). The views on the usefulness of SOEs as a policy instrument went from enthusiasm to frustration and criticism (Shapiro and Willig 1990). When overall support of state intervention grew, policymakers saw in SOEs a means to pursue policy objectives in the economy that was more direct and controllable than instruments such as taxes and subsidies, which only partially align the incentives of policymakers and private firms. After establishing an SOE, the government could appoint its managers and direct them to pursue its goals, be they the production of a manufactured good or an infrastructure investment. Experience showed, though, that there were two problems with this. First, it is not always good that policymakers have unbridled control of policy instruments, for their objectives only partially coincide with those of society. Second, SOE managers often have their own agenda and only partially pursue the goals of policymakers. Thus, beginning in the 1970s, and more forcefully in the 1980s and 1990s, “deregulation” and “privatization” gradually became the buzzwords almost everywhere. Largely, this description fits well with how SOEs were seen in Brazil. Thus, prior to the Great Depression, SOEs were almost nonexistent in Brazil, except in finance and, as of the early twentieth century, in the railway sector. With the establishment of President Getúlio Vargas’s Estado Novo (New State), the liberalism of Brazil’s First Republic (1889–​1930) gave way to a more interventionist state. Faoro (1957, 803, cited by Lazzarini 2011) quotes Getúlio criticizing the trade openness and liberal views of previous governments: I consider advisable the nationalization of certain industries and the gradual socialization of others, results possible to reach through a rigorous control of public utilities and the slow penetration of the management of private companies which have their development dependent on government favors.

Starting in the 1930s, and with greater enthusiasm in the 1940s, the number of SOEs increased considerably. Yet, by 1949 the federal government controlled only 28 SOEs, up from 20 in 1940. Of those 28 SOEs, 12 were electricity companies, 5 financial institutions, 6 transportation firms, and 5 industrial enterprises. Meanwhile, at the local state level there were 34 SOEs, including 12 financial institutions and 10 transportation companies (Pinheiro and Oliveira Fo 1991; Rezende 1980). Upon his return to the presidency in 1950, Vargas embarked on a new round of SOE creation, later extended by his successors. Overall, the federal government established 15 new SOEs in the 1950s, while local state governments created another 49 (Rezende 1980). In the 1960s, the number of SOEs continued to expand and, by the early 1970s, the state’s entrepreneurial activities had gained considerable prominence. According to Baer, Kerstenetzky, and Villela (1973, 29), Taking the top 25 companies in terms of value of sales during 1971, it was found that 8 were government firms which accounted for 31 per cent of sales. Of the 25 largest

704   Armando Castelar Pinheiro companies in terms of employment, 7 were government firms accounting for 51 per cent of employment. Finally, of the 25 largest companies in terms of assets, 17 were government firms accounting for 82 per cent of total assets.

The creation of SOEs reached its peak during the military regime (1964–​1985), more specifically in the 1970s, when the federal government established more than 70 SOEs. A similar process took place at the state and municipal level. Overall, 219 new SOEs were established in the period 1966–​1977 (Rezende 1980).1 With the rise in the number and size of SOEs, the government gradually lost control of their activities. Thus, in 1979 the Ministry of Planning created a new department—​the Special Secretariat of State Enterprise Control (SEST; Secretaria Especial de Controle de Empresas Estatais)—​with the specific goal of regaining control over federal SOEs. The first thing SEST did was a census to count and identify SOEs. Overall, it counted 505 institutions under federal control, 268 of which were SOEs. Congress had approved the creation of only 40 of those 268 enterprises. What led successive governments to expand so significantly the number of SOEs? Today, one would probably point to the many market failures that can possibly justify the existence of SOEs: natural monopolies, incomplete markets, asymmetric information, externalities, and so on. However, these were not concepts familiar to policymakers in the 1950s. Ideology and external influences were probably stronger motivations, but there is no consensus in the literature whether the establishment of SOEs resulted from a common rationale or ideology, or just from overlapping but otherwise disconnected policies. Baer, Kerstenetzky, and Villela (1973) and Cardoso (1973), for instance, argue that the rapid postwar growth in the number of SOEs was not a planned process, nor the result of a nationalistic ideology. On the other hand, for Martins (1977, 26–​27), . . . there was always an ideology of state intervention (whether in the form of statism, nationalism, or developmentism), in which the point of reference was the concept of the nation. [ . . . ] This was why, on the ideological plane, these three “isms” often appeared in an intermingled form, as interchangeable concepts [ . . . ]. It is historically inexact, however, to state, as is now frequently stated, that the business activities of the state emerged in Brazil almost accidentally and without any link to any defined political project.

I believe the truth lies somewhere in between. There was indeed a common thread linking the creation of SOEs throughout this period: the desire to industrialize Brazil, which Brazil’s elite believed to be the only means to become a developed economy. Yet, there was no unifying plan or program for that; rather, with this overarching goal in mind, governments reacted to short-​term practical problems by creating SOEs, or by resorting to other policy instruments (Pinheiro 2016). Thus, the government established most SOEs either to produce manufactured goods or to finance and supply inputs to private manufacturing firms. This was the case with the steel mills established in the 1940s and 1950s, with large infrastructure investments, and

The Rise and Fall of State Enterprises    705 with the National Development Bank (BNDES; Banco Nacional de Desenvolvimento Econômico e Social), originally conceived as an Industrial and Investment Bank (Araújo, undated). Later, in the 1970s, the government entered into joint ventures with private and foreign investors to develop the domestic basic inputs and capital goods sectors. Overall, this process followed the footsteps of several European countries, notably France, Italy, Portugal, and Spain (Toninelli 2008). SOEs were also the result of a nationalistic ideology and concerns with national security, the influence of which was reinforced by the prevailing Cold War mood. Thus, although foreign firms played an important role in Brazil’s industrialization as providers of capital and technology, the government was throughout unwilling to allow them to control sectors deemed important for national security. These included natural resources (land, oil, mining, etc.) and infrastructure. Yet, as national private firms lacked the capital and expertise to enter those activities, the solution was to assign SOEs with the responsibility to develop them. Partly related to that was the nationalization of foreign companies, notably in infrastructure. This process also reflected regulatory failures of different sorts, which in gen­ eral led those companies to invest less than the government deemed necessary to allow rapid industrialization. This was particularly significant in the case of foreign railway and electricity companies. Starting in the late 1960s, reforms in the governance (1967 Administrative Reform) and finances of the SOEs gave them the flexibility and means to establish their own subsidiaries. Petrobrás (oil and gas) and Vale (mining) were key participants in this process. Petrobrás became a conglomerate with 35 subsidiaries, with activities ranging from oil prospecting, refining, distribution, and transport, to the production of fertilizers, petrochemicals, and foreign trade. Companhia Vale do Rio Doce (CVRD), in turn, entered into reforestation, fertilizers, cellulose, and civil construction (Pinheiro and Oliveira Fo 1991). Finally, several SOEs originated from government bailouts of private companies, in most cases when they failed to honor loans with government banks, notably BNDES. Similar operations led to the creation of SOEs in Western Europe, from Germany in the 1930s to the United Kingdom in the 1970s (Toninelli 2008). What was peculiar to Brazil, though, was that these were mostly small companies, situated in sectors where there was no justification for the state to intervene: hotels, sugar mills, publishing companies, and so on. In SEST’s 1979 census, 76 of the SOEs then identified became state-​owned in that fashion.

33.3.  The Declining Fortunes of State-​O wned Enterprises Urban elites and private firms largely supported the expansion in the number and size of SOEs, seeing them as a lever to achieve their own economic goals. In particular, there

706   Armando Castelar Pinheiro was a general perception that private national companies lacked the capital and technology to substitute for SOEs in activities such as infrastructure and heavy industries. The fact that Western Europe followed a similar path also helped to legitimate this support. In a paper written in the early 1970s, Baer, Kerstenetzky, and Villela (1973, 281) sum up the prevailing view: The continuous growth in the participation of the state in economic activities in Brazil in the last three decades was almost inevitable. The Brazilian private sector is still relatively small and does not have the capacity to play an important role in the country’s enormous needs for infrastructure or in the industries which use more sophisticated technology, which are also the most dynamic—​petrochemicals, steel, transportation equipment, etc.

A few years later, though, this support began to wane. Reflecting this reversal, in early 1975 the newspaper O Estado de São Paulo published a series of articles entitled “Os caminhos da estatização” (“The Paths to Nationalization”) criticizing the expansion of SOEs. Soon after, a group of businessmen started a movement named Campanha contra a estatização (“Campaign against Nationalization”) that would also criticize what they saw as the growing ascendency of SOEs in the economy. Four main factors led to this change of mind: (1) The economic slowdown and heightened volatility that followed the first oil shock, which put an end to Brazil’s “economic miracle.” (2) The competitive pressure imposed by SOEs in sectors in which private firms previously had faced little competition, thanks to the establishment of new SOEs in the wake of the 1967 administrative reform. Thus, among the main proposals put forward by the participants in the “Campaign against Nationalization” was “the suppression of some privileges enjoyed by state-​owned companies, such as the freedom to invest funds and exemption from some taxes, limitation of their capacity to create subsidiaries, through control of diversification, prohibition to use funds arising from compulsory savings and other tax incentives” (Pessanha 1981, 95–​96). (3) The large investment projects in which SOEs embarked upon as a policy reaction to the first oil shock. Those projects led SOEs to compete with private firms, in privileged conditions, for the scarcest resources in the economy: skilled labor, low-​cost financing, and foreign exchange. (4) The exclusion of private entrepreneurs from a number of important forums, such as the Conselho de Desenvolvimento Econômico (Economic Development Council), thus greatly reducing their influence on economic policy. Pessanha (1981, 154) notes that one of the key demands of the dissatisfied businessmen was to be “heard throughout the whole process of decision on the national economy, helping to establish criteria for the activity of the state and the private sector, guiding the use of the SOEs and controlling their expansion, deciding directions for the investment of their savings, etc.”

The Rise and Fall of State Enterprises    707 The government’s decision to keep the economy expanding at a brisk pace, despite the signs that it was overheating and the significant drop in terms of trade, began pushing up inflation and the external deficit. In the late 1970s, policymakers finally became more concerned with avoiding an economic crisis than with fostering industrialization. The priority was no longer growth and import substitution, but curbing inflation and overcoming the foreign exchange crisis. Since SOEs accounted for a considerable share of domestic absorption, it would be almost impossible to stabilize the economy without bringing their expenditures and deficits under control (Werneck 1987). The government reacted by curbing the growth of SOEs, to which end it created the aforementioned SEST, in 1979. That move did not reflect a change in the prevailing views about the role of SOEs in fostering economic development. In particular, the reasons behind the creation of SEST differed from those that would later lead to privatization. The point in question at that time was not the inefficiency of SOEs, but the need to decelerate their expansion, to control aggregate demand, a task complicated by the lack of control of these companies by federal authorities. Resende (1980, 37) summarizes well the predominant view: In fact, the whole debate on the need to limit the increase in the functions of the state reflects the incapacity of the public administration to control the actions of government companies, whose decisions to expand escape the control exercised during the periodic analysis of the budget . . . to the extent that the decisions to invest in certain sectors by private enterprises are subordinated to public credit and/​or fiscal incentive schemes, control of the decisions of the privately-​controlled companies is greater than the control of the decisions of the public companies, whose capacity to mobilize funds gives them a certain independence in relation to the central power.

The economic crisis of the late 1970s and early 1980s also affected SOEs in other, less benign, ways. Starting in the mid-​1970s, the government forced SOEs to reduce the price of the goods and services they sold, in real terms, initially to mitigate the rise in inflation and, after 1982, to subsidize manufacture exports. This significantly reduced their operating margin. Also pressuring the bottom line were the large external loans that federal SOEs contracted to finance the country’s growing current account deficit and to leave more of the domestic credit to private firms. With their external obligations considerably increased, the rise in international interest rates starting in 1979, and the significant currency devaluation after 1983, seriously harmed those companies. These policies also weakened the professional ethos of SOE employees, compromising the efficiency of those companies. Other factors also reinforced this process by weakening incentives to seek greater efficiency. First, SOEs operated on a soft budget, with the central government rescuing them when losses became large. Second, they often enjoyed monopoly rents or benefited from subsidies, which hid their inefficiencies. Third, in the 1980s the top SOE managers were often political appointees, were seldom rewarded based on profit maximization, and suffered considerable turnover.

708   Armando Castelar Pinheiro Soon, SOEs would become large whirlpools of public resources; increasingly, many people saw them no longer as national champions that would bring development, but as a drag on an economy trying to escape an economic crisis. This, plus the success of privatization in Europe, pressure from multilateral organizations, and demands from foreign investors, gradually made privatization increasingly acceptable to a large segment of Brazilian society (Pinheiro and Giambiagi 2000).

33.4. Privatization The campaigns against nationalization of the mid-​1970s mentioned the sale of SOEs, but accompanied by so many misgivings about its effectiveness that the theme had only a symbolic role in the overall body of proposals. In the 1970s privatization was clearly not a priority for the private sector.2 At the end of the 1970s, the economic situation further worsened, SOEs were increasingly seen as part of the problem, and the idea of privatization began to permeate government discourse, although with a lack of practical consequence. In a message to his cabinet shortly after taking power in March 1979, President Figueiredo recommended adoption of the measures necessary “for privatization of the SOEs and services that are not strictly essential to correction of market imperfections or for meeting the needs of national security” (Palatnik and Orenstein 1979, 52). As I have outlined elsewhere (Pinheiro 2011, 259), It was not until 1981, though, that privatization was actually put on the economic policy agenda. In July of that year, a presidential decree created the Comissão Especial de Desestatização (Special Privatization Committee) and set “rules for the transfer, transformation and divestiture of companies controlled by the federal government.” The main objectives of the committee were to strengthen the private sector, limit the creation of new SOEs, and close or transfer to the private sector those SOEs whose control by the public sector was no longer necessary or justifiable. Once set up, the committee identified 140 SOEs ready to be privatized in the short term. Of these, 50 were initially put on the list for sale. However, the actual balance of this first attempt at privatization was not to reach even this number: a total of 20 companies were sold to private investors, 1 was leased, and 8 were incorporated into other public institutions. In this same period, however, 6 companies in a bankrupt situation were incorporated by the BNDES, through what were then called “hospital operations.” The companies sold in this period were, in general, cases of reprivatization, and their list did not include any of the major SOEs.

Privatization continued at the same pace in the Sarney administration (March 1985 to March 1990), despite the aggressive rhetoric reflected in the series of presidential decrees and draft bills restructuring and enlarging the privatization program. In total, 18 SOEs were sold, with an equal number transferred to the state governments, two merged into

The Rise and Fall of State Enterprises    709 other federal institutions, and four closed down. Most were small-​and medium-​size companies in sectors dominated by the private sector, and whose privatization aimed at improving the financial health of their owner, BNDESPAR, the subsidiary of BNDES responsible for capital market operations. On assessing the Brazilian privatization experience in the 1980s, the World Bank (1989, 27)  concluded that “Brazil’s first flirtation with privatization was a ‘classic example of failure.’ ” Indeed, both in speed and scope, the sale of SOEs in the 1980s fell well short of the level promised by government rhetoric. Further, BNDES was responsible for most sales, as it sought to free itself of problematic, loss-​making companies. Thus, most sales did not result from a favorable perception about privatization on the part of the government. Thus, in the 1980s privatization was not a priority for the public sector. Indeed, the 1988 Constitution was clearly a nationalizing one, establishing monopolies for SOEs in telecommunications, oil, and distribution of gas, and setting up barriers to foreign ownership in mining and electricity. Thus, at the end of the 1980s, SOEs continued to play a dominant role in the Brazilian economy, ranking among the largest firms in Brazil. In 1989, SOEs accounted for 8 of the Brazil’s 50 largest firms (ranked by total sales),3 17 of the 20 largest firms (ranked by net worth), and half of the 20 most profitable firms. Furthermore, in 1989, while the 500 largest firms employed 1,986,000 workers, the 50 largest SOEs accounted for 650,000 jobs. Finally, SOEs accounted for around 17% of total investment (1987) (Pinheiro and Oliveira Fo 1991). Yet, less than two years after the promulgation of the new constitution, the Fernando Collor administration launched the National Privatization Program (PND; Programa Nacional de Desestatização), significantly widening the scope of privatization. Different factors contributed to this turnaround, including changes in the international political scene, the further deterioration in the financial performance of SOEs, and the needs of macroeconomic policy. A noteworthy feature of the public debate about privatization in the 1990s was how the issue of national security and the concern with foreign ownership of infrastructure lost importance. This change may be largely credited to the process of democratization (1985), which reduced the influence of the military, and to the end of the Cold War (Nestor and Mahboodi 1999). The fact that President Collor was elected by popular vote, in contrast to the negotiated transition that made possible the rise of Sarney to the presidency, was also important in legitimizing this change of direction. The early 1990s also saw a more dramatic shift in economic policy, from an almost complete focus on capital accumulation—​presented as a concern to “occupy empty spaces”—​to a policy more concerned with efficiency and productivity growth. SOEs are a good instrument to accumulate capital and diversify industrial output, but they are in general less efficient than private companies. Thus, it was not a mere coincidence that the PND was launched simultaneously with trade liberalization and significant deregulation of the domestic economy, together with the ending of public monopolies in sectors such as sugar, ethanol, coffee, wheat, and so on.

710   Armando Castelar Pinheiro By that time, it had also become clear that the state had exhausted its capacity to lead the process of capital accumulation, since it was not capable of either generating a fiscal surplus, or of borrowing abroad. Finance to SOEs disappeared for another less obvious reason: government banks were (and still are) responsible for virtually all long-​term credit in Brazil. Because these banks could not collect the collateral offered by SOEs for political and legal reasons, SOEs did not bother to pay them back.4 So in the mid-​1980s the government forbade public banks to lend to SOEs, drying up their last source of finance. Privatization became then the only means through which government banks could finance the sectors in which SOE presence was massive—​being private, these companies could credibly offer their assets as collateral. Also relevant was that technological progress eroded the relevance of natural monopolies, notably in telecommunications, but also in electricity, thus allowing for the breakdown of SOEs in infrastructure before their sale. This favored the establishment of competition in sectors in which this was not viable in the past. Privatization began in the 1990s with very optimistic targets in terms of revenue and timetable, which turned out to be impossible to fulfill due to the poor financial situation of the SOEs and the complexity of these companies’ stockholders’ agreements. The SOEs were not ready for sale; they needed a long process of preparation before they could be privatized. It was only at the end of 1991 that the first company was sold under the PND. Overall, the Fernando Collor and Itamar Franco governments (1990–​1994) privatized 33 companies, collecting US$8.6 billion, 12 times as much as in the previous decade. Almost all of those companies were in manufacturing, mainly in the steel, petrochemical, and fertilizer sectors. Privatization of the public sector monopolies was not even considered at the time. Starting in 1995, two near-​simultaneous movements greatly widened the scope of privatization: the end the public sector monopolies in infrastructure and the decision of the local state governments to develop their own privatization programs. Different factors fostered these two movements. First, the success of the Real Plan in achieving price stability gave the government the political leverage to pass the necessary constitutional amendments through Congress so as to extend privatization to the telecom and gas sectors and allow foreign investors to enter the mining and electricity sectors. Second, to sustain price stability the government needed to achieve fiscal discipline, and this limited its ability to carry out the high levels of investment necessary to sustain the recovery in economic growth. Also, for fiscal reasons, economic policy limited the access of SOEs to domestic and external financing. Third, the states saw in privatization an important source of funding, which would allow them to reduce their debt and, in some cases, expand spending. The debt restructuring contracts between the states and the federal government required the states to pay part of the principal upfront, which could be done only through the sale of their assets (i.e., through privatization) (Pinheiro and Giambiagi 2000). Loans provided by BNDES facilitated this, by allowing the states to borrow against future privatization revenues.

The Rise and Fall of State Enterprises    711 Fourth, stability itself, and the improved perception of investors about Brazil’s risk and growth potential—​reflected, for example, in the large expansion in foreign direct investment—​helped to increase the value of these companies, making privatization more interesting for both the public sector and private investors. Fifth, the success of the privatizations carried out in 1991–​1994, evidenced by the companies’ increased efficiency and investment, helped to widen political support for the program. Pinheiro (1996) shows that privatization substantially improved the performance of the former SOEs, with significant increases in real sales, sales per employee, net profit, stockholders’ equity, investment, fixed assets and the ratio of investment to sales.5 However, of all the factors contributing to expanding privatization in Fernando Henrique Cardoso’s first term, the most important was that privatization was key to sustaining the Real Plan. With the large sales of 1997–​1998, Brazil attracted sizable volumes of foreign direct investment, which helped to finance the country’s rising current account deficit, reducing the need for a currency devaluation. Privatization was also instrumental in averting an explosion in government debt, in spite of the growing fiscal deficit.6 In all, the 80 SOEs privatized in 1995–​1998 provided total revenues of US$60.1 billion. This was the peak of the privatization program in Brazil, which would decelerate, although not fully stop, in the following years. In Cardoso’s second term (1999–​2002), privatization was no longer critical to sustain the Real Plan, due to the sharp currency devaluation of early 1999 and the major fiscal adjustment program implemented in the following years. Moreover, a weaker currency and higher taxes seriously harmed the government’s popularity, in turn raising the opposition to further privatization. Thus, the government was left with few SOEs that both were attractive to private investors and whose privatization was politically feasible. Thus, in 1999–​2002 the focus shifted from the sale of assets to attracting private investors to undertake infrastructure investments. In Luis Inácio Lula da Silva’s first term (2003–​2006), however, not even this light form of privatization was tolerated, and the program stopped altogether, although the new government did not try to undo any previous privatization. Nevertheless, the government’s attempt to expand infrastructure investment using existing SOEs and the sectoral ministries failed, and gradually it moved back into offering concessions to private investors, notably in the electricity and transportation sectors. This policy would stay in place during Dilma Rousseff ’s administration (2011–​2016), although regulatory risk increased considerably, requiring more credit subsidies to make the concessions attractive to private investors.

33.5.  Final Remarks Overall, privatization in Brazil has always reflected more a pragmatic response to short-​term macroeconomic problems than a dramatic shift in ideology, as was the case

712   Armando Castelar Pinheiro in the United Kingdom with Margaret Thatcher. Thus, only a minority of Brazilians supported it as part of a process to reduce state interference in the market. Most Brazilians continue to trust the state to lead the way to economic development. Thus, the Lula and Rousseff administrations created a number of SOEs and strengthened several others. It should surprise no one, then, that after 35 years of privatization, the state still owns a large share of corporate assets in Brazil. The state continues to influence the decisions of most large companies in Brazil, either directly, through ownership of large SOEs, or indirectly, through control of financial institutions that have controlling or significant participations in the capital of other firms. At the end of 2014, the Brazilian federal government directly or indirectly owned 135 SOEs, essentially the same number as in 2010 (Table 33.1).7 Of these 135 companies, 18 were in the financial sector, while the other 117 were in sectors such as oil and gas, electricity, transportation, communications, food warehousing, health, and research and development (Figure 33.1). Together, those SOEs held assets worth R$4.5 trillion—​ of which about three-​quarters were in the financial sector and one-​quarter in the nonfinancial sector—​and had a net worth of about R$595 billion, split between financial (36.5%) and nonfinancial (63.5%) SOEs.8 In 2014 Petrobrás was the SOE with the highest asset value, net worth, and net revenue. Federal SOEs employed 552,900 people (excluding outsourced workers), 41% working in the financial sector and 59% in the nonfinancial sector (Figure 33.2). The number of SOE employees grew quite considerably (11.2%) between 2010 and 2014. Among nonfinancial SOEs, the Petrobrás and Eletrobrás groups are the leading conglomerates, comprising between them a total of 67 SOEs. Together, these two groups accounted for 27% of the net sales, 45% of total assets, and 40% of the net worth of Brazil’s 20 largest companies. Of the other 50 nonfinancial SOEs, 18 (Valec, EPE, and Embrapa, for example) depend on fiscal transfers to cover their expenses, including personnel and other current expenses. The SOE financial sector comprises the Banco do Brasil Group, BNDES, CAIXA, Banco da Amazônia, Banco do Nordeste do Brasil, and the Financiadora de Estudos e Projetos. In 2013, they accounted for 45.1% of total assets, 31.4% of the equity, and 43.6% of the net profit of Brazil’s banking sector. In addition, government banks account for more than half of all the credit extended by Brazilian banks.

Table 33.1 Number of Federal SOEs (End of the Year) 2010

2011

2012

2013

2014

Direct ownership

44

44

46

48

48

Indirect ownership

94

95

95

93

87

138

139

141

141

135

Total Source: Brasil (2015).

The Rise and Fall of State Enterprises    713 Communications, 2%

Mining, 2%

Airport administration, 1%

Transportation, 3%

Regional development, 1%

Food wholesaling, 3% Manufacturing , 3% Research and development, 3% Health and social assistance, 3%

Oil and gas, 27%

Ports, 6% Commerce and services, 11%

Power, 22% Finance, 13%

Figure 33.1.  Distribution of the number of federal SOEs by sector. Source: Brasil (2015).

600.0 500.0

497.0

515.3

539.2

550.0

552.9

315.8

323.2

323.7

Thousands

400.0 300.0 200.0

289.1

299.9

208.0

215.4

223.5

226.9

229.1

2010

2011

2012

2013

2014

100.0 –

Finance

Nonfinancial sectors

Total

Figure 33.2.  Number of SOE employees.* * Does not include outsourced workers. Source: Brasil (2015).

However, the influence of the state over Brazilian companies extends beyond the control of SOEs. As Lazzarini (2011) shows, the state also operates through a network of equity participations in private companies held by BNDES and the pension funds of SOEs, both of which are controlled by the federal government. Somewhat paradoxically,

714   Armando Castelar Pinheiro the privatization process strengthened such mechanisms. This occurred because both BNDES and the pension funds were leading participants in the new ownership structure that emerged from privatization. Lazzarini (2011) also analyzes the boom of initial public offerings (IPOs) in the stock exchange to conclude that the state, through BNDES, was the leading investor in the companies that opened up their capital between 2004 and 2009. The Lula and Rousseff governments deepened the connections between BNDES, pension funds, and private groups, significantly expanding those channels of influence. This derived, in part, from the process of ownership concentration in this period, thanks to mergers and acquisitions that put a larger number of companies under the control of the same corporations. Another contributing factor was the creation of the so-​called national champions policy, as part of Lula’s and Rousseff ’s industrial plans. Also important were the myriad consortia and companies established by construction companies, SOEs, and pension funds to compete for infrastructure concessions. These partnerships allowed private firms to control large projects with little of their own capital, a means per se to compensate for increased regulatory risk. Lazzarini (2011, 35) exemplifies this type of association with the case of the hydropower plant of Belo Monte, in the Xingu River: [T]‌he auction was carried out and won, in April 2010, by a group of nine partners, including the SOE Eletrobras and several private groups (such as Serveng, Mendes Jr. and Bertin). Lately, other investors became associated with the consortium, including pension funds Petros, Funcef and Previ [ . . . ]. As usual, BNDES financed a substantial share of the construction.

According to Lazzarini (2011), if we factor in these mechanisms, we can conclude that the state’s control over corporate decisions actually expanded after privatization. And this does not consider the rising influence of BNDES on account of the significant expansion of its lending activities, and the subsidies transferred in that way, which also gave the bank much more leverage to influence business decisions. The flipside was that the private entrepreneurs that participated in this network of equity participations greatly increased their influence over the allocation of one of the most cherished resources in the economy: the long-​term, often subsidized financing extended by BNDES and the pension funds of SOEs. This was exactly what the private sector was asking for when it started the Campaign against Nationalization in the 1970s:  to restore the intertwined relationships created in the 1930s between the private and state sectors that so benefited the private entrepreneurs with the right kind of contacts or relationships. Thus, although SOEs are no longer so numerous or widespread as in the 1970s, the state still retains to this day much influence on corporate decisions. It is difficult, for instance, to enter the oil sector without some association with Petrobrás, or to invest in infrastructure without a partnership with BNDES, SOEs, or their pension funds. Indeed, in many ways it is hard to enter and prosper in the Brazilian economy without participating in the many networks that link private firms and state actors.

The Rise and Fall of State Enterprises    715

Notes 1. As Rezende (1980) points out, though, these numbers reflect in part the change in classification of firms and other public institutions that resulted from the 1967 Administrative Reform (Decreto–​Lei 200), and were not all new organizations. 2. In the view of businesspeople, one of the reasons why privatization was not a solution was the private sector’s lack of funds and the concentration of credit in government banks. In a document produced by business leaders, this issue is expressed as follows: “Either the private-​sector company acquires state-​controlled companies from the government, with funds from the public sector itself, an option which will make it extremely difficult to choose the new owners without falling into paternalism, or the already scarce funds of the private sector will be absorbed in buying existing undertakings, leading the government to fill up the newly-​formed ‘empty spaces’ with these funds” (Pessanha 1981, 105). 3. And 126 of the largest 500, down from 158 in 1980. 4. In fact, SOEs failed to pay not only the public banks, but also its SOE suppliers: the steel company would not pay its electricity supplier, which would not pay the power generator, which would not pay the oil company that supplied its fuel, and so on. And there was little creditors could do beyond applying political pressure, not least because legally a creditor cannot ask for a SOE bankruptcy. As arrears built up, every now and then these multiple defaults were “solved” by transferring debts to the Treasury, with the taxpayers picking up the tab. 5. See also Anuatti Neto (2005). 6. Carvalho (2001) shows that thanks to the use of privatization revenues to redeem public debt, in December 1999 this was 8.4% of GDP lower than it would have been without privatization. 7. Indirect ownership refers to subsidiaries of SOEs directly controlled by the federal government. 8. At the end of 2014 the real traded at R$2.6562/​US$.

References Aharoni, Y. 1986. The Evolution and Management of State Owned Enterprises. Cambridge, MA: Ballinger. Anuatti-​Neto, Francisco, Milton Barossi-​Filho, Antonio Gledson de Carvalho, and Roberto Macedo. 2005. “Os efeitos da privatização sobre o desempenho econômico e financeiro das empresas privatizadas.” Revista Brasileira de Economia 59 (2): 151–​175. Araújo, Victor Leonardo. n.d. A criação do BNDE e a controvérsia Lafer-​Jafet. Mimeo. Baer, Werner, Isaac Kerstenetzky, and Annibal V. Villela. 1973. “The Changing Role of the State in the Brazilian Economy.” World Development 1 (11)” 23–​34. Baer, W, and C. McDonald. 1998. “A Return to the Past? Brazil’s Privatization of Public Utilities: The Case of the Electric Power Sector.” Quarterly Review of Economics and Finance 38 (3): 23–​34. Bastos, P. 2006. “A construção do nacional-​desenvolvimentismo de Getúlio Vargas e a dinâmica de interação entre Estado Mercado nos setores de base.” Revista Economia 7 (4): 239–​275. Brasil. 2015. Perfil das Empresas Estatais Federais, 2014. Ministério do Planejamento Desenvolvimento e Gestão. Brasília: MP/​SE/​DEST.

716   Armando Castelar Pinheiro Cardoso, Fernando Henrique. 1973. “Associated-​Dependent Development:  Theoretical and Practical Implications.” In Authoritarian Brazil: Origins, Policies and Future, edited by Alfred Stepan, 142–​176. New Haven, CT: Yale University Press. Carvalho, M. A. 2001. “Privatização:  Aspectos fiscais e dívida pública.” MSc dissertation, Fundação Getúlio Vargas. Faoro, Raymundo. 1957. Donos do Poder: Formação do patronato político brasileiro. São Paulo: Editoria Globo. Lazzarini, Sergio. 2011. Capitalismo de Laços. Rio de Janeiro: Campus/​Elsevier. Martins, Luciano. 1977. Expansão recente do Estado no Brasil:  Seus problemas e seus atores. Mimeo. FINEP. Nestor, Stilpon, and Ladan Mahboobi. 1999. “Privatisation of Public Utilities:  The OECD Experience.” In A privatização no Brasil: O caso dos serviços de utilidade pública, edited by Armando Castelar Pinheiro and Kiichiro Fukasaku, 103–​144. Rio de Janeiro: Banco Nacional de Desenvolvimento Econômico e Social. Palatnik, Beny, and Luiz Orenstein. 1979. “Perspectivas do processo de privatização no Brasil.” Encontros com a civilização brasileira September, 43–​62. Pessanha, C. F. 1981. “Estado e economia no Brasil: A campanha contra a estatização.” PhD dissertation, Instituto Universitário de Pesquisas do Rio de Janeiro. Pinheiro, Armando Castelar. 1996. “Impactos microeconômicos da privatização.” Pesquisa e Planejamento Econômico 26 (3): 357–​398. Pinheiro, A. C. 2002. “The Brazilian Privatization Experience:  What’s Next?” University of Oxford Centre for Brazilian Studies, Working Paper No. CBS-​30-​02. Pinheiro, Armando Castelar. 2011. “Two Decades of Privatization in Brazil.” In The Economies of Argentina and Brazil: A Comparative Perspective, edited by Werner Baer and David V. Fleischer, 252–​278. Cheltenham, UK: Edward Elgar. Pinheiro, Armando Castelar. 2016. “Brazil’s Deindustrialization.” Mimeo. Pinheiro, Armando Castelar, and L. C. de Oliveira. 1991. “Brazilian Privatization: A Decade of Experience.” Discussion Paper No. 7, CEBRAP. Pinheiro, Armando Castelar, and Fabio Giambiagi. 2000. “Os antecedentes macroeconômicos e a estrutura institucional da privatização no Brasil.” In A privatização no Brasil: O caso dos serviços de utilidade pública, edited by Armando Castelar Pinheiro and Kiichiro Fukasaku, 13–​44. Rio de Janeiro: Banco Nacional de Desenvolvimento Econômico e Social. Putniņš, T. J. 2015. “Economics of State-​Owned Enterprises.” International Journal of Public Administration 38 (11): 815–​832. Radygin, A., et  al. 2015. “The State-​Owned Company:  ‘State Failure’ or ‘Market Failure’?” Russian Journal of Economics 1 (1): 55–​80. Rezende, Fernando. 1980. “A empresa pública e a intervenção do Estado na economia: ação suplementar à iniciativa privada—​perspectivas em face da conjuntural atual.” In A empresa pública no Brasil: Uma abordagem multidisciplinar, edited by Sérgio Henrique Abranches, 54–​55. Instituto de Planejamento Econômico e Social (IPEA). Shapiro, Carl, and Robert Willig. 1990. “Economic Rationale for the Scope of Privatization.” Working Paper No. 41, Department of Economics, Princeton University. Toninelli, Pierangelo. 2000. “The Rise and Fall of Public Enterprise.” In The Rise and Fall of State-​Owned Enterprise in the Western World, edited by Pierangelo Toninelli, 3–​22. Cambridge: Cambridge University Press.

The Rise and Fall of State Enterprises    717 Toninelli, Pierangelo. 2008. “From Private to Public to Private Again: A Long-​Term Perspective on Nationalization.” Análise Social 43 (189): 675–​692. Werneck, R. 1987. Empresas estatais e política macroeconômica, 124. Rio de Janeiro: Ed. Campus. World Bank. 1989. “Brazil:  Prospects for Privatization.” Report No. 7550. Available at http://​ documents.worldbank.org/​curated/​en/​377441468239129268/​Brazil-​Prospects-​for-​privatization.

Chapter 34

Antitrust a nd C om petition P ol i c y in Bra z i l Eduardo Pontual Ribeiro, Camila Pires-​A lves, and Luis Carlos D. Prado

34.1. Introduction In the last quarter of the twentieth century, antitrust legislation spread beyond the few developed countries that already had such laws by the end of World War II. From the 1990s onward, large numbers of developing countries have adopted antitrust laws.1 Competition is a dynamic concept concerning the way that economic rivalry operates and imposes the market pressure that leads economic agents either to respond by increasing efficiency or to suffer (economic) penalties for their failure to do so. But competition can also reduce the number of players in the market, leading to greater market power; moreover, as in the case of market power, collusion reduces economic efficiency and consumer welfare (Motta 2004). Thus, competition’s capacity to trigger the right stimuli for entrepreneurs depends on its operating within the right institutional environment. Competition policy is the set of public policies that aim to promote just such an institutional environment, where competition produces the right economic results—​ that is, economic growth, technological development, increasing efficiency, and consumer welfare. While antitrust policies—​ focused on fighting cartels and controlling market concentration—​are considered a subset of competition policy (Fox 2007), we follow the convention, as in Europe and in the United Kingdom, of using them as synonyms. The aspect of competition policy relating to the regulation of markets is not discussed in this chapter.

Antitrust and Competition Policy in Brazil    719 The normative goal of competition policy as considered in this chapter is to eliminate the threat of “significant impediment to effective competition” (using the European standard), or of a significant reduction or elimination of competition from business acts or practices. In many markets, competition may not be sufficient to reach socially desirable goals, due to technical or informational factors. Competition or antitrust laws do not address these market characteristics and do not provide continuous market regulation. Competition policy provides a way to affect the behavior of market players in two situations: (1) preventing mergers that might harm the competition environment through an excessive increase in market power (mergers that increase market power generate lower incentives for players to benefit consumers through lower prices, larger quantities, better products, and lower costs); and (2) punishing the economic behavior of market players who obstruct the positive effects of competition, such as in the case of cartels or other forms of business practice that amount to an abuse of market power. Put another way, competition policy does not seek to control prices and quantity directly and is not a substitute for classical regulation as in markets subject to natural monopolies. The timeliness of the two main actions mentioned in the previous paragraph differs. Respectively, (1) it is localized in time and reactive, called into action only in the case of a specific business act that may increase market concentration, namely mergers and or acquisitions in their many forms, and interfering only when the increase of market concentration may lead to welfare losses; and (2) it refers to a permanent vigilance against business practices that protect firms from competition, such as cartels or exclusionary practices (Motta 2004). A secondary path to action is competition advocacy, where rules and regulations receive an opinion on whether the intended or unintended consequences will be to enhance or diminish competition (Clark 2005). Developing countries are as hurt by abuses of market power as much as developed economies are. One can find at least as many examples of cartels, anti-​competitive behavior, and predatory market practices in developing countries as in developed ones. Therefore, independent antitrust authorities are important in order to promote an economic environment that can facilitate development-​promoting policies. International cartels are also very harmful to developing countries, and a less open market often creates opportunities for the concentration of economic power in ways that reduce competition and decrease consumer welfare. In Brazil, as in other developing economies, the creation of legislation and the setting up of independent authorities with the skills and strength to enforce antitrust policies can be seen as a significant improvement to the domestic institutional environment (Fox and Sokol 2009). The actual design of competition policy in general and antitrust policy in particular has been a key topic in the discussion in the literature across both academia and multilateral institutions (e.g., Jenny 2016; OECD 2007). This chapter considers antitrust policy in Brazil as a case of institution building in a developing economy. We begin with a historical account of the legislation that

720   Eduardo Pontual Ribeiro et al. ultimately became the modern antitrust legislation; the legal framework did not appear in a vacuum, but rather can be traced to a somewhat vague but nevertheless significant code of law. We proceed by describing the current legal framework (Law 12.520/​2011), itself an evolution of the first modern act (Law 8.884/​94). The current legal framework and the institutional design is a necessary but, certainly, not a sufficient condition for policy. The final section describes the capacity building that generated an internationally acclaimed policy, on a par with developed countries’ agencies. The central thread of the analysis is the role of the main competition policy agent, the Administrative Council for Economic Defense (Conselho Administrativo de Defesa Econômica [CADE]), a 50-​ year-​old agency applying a 20-​year-​old policy. The case of competition policy in Brazil may also be relevant to other developing countries that are interested in competition policy. Its experience highlights the legal framework, design, and capacity building required for policy legitimacy and practice.

34.2.  Historical Evolution of Early Competition and Antitrust Policy in Brazil It was only in the 1990s that modern antitrust legislation was passed in Brazil. Its inception was accompanied by economic changes such as the deregulation of the economy, phasing out of price controls in many sectors, the sale of state-​owned enterprises (SOE), and the breakup of SOE monopolies. Last but not least, its passing coincided with price stabilization after decades of high inflation and particularly the hyperinflation of 1987–​ 1994 (Baer 2014). Nevertheless, such legislation and policy goals were not previously unheard of in Brazilian law; this section highlights this historical evolution.

34.2.1. Conselho Administrativo de Defesa Econômica before Law 8.884/​94 Beginning in the 1930s, there were a number of failed attempts to approve laws against trusts in Brazil. At that time, large foreign companies, often with market power or monopolies and the ability to organize cartels in the country, were called trusts. Following the 1937 Constitution, with its concept of economic crimes against the state, Decree 869/​1938, issued under the dictatorship of the Estado Novo, declared that mergers or acts that aim to eliminate competition would be unlawful as a crime against “the people’s economy” (economia popular). Subsequently, the first law in Brazil to deal with antitrust issues in a more structured way was the Decree-​Law n.7.666, issued in 1945, some months before the end of the Estado Novo and the deposition of Getúlio

Antitrust and Competition Policy in Brazil    721 Vargas. This law authorized the government to impose sanctions on domestic and foreign firms that had links to or that organized trusts and cartels. Filled with nationalistic rhetoric, the law had some features that would be used in future antitrust laws in Brazil. This law suffered strong opposition from the press and the entrepreneurial classes, who considered it excessively interventionist and influenced too much by left-​wing ideas.2 The US Department of State saw the law as an act of economic nationalism and pressed the Vargas government to repeal it.3 The Act was actually repealed by Provisional President José Linhares, a few days after the deposition of Vargas. However, supporters of a Brazilian antitrust law were able shortly afterward to approve the article (148) in the Brazilian Constitution of 1946 that states, The law will repress any form of economic abuse, including joint ventures or groups of individual or social enterprises, whatever their nature, which are designed to dominate the domestic markets, eliminate competition and arbitrarily increase profits.

The enforcement of this article hinged on a law to be passed. This motivated Bill 122 of 1948 for a Brazilian antitrust law. However, this project was not well received in Congress and was never approved. Only in the 1960s did debate on the creation of an antitrust law in Brazil resurface. In September 1962, Article 148 of the Brazilian Constitution was bolstered with the approval of Law 4.137/​1962. This law created the CADE. This law was approved in a period of political instability and was never thoroughly enforced. Repression of cartels at the time was viewed as unfriendly to foreign companies and large local businesses; therefore, there was strong resistance to its implementation. Moreover, it was not a very well designed law. It had a strongly interventionist bias, and its objectives and instruments were not comparable to a modern antitrust legal system. During the 21 years of the military regime in Brazil (1964–​1985) there was little or no debate surrounding antitrust policy or efforts to actually implement it. The government had a preference for direct intervention either in the form of price controls, direct market regulation, or state monopolies, rather than creating a competitive economic environment in Brazil (Baer 2013). The government fostered and consolidated a large industries market division among three groups of ownership: large domestic private companies, foreign private companies, and SOEs. Within this environment there was little room for competition policies or antitrust laws. Prices were directly controlled by the SUNAB, the National Supply Superintendence (Superintedência Nacional de Abastecimento) and later by CIP, the Interministerial Price Council (Conselho Interministerial de Preço). In the late 1980s, during the First Republic, some new antitrust regulation was issued through Decree 92.323/​1986. Using wording very similar to the 1946 Constitution, and parts of the 1962 law, the decree outlined in more detail economic felonies in an administrative law related to the abuse of market power (and elimination of competition, or “arbitrary” profit increases). It also stipulated the amounts of fines and penalties. The end of the 1980s saw the creation of an important legal basis for the creation of modern antitrust legislation in Brazil: in the Constitution of 1988, Article 173, paragraph 4 states that the Law should repress “the abuse of market power that aims to the control

722   Eduardo Pontual Ribeiro et al. of markets, the elimination of competition and the arbitrary increase of profits.” At the same time, the Constitution affirmed the role of free enterprise and the social role of capital as the bedrock for Brazil’s economic organization. At the beginning of the 1990s, during the Collor government, Law 88.158/​1991 was passed, creating the Secretary of Economic Law and modernizing some articles of Law 4.137/​1962. Even so, the CADE was not yet a modern antitrust agency, and it did not have the institutional tools to enforce the antitrust law. From February 1992, the president of CADE, Ruy Coutinho, was central to the creation of a working group for the elaboration of a modern antitrust law. However, the elaboration of this new law faced strong opposition from Brazilian entrepreneurs’ organizations, such as CNI and FIESP (Dutra 2012).

34.2.2. Modern Brazilian Competition Policy under Law 8.884/​94 Law 8.884/​94, the first modern Brazilian antitrust law, was issued on June 11, 1994, at the time of the elaboration of the stabilization plan (Real Plan). It was originally intended to create the institutional environment to moderate prices in Brazil, without the use of the direct control mechanisms used since the 1960s. At that time, lawyers were directly involved in the design of the law. It was the decisive influence of Finance Minister Rubens Ricúpero that increased the role of economists in the operation of the system, and also that led to an economist, Gesner de Oliveira, being installed as the first president of the new CADE, under Law 8.884/​94.4 As such, by the mid-​1990s Brazil finally was able to create a modern antitrust system, after five decades of failed attempts. Law 8.884/​94 also organized the Brazilian Antitrust System (SBDC) and the agencies associated with it. The SBDC was composed of the Secretary of Economic Law (SDE—​ Secretaria de Direito Econômico, under the Ministry of Justice), the SEAE (Secretaria de Acompanhamento Econômico, under the Ministry of Finance), and CADE. The first two institutions had investigative roles relating to abuse of dominance, and cartels and mergers, respectively; CADE was the ruling tribunal within the system. Jenny (2016) provides a comparative analysis of the institutional design of competition authorities across the world. Using his framework as a synthesis, we can say that Law 8.884/​94 stated as its goal the suppression of acts that could reduce competition significantly, through mergers or business practices. The rulings were made under administrative law, by executive branch public officials outside of the judicial system. Nevertheless, the decisions could be challenged in the courts. The functions of SBDC in Brazil were on competition only, with no role in direct consumer protection (a separate code, the Código de Defesa do Consumidor, from 1990, oversaw this), and no role in sector regulation. In terms of organizational structure, as mentioned, investigation and rulings were the responsibilities of separate agencies. The investigative agencies had single decision-​making, while CADE, the ruling agency, deliberated in a council of six commissioners plus the president. CADE deliberated on all cases, whether merger

Antitrust and Competition Policy in Brazil    723 cases or formal investigations (SDE could independently drop a case very early on). The instruments available for CADE depended on whether the case was one of a merger, or of a cartel or business conduct (abuse of dominance). With mergers, CADE could declare the contract that regulated a merger null, or change specific clauses of contracts, or approve the merger conditionally, requiring certain steps to be taken, such as the sale of assets. With cartel/​abuse of dominance cases, CADE could impose fines (proportional to a firm’s revenues), for both firm and individuals, and ancillary penalties, such as recommending that the firm be barred from participating in public procurement. Last but not least, while at the investigative bodies decision-​making was in the hands of an agent who could be dismissed immediately by the president or the ministers of justice or finance, CADE’s commissioners and president had a fixed-​term two-​year mandate appointed by the president of Brazil and subject to approval hearings by the Senate. The mandate and the new structure afforded the autonomy necessary for antitrust policy implementation. Related to independence, CADE’s budget is negotiated with the justice minister. It does obtain funds from merger case-​handling fees, but fines levied do not revert to CADE. When Law 8.884/​94 was enacted, it was the most complete, comprehensive, and effective protection law to that date governing Brazilian competition. However, Brazil had little experience in competition policy. It was fiercely opposed by some economic experts,5 who claimed that market liberalization would suffice. At the same time, the application of antitrust policy and legal text needed to be interpreted and matured through case examples and administrative tradition. Businesses had to understand and internalize the incentives and penalties the antitrust law provided. The SBDC had also to learn how to provide credible, coherent analysis in both mergers and abuse of dominance cases. With mergers, analysis is mainly economic, with objectively measured concentration indices and a list of market contestability indicators. With abuse of dominance and cartel cases, analysis deals with both economic metrics and actions, behaviors, and so on. The first task is learning how to obtain such data, and the second task how to coherently analyze it. Nevertheless, the law’s wording, with its adjectives (“abuse,” “unjustified,” “excessive”), requires a degree of analysis (“too much concentration,” “very low contestability,” and so on, instead of simple binary terms). The gauging of what is reasonable concentration and competition and what is unlawful involves somewhat qualitative decision-​making that may be deemed arbitrary and unreasonable in courts. An important job for competition policy is to specify a methodology for rulings that is coherent, but also adaptable to market-​ specific circumstances. Over 17 years, that enterprise has been contributed to, in something of a “learning by doing” process, by the decisions of CADE, books and articles by various jurists and economists, and also the arguments brought to court by lawyers and consultants in many controversial cases. This is described in more detail in the following sections. At the same time, through the 1990s, economy-​wide conditions were ripe for the application of a competition policy. The easing of capital controls and foreign investment, the end of explicit price setting by the government in many sectors (such as energy, food,

724   Eduardo Pontual Ribeiro et al. and to some extent health), the sale of SOEs, and the end of state monopolies all helped to foster competition and the relevance of competition policy.6 Deregulation and opening of the economy may be considered a precondition for an effective competition agency. Many developing countries face significant challenges on these issues due to thin or small markets (where few firms can be profitable) and markets with histories of state monopolies. Even when privatized, if competition is not brought in or a sector is not directly regulated, competition agencies face repeated claims of abusive practices relating to such monopolies. Relative to other countries, such as Russia or India, this was a comparatively mild issue for Brazil. Anti-​competitive behavior of former state monopolies in non-​competitive environments has decreased over time, although it has not entirely disappeared (Cowie and Mattos 2005; Fox and Sokol 2009; Oliveira and Furakawa 2005). At the same time, the need to actively fostering competition after the market liberalization of the 1990s can be seen in the possibility that domestic market concentration actually surprisingly increased, as suggested by Amman and Baer (2008). Law 8.884/​94 inaugurated modern practice and analysis regarding the control of mergers and abusive practices. Firms with a large market share (above 20%) or with revenues over a certain threshold were required to file a merger for clearance at SBDC after the transaction was signed. Merger analysis was ex post. If a merger was deemed anti-​competitive, it would be blocked and reversed, or conditions would be imposed before approving it, such as the sale of part of the merged company. With cartels, administrative sanctions included fines as a proportion of revenues. In something of a novelty, the law also presented a list of business practices that would be considered anti-​ competitive and subject to fines, such as predatory pricing and tying, among others. Some of these were present in the 1960s legislation discussed earlier, and followed very closely Decree 92.323/​1986. Law 8.884/​94 generated an organized legal infrastructure for prosecution by administrative law. The law stated that CADE’s rulings could be appealed only in courts. CADE’s decisions were unrelated also to civil (private damages litigation) or criminal cases (cartels are a criminal offense). Important merger cases have been decided under Law 8.884/​94, such as Nestle-​ Garoto (2004) and Brasil Foods (2011), with recommendations of blockage and structural restrictions, respectively. The first cartel cases were seen as somewhat of a surprise, given the history of price controls by the government. In one case, still under dispute in the courts, a cartel of steel makers had, as a main piece of evidence, a joint meeting between the four largest steel makers in Brazil and SEAE representatives. Before Law 8.884, such meetings were required to arrange annual price increases in steel products. However, after Law 8.884, since the meeting took place in 1996, it was deemed a cartel.7 One characteristic of the 1994 Law was the use of different institutions on the same policy (even different ministries, such as the finance ministry [SEAE] and the justice ministry [SDE], as well as a more independent agency [CADE]). Early on, the finance and the justice ministries had to coordinate to nominate commissioners and the CADE presidency. The separation between investigative powers and ruling decision across authorities was deemed necessary to provide checks and balances and dilute power.

Antitrust and Competition Policy in Brazil    725 But early on, the overlapping of efforts proved inefficient with both SDE and SEAE providing opinions on merger cases, increasing ruling times. In 2006 an agreement (Portaria SEAE SDE 33/​2006) was reached between the finance and justice ministries to the effect that SDE would follow SEAE’s opinion on merger cases and, in practice, SEAE delegated anti-​competitive conduct evaluation to SDE. Cooperation and trust were required to reach that agreement. Despite its important role in modernizing Brazilian competition policy, Law 8.884/​94 was seen to have significant limitations, as discussed by competition authorities themselves (e.g., peer reviews jointly carried out by IDB and OECD [2005, 2010]). Some of those limitations were addressed by Law 12.529/​11, such as the overlapping institutional structure and the post-​merger review analysis, among other advances.

34.3.  Competition Policy in Brazil under Law 12.529/​11 The current Brazilian competition law, Law 12.529/​11, came into effect in May 2012, during the first mandate of President Dilma Rousseff. Continuing with Jenny’s (2016) analytical structure, used earlier in describing the competition policy system in Brazil under Law 8.884/​94, the new law did not alter the goal, type, sanctions, or functions of the competition system. It did significantly alter the organizational structure and deepen the independent status of competition policy actors. The Secretary for Economic Accompaniment (SEAE) of the Ministry of Finance, as cited before, retained its competition advocacy role, providing economic opinions on matters relevant to the promotion of competition (law, rules, regulation acts, tariff revision, etc.), but was not responsible for investigations into mergers (or cartels). SDE also was stripped of its investigative role and was removed from the SBDC. Both investigative and decision-​making roles were concentrated in a new and reorganized CADE.

34.3.1.  Institutional Design of Competition Policy in Brazil According to Law 12.529/​11, CADE is composed of three main bodies:  the General Superintendence (SG), the Administrative Tribunal (Tribunal), and the Department of Economic Studies (DEE). There is also the supportive role of the attorney general’s office (Procuradoria Geral—​ProCADE). In general, the SG has a principal investigating role within CADE, while the Tribunal is responsible for final administrative decisions. The SG is composed of eight merger and antitrust units, grouped according to their activities and relevant market specificities, organized under two deputy superintendents

726   Eduardo Pontual Ribeiro et al. and one general superintendent. The SG’s main responsibilities are as follows: (1) the uncovering and investigation of abusive practices; (2)  the analysis of merger cases; (3)  the negotiation of remedies agreements in merger cases (ACC—​Merger Control Agreement), in anti-​competitive practices cases (TCC—​Cease and Desist Agreement), and in leniency agreements. With mergers, the SG may approve a case, close it, or refer it to the Tribunal if not approved. The SG works with a two-​tier system for analysis; if deemed simple (simplificado) in a first pass (see conditions in the following), the merger case is approved in a few weeks; if deemed regular (ordinary) or complex, the Superintendence may take longer to present its opinions, requesting further information from parties and/​or competitors. The time frame for ordinary or complex cases may reach more than six months. On anti-​competitive practices, the SG has investigatory powers with a significant degree of independence from the Tribunal, being able to close administrative inquiries regarding anti-​competitive practices and to decide whether to close the case or to send it to the Tribunal for sanctions. As in Law 8.884/​94, six commissioners and one president make up the Tribunal. The Tribunal’s duties include the following: (1) issuing a final decision on anti-​competitive practices and merger cases; (2)  the approval of ACC and TCC agreements; and (3) evaluating appeals against preventive measures issued by the SG or by a commissioner. The usual procedure is that the Tribunal is called into action by a negative (no clearing) decision in a CADE by the SG.8 The Department of Economics Studies (DEE) is run by the chief economist and provides technical assistance for the general superintendence and the Tribunal in the investigation and analysis of mergers and anti-​competitive conducts. The DEE may present studies on the evolution of specific markets (such as gasoline and higher education) and on the effects of CADE’s decisions in certain markets. The creation of DEE contributed to the quality and skill in CADE’s decisions, by providing support on economic analysis. The attorney general’s office provides legal advice on cases and represents CADE at the courts in typically two instances: when CADE seeks a search warrant, or when defending CADE’s decision in a judicial appeal by firms. The new law enhanced the independence of the agency compared to the previous organization of SBDC. In addition to the president of CADE and commissioners in the Tribunal, the investigative body’s head (general superintendent) and the legal advisory head (CADE general attorney) are chosen by the president of the Republic and are approved by the Senate for a fixed two-​year term, renewable for the same length of time. The new law did not alter the funding for CADE, which depends on the Ministry of Justice budget. Regarding external oversight, Law 12.529/​11 dictates that a member of the Federal Public Prosecutor (the equivalent of the District Attorney in the US) is to be heard with regard to cartel and abuse of dominance cases. Cartels are a penal crime under Brazilian law, and such cases would be brought by the Federal Public Prosecutor to the judiciary.

Antitrust and Competition Policy in Brazil    727 The opportunity to hear (but not vote) on anti-​competitive conduct was seen as important in order to strengthen ties with officials that were more experienced in investigative procedures. The institutional reorganization provided for by Law 12.529/​11 contributed to the reduction of cases analyzed by the Tribunal. As a result, there has been both an empowerment and a rationalization of CADE’s internal structure, in which the General Superintendence, alone, deals with the majority of cases, whereas the Tribunal, once overloaded with simple but numerous cases, is now able to focus on complex cases that would have a significant impact on the economy and to create relevant jurisprudence. Such organizational changes were part of the recommendations of the joint peer reviews by the Inter-​American Development Bank (IDB) and the Organisation for Economic Co-​operation and Development (OECD) (2005, 2010), which were concerned with the vesting of too much power in the SG (investigating and approval ruling). Thus far, as in the examples of arrogation by the Tribunal, scrutiny on SG decisions seems active.

34.3.2. Competition Policy Practice: Abuse of Dominance or Anti-​Competitive Behavior Brazilian competition law defines an (administrative) infraction as firm or individual actions that may (1) limit, restrain, or in any way damage competition; (2) lead to control of a relevant market; (3) arbitrarily increase profits; or (4) be characterized as abuse of a dominant position. In order to exemplify the applicable practices that may be considered abusive, the law contains a non-​exhaustive list of potentially anti-​competitive or abusive practices. Among the potentially abusive practices listed in Law 12.529/​11, art. 36, we can identify (1) acting collusively; (2) limiting the access of or producing barriers to entry for rivals or suppliers; (3) tying sales; (4) imposing retail conditions (price, quantities, etc.); (5) pricing discrimination; (6) predatory pricing; and (7) abuse of intellectual or technological property rights or copyrights. The list does not explicitly mention exclusive dealing. A clear change in stance from the previous law was the exclusion of “abusive prices” as a competition infringement. The imposition of penalties in the form of fines may vary from 0.1% to 20% of last year’s firm or conglomerate revenues. The law considered imposing fines on executives and individuals, generating personal responsibility for firm acts. Other forms of penalty may be imposed, such as prohibition from taking loans with official financial institutions or participating in public procurement, and obligations on property rights, such as compulsory licensing of intellectual property rights, or selling assets, among other measures.

728   Eduardo Pontual Ribeiro et al.

34.3.3.  Competition Policy Practice: Merger Control Law 12.529/​11 inaugurated pre-​merger review for merger filings. A  merger (in the varying forms of horizontal or vertical arrangements)9 must be submitted for CADE’s approval when it involves one party with at least R$750.000.000 (approximately US$250 million with a USD/​BRL 3,00 exchange rate) in revenues (turnover) and another party with revenues of R$75.000.000 (US$25 million) or higher. This new higher threshold has contributed to a decrease in the of number of cases analyzed by the agency. Article 88 of the law states that “mergers, in any form, that lead to the elimination of competition in a substantial part of the relevant market, or that may create or reinforce a dominant position, or may result in relevant market dominance, will be blocked.” Two concepts are key for analysis: “market dominance,” understood in the law as the capability of a large firm to unilaterally affect market conditions, and a “relevant market,” namely, the collection of actual competitors of a firm (for an economic analysis of these concepts see Motta 2004). In general, mergers that increase market concentration significantly and to very high levels, in markets with no evidence of contestability, will be blocked or subject to merger remedies to counter the losses to consumers. There are waivers if the merger is able to generate significant and unique economic efficiency, such as (1) productivity or competitiveness growth; (2) improvements to the quality of goods and services; and (3) remarkable technological or economic development. The law states that a major part of the claimed efficiencies must reach consumers. The pre-​merger notification procedure, while placing mergers in a state of suspension after deals are signed, imposes upon CADE a maximum of 330 calendar days (240, with a possible extension of 90 days) for its final decision. Streamlining merger analysis, the SG classifies cases into simple and ordinary (regular) cases. Simple cases are those with little potential of anti-​competitive impact such as (1) market entry with no overlap between parties; (2) joint market share less than 20% (the reference for a dominant position, according to the Law 12.529/​11) for horizontal mergers and 30% for vertical mergers; (3) low concentration effects based on Herfindahl–​Hirschman index (HHI) index variation; and (4) classic or cooperative joint ventures (to act in nonvertically or nonhorizontally related markets). Simple cases are solved in a short period of time, recently specified as a 30-​day deadline.10 The SG’s statement about a case is sent to the Tribunal only in nonsimple cases, when the recommendation is to block the merger or approve with remedies or conditions. In these cases, a designated commissioner presents the case to the Tribunal for judgment and the final decision is made by CADE. In all other situations (i.e., either a simple case, or an ordinary case that is approved without restriction), the SG’s decision is final unless a commissioner arrogates the case.

34.3.4.  CADE in Figures CADE’s development in recent years has consolidated and internalized the changes generated by the new antitrust law, as the relevant numbers show.

Antitrust and Competition Policy in Brazil    729 Considering merger control, there has been a significant reduction in notified cases. Between 2008 and 2011, the average number of notified mergers per year was 615 cases. Between 2013 and 2015, when all cases were under the new law, the average number of received cases by the SG was 402 cases.11 However, there was a relative increase in blocked or restricted cases, suggesting that earlier mergers above the previous notification threshold and the new threshold were mostly not anti-​competitive, and drained resources from the SBDC. During 2015, 386 merger cases were analyzed by CADE; 376 were approved without restrictions, seven were approved with restrictions, and one was rejected by the Tribunal. Among the total of mergers approved without restrictions, 368 were reviewed solely by the General Superintendence, while eight were decided by the Tribunal. These figures have been stable over time; no more than three mergers per year were blocked between 2008 and 2016. The average time length for merger analysis in the SG was 18 days for simple cases and 82 days for ordinary (or complex) cases (CADE 2015, 70–​73). For comparison, in 2010, for example, before the new law, the average time of a merger case at SBDC (investigation at SEAE and ruling at CADE) was 156 days. In 2015, it was 28 days, showing an important increase in capacity in processing cases (CADE 2016). Regarding anti-​competitive practices analyses, the number of cases judged and the number of cases with penalties increased, showing significant efforts from the SBDC to punish competitive misbehavior among firms. In 2015, 52 cases involving anti-​competitive practices were decided by the Tribunal, and three-​quarters of those resulted in conviction. In comparison, in 2010 there were 18 cases judged, four of which resulted in fines. Over the last five years, mainly under the new Law 12.529/​11, there has been a significant increase in the number of Cease and Desist Agreements (from 11 in 2010 to 58 in 2015)  signed with CADE. The same has occurred with Leniency Agreements, since in 2012–​2015, 38 agreements were signed, whereas the four years immediately prior to the new law saw only half as many (CADE 2016). Inevitably, there has been a significant reduction in the case stock and a persistent increase in fines collected by CADE (from R$23 million in 2010 to R$524 million in 2015).12 Rulings are generally sustained when contested in courts. According to CADE (2016), approximately 72% of rulings were not overturned or altered by judicial review in the 2012–​2016 period.

34.4.  Capacity Building and the Practice of Competition Policy Given the legal infrastructure for competition and antitrust policy in Brazil, current practice reflects a more than 20-​year history of capacity building and institutional evolution. This section looks at different dimensions of this evolution, such as the transition to the current law, international support, human resources, the creation of the Chief Economist’s Office (DEE), and investigative tools in nonjudicial investigations.

730   Eduardo Pontual Ribeiro et al. Over time, the SBDC has made choices in its policy emphasis aimed at building stronger institutions. First, from 1994 to 2003, when the SBDC faced skill shortages, distrust from the business community, and a lack of investigative tools, the emphasis was on merger control (Martinez and Araújo 2013; see also testimonies in Dutra 2012). Early reversals or ineffectual outcomes of rulings in landmark cases such as Antarctica-​ Brahma (beer and beverages), Kolynos-​ Colgate (toothpaste), and Nestlé-​ Garoto (chocolates) led to procedure revisions within SBDC on merger cases. Guidelines were issued in the meantime (2001) by SEAE/​SDE, and rationalization of case handling by SEAE and SDE, as mentioned earlier, proved key to improving rulings. Second, from 2003 to 2012, there were clear improvements in the system’s investigative capacity, as well as in the use of agreements for sanctions required to approve mergers or to convict cartelists.

34.4.1.  The Transition to the Current Law (12.529/​11) After more than five years, the competition law change was signed by President Rousseff, following a vote in Congress in late 2011. As mentioned earlier, it significantly changed the SBDC and competition practice. It consolidated investigative and ruling dimensions of competition policy into a single entity (CADE), composed now of a General Superintendence (SG) for investigation and a Tribunal for ruling. Ruling would be on noncleared merger cases or recommended fines for cartels or abuse of dominance cases only, instead of all cases. Merger control became ex-​ante (pre-​merger notification system), with a limited time for rulings. The challenges of transition were significant; there was suspicion that the ex-​ante control would keep mergers in a state of suspension for a long time, increasing their financial and economic cost, and that integrating different entities and personnel, such as SEAE (merger analysis), SDE (cartel and abuse of dominance investigation), and CADE itself (ruling) would lead to corporate culture clashes and would delay the system. Furthermore, there were the infra-​legal details to be worked out in CADE’s Regimento. However, to many the transition was remarkably successful. Within a month, the new CADE integrated SDE, SEAE, and Tribunal personnel in a new, larger building. The fast-​track classification seemed to work. In 2012 CADE ruled on 731 merger cases and 69 cartel or abuse of dominance cases,13 the highest figure ever. According to Frade (2015), in 2011, before the new law, a decision on a merger case would take 155 days on average (see also section 3.4). After the new law, with the possibility of classifying the cases as either simple or ordinary/​complex, average decision-​time length reduced to a little over 200 days in complex cases (often with restrictions or even complete blocking), 65 days in ordinary cases and less than 30 days for fast-​track cases. Simple cases account for nearly 85% of merger cases presented. As seen in section 3.4, these figures for 2012–​2013 are still representative of the current time lengths. There were also various decisions of an infra-​legal nature that needed to be made, such as merger notification criteria according to firm size; dealing with non-​merger

Antitrust and Competition Policy in Brazil    731 firm contracts (contrato associativo); minority shareholdings; leniency program criteria. In around a year, most of these issues were tackled in decisions, providing fluidity for firms and their advisors. Some of these issues have been improved (as in the case of firm contracts that require CADE evaluation), but in the 2012–​2016 period all sensitive issues were addressed one way or another.

34.4.2.  Legitimacy Building through Transparency Since its early stages, CADE’s legitimacy and policy practice has been built on a transparency principle. All case materials are given extensive public access, except for commercially sensitive data. Generally, apart from court case materials (as in the United States), only summaries of decisions are publicized internationally. Tribunal sessions are transmitted live via the web, and historical voice recordings are available. All parties meeting with CADE’s officials are featured in a publicly available agenda. Transparency is enhanced by the large number of guidelines published. These guidelines provide a summary of antitrust practice at CADE. Although they may not be strictly binding, they inform relevant parties about how the analysis will be carried out. After 2012, with the new law, one can read the Horizontal Mergers Guideline (replacing the previous 2001 SEAE/​SDE guide), and Gun-​jumping, Compliance, Leniency, and Cease and Desist Agreements. Many of those are also available in English.14 Publishing guides is an approach not without its critics. It is said that it creates binding opinions and reduces the scope for flexibility in making legal interpretation changes in specific cases, and/​or in response to market developments. On the other hand, making clear how CADE sees itself reduces regulatory uncertainty, and, perhaps more important, facilitates “entrants” into the practice of competition policy advising. Teaching in universities and the training of non-​officials (law and economic consulting firms) would be much more difficult without CADE’s transparency. This transparency may also help to reduce the risk of regulatory capture, as decision-​ making rationales are explicit. Other aspects that may reduce chances of regulatory capture are that CADE does not deal with a specific sector (making firm-​specific “investments” costly in relation to potential benefits) and that in the past 10 years very few appointed commissioners have been lawyers or economists actively defending firms’ interests before appointment, thus avoiding one side of the revolving door issue that weakens regulatory strictness (Dal Bó 2006). Since 2006 most commissioners have been either scholars from universities15 or public officials. On the other side of the revolving door issue, it is also uncommon for former commissioners to go on to represent firms on mergers or cartels full-​time, although this has been the case for some middle-​ to high-​level members of SBDC. It is also worth pointing out the fact that a given firm may have an interest in CADE being “lenient” when it is proposing a merger, but, conversely might wish for a “strict” CADE in the event of mergers among competitors; as such, there is perhaps a natural balance in that it is not clear that it would be in firms’ interests to sway CADE in any particular direction, even if they were able to.

732   Eduardo Pontual Ribeiro et al.

34.4.3.  International Support and Capacity Building Brazilian competition policy actors from governmental institutions have participated extensively in international fora. Participation in international fora represents an opportunity for CADE in the area of institutional building. First, learning from positive and negative experiences abroad is a source for improvements in competition law practice. Second, CADE seeks legitimacy by presenting itself as on a par with international competition agencies. Third, it provides a platform to aid smaller competition agencies in the region. CADE’s participation has included the IDB-​OECD Competition Committee (as a guest), the International Competition Network (ICN), and UNCTAD’s Competition Division, as well as, more recently, the IDB/​World Bank-​sponsored Centro Regional de Competencia para America Latina. The OECD Competition Committee publishes documents on competition themes (minority shareholdings, vertical practices, horizontal agreements) and invites countries to send a summary of their experience on the matter. This provides an interesting exposure opportunity and involves cross-​country comparisons. Brazil is regularly invited to share its experiences and opinions. Peer reviews conducted in 2005 and 2010 (IDB and OECD 2005, 2010) originated from participation in the IDB-​OECD Latin American Competition Forum. A  foreign consultant was sent to Brazil for in-​depth interviews on and investigation into the workings and practices of competition law. This was a more detailed analysis than an earlier evaluation report in 2000. The second peer review followed up on the developments from the previous review. The recommendations were mostly integrated in the 2011 law review, since, as mentioned earlier, an earlier version of the law had been discussed since the mid-​2000s. The peer review acted as a reinforcing push for those aiming to bolster the role of competition policy (then in the form of the SBDC) in the economy and to increase its efficiency. International reviews of CADE’s workings are carried out yearly by the OECD Competition Department,16 where the most significant cases since 2010 in mergers and cartel/​abuse of dominance are summarized. The SBDC agencies have participated in the ICN since its inception. Both CADE and SEAE participate in meetings and working groups. Indeed, Brazilian agencies’ increased depth of involvement led to the hosting of the 2012 meeting in Rio de Janeiro, and a Brazilian vice presidency in 2013. CADE participates in a world ranking of competition agencies, sponsored by news service/​publishing house Global Competition Review.17 Each improvement in the rankings is extensively disseminated, both internally and externally, in an effort to show that “we’re on the right track” and to boost morale. CADE’s rise in the rankings has been impressive, from three stars to four stars in less than a decade, and now residing among the top 10 agencies according to this enforcement reputation ranking. The Brazilian agency was behind only those of France, Germany, the United States, the European Union, Japan, and South Korea in the 2015 results.

Antitrust and Competition Policy in Brazil    733 Last but not least, participation in international fora has led to closer ties with other jurisdictions for the purposes of international mergers and cartels. Handling international cases (e.g., a merger of two multinationals that have activities in and are registered in Brazil, almost always with plants) includes dealing with a number of issues. On mergers, while the evaluation of impacts is clearly independent across countries, in the case of remedies the timing and design of remedies may involve cooperation across jurisdictions. For instance, in a number of cases the remedies agreed with CADE included commitments abroad that were also agreed with jurisdictions abroad; these included the cases of Munksjö AB/​Ahlstrom Corporation; Synverse Holdings, Inc./​ WP Roaming III S.À. R.L. (MACH); and Continental AG/​Veyance. On cartel cases, there may be legal exchanges of information, particularly from those granted leniency. This raises significant legal matters that need to be addressed (at the very least, the documents’ languages, e.g., Portuguese and English), including commercially sensitive information across countries, and separate business entities. The ongoing contact across jurisdictions provided by personal meeting opportunities in the previously mentioned fora is key for successful action, according to CADE officials.18

34.4.4.  Human Resources and Capacity Building One important aspect of an agency is capacity building for staff. This was central to the consolidation of CADE in the 2000s, at least compared to SEAE. Repeatedly, the OECD peer reviews and the testimonies in Dutra (2012) pointed out the need to stabilize personnel in the investigative and ruling bodies of SBDC. CADE did not have a fixed staff, and, much like SEAE, used a large number of temporary contracts. In the second half of the first decade of the 2000s, the Tribunal staff (CADE’s role at the time) was stabilized through the extensive use of public managers or gestores (EPPG), an elite group of civil servants with very high wages,19 and a highly competitive selection process.20 This young (mainly in their late twenties or early thirties at that time) and capable staff, with a high learning capacity, was exposed to training in competition by seminars and also through an in-​house MBA. There was an active policy of support for short internships in major international agencies such as the US Federal Trade Commission and Department of Justice and Europe’s Directorate-​General for Competition. Staff were sent to attend training in negotiation techniques abroad and investigative techniques with the Federal Police. Together with the staff from SDE, mostly gestores as well, they are now the backbone of CADE, with leading roles, particularly in the SG. CADE has a long-​term view of building awareness on competition and generating interest in competition policy among potential future staff. To attract the interest of students in law and economics to the field of competition policy, CADE funds a short summer internship (known as “PinCADE”). Some of these interns become commissioners or work at CADE. Many current lawyers in competition were interns at one point in time. On the international front, CADE also funds a short-​term

734   Eduardo Pontual Ribeiro et al. international internship to disseminate knowledge, particularly to smaller or younger agencies. Yet human resources issues may remain a hindrance, as suggested by early analysis by IDB and OECD (2010). As of 2015, 90 people were allocated to the SG, the DEE, or with each commissioner in the Tribunal, handling cases directly or providing information for decision-​making. The institution has become more efficient over time, as previous numbers and a comparison by Armogoon and Lyons (2014) have shown, but there remains an evident gap and a need for human resources since the productivity gains from the institutional changes brought about by the new law have now been exhausted.

34.4.5.  Investigative Tools for Cartels and Antitrust Remedies Organizing or participating in a cartel has been considered a crime in Brazil since the 1940s (Decree 2.848/​1940 and, more recently, Law 8.137/​1990). The antitrust law of 1994 also included cartels as a felony against the economy, as in the Federal Constitution, and subject to administrative judgment and fines. The 2011 antitrust law changed little on cartel prosecution, but consolidated important tools for repression. The practice of cartel investigation provides another good example of institutional building for antitrust practice. Martinez and Araújo (2015) conclude that cartel investigation was not a priority in the first years of the Brazilian antitrust system. One of the reasons was the lack of appropriate investigative tools for collecting strong evidence for conviction. This proved a burden for the antitrust system. Once a cartel claim is presented to the administration by individuals or local public prosecutors, SBDC must provide a first evaluation of whether it is worth continuing with the investigations on its own. If the early evidence is very weak, the claim would be closed; if not, it takes on the investigation. The large numbers of cases are exemplified in fuel (gasoline but also LPG) cases. CADE (2015) recounts that from 1993 to October 2013, 166 cartel claims in the fuel sector were investigated and reached a final decision, with 132 (80% of claims) closed without a second round of investigations. Even of those cases that then continued, 42% ended in conviction, but 58% closed. A 2000 amendment to Law 8.884/​94 addressed this issue, allowing the SBDC (SDE in particular) to pursue judicially approved search warrants and begin a leniency program. Search warrants and dawn raids proved key in obtaining hard evidence in the very first case in which they were used (the “crushed rock” cartel), as the cartel had written rules and regulations and a system of central control for prices and quotas across members (Martinez and Araújo 2015, and case materials). A more committed and stable legal support division within CADE (and formerly SDE) contributed to the cartel-​fighting efforts, successfully obtaining permission for dawn raids from the judiciary and sustaining appeals to dismiss claims of violations of due legal process from accused cartel parties. CADE (2016) indicates that, at most, one-​ quarter of cases are partially overturned through appeals in courts.

Antitrust and Competition Policy in Brazil    735 The leniency program has been a clear success, recognized in international fora as an effective tool for antitrust practice, and by lawyers themselves (Calliari, Andreoli, and Alves dos Santos 2015). Over time it has evolved, and in 2012 it was enshrined in the law proper. Current practice is summarized in a guideline by CADE, published in Portuguese and English. A leniency program is a very important means of starting cases (although not the only one). An accused party must provide hard evidence of conspiracy to violate the antitrust law, involving itself and other companies and individuals in cases that CADE was not aware of; settlements may be used by parties in such cases. A firm seeking leniency must confess and plead guilty, cease the alleged activities immediately (whether as ringleader or not), and cooperate throughout the investigation. A fine will still be imposed, but with a two-​thirds reduction from regular fines. As a major benefit, immunity from criminal charges is granted. Only first-​time applicants are benefitted, providing the contributions to the case are deemed sufficient by the authority. Only after 2007 could CADE use settlements (TCC in the Brazilian acronym) in cases. Settlements are deals with authorities after a case is known to CADE and under investigation. The advantage of settlements is a shortening of the ruling process, saving resources for other investigations, and guaranteeing that parties will cease the activity, will not challenge CADE’s ruling, and will actually pay the fines levied. The 2013 bylaws on settlements require firms to plead guilty and contribute hard evidence to the case. CADE’s experience with leniency and settlements has influenced leniency programs elsewhere, including the now-​famous Lava Jato (“Car Wash”) corruption case surrounding Petrobrás and many senior politicians. The evolution of antitrust remedies is another example of institutional building.21 Remedies are restrictions put in place as conditions to be met in order for a merger to be approved, usually the sale of productive capacity or a brand to entrants, reducing market concentration or fostering market contestability. Until 2002, such remedies were generic (such as pledges of quality improvement) and were difficult to enforce. From 2002 to 2006, they were barely imposed, and very few mergers received restrictions from CADE. From 2006 to 2009, compliance failure or court reversals of important remedies (such as in the cases of White Martins/​Petrobrás—​ Gemini, Owens Corning/​Saint Gobain, and Sendas/​Pão de Açúcar) shifted policy toward the use of agreements between firms and CADE for remedies rulings. The use of agreements is not without controversy. As an agreement between firm and CADE, opponents claim that CADE trades off stricter, required penalties in return for a higher likelihood of compliance (but not certainty), thereby giving in to firms’ interests. Conversely, supporters argue that even the best and strictest remedy is innocuous and does not solve the competitive merger problem if unenforceable. As of 2015, the use of agreements has aligned with the international best practice with the use of trustees (third parties to implement or monitor imposed remedies) and cross-​ border remedies. This is exemplified by the issuing of remedies guidelines in 2017.

736   Eduardo Pontual Ribeiro et al.

34.5.  Concluding Remarks Competition policy is a reality in the Brazilian economy, with an active role in antitrust played by CADE, the recently reorganized antitrust authority. While competition policy is more general than antitrust, including competition advocacy and dimensions of market regulation, this chapter has described current competition policy and practice in Brazil from an antitrust perspective, in particular CADE’s roles in merger control and cases of abuse of dominance and cartels. The experience may be seen as positive. While there are no clear indicators of success, since the ultimate goal of competition policy centers on deterrence (both of anti-​ competitive mergers and cartels), the prominence of CADE in international competition policy discussion, the levels of fines collected, and the view from practitioners all suggest a positive trajectory. Internationally, CADE received the Antitrust Agency of the Americas (2015 and 2016) award from the Global Competition Review organization. This path, while based on laws that provide the basic foundations for policy action, has been highly dependent on institutional capacity building. In terms of foundations, modern antitrust policy law was not introduced in a vacuum. Indeed, some of the wording of Law 8.884/​94 and Law 12.529/​11 had been in the codes for at least 20 years prior, and some of the institutions (such as CADE) were 25 years old when the 1994 law was enacted. Policy improvements and the capacity for revision may be seen as a positive characteristic. The law itself received amendments in 2000 and 2007 and was under review as early as 2006. A major law review was sanctioned in 2011 (Law 12.529/​11) and took effect in 2012. A transition period focused on legitimacy building, by highlighting efficiency (the reduction in the duration of cases), with evidence of little sacrifice of the efficacy of sanctioning. Since 2012, more merger cases have been rejected and cartel fines collected than in the prior 15 years. This chapter has focused on the capacity building necessary to make the antitrust policy designed in the law actually work. Strategic policy emphases, role specialization across many actors (implying coordination), and human resource development have been crucial. In the human resource dimension, a stable, skilled, and young staff with the power to make decisions has proved very important over time. Meanwhile, the legitimacy-​building strategy has targeted both domestic and international dimensions. Internally, efficiency and transparency have been used as a means of building legitimacy in the transition to the new law in 2012. External legitimacy has been sought using international benchmarks. International engagement of this nature has also facilitated technology transfer. Transparency has consistently been a hallmark of investigations and rulings. An additional significant benefit of transparency is that the easy access to documents and decisions also has a human resources-​building dimension outside of the institution, in both economics and law schools, preparing the next generation of practitioners and officials.

Antitrust and Competition Policy in Brazil    737 Several important future opportunities and challenges must be faced. One is carrying out ex-​post evaluations of CADE decisions, so as to learn positive and negative lessons from previous experiences. Another is continuous human resources development. A third is understanding that the use of information systems and online sources of information is important for better case handling and external and internal transparency and jurisprudence development. The consolidation of antitrust as an established area of research and practice in economics and law degrees around the country would contribute to the improvement of competition policy. In this regard, the continued strengthening of the Department of Economic Studies and the institution’s connection with academia is highly desirable. Currently competition policy has only a minor role for advocacy—​that is, engaging civil society, the judiciary, firms, and the government in recognizing the gains from competition and how badly designed regulation can harm welfare by reducing competition. Pro-​market policy and regulation differ from pro-​business regulation, and competition policy agencies have a very important role in fostering the dialogue necessary for the design and implementation of pro-​competitive policies at all levels of government.

Notes 1. According to Fox (2007, 104) more than a hundred countries have adopted antitrust laws, a large share of those developing countries. 2. It is true that books published in Brazil on antitrust were mainly from left-​wing or nationalist authors, such as the historian Moniz Bandeira (1975); see also Mirow (1978). 3. Bandeira (1975, 3, footnote) presents a telegram of June 27, 1945, no. 26, from the Brazilian ambassador in Washington, who informs Vargas of this fact. 4. Details of the elaboration and approval of Law 8.884/​94 can be found in an interview with Rui Coutinho do Nascimento contained in Dutra (2012). 5. Mário Henrique Simonsen, former finance minister and a well-​known scholar, wrote an opinion article asserting that “O CADE atrapalha” (“CADE interferes and make it worse”) in 1997 (see, e.g., Dutra 2012). 6. For the liberalization process of the Brazilian economy in the 1990s, see Faucher, Chapter 7, and Pinheiro, Chapter 33 in this volume; for concentration in markets and discussion of specific markets, see also Baer (2008). 7. See OECD (2006). A more recent cartel conviction included a phone exchange between fined cartel members (see Processo Administrativo 08012.004472/​2000-​12, gasoline cartel in Bauru, SP, 17), who were longing for a return of the “good old days” of price control, because other petrol station owners were acting “impolitely” or were being “ungentlemanly” in engaging in price competition. 8. In the case that a Tribunal member disagrees with the clearance of a merger or cartel or abuse of dominance case by the SG, they may call for arrogation, reopening the case at the Tribunal. This is an example of checks and balances within the organization. 9. Resolution 10/​2014 defines what can be understood as associative contracts, as defined by the law, as all kinds of contracts for the horizontal cooperation of vertical risk sharing with a consequent dependence relation between the parties.

738   Eduardo Pontual Ribeiro et al. 10. Resolution n.16/​2016. 11. Statistics available in IDB and OCDE (2005, 2010) and CADE (2016). As a point of comparison, the equivalent American agencies received 1,663 merger cases through pre-​ merger notification in 2014, according to data available at www.ftc.gov.br. 12. The amount collected in fines by CADE in 2015 was three times more than in 2014 (R$169 million). 13. OECD Annual Country report, http://​www.oecd.org/​officialdocuments/​ publicdisplaydocumentpdf/​?cote=DAF/​COMP/​AR(2013)19&docLanguage=En. 14. See “Guias do Cade,” available at http://​www.cade.gov.br/​acesso-​a-​informacao/​ publicacoes-​institucionais/​guias_​do_​Cade/​capa-​interna. 15. Disclaimer: including the first and third authors. 16. See OECD, “Annual Reports by Competition Agencies on recent developments,” available at http://​www.oecd.org/​daf/​competition/​annualreportsbycompetitionagencies.htm. 17. An interesting academic evaluation of the rankings is given by Armoogum and Lyons (2014). 18. Examples of the dimensions of international cooperation involved in competition enforcement are outlined in Carvalho, Ragazzo, and Silveira, eds. (2014). 19. For a comparison, as of August 2016 the entry wage for the gestor career was 20% higher than the entry salary for PhD-​holding faculty in the federal university system. 20. A discussion of the characteristics and role of these gestores is given by Berst et al. (2016). 21. The topic follows Cabral and Mattos (2016) and a research project on remedies (PNUD/​ CADE TR 140678—​Num. PNUD 0000028116—​PRODOC PNUD).

References Amann, E., and W. Baer. 2008. “Neo-​liberalism and Market Concentration in Brazil:  The Emergence of a Contradiction?” The Quarterly Review of Economics and Finance 48 (2): 252–​262. Armoogun, L., and B. Lyons. 2014. “What Determines the Reputation of a Competition Agency?” Working paper. Available at https://​editorialexpress.com/​cgi-​bin/​conference/​ download.cgi?db_​name=IIOC2014&paper_​id=470. Baer, W. 2007 The Brazilian Economy:  Growth and Development, 6th edition. Boulder, CO: Lynne Rienner. Basile, J., and L. Marchesini. 2014. “Small Antitrust Cases Show How Widespread Violations in Brazil Are.” International Valor, September 15, 27. Bersch, K., S. Praça, and M. Taylor. 2016. “State Capacity and Bureaucratic Autonomy Within National States: Mapping the Archipelago of Excellence in Brazil.” In States in the Developing World, edited by A. Centeno, A. Kohli, and D. Yashar, 157–​183. Cambridge:  Cambridge University Press. Cabral, P., and C. Mattos. 2016. “Remédios em atos de concentração:  Teoria e prática do CADE.” Revista de Direito da Concorrência 4 (1): 57–​94. CADE. 2014. “Varejo de Gasolina.” Cadernos do CADE (1). Available at http://​www.cade.gov. br/​acesso-​a-​informacao/​publicacoes-​institucionais/​dee-​publicacoes-​anexos/​cadernos-​do-​ cade-​varejo-​de-​gasolina.pdf. CADE. 2015. “Guidelines: CADE’s Antitrust Leniency Program.” Available at http://​en.cade.gov. br/​topics/​publications/​guidelines/​guidelines-​cades-​antitrust-​leniency-​program-​final.pdf. CADE. 2016. “CADE em números.” CADE. http://​www.cade.gov.br/​acesso-​a-​informacao/​ institucional/​cade-​em-​numeros/​view

Antitrust and Competition Policy in Brazil    739 Calliari, M., D. Andreoli, and R. Alves dos Santos. 2015. “Leniency Agreements in Brazil.” In Overview of Competition Law in Brazil, edited by C. Zarzur, K. Katona, and M. Villela, 315–​ 324. São Paulo: IBRAC/​Editora Singular. Carvalho, V., C. Ragazzo, and P. Silveira, eds. 2014. International Cooperation and Competition Enforcement:  Brazilian and European Experiences from the Enforcers’ Perspective. New York: Wolters Kluwer. Clark, J. 2005. “Competition Advocacy: Challenges for Developing Countries.” OECD Journal of Competition Law & Policy 6 (4): 69–​79. Cowie, M., and C. Mattos. 1999. “Antitrust Review of Mergers, Acquisitions, and Joint Ventures in Brazil.” Antitrust Law Journal 67 (2): 113–​157. Cristofaro, P. P. 2015. “Judicial Review of CADE’s Decisions.” In Overview of Competition Law in Brazil, edited by C. Zarzur, K. Katona, and M. Villela, 371–​394. São Paulo: IBRAC/​Editora Singular. Da Mata, M. 1980. “Controle de preços na economia brasileira:  Aspectos institucionais e resultados.” Política e Programação Econômica 10 (3): 911–​954. Dal Bó, E. 2006. “Regulatory Capture, A  Review.” Oxford Bulletin of Economic Policy 22 (2): 203–​225. Duarte, L. M., and R. A. Santos. 2015. “Cartel Settlements in Brazil: Recent Developments and Upcoming Challenges.” In Overview of Competition Law in Brazil, edited by C. Zarzur, K. Katona, and M. Villela, 285–​314. São Paulo: IBRAC/​Editora Singular. Dutra, P. 2012. Conversando com o CADE. São Paulo: Editora Singular. Fox, E. 2007. “Economic Development, Poverty, and Antitrust: The Other Path.” Southwestern Journal of Law and Trade in the Americas 211 (13): 101–​125. Fox, E., and Sokol, D. 2009. Competition Law and Policy in Latin America. Oxford: Hart. Frade, E. 2014. “Questions and Answers with the General Superintendent.” In Overview of Competition Law in Brazil, edited by C. Zarzur, K. Katona, and M. Villela, 41–​54. São Paulo: IBRAC/​Editora Singular. IDB and OECD. 2005. Competition Law and Policy in Brazil: A Peer Review (2005). OECD. Available at http://​www.oecd.org/​daf/​competition/​brazil-​competition.htm. IDB and OECD. 2010. Competition Law and Policy in Brazil: A Peer Review (2010). OECD. Available at http://​www.oecd.org/​daf/​competition/​brazil-​competition.htm. Jenny, F. 2016. “The Institutional Design of Competition Authorities: Debates and Trends.” In Competition Law Enforcement in the BRICS and in Developing Countries: Legal and Economic Aspects, edited by F. Jenny and Y. Katsoulacos, 1–​57. Switzerland: Charn. Kovacik, W., and J. Baker. 2008. Antitrust Law in Perspective: Cases, Concepts and Problems in Competition Policy. St. Paul: Thomson/​West. Martinez, A. P., and M. T. Araújo. 2015. “Anti-​Cartel Enforcement in Brazil: Status-​Quo and Trends.” In Overview of Competition Law in Brazil, edited by C. Zarzur, K. Katona, and M. Villela, 257–​274. São Paulo: IBRAC/​Editora Singular. Mirow, K. R. 1978. A ditadura dos cartéis:  Anatomia de um subdesenvolvimento. São Paulo: Civilização Brasileira. Moniz Bandeira, J. 1975. Cartéis e desnacionalização, a experiência brasileira: 1964–​1974. São Paulo: Civilização Brasileira. Motta, M. 2004. Competition Policy. Oxford: Oxford University Press. OECD. 2006. “Roundtable on Prosecuting Cartels without Direct Evidence of Agreement:  Contribution from Brazil.” DAF/​COMP/​GF/​WD(2006)37. http://​www.oecd. org/​daf/​competition/​prosecutionandlawenforcement/​36063750.pdf

740   Eduardo Pontual Ribeiro et al. OECD. 2007. “Implementing Competition Policy in Developing Countries.” In OECD Promoting Pro-​Poor Growth Policy: Guidance for Donors. Paris: OECD. Oliveira, G., and Fujiwara, T. 2006. “Competition Policy in Developing Economies: The Case of Brazil.” Northwestern Journal of International Law & Businesses 26: 619. Pires-​ Alves, C., and T. Leandro. 2015. “Competition Policy and the Role of the Chief Economist’s Office:  The Use of Economic Evidence, the international Practices and the Brazilian Experience.” In International Cooperation and Competition Enforcement: Brazilian and European Experiences from the Enforcer’s Perspective, edited by V. Carvalho, C. Ragazzo, and P. Silveira, 17–​27. New York: Wolters-​Kluwer. Ribeiro, E. 2011. “Economic Analysis in Antitrust: The Case of Brazil.” In Competition Law Enforcement in the BRICS and in Developing Countries: Legal and Economic Aspects, edited by F. Jenny and Y. Katsoulacos, 207–​221. Berlin: Springer.

Chapter 35

C orru p tion S c a nda l s , the Evolu tion of A nt i -​ C orrup tion In st i t u t i ons , and Their Im pac t on Brazil’s Ec onomy Mariana Mota Prado and Lindsey Carson

35.1. Introduction The common definition of corruption as the misuse of public power for private benefit (Lambsdorff 2007; Rose-​Ackerman 1999) belies the phenomenon’s complex and multifaceted character. Corruption includes an expansive array of behaviors (e.g., bribery, extortion, collusion, nepotism, clientelism, fraud, embezzlement), which may be prohibited formally by law, regulation, or policy, or may “merely” violate the fundamental tenets of civil service and public trust (Nye 1967). Corruption may be petty, pursued by low-​level officials who demand or divert relatively modest sums, or grand, involving misconduct by high-​level officials with significant influence or authority over major projects, assets, or contracts. There are multiple mechanisms through which corruption can have a negative impact upon a country’s economy, as illustrated by the empirical studies summarized in Table 35.1. While the impact of corruption on a country’s economy has been widely explored, the direct and indirect impact of corruption scandals on a country’s economy has received far less attention in the academic literature. Numerous studies have examined the political and electoral impacts of corruption scandals (Balan 2011; Bowler and Karp 2004; Costas-​Perez et al 2011; Ferraz and Finan 2008), but the economic repercussions

742    Mariana Mota Prado and Lindsey Carson Table 35.1 How Corruption Damages Countries’ Economies: Review of Studies Impact of Corruption

Author

Evidence

Mauro 1995

An improvement of one-​standard deviation in Business International’s Corruption Index was associated with an increase in the overall investment rate by 2.9% of GDP.

Brunetti, Kisunko, and Weder 1998

Survey data from 51 developed and developing economies shows that corruption negatively affects investment.

Corruption reduces the ratio of overall capital investment to GDP.

Campos, Lien, and Corruption is negatively correlated with gross Pardhan 1999 investment to GDP levels, but more “predictable” corruption regimes have less negative impact on investment than those that are less predictable. Lambsdorff 2003

Corruption reduces the ratio of overall capital investment to GDP, but the ratio of investment to GDP is a poor indicator of a country’s attractiveness to investors because corruption may lower both investment and GDP, which are both affected by strong law and order institutions.

Knack and Keefer 1995

Weaker institutional environments (as measured by a variable of institutional quality that incorporates corruption as one factor) tend to attract less FDI and to have lower rates of economic growth.

Wei 2000

Data on bilateral investment flows between 12 developed countries and 45 host countries in 1990 and 1991 show that an increase in a host government’s level of corruption from that of Singapore to that of Mexico has the same negative effect on inward foreign direct investment as raising the marginal tax rate by 42 percentage points.

Habib and Zurawicki 2002

Bilateral FDI flows between 7 home and 89 host countries over 1996–​1998 shows that high levels of perceived corruption had a significant negative effect on investment inflows.

Aizenman and Spiegel 2003

Data over the period of 1990–​1999 demonstrate a robust negative relationship between the level of corruption and the ratio of FDI flows to domestic investment flows.

Corruption reduces foreign direct investment (inward) flows.

Corruption Scandals and Anti-Corruption Institutions    743 Table 35.1 Continued Impact of Corruption

Author

Evidence

Friedman et al. 2000

Data from 69 countries show that corruption increases the share of economic activity in the informal sector, resulting in diminished tax revenues as a percentage of both official and total GDP.

Johnson 2000

Firm-​level regressions for three formerly communist countries show that higher levels of corruption were correlated significantly with participation in the black market.

Shleifer and Vishny 1993

Presenting a model of the costs of corruption that takes into account the distortions caused by the need for secrecy, namely the shifting of government investments from high-​value sectors like health and education to sectors in which corruption can be more easily concealed.

Mauro 1998

Hypothesizing that the significant, negative, robust correlation between levels of corruption and public expenditures on education is attributable to the challenges government officials face in collecting large bribes in that sector.

Gupta et al. 2001

Analyzing data from four different sources for up to 120 countries during 1985–​1998 and finding a significant positive correlation between corruption and military expenditures as a share of GDP and total government spending.

Baldacci et al. 2004

Examining data from 39 low-​income countries and finding evidence that corruption reduces the effectiveness of public capital expenditures.

Gupta et al. 2002

An increase of one standard deviation in corruption boosted the Gini coefficient of income inequality by approximately 11 points (on a 100-​point scale) and lowered the growth of income among the poor by about five percentage points per annum.

You and Khagram 2005

Noting the mutually reinforcing relationship between corruption and inequality.

Corruption is correlated with informal economies and diminished tax revenues.

Corruption generates inefficient allocation of public funds.

Corruption increases inequality.

(continued)

744    Mariana Mota Prado and Lindsey Carson Table 35.1 Continued Impact of Corruption

Author

Evidence

Mauro 1995

See supra.

Ehrlich and Liu 1999

Hypothesizing that corruption undermines economic growth by consuming financial resources that could otherwise be used for production or investments in human capital development.

Mo 2001

Evidence from 46 countries indicates that a 1% increase in the corruption level reduces a country’s growth rate by 0.72%, with the majority of the effect attributable to corruption’s effect on political instability.

Swaleheen 2011

Corruption has a direct, negative effect on economic growth such that a one standard deviation change in the level of corruption changes the rate of growth of real per capita GDP by 0.12 percentage point, and noting that the impact of corruption is nonlinear, with the largest effects witnessed in countries with low levels of corruption.

Corruption has a negative impact on economic growth.

Source: Prepared by the authors.

of corruption scandals largely remain a black box in scholarly analyses. A notable exception is a 2015 study estimating that the recent Petrobrás scandal and the associated retraction in the construction sector may have caused production losses equivalent to 2% of Brazil’s gross domestic product GDP in 2014 (Oliveira et al. 2015). While the study was premised on a familiar hypothesis that corruption scandals can have a negative impact upon a country’s economy by reducing investments, the magnitude of the authors’ results may reflect the dominance of Petrobrás in the construction sector and the large share of the Brazilian economy that sector represents, as well as the impact of corruption scandals on economic indicators. Thus, while interesting, the Petrobrás case is unlikely to be very representative of the different ways in which scandals may harm a country’s economy. The absence of further analysis seems particularly relevant in the case of Brazil, a country that has grappled with corruption for most of its political history, and in which the issue has assumed a particularly prominent position in the country’s politics since its return to democracy in 1988. Numerous scandals at the federal, state, and municipal levels and across all branches of government confirm that corruption remains entrenched in the country’s political system.

Corruption Scandals and Anti-Corruption Institutions    745 While the existing literature suggests that, under certain circumstances, citizens respond politically to scandals at the ballot box, we consider whether other types of responses to scandals may have direct or indirect impact on a country’s economy. More specifically, we develop a three-​part analysis. First, we suggest that a corruption scandal can create a window of opportunity for reforms, such that a government’s action or inaction in the aftermath of a scandal may signify its commitment to institutional change. We then assess the impacts of past scandals and responsive institutional reforms on perceptions of corruption. On the one hand, by providing evidence about the existence of corruption in the country, scandals may worsen perception indicators. On the other hand, the ameliorative measures taken in the aftermath of such scandals may generate optimism about the government’s ability to deal with the problem, which may, in turn, positively affect perception indicators. In the third part, we hypothesize how foreign and domestic investors may react to government reforms as well as perception-​based indicators of corruption. Considering that higher levels of corruption are generally associated with negative impacts on a country’s economy, one cannot help but wonder if such scandals may be generating short-​term pain, by scaring off investors, but long-​ term gain, by increasing the country’s governance system and reducing overall levels of corruption (Kaufmann et al. 2006, 2009). This chapter shows that some corruption scandals have had a significant impact on institutional reforms and corruption indicators in Brazil. The first two sections provide qualitative and quantitative data mapping these connections. The last section offers some hypotheses about the potential economic impacts of these observable effects of scandals on Brazil’s economy in the short and long term. Our study introduces a fundamental hypothesis that we suggest should guide future research. Specifically, given the current paucity of scholarship on the topic, we emphasize the need for targeted research and analysis on the direct and indirect impact of corruption scandals on a country’s economy.

35.2.  Corruption Scandals and Government Responses (1992–​2016) Although Brazil has experienced impressive economic expansion since the turn of the millennium (OECD 2015; World Bank 2016), strong evidence indicates that the net effect of corruption on the country’s economic development has been negative (Iquiapaza and Amaral 2007). Silva, Garcia, and Bandeira (2001) estimate that if levels of corruption in Brazil had been as low as those in Denmark (the least corrupt country in their sample), per worker incomes would have been 43% higher in 1998, representing a per capita income loss of US$2,840.81. Recent studies estimate that corruption consumes between 1.4% (FIESP 2010) and 5% (Época 2008) of the country’s GDP, translating into economic losses of between £10.5 billion and £32.9 billion each year.

746    Mariana Mota Prado and Lindsey Carson The democratic constitution enacted in 1988 laid the groundwork for the development of Brazil’s modern web of accountability institutions, particularly in the areas of oversight and investigation. Under the Constitution, the Ministério Público Federal (Federal Public Prosecutors’ Office, MPF) gained independence from the executive branch, emerging as the de facto “fourth branch of government,” empowered to act in the defense of the public. The Constitution also conferred greater powers and responsibilities on the Tribunal de Contas da União (National Accounting Tribunal or Federal Audit Court, TCU) and guaranteed full freedom of the press. In spite of, or because of, these and other strong oversight and investigative institutions, enforcement and audit authorities have discovered and exposed corruption scandals during the terms of each of Brazil’s first six post-​authoritarian presidents. Each scandal has prompted further governmental reforms to increase transparency and accountability.

35.2.1. Collorgate (1992) In June 1992, Congress launched a Comissão Parlamentar Inquérito (Parliamentary Investigation Commission, CPI) to investigate public allegations by Pedro Collor de Melo that his brother, Brazilian president Fernando Collor de Melo (“Collor”), was involved in an elaborated extortion and influence-​peddling scheme orchestrated by his former campaign manager, Paulo César (“PC”) Farias. Assisted by leads from investigative journalists, in August 1992, the CPI announced that it had uncovered conclusive evidence implicating Collor in the multi-​million-​dollar scheme. On September 1, 1992, the Brazilian bar and press associations submitted a formal request to open impeachment proceedings, and, on September 29, 1992, Collor was formally impeached by the Chamber of Deputies by a vote of 441 to 38. Although he resigned on December 29, Collor was formally impeached by the Senate the following day. While his Senate impeachment stripped Collor of his ability to participate in politics for eight years, in 2006 he was elected to the Brazilian Senate, where he currently chairs the Infrastructure Commission. Collorgate exposed gaps in the existing legal framework that prompted swift legislative action. In June 1992, just as the scandal was erupting, Congress expanded the MPF’s authority to act as a horizontal accountability body through the Administrative Improbity Law (Law 8429/​92). This law strengthened the MPF’s role as the primary enforcer of political law and protector of collective interests by granting it enhanced authority to act against corruption and the misuse of public funds (Arante 2011). The following year, it passed the Government Procurement Act (Law 8886/​93), which established rules on public bidding. The new president, Itamar Franco (Collor’s former vice president), created a code of conduct for civil servants, and also worked to form the Secretariat of Internal Control in the Finance Ministry, designed to oversee the spending of the executive. In 2002, the Secretariat would be merged with the newly created and more powerful Federal Comptroller’s Office (CGU) (Praça and Taylor 2014, 33).

Corruption Scandals and Anti-Corruption Institutions    747

35.2.2.  Budgetgate (Anões do Orçamento) (1993) Less than a year after Collor’s resignation, the Brazilian political system was rocked by yet another scandal. Under suspicion for murdering his wife, a former staff aide to the Comissão Mista do Orçamento (Joint Budget Committee of Congress, CMO) revealed that, since 1989, members of the Committee had been receiving kickbacks and bribes for approving budget amendments that provided benefits to construction companies, bogus nonprofits, and municipal governments. In his testimony before the CPI, the aide implicated 38 sitting government officials in the anões do orçamento (“budget dwarves,” coined after the short stature of the legislators involved) scandal. The Federal Police (Departamento da Polícia Federal, DFP) soon launched their own criminal investigation into the budget process. Eventually, 19 legislators were expelled from Congress. The scandal highlighted the lack of transparency and overconcentration of power in the budget process and triggered significant reforms in budgetary institutions. Based on recommendations from the CPI, in 1995 Congress implemented new procedural and reporting requirements for the CMO, increased membership on the Committee, limited the powers of the CMO’s leaders, and reduced the number of permitted individual amendments (Melo 2013; Samuels 2002, 325). Additionally, the scandal led to new legislation in 1994 that rendered any member of Congress who had resigned to avoid the consequences of an ongoing corruption investigation ineligible to be on the CMO (Praça and Taylor 2014, 33).

35.2.3.  The Case of the São Paulo Regional Labor Court (1998) In 1998, it was revealed that while R$226 million had been transferred from the federal treasury to build the São Paulo Regional Labor Court of the Second Region (TRT) since 1992, only R$63 million had been spent on actual construction, leaving R$169 million unaccounted for (Davis et al. 2015, 666). While the TCU had identified problems with the court’s procurement process within a year of the bid selection, its calls to halt the project went ignored, allowing construction to begin in 1994 (Taylor 2009, 156). After public revelation of the overspending, the ensuing CPI revealed that the bulk of the funds had been embezzled through a scheme masterminded by Judge Nicolau dos Santos Neto (nicknamed “Lalau”) with the assistance of Senator Luiz Estevão. It was also revealed that Estevão was one of the owners of the company contracted to build the TRT courthouse, and that the engineer in charge of supervising the construction progress received the equivalent of US$42,000 via a related company (Davis et al. 2015, 679). In June 2000, Estevão became the first senator to be impeached; after being stripped of his parliamentary immunity, he was also charged with criminal embezzlement. After a lengthy criminal process involving 34 appeals, he began serving a 31-​year prison sentence in March 2015 (Richter 2016).

748    Mariana Mota Prado and Lindsey Carson Meanwhile, Lalau absconded overseas, putting much of the embezzled money in foreign accounts and assets, including an apartment in Miami and bank accounts in Switzerland, the Cayman Islands, the Bahamas, and Panama (Davis et al. 2015, 682). Some of these assets were repatriated through legal actions in the United States and Switzerland, but much of it remains hidden in other jurisdictions. However, after nine months as a fugitive, Lalau returned to Brazil, where he was convicted and sentenced to prison in 2002. This scandal highlighted procedural and institutional deficiencies in the TCU that had first been exposed during Collorgate. While a proposal during the Collorgate scandal to strengthen the oversight of federal funds by transforming TCU into an independent technical accounting body called Auditoria Geral da União (General Accounting Office) was never realized (Fleischer 2000, 104), the Lalau scandal underscored the need for internal control mechanisms and enhanced coordination. In 2001, President Cardoso established the Corregedoria Geral da União (later strengthened and renamed the Controladoria Geral da União, CGU, under Lula), not only as a response to the scandal, but also as a strategic reform to address long-​standing deficits in the country’s anti-​corruption architecture (Loureiro et al. 2012, 57).

36.2.4.  The Case of SUDAM and SUDENE (2001) In 2001, investigations conducted by the Federal Police, federal prosecutors, the TCU, a congressional CPI, and the press revealed that R$2 billion had been embezzled from each of two federal regional development programs, the Superintendência para o Desenvolvimento da Amazônia (Amazonian Development Superintendency, SUDAM) and the Superintendência para o Desenvolvimento da Nordeste (Northeastern Development Superintendency, SUDENE). President Cardoso abolished both agencies, transferring their powers to newly created bodies. After being implicated in the embezzlement scheme, the leader of the Senate, Jader Barbalho, resigned first his leadership position, then his office, and was briefly arrested and imprisoned before being released. Because he resigned his post before being expelled, he retained full political rights; in 2002, he was elected to the Chamber of Deputies and was re-​elected to the Brazilian Senate in 2011. In December 2001, as a response to this scandal, Congress removed the requirement for congressional consent for the Supreme Court to proceed with cases against elected politicians (Constitutional Amendment n. 35) (Melo 2013, 18). The Amendment prevented cases from being held up due to corporativism, a common practice in Congress before the Amendment.

35.2.5.  Operation Anaconda (2003) In 2003, a joint operation (code-​named “Anaconda”) organized by the MP and the Federal Police exposed a syndicate of lawyers, federal judges, and police in a scheme

Corruption Scandals and Anti-Corruption Institutions    749 that involved selling judicial decisions to criminals. At the time, it had become clear that despite the fact that a number of judges were living overtly wealthy lifestyles, none of the other judges or court officials had taken any explicit steps to report them or hold them to account (Praça and Taylor 2014, 38). The lack of internal accountability revealed by the scandal, along with growing popular dissatisfaction with the functioning of judicial institutions and processes, prompted President Inácio Lula da Silva in 2003 to create a new position, Secretaria de Reforma do Judiciário (Office for Judicial Reform, SRJ) in the Ministry of Justice. The SRJ is charged with creating, promoting, coordinating, and managing the process of reforming the judiciary and with promoting communication and cooperation among the branches of government as well as judges, public prosecutors, public defenders, lawyers, law experts and civil society. In 2004, Constitutional Amendment 45/​2004 was enacted to modernize the administration of justice. Among other actions, the Amendment established the CNJ as an administrative oversight body to provide guidance to the judiciary, as well as investigate and punish judicial officials in violation of the law or ethical rules (Prado 2010). President Lula also took broader action against corruption, establishing by decree in 2003 the Council on Public Transparency and Combating Corruption (Decree 4,923/​ 03) as an advisory body to the Controladoria-​Geral da União (Office of the Comptroller General of the Union, CGU). The members of the Council include representatives of 10 public entities, including the CGU, the MPF, and the TCU, and 10 civil society organizations, including the Brazilian Bar Association and the Brazilian Press Association. Its purpose is to discuss and recommend measures to improve control mechanisms especially over public resources, to promote transparency in the public administration, and to combat corruption and impunity.

35.2.6.  Big Monthly Stipend or Mensalão Scandal (2005) The mensalão (“big monthly stipend”) scandal involved a series of regular pay-​ offs from the ruling Partido dos Trabalhadores (Workers’ Party, PT) to allies in the Chamber of Deputies in exchange for support for their legislative agenda. The scandal broke in 2005 when the president of an allied party, Roberto Jefferson, revealed the R$30,000 (US$12,000, according to the exchange rate at the time) per month payments in an interview with the Brazilian newspaper Folha de São Paulo. The funds were alleged to have originated in the advertising budgets of state-​owned enterprises and were transferred via fake contracts with corrupt advertising agencies. A  series of intersecting CPIs were launched to investigate the payments scandal, as well as related and overlapping alleged wrongdoing related to the Postal Service and the regulation of gambling. In 2005, Congress expelled Jefferson as well as the alleged mastermind of the scandal, President Lula’s former chief of staff, José Dirceu, while several others legislative members resigned their seats preemptively. In total, the congressional committee investigations named 18 deputies (and one former deputy) who had received the mensalão payments.

750    Mariana Mota Prado and Lindsey Carson At the request of President Lula, the Procurador Geral da República (Chief Public Prosecutor) launched an independent criminal inquiry into the affair, and, in May 2006, asked the Supremo Tribunal Federal (Supreme Court, STF) to initiate criminal proceedings against 40 individuals. In November 2012, the STF found 25 individuals guilty of crimes including embezzlement, money laundering, corruption, and the misuse of public funds. Arrest warrants were issued for 12 of the 25 in November 2013. The convictions handed down in the mensalão case were widely heralded as a landmark for Brazilian justice and a blow to the culture of impunity (The Economist 2013). While the degree to which the STF’s actions signify a movement toward real reform in the judicial process remains unanswered, the mensalão has triggered campaign finance reforms: the Superior Tribunal for Electoral Law (TSE—​Tribunal Superior Eleitoral) has enacted a regulation requiring parties to fill out an online declaration of campaign contributions that will be publicly available at TSE’s website before elections (Resolução-​ TSE 23.376/​12) (Melo, 2013, 17). Although the system, known as Sistema de Prestação de Contas Eleitorais (Accounting System of Campaign Financing, SPCE), is likely to create more transparency, it is unlikely to curtail off-​books campaign contributions, a widespread practice in Brazil.

35.2.7.  The “Car Wash” Scandal, or the Petrobrás Case (2014) The ongoing Petrobrás scandal involves massive alleged malfeasance by corporate, as well as political, elites surrounding the enormous state-​run oil company Petrobrás. The scandal began in March 2014 with a Federal Police investigation into a money-​ laundering scheme, in which investigators were alerted to repeated inexplicable money transfers between Petrobrás and a number of companies with whom they did business. Over the following year, evidence came to light suggesting that Petrobrás directors and managers were receiving enormous bribes to award contracts to third-​party corporations at greatly inflated prices. These Petrobrás directors and senior employees were in turn often appointed by members of the ruling political coalition, who were receiving kickbacks and campaign donations from the Petrobrás beneficiaries as well as the outside contractors, suggesting that these individuals may have been appointed based on their willingness to participate in or at least tacitly condone the scheme. The scope of the corruption is unprecedented: politicians and Petrobrás employees received hundreds of millions (if not billions) of dollars in kickbacks between 2004 and 2012. Political kickbacks were often set at 3% of the value of the contracts awarded by Petrobrás (Almeida and Zagaris 2015, 87–​90). Indeed, the name “Car Wash” comes from the fact that a black market currency dealer involved in the scheme, who would turn into one of the investigation’s first informants, used a car wash in Brasília to launder some of the money going to politicians. At least nine of Brazil’s largest construction companies have been implicated as participating in the scheme (Costas 2014), and 25 companies are under investigation. However, this may only be the beginning of a larger corruption probe, as evidence has been found suggesting that as many as 170 private companies

Corruption Scandals and Anti-Corruption Institutions    751 are involved in political corruption (Almeida and Zagaris 2015, 94). As of November 2016, 52 criminal charges have been laid against 245 people (MPF 2016), and at least 50 politicians have been investigated. These politicians include former presidents Lula and Collor de Mello, the Senate president, and the president of the House of Representatives (Almeida and Zagaris 2015, 88). The effects of the investigation have been far reaching. Petrobrás has been estimated to have lost US$31.5 billion in the scheme (Almeida and Zagaris 2015, 88), and it has likely contributed to Brazil’s recent economic troubles. The Petrobrás scandal was also a factor in the recent impeachment of President Dilma Rousseff, and suggestions have been made that the impeachment was motivated by the need to pass an amnesty bill through a favorable government to protect corrupt politicians from investigation and prosecution (Greenwald 2016). Whether or not this is true, the political and economic ramifications of the revelations to date will likely be felt for years to come, and more revelations may yet come to light as the investigation continues. One important institutional ramification of the Car Wash scandal has been the STF decision striking down as unconstitutional a law that allowed donations to political election campaigns from corporations (Kiernan 2015). The election financing law was challenged in the Court by the national bar association of Brazil in what is largely viewed as an attempt, following the Petrobrás revelations, to limit crony capitalism in the country and prevent the corporate capture of politicians (Kiernan 2015). It may also prevent one avenue of laundering bribe money via on-​the-​books campaign donations. Brazilian election campaigns have historically been primarily funded by campaign donations, with as much as 76% of campaign financing attributed to corporate donations in the 2014 elections (TSE 2015). Before the STF ruling, companies were allowed to donate up to 2% of their previous year’s gross revenue to candidates or party funds, and with the most expensive elections in the world outside of the United States, large corporate donations proved key to electoral success (Melo 2013). It appears that corporate donations are a sound financial investment for corporate interests so long as the targeted candidate emerges victorious. Based on information from the 2006 federal deputy election, Boas, Hidalgo, and Richardson (2013) find that public works firms that donated to successful candidates from the then-​ruling PT received, on average, at least 14 times the value of their electoral contributions in additional government contracts in the ensuing legislative term. Musacchio and Lazzarini (2014, 274–​275) also show that firms that donated to the campaign of successful candidates are more likely to get Brazilian Development Bank (BNDES) financing than other firms (regardless of the party). By removing the ability for corporations to fund election campaigns, the hope is that the close relationship between individual candidates and corporations will be weakened, and thus policy is less likely to be controlled by corporate interests. Following the decision, campaign finance laws allow individuals to donate up to 10% of their earnings in the previous year. Political candidates can also receive public funding to fund their campaigns. Another institutional change that took place in the wake of the Car Wash scandal was the February 2016 decision by the STF to execute jail sentences immediately (Nicholson 2016). The decision reverts the prevailing jurisprudence, which determined that

752    Mariana Mota Prado and Lindsey Carson execution of jail sentences would only start after all grounds for appeal had been exhausted. This was deemed one of the obstacles to effective punishment in Brazil, since these appeals could take years and often run against the statute of limitations, forcing courts to close the case without imposing any effective punishment (Prado and Carson 2016). While the decision was made in the case of an armed robber, it is perceived to have wide implications for corruption cases in general, and the Car Wash scandal in particular.

35.3.  Scandals and Corruption Indicators Corruption’s complex and clandestine nature presents inherent challenges to assessing its current and historical levels in any environment, including Brazil (Golden and Picci 2005; Knack 2007). In the absence of any comprehensive, direct mechanism to measure the prevalence and magnitude of corruption in a given environment, researchers and policymakers have developed a series of indirect indices which can be grouped into three broad categories:  (1) objective data related to enforcement actions and audit reports; (2) experience-​based surveys; and (3) perception-​based surveys. This section will focus on the latter, as they are the ones most directly affected by scandals. The relationships among scandals, perceived levels of corruption, and the aggressiveness of government anti-​corruption activities exposes one of the primary flaws of perception-​oriented indicators: their vulnerability to echo and to reinforce misinformation or biases among respondents and the public at large. Perceptions are likely to be based on observable reports or developments that appear to reflect the level of corruption within a society, such as muck-​raking media stories or the announcement of new or emboldened governmental anti-​corruption campaigns (Treisman 2007, 215). Given the susceptibility of perception-​based surveys to the “noise” concerning and surrounding corruption, a vibrant and free press that reports aggressively on such activities may trigger one of two potential alternative results: (1) participants will overestimate the prevalence of corruption in society, or (2) reports of governmental action will produce (possibly undue) confidence concerning the status of the battle against corruption. In the first scenario, respondents may interpret media reports or government initiatives as signs of worsening corruption, while, in reality, they may be mere manifestations of heightened awareness of or commitment to reducing existing levels. In the second scenario, the public may have an overly optimistic perception of the government’s commitment to reform or the speed or significance of any official responses to the scandals. Public perceptions of corruption are relevant to a country’s economy because decisions to invest or start or expand a business are predicated on expected costs and benefits, expectations that are themselves derived from perceptions of all aspects of the market, including the level of corruption; thus, the perceived level of corruption in a

Corruption Scandals and Anti-Corruption Institutions    753 country is likely to affect its economic development, regardless of the corruption reality (Dow Jones and MetricStream 2016; Control Risks Group Limited 2006).

35.3.1.  Perception-​Based Surveys on Levels of Corruption Perception-​based indices in Brazil exhibit significant variance and appear to be particularly susceptible to change based on highly publicized corruption scandals at the federal level. More specifically, these indicators show significant variation as a result of the mensalão scandal in 2005, and the Car Wash scandal in 2014. In the most widely cited perception-​based indices—​Transparency International’s Corruption Perceptions Index (CPI) and the World Bank’s Control of Corruption indicator (CC), Brazil generally ranks around the median of surveyed countries, and its scores have remained relatively stable since the transition to democracy, despite numerous public scandals and a plethora of anti-​corruption reforms. One can see, however, significant drops in the CPI in 2006 and 2015 (Figure 35.1).1 Similarly, the CC indicator has fallen considerably over the past two years in the wake of the Car Wash scandal, after a significant dip around the time of the mensalão scandal (Figure 35.2). 44 42 40 38 36 34 32 30

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

Figure 35.1.  Corruption perceptions index, Brazil. Produced by the authors. Data source: Transparency International.

75 70 65 60 55 50 45 40 35 30 25

1996 1998 2000 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Figure 35.2.  Brazil’s global percentile rank, World Bank, Control of Corruption Indicator. Produced by the authors. Data source: World Bank Governance Indicators.

754    Mariana Mota Prado and Lindsey Carson These variations show that, in the wake of a revelation of a major scandal, the public has been likely to seize upon that news as evidence of a significant increase in corruption, regardless of any changes in the actual level of corruption in the country. This experience is not unique to Brazil. Such trends were detected in Germany and Ireland in 2000, following public revelations of political corruption in those countries (Lambsdorff 2000). Thus, scandals may actually worsen, or at least keep steady, the country’s perception-​based indicators, even if such scandals are the result of governmental action against corruption that are, in fact, deterring and reducing corruption effectively. If, as Melo (2013) contends, this model holds for Brazil, the country’s persistently mediocre perception-​based indicators may be attributable to the greater visibility of corruption resulting from the effective functioning of the country’s democratic institutions, including the press.

35.3.2.  Perception Based Surveys on Anti-​Corruption Efforts Perception-​based surveys also reveal how members and segments of the public view the government’s anti-​corruption efforts in the wake of scandals. Specifically, perception-​ based indicators may capture overly optimistic attitudes generated by well-​publicized reports of individual incidents of government reform and anti-​corruption activity in the wake of scandals. While such developments are certainly auspicious, it is unclear the extent to which they reflect a true turning of the tide of corruption in the country, as opposed to isolated events in an otherwise unchanged institutional landscape. Fifty-​six percent of Brazilian respondents in the 2013 Global Corruption Barometer replied that the government’s actions in the fight against corruption were ineffective, while only 23% said they were effective. Since then, the country has seen arrests in connection with the mensalão case at the end of 2013, which included high level politicians, as well as the arrests in connection with the Car Wash case. In Transparency International’s latest Putting Corruption Out of Business survey of 3,000 business people in 30 diverse economies around the world, 38% of Brazilians listed “corruption and bribery-​related crimes are not prosecuted” as the top barrier to stopping bribery and corruption in the country’s private sector, compared with 28% of total survey participants; only five countries—​Senegal, Turkey, Hungary, Nigeria, and the Philippines—​had a higher percentage of respondents listing lack of prosecution as the main obstacle (Transparency International 2012). In contrast, in some cases, even drastic changes in enforcement action do not seem to have triggered significant change in perceptions. For example, in Bertelsmann Stiftung’s BTI Transformation Index, Brazil received an 8 out of 10 on expert assessments of the extent to which “public officeholders who abuse their positions [are] prosecuted or penalized” both in 2014 and in 2016 (Bertelsmann Stiftung 2014, 2016). In the qualitative evaluation, the respondents noted the recent increases in the number of public servants and elected officeholders suspended or removed from office, forced resignations of several ministers from President Dilma Rousseff ’s administration, and the much-​heralded

Corruption Scandals and Anti-Corruption Institutions    755 convictions and sentences handed down in the mensalão scandal (Bertelsmann Stiftung 2016, 9). The evaluation also noted the more recent and much larger Petrobrás (Car Wash) scandal that has rocked Brazilian politics since 2014, as described earlier in this chapter, citing dozens of arrests and the investigations of many politicians and major corporations.

35.3.3.  Perception-​Based Surveys on Institutional Corruption Scandals and responsive governmental measures also affect perceptions of corruption and general trust on institutions. Some surveys suggest that Brazilians’ perception of the judiciary has changed significantly in the wake of the mensalão and Car Wash scandals. For instance, the Brazilian National Transport Confederation reported that the percentage of citizens saying that they always or usually trust the judiciary has increased from 28% of the respondents in October 2013 to 34.3% in October 2015 (Nicholson 2016). The 2013 Global Corruption Barometer reveals that Brazilians perceive stark differences in the extent of corruption in various institutions. Only 50% of respondents from Brazil considered the judiciary corrupt or extremely corrupt; the figure leapt to 81% when asked about political parties (Transparency International 2013b; see Table 35.2). This number contrasts with the 2016 International Bar Association (IBA) Judiciary Integrity Initiative, which found that only 27% of Brazilians think that corruption occurs in the judicial system (IBA 2016). The difference between the IBA report and the Barometer may be attributable to the active stance taken by the Brazilian courts in

Table 35.2 Global Corruption Barometer, 2013 (Brazil) Percent Who Consider the Institution “Corrupt” or Extremely “Corrupt” Political parties

81%

Legislature

72%

Police

70%

Medical and health services

55%

Judiciary

50%

Public officials/​civil servants

46%

Media

38%

Business/​private sector

35%

Nongovernmental organizations (NGOs)

33%

Religious bodies

31%

Military

30%

Prepared by the authors. Data source: Transparency International.

756    Mariana Mota Prado and Lindsey Carson the last three years, in connection with the mensalão and Car Wash scandals. However, it could also be related to the fact that Transparency International surveys the general public, whereas IBA consults civil society and legal professionals (judges, lawyers, and prosecutors) (Nicholson 2016). In contrast to the judiciary, according to the Global Corruption Barometer, the legislature places second on the list of public institutions perceived to be most corrupt by Brazilian citizens. The low standing of Congress may reflect ubiquitous and incessant investigations into and evidence of alleged corruption among its members. Members of Congress have featured prominently in three of the country’s most notorious corruption scandals over the past two decades—​the anões do orçamento (“budget dwarves”) scandal in 1993–​1994, the sanguessuga (“bloodsucker”) scandal between 2004 and 2006, and the mensalão affair, which began in 2005 but was only decided in 2013. However, the judiciary has not been immune to scandals. The direct control the Brazilian judiciary exercises over its own budget and the absence of effective accountability mechanisms for the use of such funds creates opportunities for embezzlement. The most prominent example of this problem was the scandals involved the São Paulo Regional Labor Court (TRT) in 1998, described earlier. In a more recent case, officials at the Labor Court in the state of Rondônia have come under investigation for allegedly transferring an estimated R$2 billion (US$1 billion) from the judicial budget to private hands. Nepotism, once a common and even legal practice, also continues to pervade the Brazilian judiciary. In a particularly egregious case from the late 1990s, Severino Marcondes Meira, a labor judge, placed 63 relatives on the court payroll. In 2013, a judge in the appeals court of Mato Grosso was sanctioned for including his two sons on the payroll, despite the fact that they did not provide any services to the judiciary. Finally, the sale of judicial opinions has been uncovered in a number of cases, including the large Anaconda Operation, described earlier. And in December 2013, the National Council of Justice (Conselho Nacional de Justiça, CNJ) initiated a disciplinary investigation in response to evidence that a judge in the state of Tocantins, José Liberato Costa Póvoa, exchanged favorable decisions for cash (prices allegedly ranged from US$5,000 to US$25,000). The discrepancy between the perceptions of the legislature and the judiciary, despite the significant number of scandals in which both have been involved, suggests that there may be other factors, in addition to information about involvement in wrongdoing, driving institutional perceptions. The recent changes in perception-​based indicators about the judiciary may suggest that adopting highly visible and harsh measures against corruption may influence such public opinion. In contrast, the same strategy does not seem to have been so effective for the police force, ranked as the third most corrupt institution in the country despite being directly involved in many corruption investigations. In contrast with the MPF, the Federal Police force is weakly institutionalized: it is subordinated to the Ministry of Justice (part of the executive branch), it offers relatively low prestige (especially in comparison to other options for young lawyers such as those offered by the MPF or the judiciary), and it has been more prone to corruption and abuse of power (Arantes 2011). It was only in the late 1990s that the MPF started

Corruption Scandals and Anti-Corruption Institutions    757 conceiving of its mission as combating corruption and organized crime. This happened alongside attempts at the federal government to strengthen the Federal Police, which have translated both in personnel and budget increases (Arantes 2011). This contributed to an increase in the number of investigations of acts of corruption between 2005 and 2009 (MESICIC 2012, 27). There was also an increase in the number of “operations,” which are defined as “the carrying out of arrest or search and seizure warrants, issued by courts after a period of investigation that may last weeks or months and almost always includes participation by the MP or other bodies such as the Revenue Service, the Social Security Ministry, the state police, or regulatory agencies” (Arantes 2011, 200). While some investigations may lead to operations, others will just lead to criminal charges. The operations, however, are very visible, often being described by catchy nicknames that facilitate publicity. They are accompanied by wide press coverage. Both the operations and the number of people imprisoned have significantly increased since 2006 (MESICIC 2012, 26). However, increases in enforcement action and sanctions are sometimes interpreted as evidence that bureaucratic waste and corruption have spiked dramatically, not that enforcement has improved. Indeed, increased enforcement action and sanctions for corruption-​related activities at the federal level in the past two decades does not seem to have positively influenced perception-​based surveys. For example, in 1999, the Tribunal de Contas da União (Federal Accounting Tribunal, TCU), Brazil’s leading government auditing institution, convicted 845 officeholders for irregularities and misconduct in public administration, with the number of convictions climbing to 1,574 in 2007; over the same period, audits performed by the TCU revealed that the percentage of cost-​intensive government programs showing severe irregularities rose from 32% to 77% (Speck 2011, 140). Similarly, reports that investigations initiated by the MPF on corruption and administrative impropriety more than tripled between 2007 and 2011 (MESICIC 2012, 36) may reveal “more on the traits of how the criminal code is applied than on the crime in question” (Speck 2000, 11). This is corroborated by empirical and anecdotal evidence suggesting that much progress has been made in professionalizing the bureaucracy and reducing corruption at the federal level. However, this may be not be captured by perception-​based surveys.

35.4.  The Short-​and Long-​Term Impact of Scandals on the Brazilian Economy Scandals have the potential to influence individual behavior, affecting citizens and investors alike. As shown in Table 35.1, numerous studies have confirmed the negative and significant correlation between perceived levels of corruption and overall investment, FDI flows, economic growth, and inequality. Explanations for the multitude of impacts of perceptions of corruption on a country’s economy generally, and investments

758    Mariana Mota Prado and Lindsey Carson levels in particular, often focus on the lack of credibility, inefficiencies, and uncertainties corruption creates in businesses’ operating environment (Brunetti, Kisunko, and Weder 1998; Habib and Zurawicki 2002; Wei 1997; Zurawicki and Habib 2010). Without minimizing the countless individual and societal factors that contribute to perceptions of corruption, multiple studies have shown that greater corruption-​ related coverage in the media is associated with an increase in a country’s perception-​ based corruption indicator (Kenny et  al. 2011; Knack 2007; Treisman 2007). While the finding that heightened public awareness of and information about corruption leads to increases in perception-​based corruption indicators may seem tautological, it helps to explain how corruption scandals may have negative effects on investment in a country’s economy. At the level of the individual, corruption scandals create the specter of uncertainty and instability in the larger political and economic environment, making investments less attractive. At the macro level, perceptions of corruption can undermine confidence in an economy and lead to downgrades by credit rating agencies, further discouraging investment. For instance, Brazil’s sovereign bonds were downgraded to junk by S&P in the aftermath of the Car Wash scandal, reducing the government’s ability to obtain credit in international markets. Highly publicized scandals also have the potential to spur institutional change, which may or may not strengthen a country’s ability to fight corruption. A country’s demonstrated performance in fighting corruption can, in turn, improve a country’s governance indicators, potentially influencing positively investors’ decisions. Thus, there are two sets of possible economic consequences of corruption scandals:  short-​term consequences, which involve institutional reforms and investors’ decisions in the aftermath of scandals; and long-​term consequences, which involve the effectiveness of reforms implemented, potentially in a rushed and ad hoc basis, and investors’ responses to the changes of governance indicators in the wake of such reforms (which could improve or worsen). While scandals have visible negative economic impacts in the short term, the long-​term consequences of scandals present as more complex and undetermined. As discussed in section 35.2, Brazil has followed a very clear pattern of implementing institutional reforms in response to large corruption scandals. While these institutional changes usually attempt to address apparent deficiencies, the sense of urgency may not favor reforms informed by careful and balanced analysis of policy alternatives to address the identified problems. Indeed, in the rush to respond promptly to the acrimony generated by a scandal, some of these measures may be guided by political calculations as to what is perceived to be an effective solution, rather than being based on an impartial, evidence-​based assessment of its potential effectiveness. As a result, there are significant uncertainties regarding the ultimate significance of such knee-​jerk reforms. An example is the restriction on corporate donations to political campaigns imposed shortly after the Car Wash scandal. There are fears that the reforms may have no effect on corruption, as corporations may be able to channel funds through individual donors or simply keep political “contributions” under the table. Another concern is that the new rules may merely increase the ability of incumbent politicians to retain their positions,

Corruption Scandals and Anti-Corruption Institutions    759 given the difficulty faced by potential challengers to raise sufficient funds to run a successful campaign (Kiernan 2015). Early indicators from the 2016 municipal elections suggest that it may have benefited rich candidates who can fund their own campaigns, as well as those supported by evangelical churches that can rally large numbers of individual donations (Boadle 2016). Whether or not this has had a corresponding effect on reducing or increasing corruption remains to be seen. In addition to the impact of individual reforms on corruption, one can also speculate about the overall impact of all these subsequent yet organic changes in the country’s ability to effectively combat corruption. The primary question is whether these cumulative changes eventually reduce corruption. While there is no conclusive evidence in either direction, the literature on the topic has been mostly positive, arguing that Brazil’s institutional changes since democratization have been effective in the fight against corruption. For example, Praça and Taylor (2014) assert that the slow march of institutional changes, such as those in response to the corruption scandals detailed in section 35.2, have created an environment in which accountability institutions now operate in a complex ecosystem that exists somewhat autonomously of the central bureaucracy and career politicians. This ecosystem consists of a “critical mass” of linkages between accountability institutions, allowing such institutions to cooperate, compete, reinforce each other, improve each other, and defend each other from political attacks. In this way, Brazilian accountability institutions have endogenously grown stronger through incremental change. While the authors note that institutional linkages can at times be fraught, they point to the increase in accountability operations, and especially joint operations, of Brazilian accountability institutions as indicators that these institutions are gaining strength and independence from those they investigate. Along similar lines, Prado and Carson (2016) argue that these organic and uncoordinated institutional changes have resulted in institutional multiplicity, which can be defined as the existence of multiple institutions with overlapping functions and purposes. Institutional multiplicity gives users of institutions the option of choosing the more honest institutions, while also limiting the ability of corrupt actors to monopolize the accountability process, creating reform without the necessity of collective action. In the specific case of Brazil, the patchwork of reforms over the past few decades has allowed the accountability institutions to escape capture by corrupt political forces, and also has created a situation in which actors seeking punishment for corruption can effectively try different institutions until the desired result is achieved. The authors suggest that some of Brazil’s recent success in prosecuting corruption may be attributable to the multitude of avenues for oversight, investigation, and punishment that institutional multiplicity has created over the past few decades. By asking a similar question about the interaction between domestic and institutional institutions, Davis, Jorge, and Machado (2015) conclude that international anti-​corruption institutions can similarly function as complements to domestic anti-​ corruption institutions, generating what they call “institutional complementarity.” They review an example from Argentina and from Brazil, including the TRT court construction case, and find that international institutions have been somewhat effective

760    Mariana Mota Prado and Lindsey Carson in supplementing domestic anti-​ corruption institutions. International institutions can relieve some of the burden of over-​matched domestic institutions, can assist in supplying key information or expertise to a prosecution, and can provide a method of circumventing corrupt or ineffective domestic institutions without being impeded by domestic politics. In performing these roles, foreign institutions may also directly or indirectly improve local institutions by demonstrating best practices, by providing additional resources, by identifying and publicizing local corruption internationally, or by providing constructive competition in the choice of proceedings. The authors also note some of the problems of institutional complementary, such as increased costs and duplication of efforts, as well as the uncertainty created when businesses might be faced with multiple processes in different jurisdictions. There is also the danger that international anti-​corruption institutions will simply replace and undermine domestic ones, rather than improve them. The authors, however, suggest that despite the inherent limitations of international institutional complements, they can serve a valuable function in ensuring the success of anti-​corruption efforts. Following the same caveats made by Davis about the negative impact of institutional complementarity, Winters (2015) argues that Brazil’s Clean Company Act is likely to be a failure. The author notes that the Act does not create a single new administrative mechanism to prosecute corruption, but rather completely decentralizes enforcement, as a host of agencies at various levels of government are empowered to pursue corruption investigations and impose sanctions. The author contrasts this with the centralized anti-​corruption institutions of the United States and the United Kingdom, and identifies a number of reasons why institutional multiplicity may prove problematic. These include a lack of consistency in procedures and interpretations of the law, leading to uncertainty for businesses and the opportunity for agencies and local governments to manipulate the law to their advantage. Smaller jurisdictions may also lack the technical knowledge to properly prosecute, potentially leading to liability imposed on innocents. Additionally, in the likely event of corruption spanning multiple levels of government, there is no clear mechanism for inter-​agency cooperation or procedures for determining who might lead the prosecution. Finally, and perhaps most important, the multiplicity may actually increase opportunities for corruption by placing enforcement power in the hands of leaders of smaller agencies and local governments who may have been involved in the corruption being pursued, or who may use their new enforcement power frivolously to collect fines, or use the threat of it to coerce bribes. The author therefore suggests that the Act should be amended to centralize enforcement power in a handful of closely related agencies. In contrast, Prado, Carson, and Correa (2016) argue that institutional multiplicity seems to be one of the strengths of the Act. By creating administrative mechanisms to prosecute corruption through a variety of agencies (in addition to the existing, and problematic, judicial processes), it allows actors seeking prosecution for corruption to effectively shop different institutions until one that achieves the desired anti-​corruption

Corruption Scandals and Anti-Corruption Institutions    761 effect can be found. It also creates the conditions by which agencies can collaborate, compete, and share knowledge, allowing for institutions to improve each other and hold each other accountable. The combination of these features prevents any one set of problematic institutional behaviors, such as endemic corruption, from becoming normalized across a society. The authors acknowledge that institutional multiplicity may exacerbate problems with interpretive and procedural inconsistency, lack of agency coordination, and limited expertise within diffuse agencies. The process may also simply be inefficient in that it may lead to a duplication of effort and an additional strain on state resources. However, despite these potential flaws, the authors see institutional multiplicity as an effective way of achieving anti-​corruption effects incrementally, and overcoming well-​ known problems in existing institutions, such as the judiciary, historically charged with enforcing corruption laws. In summary, there is much uncertainty about the long-​term impact of recent reforms, which seem to have generated institutional multiplicity intentionally (as seems to be the case of Brazil’s Clean Company Act) or unintentionally (as Brazil’s myriad of accountability institutions). While institutional multiplicity and complementarity may create uncertainty for businesses in the short term, it has the potential to improve governance indicators and effectively reduce corruption in the long term. Whether this will indeed be the case, remains to be seen.

35.5. Conclusion In the decades since democratization, corruption scandals have triggered significant institutional change in Brazil, although published reports of graft, embezzlement, and abuse of public power have not been the sole driver of change. On the contrary, Brazil’s Clean Company Act, enacted in 2013 and in effect since January 2014, is an example of a major institutional reform that seems to be mostly motivated by Brazil’s signature of the Organisation for Economic Co-​operation and Development (OECD) Convention on Combating Bribery of Foreign Public Officials in International Business Transactions, which requires that signatories impose liability for corrupt acts on corporations and other legal persons in accordance with the signatories’ legal principles (Prado, Carson, and Correa 2016, 114). By the same token, we do not suggest that scandals are the primary or sole driver of institutional change. Some authors would argue that while scandals create a window of opportunity for institutional change, certain preconditions are required for such a window to lead to effective change. Indeed, Alston et al. (2016) argue that most countries continue on auto-​pilot, with little meaningful development or institutional change, until shocks to the system, such as major corruption scandals, disrupt the dominant beliefs and networks of power. These shocks create windows of opportunity that can be used to quickly alter institutions at the constitutional level, based upon new sets of

762    Mariana Mota Prado and Lindsey Carson societal beliefs (such as the rule of law), but only if the necessary leadership emerges to push society toward the right set of beliefs. Such a leader, the authors argue, must recognize both the problem creating the shock and its solution, must be a deft strategist and manipulator of existing power structures, and must have moral authority in the eyes of the public. Of course, such windows of opportunity may also be used for regression, or can be missed entirely. However, the authors argue that Brazil has made effective use of its opportunities over the past few decades. They believe that Brazil has made considerable strides toward the necessary institutions and beliefs in the rule of law, social inclusion, and fiscal responsibility to make the most of its opportunities and undergo the transition to being a developed nation, even if Brazil’s economic growth has been slow and its performance on development indicators mediocre. Regardless of what ultimately drives institutional change, and the role that scandals play in it, it seems clear that scandals not only have a direct impact on individual behavior, but could also have a long-​term impact on a country’s governance system and overall levels of corruption. The impact of scandals on political behavior has been widely explored in the academic literature, while the same impact on business decisions by individual companies is just assumed, but not carefully explored. This chapter suggests at least two mechanisms that can affect business decisions in the aftermath of corruption scandals:  immediate institutional reform in response to deficiencies revealed by the uncovered corruption scheme, and perception-​based indicators. On the one hand, scandals can worsen the perception of corruption in a country, thus negatively affecting a country’s ability to attract investment. On the other hand, the institutional responses to corruption may generate positive perceptions about the country’s capacity to combat corruption, which can improve indicators and help attract investments. While harder to measure, there are yet two other dimensions in which scandals may affect a country’s ability to attract investment. One is the direct impact of scandals on investors’ behavior, which may or may not be mediated by perception-​based surveys. One could imagine, for instance, that the simple access to a report about a country being engulfed in a massive corruption scandal may scare off investors, independent of the impact of the scandal on perception-​based surveys and other indicators. A more indirect effect is related to the institutional transformations that take place in response to scandals. These reforms can generate at least three possible effects: (1) they can be ineffective, thus merely keeping constant the country’s ability to fight corruption and therefore not having any impact on investment decisions; (2) they can be very effective, enhancing the institutional framework for effectively combating corruption and ultimately having—​even if in the long term only—​a positive impact on the economy; or (3) they can be very effective in imposing sanctions, but their ad hoc and incremental nature may create a decentralized system and generate a lot of uncertainty for companies, thus ultimately having a negative impact on a country’s investment prospects. These hypotheses suggest that more attention should be devoted to exploring the direct and indirect impact of corruption scandals on a country’s economy.

Corruption Scandals and Anti-Corruption Institutions    763

Note 1. There are no data about the impact of Collorgate on the CPI because the indicator only started being produced in 1995.

References Aidt, T. S. 2009. “Corruption, Institutions, and Economic Development.” Oxford Review of Economic Policy 25 (2): 271–​291. Aizenman, J., and M. M. Spiegel. 2003. “Institutional Efficiency, Monitoring Costs, and the Investment Share of FDI.” Federal Reserve Bank of San Francisco Working Paper No. 2003-​06. Alencar, C. H. R., and I T. Gico, Jr. 2011. “When Crime Pays: Measuring Judicial Performance against Corruption in Brazil.” Law and Business Review of the Americas 17 (3): 415–​434. Almeida, M. A., and B. Zagaris. 2015. “Political Capture in the Petrobras Corruption Scandal: The Sad Tale of an Oil Giant.” Fletcher Forum Journal 39: 87–​100. Alston, L., M. Mello, B. Mueller, and C. Pereira. 2016. Brazil in Transition: Beliefs, Leadership and Institutional Change. Princeton, NJ: Princeton University Press. Amarribo Brasil. 2012. UN Convention against Corruption Civil Society Review: Brazil 2012. Available at http://​www.transparency.org/​files/​content/​publication/​206-​brazil-​full-​report. pdf, accessed September 22, 2013. Arantes, R. 2011. “The Federal Police and the Ministério Público.” In Corruption and Democracy in Brazil: The Struggle for Accountability, edited by T. J. Power and M. M. Taylor, Chapter 8, 218–​250. Notre Dame, IN: University of Notre Dame Press. Avritzer, L., 2008. “Índices de percepção da corrupção.” In Corrupção: ensaios e críticas, edited by L. Avritzer, N. Bignotto, J. Guimarães, and H. M.  M. Starling, 505–​510. Belo Horizonte: Editora UFMG. Avritzer, L. 2011. “O real combate à corrupção.” Carta Capital, September 4. Available at http://​ www.cartacapital.com.br/​sociedade/​combate-​a-​corrupcao-​e-​programa-​de-​governo. Balán, M. 2011. “Competition by Denunciation:  The Political Dynamics of Corruption Scandals in Argentina and Chile.” Comparative Politics 43 (4): 459–​478. Baldacci, E., Arye L. Hillman, and N. C. Kojo, N.C. 2004. “Growth, Governance, and Fiscal Transmission Policies in Low-​Income Countries.” European Journal of Political Economy 20: 517–​549. Bardhan, P. 1997. “Corruption and Development:  A Review of Issues.” Journal of Economic Literature 35 (3): 1320–​1346. Bertelsmann Stiftung. 2014. BTI 2014: Brazil Country Report. Gütersloh: Bertelsmann Stiftung. Available at http://​www.bti-​project.de/​uploads/​tx_​jpdownloads/​BTI_​2014_​Brazil.pdf. Bertelsmann Stiftung. 2016. BTI 2016: Brazil Country Report. Gütersloh: Bertelsmann Stiftung. Available at https://​www.bti-​project.org/​fileadmin/​files/​BTI/​Downloads/​Reports/​2016/​ pdf/​BTI_​2016_​Brazil.pdf. Boadle, A. 2016. “Millionaires, Evangelicals Benefit from Brazil Campaign Funds Ban.” Reuters, September 29. Available at http://​www.reuters.com/​article/​us-​brazil-​politics-​election​idUSKCN11Z2NO. Boas, T. C., F. D. Hidalgo, and N. P. Richardson. 2013. “The Spoils of Victory:  Campaign Donations and Government Contracts in Brazil.” Available at http://​people.bu.edu/​tboas/​ political_​investment.pdf.

764    Mariana Mota Prado and Lindsey Carson Bowler, S., and J. A. Karp. 2004. “Politicians, Scandals, and Trust in Government.” Political Behavior 26: 271. Brollo, F. 2009. “Who Is Punishing Corrupt Politicians—​Voters or the Central Government? Evidence from the Brazilian Anti-​corruption Program.” Available at http://​www.bu.edu/​ econ/​files/​2012/​11/​dp168.pdf. Brunetti, A., Gregory Kisunko, and Beatrice Weder. 1998. “Credibility of Rules and Economic Growth: Evidence from a Worldwide Survey of the Private Sector.” World Bank Economic Review 12: 353. Bulla, B. 2010. “CNJ age contra corrupção no Judiciário e gera reação corporativa de juízes.” Última Instância. Available at http://​ultimainstancia.uol.com.br/​conteudo/​noticias/​49686/​ cnj+age+contra+corrupcao+no+judiciario+e+gera+reacao+corporativa+de+juizes.shtml. Campos, J. E, Donald Lien, and Sanjay Pradhan. 1999. “The Impact of Corruption on Investment: Predictability Matters.” World Development 27: 1059. CGU. 2011. Orientações para implantação de unidades de corregedoria nos órgãos e entidades do poder executivo federal. Brasília: Controladoria Geral do União. CGU. 2014. Programa de fiscalização a partir de sorteios públicos [Random Audits Program]. Available at http://​www.cgu.gov.br/​english/​AreaAuditoriaFiscalizacao/​ ExecucaoProgramasGoverno/​Sorteios/​index.asp (English); http://​www.cgu.gov.br/​ controleinterno/​AvaliacaoExecucaoProgramasGoverno/​Sorteios/​index.asp (Portuguese). Control Risks Group Limited. 2006. International Business Attitudes to Corruption—​Survey 2006. Available at http://​www.csr-​asia.com/​summit07/​presentations/​corruption_​survey_​JB.pdf. Costas, R. 2014. “Petrobras Scandal: Brazil’s Energy Giant under Pressure. BBC. November 21. Available at http://​www.bbc.com/​news/​business-​30129184. Costas-​Pérez, E., A. Solé-​Ollé, and P. Sorribas-​Navarro. 2012. “Corruption Scandals, Voter Information, and Accountability.” European Journal of Political Economy 28 (4): 469–​484. Davis, K. E., G. Jorge, and M. R Machado. 2015. “Transnational Anticorruption Law in Action: Cases from Argentina to Brazil.” Law and Social Inquiry 40 (3): 664–​699. Dow Jones and MetricStream. 2016. Global Anti-​ Corruption Survey Results 2016. Available at https://​dow-​jones-​2097.docs.contently.com/​v/​ 2016-​dow-​jones-​and-​metricstream-​global-​anti-​corruption-​survey. Edelman. 2016. 2016 Edelman Trust Barometer:  Global Results. Available at http://​www.edelman.com/​insights/​intellectual-​property/​2016-​edelman-​trust-​barometer/​global-​results/​. Época. 2008. Por que o Brasil pode vencer a corrupção: Special “Debate” Supplement, July 22. Available at http://​revistaepoca.globo.com/​Revista/​Epoca/​1,,EMI5188-​15273,00.html Ehrlich, I., and F. T. Liu. 1999. “Bureaucratic Corruption and Endogenous Economic Growth.” Journal of Political Economy 107 (S6): S270–​S293. Federação das Indústrias do Estado de São Paulo (FIESP). 2010. “Relatório corrupção: Custos econômicos e propostas de combate.” Available at http://​www.fiesp.com.br/​indices-​ pesquisas- ​ e -​ publicacoes/​ relatorio-​ c orrupcao- ​ c ustos- ​ e conomicos- ​ e - ​ propostas- ​ d e-​ combate/​. Ferraz, C., and F. Finan. 2008. “Exposing Corrupt Politicians: The Effects of Brazil’s Publicly Released Audits on Electoral Outcomes.” Quarterly Journal of Economics 123 (2): 703–​745. Ferraz, C., and F. Finan. 2011. “Electoral Accountability and Corruption: Evidence from the Audits of Local Governments.” American Economic Review 101 (4): 1274–​1311. Fleischer, D. 2000. “Além do Collorgate: Perspectivas de consolidação democrática no Brasil via reformas políticas.” In Corrupção e reforma política no Brasil: O impacto do impeachment de Collor, edited by K. Rosenn and R. Downes, 81–​110. Rio de Janeiro: Editora FGV.

Corruption Scandals and Anti-Corruption Institutions    765 Friedman, E., S. Johnson, D. Kaufmann, and P. Zoido-​Labaton. 2000. “Dodging the Grabbing Hand: The Determinants of Unofficial Activity in 69 Countries.” Journal of Public Economics 76: 459–​493. Global Integrity. 2009. Global Integrity Scorecard:  Brazil 2009. Washington, DC:  Global Integrity. Available at http://​report.globalintegrity.org/​reportPDFS/​2009/​Brazil.pdf, accessed September 22, 2013. Golden, M. A., and L. Picci. 2005. “Proposal for a New Measure of Corruption, Illustrated with Italian Data.” Economics and Politics 17 (1): 37–​75. Gray, C. W., and D. Kaufmann. 1998. “Corruption and Development.” Finance and Development 35 (1): 7–​10. Greenwald, G. 2016. “In Brazil, Major New Corruption Scandals Engulf the Faction That Impeached Dilma.” The Intercept, November 25. Available at https://​theintercept.com/​ 2016/​11/​25/​in-​brazil-​major-​new-​corruption-​scandals-​engulf-​the-​faction-​that-​impeached-​ dilma/​. Gupta, S., H. Davoodi, and R. Alonso-​Terme, R. 2002. “Does Corruption Affect Income Inequality and Poverty?” Economics of Governance 3: 23–​45. Gupta, S., L. de Mello, and R. Sharan. 2001. “Corruption and the Provision of Health Care and Education Services.” In The Political Economy of Corruption, edited by A. K. Jain, 111–​141. London: Routledge. Habib, M., and L. Zurawicki. 2002. “Corruption and Foreign Direct Investment.” Journal of International Business Studies 33 (2): 291–​307. IBA. 2016. The International Bar Association Judicial Integrity Initiative:  Judicial Systems and Corruption. Available at http://​www.ibanet.org/​Legal_​Projects_​Team/​ judicialintegrityinitiative.aspx. Iquiapaza, R., and H. Amaral. 2007. “Reflections on the Impact of Corruption on Economic Development:  A Literature Review in the Brazilian Economy.” MPRA Paper No. 1818. Available at http://​mpra.ub.uni-​muenchen.de/​1818/​. Johnson, S., D. Kaufmann, J. McMillan, and C. Woodruff. 2000. “Why Do Firms Hide? Bribes and Unofficial Activity after Communism.” Journal of Public Economics 76: 495–​520. Kaufmann, D., A. Kraay, and M. Mastruzzi. 2006. “Measuring Governance Using Cross-​ Country Perceptions Data.” In The International Handbook on the Economics of Corruption, edited by S. Rose-​Ackerman, 52–​104. Cheltenham: Edward Elgar. Kaufmann, D., A. Kraay, and M. Mastruzzi. 2009. “Governance Matters VIII: Aggregate and Individual Governance Indicators 1996–​2008.” World Bank Policy Research Working Paper No. 4978. Kenny, C., M. Klein, and M. Sztajerowska. 2011. “A Trio of Perspectives on Corruption: Bias, Speed Money, and ‘Grand Theft Infrastructure.’” World Bank Policy Research Working Paper No. 5889. Available at https://​openknowledge.worldbank.org/​handle/​10986/​3657. Kiernan, P. 2016. “Brazil Supreme Court Bans Corporate Donations to Politicians and Parties.” The Wall Street Journal, September18. Available at http://​www.wsj.com/​articles/​ brazil-​supreme-​court-​bans-​corporate-​donations-​to-​politicians-​and-​parties-​1442616836. Knack, S. 2007. “Measuring Corruption:  A Critique of Indicators in Eastern Europe and Central Asia.” Journal of Public Policy 27 (3): 255–​291. Knack, S., and P. Keefer. 1995. “Institutions and Economic Performance: Cross-​Country Tests Using Alternative Institutional Measures.” Economics and Politics 7 (3): 207–​227. Lambsdorff, J. G. 2000. “The Precision and Regional Comparison of Perceived Levels of Corruption:  Interpreting the Results.” Background Paper to the 2000 Corruption

766    Mariana Mota Prado and Lindsey Carson Perceptions Index. Transparency International and Göttingen University. Available at http://​www.icgg.org/​downloads/​2000_​CPI_​Precision.pdf. Lambsdorff, J. G. 2003. “How Corruption Affects Persistent Capital Flows.” Economics of Governance 4 (3): 229–​243. Lambsdorff, J. G. 2007. The Institutional Economics of Corruption and Reform: Theory, Evidence, and Policy. Cambridge: Cambridge University Press. Li, H., L. C. Xu, and H.-​F. Zou. 2000. “Corruption, Income Distribution, and Growth.” Economics and Politics 12 (2): 155–​182. Loureiro, M. R., F. L. Abrúcio, C. Olivieri, and M. A. Teixeira. 2012. “Do controle interno ao controle social: A múltipla atuação da CGU na Democracia Brasileira.” Cadernos Gestão Pública e Cidadania, São Paulo 17 (60): 54–​67. Mauro, P. 1995. “Corruption and Growth.” Quarterly Journal of Economics 110 (3): 681–​7 12. Mauro, P. 1998. “Corruption and the Composition of Public Expenditure.” Journal of Public Economics 69: 263–​279. Mechanism for Follow-​Up on the Implementation of the Inter-​American Convention against Corruption (MESICIC). 2012. Federative Republic of Brazil Final Report. Melo, M. A. 2013. “Brazil:  Democracy and Corruption.” In Democracy Consensus Project Seminar. Legatum Institute, Rio de Janeiro, May 15, 40. Melo, M. A., C. Pereira, and C. M. Figueiredo. 2009. “Political and Institutional Checks on Corruption:  Explaining the Performance of Brazilian Audit Institutions.” Comparative Political Studies 42: 1217–​1244. Miranda, R. F., C. Silva, and F. de Souza Freire. 2016. “Corruption in Public Management in Brazil:  A Hidden, Regional Perspective.” In Corruption, Economic Growth, and Globalization, edited by A. A. C. Teixeira, C. Pimenta, A. Maia, and J. A. Moreira, 150–​170. New York: Routledge. Mo, P. H. 2001. “Corruption and Economic Growth.” Journal of Comparative Economics 29: 66–​79. MPF (Ministério Público Federal). 2016. “A Lava Jato em números.” Available at http://​lavajato. mpf.mp.br/​atuacao-​na-​1a-​instancia/​resultados/​a-​lava-​jato-​em-​numeros-​1. Musacchio, A., and S. G. Lazzarini. 2014. Reinventing State Capitalism: Leviathan in Business, Brazil and Beyond. Cambridge, MA:  Harvard University Press. Introduction available at http://​people.hbs.edu/​amusacchio/​downloads/​reinventing_​state_​capitalism.pdf. Nicholson, B. 2016. “Brazil’s Fight against Corruption.” IBA Global Insight, April–​May. Available at http://​www.ibanet.org/​Article/​Detail.aspx?ArticleUid=A3D435C5-​FB84-​48F2-​ 9BE5-​E03D7B312F1A. Nye, J. S. 1967. “Corruption and Political Development: A Cost-​Benefit Analysis.” American Review of Political Science 61 (2): 417–​427. OECD (Organisation for Economic Co-​operation and Development). 2011. Convention on Combating Bribery of Foreign Public Officials in International Business Transactions. https://​www.oecd.org/​daf/​anti-​bribery/​ConvCombatBribery_​ENG.pdf OECD (Organisation for Economic Co-​operation and Development). 2015. Data. Available at https://​data.oecd.org/​gdp/​gross-​domestic-​product-​gdp.htm#indicator-​chart. Oliveira, G. J., J. G. Rodas, D. A. Guimaraes, F. S. Marcato, A. Curi, R. F. Pastore, and T. Farina. 2015. “Working Paper on Brazilian Cartel Investigations, the ‘Operation Car Wash’ (Operação Lava Jato) and Its Impacts on the Economy.” Group of Infrastructure Economics and Environmental Solutions; CEDES—​Center of Studies on Economic and Social Law. April. Available at SSRN, https://​ssrn.com/​abstract=2710438.

Corruption Scandals and Anti-Corruption Institutions    767 Power, T. J., and M. M. Taylor, eds. 2011. Corruption and Democracy in Brazil: The Struggle for Accountability. Notre Dame, IN: University of Notre Dame Press. Praça, S., and M. M. Taylor. 2014. “Inching towards Accountability: The Evolution of Brazil’s Anticorruption Institutions, 1985–​2010.” Latin American Politics and Society 56 (2): 27–​48. Prado, M. M., and L. Carson. 2016. “Using Institutional Multiplicity to Address Corruption as a Collective Action Problem: Lessons from the Brazilian Case.” The Quarterly Review of Economics and Finance 62: 56–​65. Prado, M. M., L. Carson, and I. Correa. 2016. “The Brazilian Clean Company Act:  Using Institutional Multiplicity for Effective Punishment.” Osgoode Hall Law Journal 53: 107–​163. Richter, A. 2016. “Ministério Público pede prisão do ex-​senador Luiz Estevão.” Agência Brasil, February 23. Available at http://​agenciabrasil.ebc.com.br/​geral/​noticia/​2016-​02/​ ministerio-​publico-​pede-​prisao-​do-​ex-​senador-​luiz-​estevao. Rose-​Ackerman, S. 1999. Corruption and Government:  Causes, Consequences, and Reform. Cambridge: Cambridge University Press. Sadek, M. T. 2008. “Ministério Público.” In Corrupção: Ensaios e críticas, edited by L. Avritzer, N. Bignotto, J. Guimarães, and H. M. M. Starling, 454–​460. Belo Horizonte: Editora UFMG. Sadek, M. T., and R. B. Cavalcanti. 2003. “The New Brazilian Public Prosecution: An Agent of Accountability.” In Democratic Accountability in Latin America, edited by S. Mainwaring and C. Welna, 201–​227. Oxford: Oxford University Press. Shleifer, A., and R. W. Vishny. 1993. “Corruption.” Quarterly Journal of Economics 108 (3): 599–​617. Silva, M. F. G., F. Garcia, and A. C. Bandeira. 2001. “How Does Corruption Hurt Growth? Evidences about the Effects of Corruption on Factors Productivity and Per Capita Income.” Available at http://​bibliotecadigital.fgv.br/​dspace/​bitstream/​10438/​1953/​1/​TD103.pdf. Speck, B. W. 2000. “Mensurando a corrupção: Uma revisão de dados provenientes de pesquisas empíricas.” In Cadernos Adenauer 10:  Os custos da corrupção, 1–​40. São Paulo:  Fundção Konrad Adenauer. Speck, B. W. 2011. “Auditing Institutions.” In Corruption and Democracy in Brazil: The Struggle for Accountability, edited by T. J. Power and M. M. Taylor, Chapter 6, 127–​161. Notre Dame, IN: University of Notre Dame Press. Swaleheen, M. 2011. “Economic Growth with Endogenous Corruption: An Empirical Study.” Public Choice 146: 23–​41. Taylor, M. M. 2009. “Corruption, Accountability Reforms and Democracy in Brazil.” In Corruption and Democracy in Latin America, edited by C. H. Blake and S. Morris, 150–​168. Pittsburgh: University of Pittsburgh Press. Taylor, M. M., and Vinícius C. Buranelli. 2007. “Ending Up in Pizza:  Accountability as a Problem of Institutional Arrangement in Brazil.” Latin American Politics and Society 49 (1): 59–​87. Transparency International. 2007. Global Corruption Report—​Corruption in Judicial Systems. https://​www.transparency.org/​whatwedo/​publication/​global_​corruption_​report_​2007_​ corruption_​and_​judicial_​systems Transparency International. 2011a. Bribe Payers’ Index. https://​www.transparency.org/​ whatwedo/​publication/​bpi_​2011 Transparency International. 2011b. Global Corruption Barometer 2010/​2011. https://​www.transparency.org/​gcb201011. Transparency International. 2012. Putting Corruption Out of Business. Data available at http://​ www.transparency.org/​research/​bps2011.

768    Mariana Mota Prado and Lindsey Carson Transparency International. 2013a. Corruption Perceptions Index. https://​www.transparency. org/​cpi2013. Transparency International. 2013b. Global Corruption Barometer 2013. https://​www.transparency.org/​gcb2013. Treisman, D. 2000. “The Causes of Corruption:  A Cross-​National Study.” Journal of Public Economics 76: 399–​457. Treisman, D. 2007. “What Have We Learned about the Causes of Corruption from Ten Years of Cross-​National Empirical Research?” Annual Review of Political Science 10: 211–​244. TSE. 2015. “President of TSE Talks about Electoral Campaign Funding at an Event in the Dominican Republic.” Tribunal Superior Eleitoral, January 30. Available at http://​www.tse. jus.br/​imprensa/​noticias-​tse/​2015/​Janeiro/​presidente-​do-​tse-​fala-​sobre-​financiamento-​de-​ campanha-​eleitoral-​em-​evento-​na-​republica-​dominicana. Wei, S.-​J. 2000. “How Taxing Is Corruption on International Investors?” Review of Economics and Statistics 1 (11): 1–​11. Welchans Research Group. 2014. “Dow Jones Anti-​ Corruption Survey Results 2014.” Available at http://​images.dowjones.com/​company/​wp-​content/​uploads/​sites/​15/​2014/​04/​ Anti-​Corruption-​Survey-​Results-​2014.pdf Winters, M. A. 2015. “Too Many Cooks in the Kitchen: Battling Corporate Corruption in Brazil and the Problems with a Decentralized Enforcement Model.” Richmond Journal of Global Law and Business 13 (4): 651–​700. World Bank. 2000. “Anti-​ corruption in Transition.” Available at http://​siteresources. worldbank.org/​INTWBIGOVANTCOR/​Resources/​contribution.pdf. World Bank. 2009. “Enterprise Survey: Country Report on Brazil.” Available at http://​www. enterprisesurveys.org/​~/​media/​FPDKM/​EnterpriseSurveys/​Documents/​Profiles/​English/​ brazil-​2009.pdf. World Bank. 2016. “Development Indicators.” Available at http://​databank.worldbank.org/​ data/​home.aspx. World Economic Forum. 2013. The Global Competitiveness Report 2013–​2014: Full Data Edition. Available at http://​www3.weforum.org/​docs/​WEF_​GlobalCompetitivenessReport_​2013-​14.pdf. You, J., and S. Khagram. 2005. “A Comparative Study of Inequality and Corruption.” American Sociological Review 70 (1): 136–​157. Zurawicki, L., and M. Habib. 2010. “Corruption and Foreign Direct Investment: What Have We Learned?” International Business and Economics Research Journal 9 (7): 1–​10.

Index

Page numbers followed by f and t indicate figures and tables, respectively. Numbers followed by n indicate notes ABC (Agência Brasileira de Cooperação, Brazilian Cooperation Agency), 536, 538–41, 543, 548 Abertura Comercial (Trade Opening), 98 ABM Amro, 216n32 abusive practices, 727 ACC (Merger Control Agreement), 726 Acre, 305–6n17 action plan for combating deforestation in the Amazon (PPCDAm), 583, 585 activity fronts, 291 Administrative Council for Economic Defense. See Conselho Administrativo de Defesa Econômica (CADE) Aerospace Technology Centre (CTA), 690–91 AES (Applied Energy Services), 137–38 Afghanistan, 662n2 Africa, 149 imports from Brazil, 408, 409t share of world exports, 407, 407t slave trade, 19–21, 27–28, 28t, 36n12, 268 South-South cooperation, 535–51 Africa-Brazil Program for Social Development, 543–44 African Union (AU), 543 Agência Brasileira de Cooperação (ABC, Brazilian Cooperation Agency), 536, 538–41, 543, 548 Agência Nacional de Saúde (ANS, National Health Agency), 601–2, 610 Agency for Studies and Projects (Financiadora de Estudos e Projetos, FINEP), 505, 689–90, 712 AGF program, 278–79, 282 agribusiness, 284–85 Certificate of Agribusiness’s Receivable Assets (CRA), 282, 285 Notes of Agribusiness Credit (LCA), 282

agricultural frontier, 291–92, 302–4 agricultural research, 312, 314–15, 319–24, 332n12, 406 agricultural technology, 12, 272, 295 agricultural trade, 410–12, 652–54, 653t Agricultural Warrant (WA), 282 agriculture, 41–42, 266–87. See also specific crops activity fronts, 291 areas of cultivated land, 270–72, 270t average earnings, 566 average hours worked per week (AHWPW), 568 census indicators, 272–73, 274t climate change and, 587 colonial cycles, 268–69 competitiveness, 411 croplands, 272–73, 274t crop production, 276, 276f, 277, 277t, 297 demand for products, 328, 329t diversification, 272–75 employment, 564–65t, 566 expansion of, 275–85, 292–94, 297, 311 exports, 21, 266, 271t, 272–73, 281–85, 281f, 311, 635, 635t family farming, 283 foreign direct investment in, 674, 675t foreign-market-oriented, 272, 278, 311 GDP growth, 297 government expenditures, 282–83 historical perspectives, 20–21, 32–34, 42–43, 58n44, 268–69, 272, 291–97 imports, 281–82, 281f incentives for, 334n24 informal workers, 557–61, 559f inland, 272–75 land distribution, 289, 290t

770   index agriculture (cont.) low pay and wage inequality, 567, 567t mechanization, 313 modernization, 292–93, 296–304, 309–37 productivity, 325–28, 326t, 343, 343f, 346, 347f, 411 real agricultural product, 92–93 regional disparities, 426–29, 427t, 429–30 science-based, 315–28 share of GDP, 93–94, 93t, 102, 266, 340–41, 341f share of labor, 93–94, 93t, 102 slave labor, 28, 28t, 32 small producers, 298–300 storage capacity, 285 support for, 298 sustainability of, 325–27 traditional, 288–308 tropical, 293–94 types of, 327–28 union density, 568 value chains, 326 working poverty, 566 world exports, 407, 407t Agronomic Institute, 273 Agropensa (Embrapa’s Strategic Intelligence System), 320 agroprocessing, 281–82, 281f Ahlstrom Corporation, 733 aircraft industry, 675, 690–91 air pollution, 578, 582, 590n7 airports contemporary challenges, 384–85 PAC program, 386–87 Projeto Crescer projects, 389 SOEs, 712, 713f Alcohol and Sugar Institute (IAA), 117 Alfonsin, Raul, 648 Alliance for Progress, 67–68 altruism, 548 Alves, Eliseu, 10, 301 Amann, Edmund, 12–13 Amazon, 581, 583–86, 585f, 590n8 Amazonas, 305–6n17, 434, 438–39 Amazônia, 303–4 AMFORP (American and Foreign Power Company), 358

Amorim, Celso, 641n30 Anaconda Operation, 748–49, 756 Ananias, Patrus, 543–44 Andean Community, 417n7 ANEEL, 139 Anglo-Saxon countries, 562, 563t Angola, 187, 536, 541, 691–92 Angra I, 365 Angra II, 365 animal hides, 43t anões do orçamento (budget dwarves), 747, 756 Anonymous Societies, 202 ANP, 139 ANPEC (Brazilian Association of Graduate Programs in Economics), 81 ANS (Agência Nacional de Saúde, National Health Agency), 601–2, 610 Antarctica-Brahma, 730 anti-competitive behavior, 727 anti-corruption efforts, 754–55, 758–59, 762 anti-corruption institutions, 741–68 anti-dumping investigations, 403 anti-globalization, 656–61 anti-poverty transfers, 511–34 Antitrust Agency of the Americas award, 736 anti-trust policy, 718–40 apparel (clothing) main exports, 635, 635t production index, 257t Applied Energy Services (AES), 137–38 Aracruz, 138–39, 688 Arbache, Jorge, 11–12 Argentina, 9, 130, 142n5, 541, 691–92 anti-corruption efforts, 759–60 Brazilian MNCs in, 692–93 Brazilian population, 655–56, 656t comparison with Brazil, 128 crisis of 2001–2002, 209, 237 economic growth before WWI, 53–54 export share, 651f health expenditures, 609 immigration from, 654 import substitution, 76–78 investment in, 187 literacy, 490 per capita GDP, 50, 56n20

index   771 per capita income growth, 156t publications, 506 railroads, 379 school enrollment, 50 services sector, 563t, 569 trade restrictiveness, 652–54, 653t Treaty of Asunción, 647t urbanization rate, 50 Armenia, 640n24 Arrow, Kenneth, 593 Arroyo Abad, Leticia, 47 ASEAN (Association of Southern Asian economies), 626, 640n24 Asia, 149, 166–67, 407, 407t Asian Tigers, 149. See also specific countries Asia Pacific Trade Agreement, 626 assassinations, 305–6n17 Association of Southern Asian Economies (ASEAN), 626, 640n24 associative contracts, 737n9 Attorney General. See Procuradoria Geral AU (African Union), 543 Auditoria Geral da União (General Accounting Office), 748 Australia cooperation with, 318 economic growth before WWI, 53 exports, 171n17, 651f migrant population, 654–56, 655t per capita income growth, 156t services sector, 563t trade restrictiveness, 652–54, 653t Austria, 341–42, 341f, 563t, 692–93 Austrian School, 126–27 automobiles, 1–2, 52, 236, 359, 398–99 ethanol engines, 363–65 flex fuel vehicles, 364 foreign direct investment in, 674 performance indicators, 251 production index, 257t, 260 requirements for fuel efficiency, 374 tax rebates, 404 Auxilio Gás, 519 average hours worked per week (AHWPW), 568 Ayrton Senna Institute, 500 Azzoni, Carlos, 13

Bacha, Carlos, 10 Baer, Werner, 6–7, 13, 72, 84n42, 103n8, 612n1, 645 Bahia agriculture, 272, 283–84, 314 coffee production, 279, 279t, 280, 284 colonial period, 20–21, 23, 29, 31–32 credit before banks, 200 first private banks, 202 petroleum, 359 slavery, 29 sugar industry, 31–32 bail-out operations, 184 Bair, Scott, 640n23 balance of payments, 404, 412–14, 412f Banco Central do Brasil (BCB, Central Bank of Brazil), 109, 112–13, 206–8, 214, 222 Focus Report, 228 PROER program, 209 PROES program, 209–10 Banco Comercial do Rio de Janeiro, 202 Banco da Amazônia, 193n6, 712 Banco da República do Brasil, 203 Banco do Brasil (BB), 110, 112–13, 133, 201–7, 211, 215n15, 215–16n29, 396, 688 Banco do Brasil Group, 712 Banco do Ceará, 202 Banco do Nordeste do Brasil (Bank of the Northeast of Brazil), 193n6, 458–59, 462n5, 712 Banco Nacional da Habitação (BNH, National Bank of Housing), 215n21 Banco Nacional de Desenvolvimento Econômico (BNDE, National Bank of Economic Development), 66–67, 70, 92, 132, 143n14, 216n37 Banco Nacional de Desenvolvimento Econômico e Social (BNDES, National Bank for Economic and Social Development), 80–82, 100, 103n3, 126, 132–33, 137–40, 143n14, 177–97, 216n37, 387, 549, 705, 708–14 disbursements, 177, 182–83, 182f, 186–90, 194n14, 194n24, 211–12, 212f establishment, 181, 193n7 fiscal costs, 190–92, 194n23 investment maintenance program (PSI), 186, 190

772   index Banco Nacional de Desenvolvimento Econômico e Social (BNDES, National Bank for Economic and Social Development) (cont.) lending, 216n39, 389, 688–90, 710 private companies, 713–14 subsidies, 178, 186, 190–93, 191f, 404 total assets, 177, 193n4 Banco Regional de Desenvolvimento do Extremo Sul, 193n6 Banco Santander, 210, 216n32 Bandeira, Moniz, 737n2 Bandeirante aircraft, 690 bandeirantes (fortune hunters), 269 bandeiras (expeditions), 22 Banespa (São Paulo State Bank), 210 Bank for International Settlements (BIS), 210–11 banking system, 45, 56n17, 198–220 central banks, 106 developmental banks, 178–81 financial fragility, 212–13, 213f foreign banks, 209f, 210–11, 211f non-earmarked credit operations, 209f, 210 private banks, 211, 211f public banks, 211, 211f reform and restructuring, 109, 208–11 total earmarked credit, 211–12, 212f total lending, 211, 211f, 212 Bank of Brazil, 45 Bank of England, 112–13 Banks of Royal Credit, 202 Baptists, 499 Barbalho, Jader, 748 Barbosa, Rui, 203, 458 Barrientos, Armando, 12–13 Baumol cost disease, 355n19 BB. See Banco do Brasil BCB. See Banco Central do Brasil BDC (Business Development Bank of Canada), 181 beans crop production, 270–72, 270t, 277, 277t, 311, 322–23 domestic supply, 301 imports, 276 productivity gains, 325–26, 326t

beef cattle, 325, 326t, 411 Belarus, 369–70, 370t, 640n24 Belgium, 651f, 655–56, 656t Belo Horizonte, 424, 590n7 Belo Monte hydropower plant, 367, 580–81, 714 Benchmark Interest Rate (Taxa Referencial, TR), 120 Benefício da Prestação Continuada (BPC, Continued Provision Benefit), 474, 511, 515f, 516–18, 526–30, 531n12 main components, 520–21, 520t transfers, 517, 517f, 518, 522, 523f Beneficio de Superação da Extrema Pobreza, 521 Benin, 541 Bergstrand, Jeffrey, 640n23 Bertelsmann Stiftung, 754–55 beverage industry. See food and beverages BG group, 692 Big Mac index, 262n10 big monthly stipend (mensalão) scandal, 227, 749–50, 754–56 bill of exchange (letra de câmbio), 205 biodiesel production, 361 biofuels, 374 biomass, 374, 375t BIS (Bank for International Settlements), 210–11 “black beans and rice” economic policy, 115 black book, 314 blacks, 555t, 556 bloodsucker (sanguessuga) scandal, 756 BNDE (Banco Nacional de Desenvolvimento Econômico, National Bank of Economic Development), 66–67, 70, 92, 132, 143n14, 216n37 BNDES. See Banco Nacional de Desenvolvimento Econômico e Social BNDES-Ex (BNDES Exportações), 689–90 BNDESPar, 137 BNDESPAR, 177, 184, 709 BNH (Banco Nacional da Habitação, National Bank of Housing), 215n21 Boianovsky, Mauro, 65 Bolivia, 417n7, 536, 626, 654–56, 656t, 691–93 Bolsa Alimentação, 519 Bolsa Escola (School Stipend), 474, 515f, 518–19, 527–29, 531n12

index   773 Bolsa Familía (PBF, Family Stipend), 13, 248, 262n11, 474–76, 511–14, 515f, 518–22, 527–30, 560 Conditional Cash Transfer (CCT) program, 2, 10 costs of, 190 main components, 520t, 521 number of participating households, 519, 519f target households, 531n14 transfers, 520t, 521, 524, 524f, 531n17 books, 36n9 bourgeoisie, state, 128 BPC. See Benefício da Prestação Continuada Bradesco, 133 Brady Plan, 223 Bragantina, 584–85 Branco, Castello, 108 brand identity, 682–83 Brasil Carinhoso, 512, 515f Brasil Foods, 724 Brasilía, 7, 690–91 Brasil Sem Miséria, 512 Braskem, 695–96 Braudel, Fernand, 64 Brazil. See also BRICS (Brazil, Russia, India, China, South Africa) countries Administrative Council for Economic Defense (Conselho Administrativo de Defesa Econômica, CADE), 732–33 agriculture, 266–308, 290t, 309–37 anti-corruption efforts and institutions, 741–68 antitrust and competition policy, 718–40 automotive regime, 236 balance of payments, 404 banking system, 45, 56n17, 198–220 BTI Transformation Index ranking, 754–55 capital inflow, 412f, 413 Civil Code, 581 Clean Company Act, 760–61 Código de Defesa do Consumidor, 722 coffee production, 279–80, 279t. See also coffee colonial period, 17–39, 70–71, 199–203, 268–69 Constitution, 119, 129–30, 206–7, 296, 305n5, 474, 511–12, 529, 542, 581, 597, 610, 721–22, 746–49

contemporary, 382–85, 650–56 Control of Corruption (CC) index ranking, 753, 753f cooperation with Africa, 535–51 corruption, 741–68, 753f current account, 233–34, 233f, 412, 412f debt, 98–99, 639n9 demand for agricultural products, 328, 329t Departamento Nacional de Obras contra a Seca (DNOCS), 457 desenvolvimento (development), 5–10, 12–13 developmentalism, 6–7, 63–88, 141, 185, 232, 378, 451–52 economic growth, 154–57, 155f, 161–68, 162f, 246–49, 377 economic integration, 433–34 economic miracle (1967–1974), 96, 147–48, 207, 401, 401t, 468 economic performance (1967–1980), 401, 401t economy, 1–174, 639n1, 650–56, 699–768 education, 489–510 employment, 570–71 energy, 358–76 energy consumption, 369, 371, 372t, 373f energy supply, 360–61, 360f, 361f, 362f, 365 Estado Nôvo (New State), 6, 91, 130–31, 214–15n13, 703, 720–21 European settlers, 19 export price index, 649–50, 649f export revenue, 649–50, 649t exports, 1–2, 41–44, 43t, 57n27, 90, 406–10, 407t, 409t, 415, 631t, 632, 635, 635t, 650–52, 651f, 651t external economic policy, 646–50, 647t Extraordinary Ministry for Zero Hunger, 519 Federal Police (Departamento da Polícia Federal, DFP), 747–49, 756–57 Federal Public Prosecutors’ Office (Ministério Público Federal, MPF), 726, 746, 749, 756–57 financial policy, 71 financial sector, 118–19 financial system, 204–8 First Republic, 40–45, 50–52, 58n39, 203–4, 269–72, 703, 721–22

774   index Brazil. See also BRICS (Brazil, Russia, India, China, South Africa) countries (cont.) foreign debt, 45–46, 229, 233–34, 233f, 400–401, 401t, 413–14, 413f foreign direct investment, 1, 8, 379, 386, 398, 415, 636–37, 637t, 664–81, 674t, 675t, 676t, 678f, 679, 685–89, 685f, 686t foreign exchange reserves, 91, 262n6 foreign policy, 535–36, 538–39 GDP, 17, 56nn20–21, 92–94, 93t, 150, 150f, 577, 631–32, 631t, 641n35 Geisel government, 116, 222 General Accounting Office (Auditoria Geral da União), 748 geography, 423–24, 594–95 global role, 1 global trade, 408 health economics, 593–618 health inequalities, 602–7, 603f health resources, 597–99 health system, 596–602 historical perspectives, 15–174 hospital beds, 597–98, 598t hospitals, 598–99, 598t imperial period, 40–41, 44–50, 56n19, 269–72 imports, 90 import-substitution industrialization, 7–8, 77–78, 89–104 income distribution, 467–88 income inequality, 577 industrial growth, 89–90 industrial policy, 100–101 infrastructure, 377–93 inland occupation, 272–75 Institutional Act No. 5, 468 internal debt (dívida interna), 207 international prospects, 645–63 international support, 732–33 Joint Budget Committee of Congress (Comissão Mista do Orçamento, CMO), 747 judicial districts (comarcas), 23 labor force, 92–94, 93t labor market, 552–76, 555t, 571t labor productivity, 338–39, 339f, 339t land distribution, 288–308, 290t

Land Law, 44 land tenure, 289, 290t Law 8.844/94, 722–25, 736, 737n4 Law 12.529/11, 725–31, 736 Law 8886/93 (Government Procurement Act), 746 Law of Economic Readjustment (Lei do Reajustamento Econômico), 205 Law of Fiscal Responsibility (Lei de Responsabilidade Fiscal), 134, 143n15, 435 Law of Similars (Lei de Similaridade Nacional), 647t Lei da Cláusula-Ouro (Gold Clause Law), 198, 205 Lei de Usura (Usury Law), 198, 205 Lei Orgânica da Assistência Social (LOAS), 517 literacy rates, 44, 50, 55n13 long-term growth, 52–54, 147 macroeconomics, 175–240, 623 manufacturing, 243–65, 253t, 338–57 merchandise exports, 44–45 military government (1964–1985), 206–8, 275–80, 276f, 399, 483, 721 Ministério de Previdência Social, 516, 520t Ministry of Agrarian and Social Development (Ministério do Desenvolvimento Social e Agrário, MDSA), 283, 305n12, 540–41 Ministry of Agrarian Development (Ministério do Desenvolvimento Agrário, MDA), 282, 300, 305n12, 530n3, 540 Ministry of Agriculture, 314, 332n8 Ministry of Agriculture, Livestock and Food Supply (MAPA), 282 Ministry of Education, 491, 505, 518–19 Ministry of External Relations (Ministério das Relações Exteriores, MRE) (Itamaraty), 536, 539–42 Ministry of Finance, 722, 724–25, 746 Ministry of Health, 519, 541, 544, 599, 610, 611f Ministry of Justice, 722–25, 734, 749 Ministry of Labor and Social Assistance, 531n13 Ministry of Mines and Energy, 519 Ministry of Planning, 505, 547, 704–5, 707

index   775 Ministry of Science, Technology and Innovation (Ministério de Ciência, Technólogia e Inovação, MCTI), 541 Ministry of Social Development (Ministério do Desenvolvimento Social, MDS), 283, 512, 530n3, 540, 543 Ministry of Social Development and Fighting Hunger, 305n12, 542, 547 Ministry of Social Development and Zero Hunger, 519 Ministry of the Environment, 578 National Agricultural Research System (SNPA), 315, 332n14 National Bank for Economic and Social Development. See Banco Nacional de Desenvolvimento Econômico e Social (BNDES) National Bank of Economic Development (Banco Nacional de Desenvolvimento Econômico, BNDE), 66–67, 70, 92, 132, 143n14, 216n37 National Bank of Housing (Banco Nacional da Habitação, BNH), 215n21 National Coffee Council, 90 National Confederation of Industries (Confederação Nacional das Industrias, CNI), 66, 722 National Council of Justice (Conselho Nacional de Justiça, CNJ), 749, 756 National Council on Food Security (Conselho Nacional de Segurança Alimentar e Nutricional, CONSEA), 542, 545 National Development Plan. See Programa Nacional de Desestatização (PND, National Privatization Program) National Energy Plan 2030, 372 National Household Sample Survey (Pesquisa Nacional por Amostra de Domicílios, PNAD), 470, 478–79, 484n1, 514, 604–7, 606t National Institute for Space Research (INPE), 581–83 National Institute of Social Security (Instituto Nacional de Previdência Social, INPS), 515 National Monetary Council (Conselho Monetário Nacional, CMN), 109, 206, 535, 583

National Privatization Program. See Programa Nacional de Desestatização (PND) National Program for the Strengthening of Family Farming (Programa Nacional de Fortalecimento da Agricultura Familiar, PRONAF), 546 National Rural Credit System (NSRC or SNCR), 278, 295, 313 National School Nutrition Program (Programa Nacional de Alimentação Escolar, PNAE), 542 National Security Council, 67 National Service for Industrial Apprenticeship (Serviço Nacional de Aprendizagem Industrial, SENAI), 498, 539, 541 National Steel Company (Companhia Siderurgica Nacional, CSN), 6, 91, 99, 116–17, 688 National Supply Superintendence (Superintedência Nacional de Abastecimento, SUNAB), 721 National System of Innovation, 689–90 National Transport Confederation, 755 National Treasury, 186, 190–93, 191f, 202, 206 National Treasury Bond (BTN), 115, 120 National Treasury Indexed Liabilities (ORTN), 109–11 National Treasury Liability (OTN), 111, 115 National Water Supply and Sanitation System (PLANASA), 384, 391n18 overseas development assistance, 536 per capita income, 150–54, 150f, 151f, 152t, 153f, 156t, 157, 157t Polity IV rank, 631t, 632–33 population, 19–20, 41, 42t, 248–49, 312, 424, 639n7, 654–56, 655t population density, 424, 425f population outside Brazil, 655–56, 656t poverty, 511–34, 530n1 Presidential Decree 6, 321, 583 productive sectors, 241–419 Provisional Measure 1.801-10, 600 public sector, 45–46 railway network, 57n32

776   index Brazil. See also BRICS (Brazil, Russia, India, China, South Africa) countries (cont.) regional disparities, 13, 73, 94–95, 408–10, 423–45, 424f, 427t, 476. See also specific regions regional policies, 457–60 savings rate, 388 school enrollment, 50, 490, 492, 496, 500–502 second National Privatization Program (PND II), 7–8, 97, 188, 229, 244–45 Secretaria de Acompanhamento Econômico (SEAE), 722, 724–25, 733 Secretaria de Direito Econômico (SDE, Secretary of Economic Law), 722–25, 734 Secretaria de Reforma do Judiciário (SRJ, Office for Judicial Reform), 749 Secretaria Especial de Controle de Empresas Estatais (SEST, Special Secretariat of State Enterprise Control), 704–5, 707 Secretariat of Internal Control, 746 services, 338–57, 563t social debt, 514–15 socio-demography, 594–95, 594t socioeconomics, 465–618 South-South cooperation, 535–51 state enterprises, 701–17 state role, 699–768 structuralism, 63–88 tariffs, 634t, 636 tax system, 408 technology made in, 506 Temer administration, 389 territory, 22, 423 trade, 394–419, 633–35, 634t, 645–63 trade agreements, 417n7 trade restrictiveness, 652–54, 653t unemployment, 571, 571t water and sanitation, 383–84, 383t “Brazil-Africa Dialogue on Food Security, Fight Against Hunger and Rural Development” conference, 545 Brazil cost, 415–16 Brazilian Agricultural Research Corporation. See Empresa Brasileira de Pesquisa Agropecuária (Embrapa) Brazilian Agricultural Research System, 314

Brazilian and Portuguese Bank (English Bank of Rio de Janeiro), 202 Brazilian Antitrust System (SBDC), 722–23, 730, 734 Brazilian Association of Graduate Programs in Economics (ANPEC), 81 Brazilian Bar Association, 749 Brazilian Cooperation Agency (Agência Brasileira de Cooperação, ABC), 536, 538–41, 543, 548 Brazilian Development Bank. See Banco Nacional de Desenvolvimento Econômico e Social Brazilian Development Model, 10 Brazilian Enterprise for Rural Technical Assistance and Extension (EMBRATER), 279 Brazilian Institute for the Environment and Renewable Natural Resources (IBAMA), 578 Brazilian Institute of Coffee (IBC), 117 Brazilian Institute of Geography and Statistics (IBGE), 148, 300, 590n7 Brazilian multinational corporations, 12, 682–97, 687t Brazilian Press Association, 749 Brazil Learning Initiative for a World Without Poverty, 547 Brazil–United States Joint Commission (CMBEU), 181 Brazil Without Extreme Poverty, 542–43 brazilwood (Pau Brasil), 20, 71, 267–68, 440n3 Bresser-Pereira, Luiz Carlos, 9, 80–81, 113–14, 124 Bresser Stabilization Plan, 106, 113–14, 114f, 121 Brexit, 657, 659 BRF, 687t BRICS (Brazil, Russia, India, China, South Africa) countries, 2, 10, 621–44. See also specific member countries annual meetings, 626–28, 640n16, 641n25 asymmetries, 630–37, 631t, 634t, 637t, 641n34 Contingent Reserve Agreement (CRA), 549, 628–29 cooperation with, 535–36 current cooperative measures, 629–30 economic interdependency, 624–26, 625t FDI asymmetries, 636–37, 637t, 642n45

index   777 future directions, 637–39 general trends, 623, 624t intra-BRICS trade, 633–34, 634t macro asymmetries, 630–33, 631t main exports, 635, 635t micro asymmetries, 633–37, 634t New Development Bank (NDB), 2, 549, 627–29 objectives, 621–22, 630 poverty reduction, 530n5 Brito, João Rodrigues de, 201 BTI Transformation Index, 754–55 BTN (National Treasury Bond), 115, 120 Buarque, Cristovam, 530n10 budget dwarves (anões do orçamento), 747, 756 Budgetgate, 747 Buenos Aires Plan of Action (Plano de Ação de Buenos Aires, PABA), 537 Build-Own-Operate, Build-Operate-Transfer, and Build-Own-Operate-Transfer arrangements, 387 Bulhões, Otávio Gouvêa de, 65–66, 108 Burkina Faso, 541 Burundi, 662n2 Business Development Bank of Canada (BDC), 181 business services, 564–65t, 566 Bye, Maurice, 64 cacao exports, 43t CACEX (Carteira de Comercio Exterior), 143n13, 402 Cadastro Único (Single Registry system), 521, 542, 550n1 CADE. See Conselho Administrativo de Defesa Econômica CAIXA, 712 Caixa de Conversão (Conversion Office), 51, 58n36 Caixa de Estabilização (Stabilization Office), 52, 58n41 Caixa de Mobilização Bancária, 205 Caixa Economica (Savings Bank), 201–2 Caixa Economica da Corte (Court’s Savings Bank), 201–2 Caixa Econômica Federal, 211, 215n15, 215–16n29, 520t

Cameroon, 545–46 Campaign against Nationalization (Campanha contra a estatização), 706, 714 campaign finance reform, 750–51, 758–59 Campanha contra a estatização (Campaign against Nationalization), 706 Campino, Antonio, 613n1 Campos, Roberto de Oliveira, 70, 108, 140, 143n21, 239n10 Campos Basin oil fields, 691 Canada, 11, 156t, 171n17, 318, 650–56, 651f, 651t, 653t, 655t, 656t Canico, 692 Cano, Wilson, 81 capacity building, 729–35 capacity utilization, 163t CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, Higher Education Coordinagion Agency), 505, 509n16, 541 Cape Verde, 536, 541, 544 capital, 660 gross fixed capital formation (GFCF), 246, 248, 258 human, 317, 431 capital goods, 96–97, 103n4 capitalism, 57n26, 127 crony, 751 popular, 133 state, 141, 143n14 capitation, 27 captaincies (capitanias, capitãos hereditários), 20, 22–23, 25, 268 Cardoso, Fernando Henrique, 8–9, 74–75, 84n40, 100, 130–33, 223, 234, 386, 469, 640n13, 748 Plano Real (Real Plan), 9–11, 99, 105–6, 121, 121n3, 125, 133–34, 209, 214, 223, 239n4, 296–97, 470, 647t, 710–11 social development strategy, 531n12 Caribbean, 545, 562, 563t Carson, Lindsey, 14 Cartão Alimentação, 519 Carteira de Comercio Exterior (CACEX), 143n13, 402 cartels, 730, 734–35 Car Wash scandal, 3, 14, 101, 387, 548, 686, 735, 744, 750–55, 758–59

778   index CASA, 691 Cash Transfer Program (Cape Verde), 544 Castro, Antonio Barros de, 80, 186, 305n3 Catholic Church, 200 Catholic schools, 80, 499, 502 cattle industry, 297, 334n23 caudilhos (chieftains), 23 Cavalcante, Ricardo, 11 Cavalcanti, Tiago, 613n1 CC (Control of Corruption) index, 753, 753f CCT (Conditional Cash Transfer) programs, 2, 10, 537 CDA (Certificate of Agricultural Deposit), 282 CDCA (Certificate of Agricultural Credit Rights), 282, 285 Cease and Desist Agreements, 726 Cédula de Produto Rural (CPR, Rural Product Note), 279 CEMIG, 359 CENPES, 368–69 Center-South Brazil, 73–74 Center-West Brazil agriculture, 277–78, 284 income, 476, 477t land disputes, 305–6n17 population, 424, 425f, 476, 477t soybean production, 280, 280t, 284 water and sanitation, 383–84, 383t Central African Republic, 662n2 Central America, 407, 407t central bank(s), 106 Central Bank of Brazil. See Banco Central do Brasil (BCB) Central Brazil, 292 Centro Regional de Competencia para America Latina, 732 CEPAL. See Comisión Económica para América Latina y el Caribe Cepel, 690 Cerrado, 278, 292–96, 303, 311, 324–25 Certificate of Agribusiness’s Receivable Assets (CRA), 282, 285 Certificate of Agricultural Credit Rights (CDCA), 282, 285 Certificate of Agricultural Deposit (CDA), 282 CESP, 359 CET (common external tariff), 402

CGE (computable general equilibrium) models, 436, 587 CGFome (Coordenação Geral de Ações Internacionais de Combate à Fome, General Coordination for International Action Against Hunger), 541–42, 545 CGU (Controladoria Geral da União, Comptroller General of the Union), 746, 748–49 Chad, 541 Chaebols, 684–85 challenges, 11–14 Chang, Ha-Joon, 127 Changing Production Patterns with Social Equity (CEPAL), 79–80 chemicals, 635, 635t, 674 Chicago School, 126–27 chicken flight, 125 chieftains (caudilhos), 23 child labor, 127, 512 Chile, 130 Brazilian MNCs in, 693 comparison with Brazil, 128 contribution of services to manufacturing in, 346 cooperation with Brazil, 537 economic growth, 54 export share, 650–52, 651f, 651t GDP, 56n20 import substitution, 76–78 PSE/FGR, 283 trade, 417n7, 652–54, 653t trade agreements, 626 urbanization rate, 50 water and sanitation, 383 Chimeli, Ariaster B., 13 China, 149, 549. See also BRICS (Brazil, Russia, India, China, South Africa) countries demand for agricultural products, 328, 329t economy, 5, 639n1, 641n32 environmental quality, 582 exports, 631t, 632, 635, 635t, 641n36, 650–52, 651f foreign direct investment in, 636–37, 637t, 673, 682 GDP, 631–32, 631t greenfields, 671

index   779 growth rate, 150–51, 155, 169n4, 170n7 Harbin Embraer, 675 IMF quota, 638, 642n47 imports from Brazil, 408, 409t, 410–11 income growth, 156t income per capita, 341–42, 341f, 370, 371t labor productivity, 338, 339t MNCs from, 688 oil consumption, 370, 371t Polity IV rank, 631t, 632–33 poverty reduction, 530n5 PSE/FGR, 283 services sector, 342–43, 346, 563t, 569 tariffs, 634t, 636 trade, 1–2, 396, 397f, 410, 417n8, 633–35, 634t, 649–50, 658 trade restrictiveness, 652–54, 653t China Development Bank, 193n4 China effect, 221, 411, 649–50 CIDE (Contribution of Intervention in the Economic Domain), 344–45 CIP (Conselho Interministerial de Preço, Interministerial Price Council), 721 Citizen Constitution, 119 Clean Company Act, 760–61 climate change, 438–39, 457, 587–88 Clinton, Bill, 640n13 clothing (apparel) main exports, 635, 635t production index, 257t CLT (Consolidação das Leis do Trabalho), 316 CMBEU (Brazil–United States Joint Commission), 181 CMN (Conselho Monetário Nacional, National Monetary Council), 109, 206, 535, 583 CMO (Comissão Mista do Orçamento, Joint Budget Committee of Congress), 747 CNI (Confederação Nacional das Industrias, National Confederation of Industries), 66, 722 CNJ (Conselho Nacional de Justiça, National Council of Justice), 749, 756 CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico), 505 coal and coke, 360f, 361, 361f, 362f, 365 COBRADI (Relatório sobre Cooperaçāo Brasileira para o Desenvolvimento

International, Report on Brazilian Cooperation for International Development), 543 cocoa exports, 271t, 272, 311 CODEVASF, 457–58 Codigo das Aguas (Water Code), 380 Código de Defesa do Consumidor, 722 Coelho, Duarte, 22 Coes, Donald, 9 coffee boom, 379 coffee culture, 280 coffee exports, 43t, 52, 94, 96–97, 271t, 272–73, 297, 311, 395, 399, 399f, 407, 411 coffee plantations, 267, 269–73, 274t, 284 coffee prices, 90, 272, 396 coffee production, 55n8, 270–80, 270t, 277t, 279t, 297, 426 Furtado on, 71–72 government support, 396 historical perspectives on, 29, 41–45, 52, 56n17, 58n37, 89–90 slave labor, 29 coffee trade, 396, 397f Collor de Melo, Fernando, 8, 98, 125, 130, 132, 648, 709, 746 Collor de Melo, Pedro, 746 Collorgate, 746 Collor Plan (New Brazil Plan), 106, 116–21, 120f, 122n6, 223 Collor Plan II, 120 Colombia, 554, 691–92 Brazilian MNCs in, 693 cooperation with, 537 foreign direct investment in, 667–68 health expenditures, 609 income per capita and oil consumption, 370, 371t minimum wage, 556 PSE/FGR, 283 colonialism, internal, 73, 75, 83–84n27, 84n33 colonial period, 17–39 agricultural cycles, 268–69 credit before banks, 199–201 The Economy of Colonial Brazil (Furtado), 70–71 first banks, 201–3 slavery, 27–30, 28t, 35

780   index comarcas (judicial districts), 23 Comisión Económica para América Latina y el Caribe (CEPAL, Economic Commission for Latin America and the Carribbean), 65–67, 70, 73, 75–78, 84n37, 95, 396, 671 Changing Production Patterns with Social Equity, 79–80 Economic Survey of Latin America: 1949, 82n15, 83n16 Comissão Especial de Desestatização (Special Privatization Committee), 708 Comissão Mista do Orçamento (CMO, Joint Budget Committee of Congress), 747 commerce, 712, 713f Commercial Code, 202 commercial farming, 328 commodity boom (2006–2010), 221–22 commodity exports, 96–97, 225, 226f, 399, 407–8 Common Economic Zone, 640n24 common external tariff (CET), 402 Common Market of the Southern Cone (MERCOSUR or MERCOSUL), 1–2, 98–99, 435, 633, 638–39, 647t, 648 common external tariff (CET), 402 imports from Brazil, 408–10, 409t trade agreements, 417n7, 626 communications labor market, 564–65t SOEs, 712, 713f communism, 67, 82n14 community health services, 596–97 Community of Portuguese Language Countries, 535–36 Companhia de Obras e Infraestrutura, 194n18 Companhia Siderurgica Nacional (CSN, National Steel Company), 6, 91, 99, 116–17, 688 Companhia Vale do Rio Doce (CVRD), 705 competition policy, 718–40 competitiveness agricultural, 411 Global Competitiveness Report (WEF), 340 manufacturing, 254–58, 260–61, 346–51 regional disparities, 429–32 services, 346–51

Comptroller General of the Union (Controladoria Geral da União, CGU), 746, 748–49 computable general equilibrium (CGE) models, 436, 587 computers, 257t, 260 Conditional Cash Transfer (CCT) programs, 2, 10, 537 Confederação Nacional das Industrias (CNI, National Confederation of Industries), 66, 722 confisco cambial (exchange rate confiscation), 84n29, 231–32 Conjuntura Economica (Business Cycle), 65–66 CONSEA (Conselho Nacional de Segurança Alimentar e Nutricional, National Council on Food Security), 542, 545 Conselho Administrativo de Defesa Econômica (CADE, Administrative Council for Economic Defense), 720–24, 732–33 Administrative Tribunal (Tribunal), 725–26, 728, 730, 733–34 agreements, 731, 735 Antitrust Agency of the Americas award, 736 capacity building, 729–35 cartel or abuse of dominance cases, 730, 734 Cease and Desist Agreements, 729 checks and balances, 737n8 competition cases judged, 729 composition of, 725 Department of Economic Studies (DEE), 725–26, 734 development of, 728–29 fines collected, 738n12 funding, 726 General Superintendence (SG), 725–28, 730, 734 guidelines for remedies, 735 Horizontal Mergers Guideline, 731 human resources, 733–34 internships, 733–34 investigative tools, 734–35 merger cases, 730 mergers approved by, 729 oversight, 726–27 Procuradoria Geral (ProCADE), 725

index   781 Regimento, 730 settlements, 726, 735 transparency, 731 Conselho de Desenvolvimento Econômico (Economic Development Council), 706 Conselho Interministerial de Preço (CIP, Interministerial Price Council), 721 Conselho Monetário Nacional (CMN, National Monetary Council), 109, 206, 535, 583 Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), 505 Conselho Nacional de Justiça (CNJ, National Council of Justice), 749, 756 Conselho Nacional de Segurança Alimentar e Nutricional (CONSEA, National Council on Food Security), 542, 545 conservation, 585 conservative modernization, 293 Consolidação das Leis do Trabalho (CLT), 316 Constitution (Brazil), 119, 129–30, 206–7, 296, 305n5, 474, 511–12, 529, 542, 581, 597, 610, 721–22, 746–49 construction, 557–61, 559f, 564–65t, 566–67, 567t constructivism, 499 Construtora Andrade Gutierrez S.A., 187 Construtora Norberto Odebrecht S.A., 187 consumer goods, 103n4 consumption household, 247 immediate, 222 consumption function, 69 Conta Movimento, 112–13 Continental AG/Veyance, 733 Contingent Reserve Agreement (CRA), 549, 628–29 Continued Provision Benefit (Benefício da Prestação Continuada, BPC), 474 Contrato Soja Verde (Green Soybean contract), 279 Controladoria Geral da União (CGU, Comptroller General of the Union), 746, 748–49 Control of Corruption (CC) index, 753, 753f Convention on Combating Bribery of Foreign Public Officials in International Business Transactions, 761

Conversion Office (Caixa de Conversão), 51, 58n36 converted Jews (cristãos-novos, New Christians), 19, 22, 268 cooperation horizontal, 547 international, 318–19 South-South, 535–51 triangular, 543, 547–48 virtual, 547 Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES, Higher Education Coordination Agency), 505, 509n16, 541 Coordenação Geral de Ações Internacionais de Combate à Fome (CGFome, General Coordination for International Action Against Hunger), 541–42, 545 corn production, 277, 277t, 278, 297, 311, 411 areas of cultivated land, 270–72, 270t productivity gains, 325–26, 326t Correa, Sezerdelo, 458 correção monetária (monetary correction), 183 Corregedoria Geral da União, 748 corruption, 387–89, 741–68 impact of, 741–45, 742–43t, 757–61 institutional, 755–57, 755t perception-based surveys of, 753–57, 755t Corruption Perceptions Index (CPI), 753, 753f Cortes, Gustavo, 11 Costa Póvoa, José Liberato, 756 Costa Rica, 563t cost services, 348–51, 349f Côte d’Ivoire, 545 cotton cultivation, 270–72, 270t, 277, 277t cotton exports, 43t, 271t, 272–73, 297, 311 Council for Economic and Social Development, 136 Council on Public Transparency and Combating Corruption, 749 Court’s Savings Bank (Caixa Economica da Corte), 201–2 Coutinho, Luciano, 81, 186, 194n11 Coutinho, Ruy, 722 CPI (Corruption Perceptions Index), 753, 753f Cpqd, 690

782   index CPR (Cédula de Produto Rural, Rural Product Note), 279, 285 CRA (Certificate of Agribusiness’s Receivable Assets), 282, 285 CRA (Contingent Reserve Agreement), 549, 628–29 crawling peg (minidesvalorações) policy, 647, 647t credit, 262n8 agribusiness, 282 before banks, 199–201 BNDES-subsidized, 137 domestic, 163t, 164 rural, 278, 312–13, 332n5 total earmarked credit, 211–12, 212f credit conditions, 247–48, 262nn8–9 credit operations balance, 247–48 cristãos-novos (converted Jews), 19, 22, 268 crony capitalism, 751 crops, 270–72, 270t, 276, 276f, 295–96 crowding-out effects, 187 “crushed rock” cartel, 734 cruzado, 111 cruzado novo, 115 Cruzado Plan, 105–6, 111–13, 113f, 121, 223, 401, 469 Cruzado II, 113 cruzeiro, 91, 96, 117, 417n2 CSN (Companhia Siderurgica Nacional, National Steel Company), 6, 91, 99, 116–17, 688 CTA (Aerospace Technology Centre), 690–91 Cuba, 42, 54, 67, 187, 541 GDP, 56n21 Mariel Port, 194n16 services employment, 563t South-South cooperation, 536 cultural ties, 548 currency, 45 appreciation, 248 cruzado, 111 cruzado novo, 115 cruzeiro, 91, 96, 117, 417n2 devaluations, 262n10, 396, 404 historical perspective on, 51 milréis, 45, 58n41 real, 99, 236–37, 715n8 current account balance, Brazilian, 163–64, 163t 1930–2015, 412–14, 412f

1967–1980, 401, 401t deficits, 233–34, 233f Current Content, 505 Curitiba, 590n7 customs unions (CUs), 640n22 Cyrillo, Denise, 612n1 decentralization, 588 deforestation, 13–14, 578, 581, 583–86, 585f, 590n4, 590n8 deindustrialization, 101–3, 243, 245, 262n3 Delfim Netto model, 231–32 demand for agricultural products, 328, 329t for manufacturing, 249–54, 249f, 250f demand-distorted growth, 69 demand-driven growth, 69 democracy, 631t, 632–33, 641n37 Democratic Popular Pact, 229–30 democratization, 469–70 demographics, 248–49, 424 historical perspective on, 19–20 and per capita income, 432–33 socio-demography, 556, 594–95, 594t demonstration effects, 69 de Moura Castro, Claudio, 13 Denmark, 58n46, 369–70, 370t, 563t, 745 dependent-development, 125 dependent economies, 71, 450–57 Depression. See Great Depression deregulation, 703 desenvolvimentismo (developmentalism), 6–7, 63–88 desenvolvimento (development), 5–10, 12–13 Desenvolvimento e Conjuntura (Development and the Business Cycle), 66 Desenvolvimento e subdesenvolvimento (Development and Underdevelopment) (Furtado), 67–68, 75, 84n32 devaluation currency, 262n10, 396, 404 internal, 238 development dependent-development, 125 desenvolvimento, 5–10, 12–13 educational, 489–510 environment and, 578–82

index   783 fazer o bolo crescer (make the cake grow) non-theory of, 139–42 human, 518–19, 529 labor market, 552–76 Productive Development Policy, 186 social, 535–51 state-led, 135, 142n12 development agencies (agências de fomento), 193n6 developmentalism, 185, 378 desenvolvimentismo, 6–7, 63–88 new, 141, 232 development banks, 178–81, 193n6 DFP (Departamento da Polícia Federal, Federal Police), 747–49, 756–57 Diamantina, Brazil, 358 diamonds, 23 Dirceu, José, 749 disability transfers, 516–21, 517f, 520t, 529 discrimination, 137, 143n17 distance education, 502 Distance to Default (Z-Score), 217n42 distensão, 131 distributive services employment in, 562, 563t, 564–65t, 566–70 low pay and wage inequality, 567, 567t Distrito Federal (Federal District), 431, 434, 480, 590n7, 654 dívida interna (internal debt), 207 dizimeiros, 27 DNOCS (Departamento Nacional de Obras contra a Seca), 457 do Couto e Silva, Golbery, 131 Doha round, 132 Doing Business survey, 340 domestic credit, 163t, 164 domestic services, 564–65t, 566, 569 Dominican Republic, 187 drought, 367, 457 Duesenberry, James, 69 dumping, 403, 417n8 Dutch disease, 57n25, 222, 231–33, 235–37 Dutch West India Company, 267–68 earnings average, 566, 571, 571t differences, 478–80

East Asia, 342 East Timor, 537, 541, 662n2 ECLAC (Economic Commission for Latin America and the Carribbean). See Comisión Económica para América Latina y el Caribe (CEPAL) A economia brasileira (The Brazilian Economy) (Furtado), 70, 83n23 Economia da Saúde (IPEA), 593 Economia de La Salud—Perspectivas para América Latina (Panamerican Health Organization), 593 Uma economia dependente (A Dependent Economy), 70, 83n23 economia popular (the people’s economy), 720–21 Economic Bulletin, 77–78 Economic Commission for Latin America and the Carribbean (ECLAC). See Comisión Económica para América Latina y el Caribe (CEPAL) The Economic Development of Latin America (Furtado), 68 The Economic Development of Latin America and Its Principal Problems (Prebisch), 68 The Economic Growth of Brazil (Furtado), 67, 70–72, 75, 82n12, 83n23 economic miracle (milagre econômico) (1967–1974), 96, 147–48, 207, 401, 401t, 468, 647t economic policy “black beans and rice” policy, 115 external economic policy, 646–50 macroeconomic, 175–240 Economic Readjustment Law (Lei do Reajustamento Econômico), 205 economic structuralism. See structuralism Economic Survey of Latin America: 1949 (CEPAL), 82n15, 83n16 economy. See also specific sectors Brazilian, 1–174, 162f, 645–63, 741–44, 743t corruption and, 741–68, 743t dependent economies, 71 international prospects, 645–63 open, 650–56 simulations, 157–60, 158t, 159f, 160f state role, 699–768 world, 1, 262n1, 621–97

784   index The Economy of Colonial Brazil (Furtado), 70–71 Ecuador, 693 Edison, Thomas Alva, 358 education, Brazilian, 489–510 Base Curricular, 495 employment in, 564–65t evaluations, 492–93, 504, 509n14 Exame Nacional do Ensino Médio (ENEM), 492–93 higher, 490, 495–96, 500–506, 509n14 importance of, 478–79 primary schools (ensino fundamental), 490–92, 508n2 regional disparities, 428f, 429, 460, 461f secondary schools (ensino médio), 490–92, 495–96, 508n2 Sistema de Avaliação da Educação Básica (SAEB), 492 teaching systems (sistemas de ensino), 500 technical, 497–99, 508n11 EFTA (European Free Trade Agreement), 640n24 EGF program, 278–79 EKC (environmental Kuznets curve) hypothesis, 577 elderly population, 594t, 595 Eldorado Florestal, 284 electrical equipment exports, 635, 635t electricity, 358–59, 371, 375, 380 domestic supply, 360f, 361, 361f, 362f new model, 367 PAC program, 386–87 potential expansion, 374, 375t privatization of, 184 rural, 284–85 Electrobras, 126 Electropaulo, 137–38 Eletrobrás, 239n10, 359, 690, 712 Embraer, 84n42, 126, 133, 187, 683, 688–90 foreign assets, 687t Harbin Embraer, 675 Embrapa. See Empresa Brasileira de Pesquisa Agropecuária EMBRATER (Brazilian Enterprise for Rural Technical Assistance and Extension), 279

emerging market-based multinational corporations (EMNCs), 684–85, 694 emigration, 654 EMNCs (emerging market-based multinational corporations), 684–85, 694 employment, 554, 555t, 571 average hours worked per week (AHWPW), 568 formalization, 557–61, 559f formal occupations, 247 full, 659 manufacturing industry participation, 244f, 246 modernization and, 298 new jobs, 247, 553 service sector, 561–70 total national jobs, 247 Empresa Brasileira de Pesquisa Agropecuária (Embrapa, Brazilian Agricultural Research Corporation), 278–79, 286, 294–96, 303, 309–37, 406, 541, 712 Agropensa (Strategic Intelligence System), 320 establishment, 314, 332n8 human resources, 317, 333n17 Labex Program, 318–19 support for ProSavana, 546 “Vision 2014–2034,” 320–22, 321f Empresa de Pesquisa Energética (EPE), 367 ENADE (Exame Nacional de Desempenho de Estudantes), 504, 508n7 Encilhamento, 45, 50, 203 ENEM (Exame Nacional do Ensino Médio), 492–93 energy, in Brazil, 358–76 consumption, 369–71, 370t, 371t, 372t, 373f domestic supply, 360–61, 360f, 361f, 362f, 365 hydroelectric, 365–66, 366f, 367 nuclear, 360f, 362f, 365 PAC program, 386–87 planning, 369–72, 372t, 373f renewable, 360f, 362f thermoelectric, 367, 368f total primary energy supply (TPES), 369 Engenho do Conde, 36n17 engenhos (sugar mills), 21–22, 25, 31–32, 36n15 Engenho Salgado, 31

index   785 engineering services exports, 187 England, 180. See also United Kingdom English Bank of Rio de Janeiro, 202 ensino fundamental (primary schools), 490–92, 508n2 ensino médio (secondary schools), 490–92, 495–96, 508n2 Enterprise Survey (World Bank), 340 entrepreneurs, 71, 294 environmental economics, 582–89 environmental issues, 13–14, 577–92 environmental Kuznets curve (EKC) hypothesis, 577 EPE (Empresa de Pesquisa Energética), 367, 712 EPPG (public managers or gestores), 733, 738nn19–20 Equalizing Premium Paid to Growers program (PEPRO), 283, 285 Equatorial Guinea, 692 Erber, Fabio, 81 Escada, 32 Escola Superior de Agricultura de Lavras, 314 Escola Superior de Agricultura Luiz de Queiroz, 314 Espírito Santo coffee production, 275, 279, 279t, 280, 284 oil fields, 691 population and reported income, 476, 477t trade with other states, 434 O Estado de São Paulo, 705 Estado Nôvo (New State), 6, 91, 130–31, 214–15n13, 703, 720–21 Estevão, Luiz, 747 Estudos Economicos (Economic Studies), 66 ethanol domestic energy supply, 361f, 362 energy curve, 364–65, 364f environmental issues, 586–87, 590n10 production, 362–65, 374, 586, 590n10 ethanol consumption, 364 ethanol prices, 363 ethanol vehicles, 363–65 Ethiopia, 545 ethnic differences, 478–80 Eurasian Economic Union, 640n24 Europe. See also specific countries immigration from, 654

incentives for farmers, 334n24 railway network, 382 services sector, 562, 563t share of world exports, 407, 407t SOEs, 705 trade, 658 European Economic Community, 68 European Free Trade Agreement (EFTA), 640n24 European settlers, 19, 43–44 European Union (EU), 416 general trends, 623, 624t PSE/FGR, 283 support to agricultural producers, 298 trade, 408, 409t, 410, 652–54, 653t Eurozone, 245–46 Evangelista de Souza, Irineu, 202 Everhardt, Allison, 391n26 Exame Nacional de Desempenho de Estudantes (ENADE), 504, 508n7 Exame Nacional do Ensino Médio (ENEM), 492–93 exchange rates, 56n16, 58n41, 71, 91, 99–100, 112, 115, 185, 224, 248 confiscation (confisco cambial), 231–32 effective, 239n6, 416 floating, 405–6 historical perspective on, 646–50, 647t industrial equilibrium effective, 224, 224f, 234–35 interest rate–exchange rate trap, 221–40 Normative Instruction 70, 398 overvalued, 398 real, 405, 412–14, 413f real effective, 224, 224f, 234–35 trade performance and, 405–12 expansion, economic, 9–10 expansion, territorial, 22 expectant mothers and infants, 519 exports Brazilian, 225, 226f, 281–85, 399–401, 399f, 407–8, 409t, 411, 649–50, 649f, 649t commodity, 96–97, 225, 226f, 399 export/GDP ratios, 48, 57n27, 94, 415 historical perspective on, 12, 41–43, 48, 55n2, 57n27, 90 nontraditional, 12

786   index exports (cont.) price index, 649–50, 649f world, 406–7, 407t, 650–52, 651f ex-tarifário, 417n5 factor proportions theory, 658 FAF (Financial Investment Funds), 120 Fajnzylber, Fernando, 78–80 family farming, 283, 546 Family Health program (Programa Saúde da Familia, PSF), 527, 597 Family Stipend. See Bolsa Familía family subsidies, 519 A Fantasia Organizada (Furtado), 82n15 Faria, Vilmar, 531n12 Farias, Paulo César (“PC”), 746 farm areas, 278 farming. See also agriculture commercial, 328 entrepreneurial, 294 family, 283, 546 incentives for farmers, 334n24 income per farms, 320 subsistence, 327–28 FAT (Fundo de Amparo ao Trabalhador, Worker Support Fund), 138 Faucher, Philippe, 8 fazer o bolo crescer (make the cake grow) non-theory of development, 139–42 FDI. See foreign direct investment Federal Comptroller. See Comptroller General of the Union (Controladoria Geral da União, CGU) Federal District (Distrito Federal), 431, 434, 480, 590n7, 654 Federal Police (Departamento da Polícia Federal, DFP), 747–49, 756–57 Federal Public Prosecutor (Ministério Público Federal, MPF), 726, 746, 749, 756–57 Federal Railroad Corporation (RFSSA), 380–81 Federal University of Rio de Janeiro, 80 Fernandez, Ramon Garcia, 81 Ferraz, Joao Carlos, 81 Ferrer, Aldo, 70 fertility rate, 594t

FGTS (Fundo de Garantia do Tempo de Serviço, Length of Service Guarantee Fund), 110, 133, 207 FGV (Fundação Getúlio Vargas), 65–66, 80–81 Fiat, 363–64 Fibria, 284 FIES (Fundo de Financiamento Estudantil), 503 FIESP, 389, 722 FIFA World Cup, 384–85 Figueiredo, João, 469, 708 FINAME (Financing Agency for Machinery and Equipment), 177, 689–90 FINAMEX (Financiadora de Maquinas e Equipamentos), 689–90 financeiras (Societies of Credit, Funding, and Investment; Sociedades de Crédito, Financiamento e Investimento), 215n20 Financiadora de Estudos e Projetos (FINEP, Agency for Studies and Projects), 505, 689–90, 712 Financial Investment Funds (FAF), 120 financial reform, 109 financial repression, 179 financial sector, 213–14 Brazilian MNCs, 688 domestic credit provided by, 163t, 164 employment, 564–65t foreign direct investment in, 674 inflationary period (1986–2005), 208 SOEs, 712, 713f value added, 118, 118t financial system(s) Brazilian, 204–8 types of, 180 Financial System for Housing (Sistema Financeiro da Habitação, SFH), 207, 215n21 financing health care, 608–9 for national champions, 138–39 Financing Agency for Machinery and Equipment (FINAME), 177, 689–90 Fincef, 137 Finep, 509n15 FINEP (Financiadora de Estudos e Projetos, Agency for Studies and Projects), 505, 689–90, 712

index   787 Finland, 563t Fiocruz (Oswaldo Cruz Foundation), 541, 544 Fiscal Responsibility Law (Lei de Responsabilidade Fiscal), 134, 143n15, 435 flex fuel vehicles, 364 Folha da São Paulo, 749 Fome Zero (Zero Hunger) initiative, 519, 522, 542 Food Acquisition Program (PAA), 282, 542, 545 Food and Agriculture Organization (FAO), 535–37, 541, 544–45 food and beverages Brazilian MNCs, 687–88, 687t exports, 635, 635t growth rates, 91 historic perspective on, 51 imports, 275–80, 311 performance indicators, 251 productivity, 257t SOEs, 712, 713f food and catering services, 354n7 food costs, 84n31 food prices, 327, 327f Ford, 52, 363–64 Ford Foundation, 81, 333n18, 504–5 foreclosure restrictions, 200 foreign banks, 202, 209f, 210–11, 211f foreign-born population, 654 foreign debt, Brazilian, 229, 233–34, 233f, 400–401, 401t gross external debt/export ratio, 413–14, 413f foreign direct investment (FDI) in Brazil, 1, 8, 44, 379, 386, 398, 415, 664–81, 674t, 675t, 676t, 678f, 679 corruption and, 741–44, 742t functions of, 665–69, 671–72 global, 682–83 Hypothesis of the Three Sectors for, 677, 677f intra-BRICS, 624–26, 625t, 636–37, 637t, 642n45 inward, 1, 684, 741–44, 742t in Latin America, 670–73 outward (OFDI), 683–89, 685f, 686t foreign exchange reserves, 91, 262n6 foreign investment, 386, 684 foreign-market-oriented agriculture, 272, 278, 311 foreign savings, 222, 233, 237

forests, 284 Formação econômica do Brasil (Furtado), 103n1 Fortaleza, 446 fortune hunters (bandeirantes), 269 fossil fuels, 367 Fox, Vicente, 130 Foz do Iguaçu, 508n5 France Brazilian population, 655–56, 656t colonization of Brazil, 33 exports, 651f income growth, 156t industrial development, 180 migrant population, 654–56, 655t secondary schools, 495 services employment, 563t Franco, Itamar, 8–9, 80, 121n3, 223, 746 Frank, Andre Gunder, 74–75 Free Trade Area of the Americas, 132 free trade areas (FTAs), 435, 626, 640n22, 640n24 Freyre, Gilberto, 21, 82n14 Fribol, 138–39 Friedman, Milton, 124, 126–27, 142n4 FTAs (free trade areas), 435, 626, 640n22, 640n24 Fuglie, Keith, 334n26 Funaro, Dilson, 113 Fundação Getúlio Vargas (FGV), 65–66, 80–81 FUNDEB (Fundo de Desenvolvimento do Ensino Básico), 491, 508n4 Fundo de Amparo ao Trabalhador (FAT, Worker Support Fund), 138 Fundo de Desenvolvimento do Ensino Básico (FUNDEB), 491, 508n4 Fundo de Financiamento Estudantil (FIES), 503 Fundo de Garantia do Tempo de Serviço (FGTS, Length of Service Guarantee Fund), 110, 133, 207 Furtado, Celso, 63–76, 81, 82n3, 82n10, 82n14, 83n19, 458, 646 Desenvolvimento e subdesenvolvimento (Development and Underdevelopment), 67–68, 75, 84n32

788   index Furtado, Celso (cont.) A economia brasileira (The Brazilian Economy), 70, 83n23 Uma economia dependente (A Dependent Economy), 70, 83n23 The Economic Development of Latin America, 68 The Economic Growth of Brazil, 67, 70–72, 75, 82n12, 83n23 The Economy of Colonial Brazil, 70–71 A Fantasia Organizada, 82n15 Formação econômica do Brasil, 103n1 “General Characteristics,” 72, 82n15 Operation Northeast, 75 “Sketch of a Program of Development for Brazil,” 66 Triennial Plan, 108 future directions, 14, 506–8, 547–50, 637–39 G20 countries, 155, 156t, 157, 157t, 170n7. See also specific countries Gabon, 692 Galeão airport, 384–85 gambling, 749 gas. See oil and gas gasohol, 364 GATT (General Agreement on Tariffs and Trade), 636, 641–42n40 GDP. See gross domestic product GE Appliances, 682 Gemini, 735 gender differences, 478–80 General Accounting Office (Auditoria Geral da União), 748 General Agreement on Tariffs and Trade (GATT), 636, 641–42n40 “General Characteristics” (Furtado), 72, 82n15 General Inspectorate of Banks (Inspetoria Geral dos Bancos), 204 General Motors, 52, 363–64 genetic research, 334n28 genome editing, 334n28 geographic information system (GIS) transportation network model, 436 geography, 49, 54, 423–24, 447–49, 594–95 and per capita income, 432–33 territorial expansion, 22

Gerdau, 687t German immigrants, 43–44 Germany, 117, 365 Brazilian MNCs in, 692–93 Brazilian population, 655–56, 656t cooperation with, 318, 536–37 corruption, 754 export share, 650–52, 651f, 651t income per capita, 156t, 341–42, 341f migrant population, 654–56, 655t secondary schools, 495 services sector, 341–42, 341f, 563t SOEs, 705 germination points (growth poles), 70 gestores or public managers (EPPG), 733, 738nn19–20 GFCF (gross fixed capital formation), 246, 248, 258 Ghana, 542–46 Gini index, 470–72, 471f, 472f, 473 Glass-Steagall Act, 215n25 Global Competition Review, 732, 736 Global Competitiveness Report (WEF), 340 Global Corruption Barometer, 754–56, 755t global economy, 1, 262n1, 621–97 globalization, 53, 645–50 anti-globalization, 656–61 happy, 135 global trade, 408, 623, 624t Goiás, 277 Gold Clause Law (Lei da Cláusula-Ouro), 198, 205 gold excavation, 20–24, 28–29, 28t, 35–36n6, 426 Goldsmith, Raymond, 45 Gomes, Maia, 462n4 Gonzalez Casanova, Pablo, 84n33 Goulart, Joao, 67–68, 82n10, 108 Government Procurement Act (Law 8886/93), 746 graduate schools, 504–6 grains and oil Seeds,311, 331–32n4 Granger causality test, 375n1 Great Britain. See United Kingdom Great Depression, 68–69, 72, 198, 213–14, 406, 702–3 trade during, 395–97, 397f

index   789 Great Divergence, 53 Great Recession, 211–13, 406 Great Specialization, 53–54 greenfields, 671, 692 Green Soybean contract (Contrato Soja Verde), 279 gross domestic product (GDP), 577 agricultural, 297 Brazilian (1967–1980), 401, 401t export/GDP ratios, 48, 57n27, 94, 415 general trends, 623, 624t global, 262n1 growth rates, 92–93, 95–97, 238n2, 247, 258, 262n1, 397, 401, 401t, 404, 623, 624t import/GDP ratios, 94, 96–97, 98t income per capita and, 341–42, 341f industrial, 401, 401t manufacturing industry participation, 244–46, 244f per capita, 41, 42t, 46–47, 50, 55n6, 56n20, 56nn20–21, 148–50, 150f, 153, 156, 170n5, 221–22, 246, 369–70, 370f, 370t, 371t, 577 regional disparities, 426–29, 427t sectoral, 93–94, 93t, 102, 340–42, 341f simulations, 158–59, 159f TPES per capita vs GDP per capita, 369, 370f trade/GDP ratio, 132 gross fixed capital formation (GFCF), 246, 248, 258 growth, economic, 11–12, 40–41, 50–52, 147–74, 262n8, 377 1967–1980, 401, 401t during accelerations, 161–62, 162f annual per capita rate, 414 average rates, 155, 155f during collapses, 161–62, 162f corruption and, 741–44, 744t demand-distorted, 69 demand-driven, 69 determinants of, 246–49 export-led, 281–85 foreign direct investment and, 665–69 GDP growth, 156, 221–22, 238n2, 246–47, 397, 401, 401t, 404 general trends, 623, 624t income growth, 152–54, 152t, 153f, 155, 156t, 157, 157t, 334n27

infrastructure link, 385–86 jobs-rich, 553 long-term, 52–54, 147, 153–54, 153f, 157–60, 158t, 159f, 160f resumed, 221 simulations, 157–60, 158t, 159f, 160f volatility, 154–57, 155f growth, industrial, 397–401, 401t growth, productivity, 12, 354n6 Growth Acceleration Program (Programa de Crescimento Acelerado, PAC), 377–78, 386–89 growth hypothesis, 369 Grupo Misto CEPAL-BNDE, 66, 70 Guanabara, 82n14 Guanabara Bay, 577–78, 580 Guarantee Program for Agricultural Activity (PROAGRO), 279, 282 Guarulhos Airport, 384–85, 391n19 Guatemala, 537 Gudin, Eugenio, 66 Guinea Bissau, 536, 541 Haddad, Eduardo, 13 Haddad, Fernando, 493 Haiti, 537, 541–42, 548 Hall, Anthony, 10 Harbin Embraer, 675 Harrod-Domar model, 110 health care community services, 596–97 employment in, 564–65t hospital beds, 597–98, 598t inequalities in outcomes, 605, 606t, 613n11 inequalities in use, 607, 607t judicialization of, 610–12 lawsuits regarding, 610–11, 611t preventive, 607, 607t private, 599–602, 609–10 public expenditures, 599, 609 public system, 608–9 regional disparities, 429 regulation of, 609–10 right of access to, 610, 612 SOEs, 712, 713f universal, 608–9 (see also Unified Health System)

790   index health economics, 593–618 health inequalities, 602–7, 603f, 613n11 health insurance, 600–602, 600t, 601t regional disparities, 429 universal, 608 (see also Unified Health System) high and medium-high technology (HT-MHT) goods, 243, 250–57, 250f, 253t higher education, 490, 495–96, 500–504 evaluations, 504, 509n14 Humboldt model, 500–501 Higher Education Coordination Agency (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, CAPES), 505 highways, 184, 275, 382 PAC program, 386–87 Projeto Crescer projects, 389 Hirschman, A., 78, 84n39 Hirschman trap, 390 historical perspectives, 15–174 colonial period, 17–39, 199–203, 268–69 eighteenth century, 23–24, 490 nineteenth century, 40–62, 654 twentieth century, 40–62, 654, 701 twenty-first century, 246–49 on competition and antitrust policy, 720–25 on education, 489–90 on banking system, 198–220 on BNDES, 181–86 on external trade and exchange rate policies, 646–50, 647t on energy industry, 358–76 on infrastructure, 379–81 on Northeast, 448–50 on trade policy, 394–419 HIV/AIDS, 544 Hobsbawm, Eric, 40 Hoffmann, Rodolfo, 12–13 Holanda, Buarque de, 25 Holanda, Fernando, 8 Holy Houses of Mercy (Santas Casas de Misericordia), 29, 200, 597 Hong Kong, 149 Horizontal Mergers Guideline, 731 hospitals, 597–99, 598t, 708 household budgets, 342–43

Household Budget Survey (Pesquisa de Orçamentos Familiares, POF), 471, 527, 530n6 household consumption, 247–48, 258 household income distribution, 470–71, 471f, 480–81 housing, 386–87 HSBC, 216n32 HT-MHT (high and medium-high technology) goods, 243, 250–57, 250f, 253t Hu Jintao, 641n29 human capital, 317, 431 human development, 518–19, 529 Human Development Index, 428f, 429 humanitarian assistance, 546 human resources, 317 Humboldt model, 500–501 Hungary, 754 hydroelectric energy, 358, 374, 375t, 381 Belo Monte project, 367 Codigo das Aguas (Water Code), 380 domestic supply, 360f, 361, 361f, 362f environmental issues, 580–81 evolution of, 365–66, 366f potential, 374, 375t Projeto Crescer, 389 Hymer, Stephen, 683–84 hyperinflation, 107–8, 118–19 Hypothesis of the Three Sectors, 677, 677f IAA (Alcohol and Sugar Institute), 117 IADB (Inter-American Development Bank), 178, 501, 727, 732 IBA (International Bar Association), 755–56 IBAMA (Brazilian Institute for the Environment and Renewable Natural Resources), 578, 590n8 IBC (Brazilian Institute of Coffee), 117 IBGE (Brazilian Institute of Geography and Statistics), 148, 300, 590n7 IBRD (International Bank for Reconstruction and Development), 178, 381 IDB (Inter-American Development Bank), 178, 501, 727, 732 IDCPs (integrated conservation and development projects), 585

index   791 IFDI (inward foreign direct investment), 1, 684 IMF. See International Monetary Fund immigration, 43–44, 55n10, 654 Imperial Agronomical Institute of Campinas, 272 Imperial Institute of Agriculture, 272 imperial period, 40–41, 44–50, 56n19, 269–72 import penetration coefficient (IPC), 254 imports food, 275–80 GDP ratios, 94, 96–97, 98t liberalization of, 401–4 tariffs on, 402 import-substitution industrialization, 6–8, 89–104, 292, 397–99, 686, 690 agricultural modernization driven by, 311–12 crisis, 75–81, 95–97 early, 182–83 horizontal, 76–78 vertical, 76 Inco, 692 income components of, 474, 475f middle-income trap, 147, 151, 414 oil consumption and, 369–70, 370t, 371t regional disparities, 428f, 429, 432–35, 455, 455t skin color and, 479 income distribution, 238, 467–88 Gini and Theil’s T indices for, 470–72, 471f, 472f Kernel density curves, 481, 482f polarization, 481, 482f income inequality, 164, 470–76, 471f, 472f, 577 corruption and, 741–44, 743t, 745 polarization, 481, 482f recent crisis (2014–2016), 482–83 regional, 439 income per capita, 147, 150, 150f, 151f, 234, 432–35 growth, 152–55, 152t, 153f, 156t, 157, 157t and share of services in GDP, 341–42, 341f simulations, 159, 160f income per farm, 320 income transfers, 474–76, 522–23, 523f

The Incomplete Industrialization of Latin America (Fajnzylber), 78 indexed money, 118 India, 506. See also BRICS (Brazil, Russia, India, China, South Africa) countries demand for agricultural products, 328, 329t economic growth, 155, 169n4, 553, 631–32, 631t, 639n1, 641n32 environmental issues, 590n2 exports, 631t, 632, 651f foreign direct investment in, 636–37, 637t GDP, 631–32, 631t income growth, 156t labor productivity, 338, 339t migrant population, 654–56, 655t multinational corporations from, 688 Polity IV rank, 631t, 632–33 poverty reduction, 530n5 share of household budgets allocated to services, 342–43 tariffs, 634t, 636 trade, 417n8, 633–35, 634t, 652–54, 653t trade agreements, 404, 626, 640n24 Indians, 19–20, 27–30 Indonesia, 155, 156t, 169n4, 642n47 industrial density, 355n18 industrial equilibrium, 224, 224f, 231, 234–35 industrial growth, 397, 399, 401, 401t industrialization, 180 deindustrialization, 101–3, 243, 245 historical perspective on, 17–18, 51, 58n38, 58n44, 74, 89–90 import-substitution, 6–8, 75–81, 89–104, 397–99, 686, 690 service, 561–62 state-led, 701 industrial parks, 257–58 industrial policy, 100–101, 136–39, 181, 227, 404 industrial water use, 588–89 industry employment, 564–65t, 566–68 exports, 411 foreign direct investment in, 671–72, 674–75, 675t, 678–79 informal workers, 557–61, 559f low pay and wage inequality, 567, 567t share of GDP, 93–94, 93t, 102

792   index industry (cont.) share of labor, 93–94, 93t, 102 working hours, 568 working poverty, 566 infant assistance, 519 infant industry argument, 101 infant mortality, 595, 596t infectious diseases, 613n11 inflation, 398–99 1960–1990, 105–23 1967–1980, 401, 401t 1986–2005, 208–11 during Bresser Plan, 114, 114f chronic, 107–8 during Collor Plan, 119–20, 120f control of, 470–73, 483 (see also stabilization) during Cruzado Plan, 113, 113f current, 11 as the “dragon,” 125 hyperinflation, 107–8, 118–19 inertial, 110, 239n3 during PAEG Plan, 110, 111f services vs CPI inflation, 345–46 during Summer Plan, 115, 116f targeting, 99–100 inflationary period (1986–2005), 208–11 inflation taxes, 106 informal economy, 741–44, 743t informal workers, 557–61, 559f, 571, 571t, 572n3 social assistance for informal rural workers, 514–16, 515f, 520–21, 520t, 529 information and communications technology (ICT), 348 infrastructure, 14, 345, 377–93 investment in, 187, 377–78 PAC program, 386–87 privatization of, 184 Projeto Crescer projects, 389 regional disparities, 430–31 initial public offerings (IPOs), 504, 714 innovation, 332n15, 506, 689–90 INPE (National Institute for Space Research), 581–83 INPS (Instituto Nacional de Previdência Social, National Institute of Social Security), 515

Inspetoria Geral dos Bancos (General Inspectorate of Banks), 204 Institute for Technological Research (Instituto Tecnológica de Aeronáutica, IPT), 690, 693 institutionalism, 456 Instituto Agronômico de Campinas, 314 Instituto de Pesquisa Econômica Aplicada (IPEA, Institute for Applied Economic Research), 101, 538, 540, 547, 593 Instituto Nacional de Previdência Social (INPS, National Institute of Social Security), 515 Instituto Tecnológica de Aeronáutica (IPT, Institute for Technological Research), 690, 693 Instituto Unibanco, 500 insurance services. See also health insurance employment in, 564–65t integrated conservation and development projects (IDCPs), 585 intellectual property, 682–83 Inter-American Development Bank (IADB or IDB), 178, 501, 727, 732 interest rates, 99–100, 134, 200, 221–40 benchmark (Taxa Referencial, TR), 120 long-term (Taxa de Juros de Longo Prazo, TJLP), 138, 191, 194n23 real, 230, 248 intermediate goods, 96–97, 103n4 Interministerial Price Council (Conselho Interministerial de Preço, CIP), 721 internal colonialism, 73, 75, 83–84n27, 84n33 internal debt (dívida interna), 207 internalization, 683–84 International Bank for Reconstruction and Development (IBRD), 178, 381 International Bar Association (IBA), 755–56 International Competition Network (ICN), 732 international cooperation scientific, 318–19, 333n19 South-South, 535–51 internationalization, 671, 689–93 International Monetary Fund (IMF), 8, 11, 98, 142n5, 210–11, 622 Governance and Quota reform, 627–29 quota shares, 638, 642n47

index   793 International Poverty Centre for Inclusive Growth (IPC-IG), 540, 543–44, 547 international production, 683–84 Internet access, 345 interventionism, 135–36 Introduction to Economics (Barros de Castro and Lessa), 80 investment, 138–39, 187–88. See also foreign direct investment in agricultural R&D, 315, 319 corruption and, 741–44, 742t in infrastructure, 187, 377–78 overseas development assistance, 536 investment maintenance program (PSI), 186, 190 investment rates, 248 involution, 71 inward foreign direct investment (IFDI), 1, 684 IPC (import penetration coefficient), 254 IPC-IG (International Poverty Centre for Inclusive Growth), 540, 543–44, 547 IPEA (Instituto de Pesquisa Econômica Aplicada, Institute for Applied Economic Research), 101, 538, 540, 547, 593 IPOs (initial public offerings), 504, 714 IPT (Instituto Tecnológica de Aeronáutica, Institute for Technological Research), 690, 693 Iran, 369–70, 370t Iraq, 691 Ireland, 369–70, 370t, 563t, 754 iron ore exports, 332n10 ISI. See import-substitution industrialization ISMART, 500 Israel, 369–70, 370t Italy, 156t, 650–56, 651f, 651t, 655t, 656t Itamaracá, 31 Itamaraty (Ministery of External Relations, Ministério das Relações Exteriores, MRE), 536, 539–42 Jaguar Land Rover, 682 Japan, 78, 149 Brazilian MNCs in, 692–93 Brazilian population, 655–56, 656t cooperation with, 537 export share, 651f imports from Brazil, 408, 409t

income growth, 156t industrial development, 180 infrastructure spending, 385 PSE/FGR, 283 support to agricultural producers, 298 trade with, 408, 652–54, 653t Japanese immigrants, 43–44 Japan Financial Corporation, 193n4 JBS, 683, 687t Jefferson, Roberto, 749 Jesuit missionaries, 22–23 Jews, converted (cristãos-novos), 19, 22, 268 João III, 22 João VI, 201, 314 Joint Brazil–United States Economic Development Commission, 92 José I, 23–24 Judiciary Integrity Initiative (IBA), 755–56 Juízo dos Órfãos (Orphans’ Judge), 200 Julião, Francisco, 67 KACO, 695 Kahn, Richard, 67 Kaldor, Nicholas, 67 Kazakhstan, 640n24 Kenya, 542, 544–46 Keynes, John Maynard, 66, 70 KfW, 193n4 King’s College, Cambridge, 67 Kiribati, 662n2 knowledge sharing, 547 Kolynos-Colgate, 730 Korean War, 76, 396, 397f Kubitschek, Juscelino, 7, 67, 81, 92, 108, 183 Kyrgyz Republic, 640n24 Labex Program (Embrapa), 318–19 Labex-USA, 333n19 labor market, 552–76, 564–65t. See also employment child labor, 127, 512 formalization, 557–61, 571, 571t historical perspectives on, 48, 230 recent changes, 554–57, 555t sectoral distribution, 93–94, 93t, 102–3 slave labor, 28, 28t, 29, 32 social assistance and, 528–29

794   index labor productivity, 101, 189, 338–39, 339f, 339t, 343, 343f, 355n16 labor unions, 568, 573n11 Lacera, Carlos, 82n14 Lagoa Rodrigo de Freitas, 580 Lagos, 130 LAIA (Latin American Integration Association), 417n7, 626 Lampreia, Luiz Felipe, 627, 630, 639n6 land disputes, 303, 305–6n17 land distribution, 30–34, 44, 55n11, 288–308, 290t land grants (sesmarias), 30–33, 44, 268 landholdings (latifundias), 268 Land Law (Brazil), 44 land tenure, 289, 290t latifundias (landholdings), 268 latifundistas (large landowners), 293 Latin America, 149, 545. See also specific countries economic growth, 155–56, 156t economic strategy, 166–67 exports, 79–80 foreign direct investment in, 670–73 imports from Brazil, 408–10, 409t income, 150, 150f, 151f service sector, 562, 563t, 569 trade with, 404, 410 Latin American Competition Forum, 732 Latin American Free Trade Area, 77 Latin American Integration Association (LAIA), 417n7, 626 Lava Jato scandal. See Car Wash scandal lavradores (sugarcane planters), 31, 36n16 LCA (Notes of Agribusiness Credit), 282 leadership, 548. See also specific individuals LEAP (Livelihood Empowerment Against Poverty), 543–44 Lebanon, 417n7 Leff, Nathaniel, 47, 54 legitimacy, 731 Length of Service Guarantee Fund (Fundo de Garantia do Tempo de Serviço, FGTS), 110, 133, 207 Leontieff, Vassily, 70 Leroy-Beaulieu, Paul, 56n23 Lessa, Carlos Francisco, 80–81, 186

letras de câmbio (bills of exchange), 205 letras imobiliárias (real estate bills of exchange), 207 Levi-Strauss, Claude, 64–65 Lewis, W. Arthur, 68 Lewis model, 48, 57n26 Lewis turning point, 230, 237 LFTs (treasury financial bills), 112 LG, 684–85 liberalism, 703 liberalization, 296 distensão, 131 import, 398, 401–4 trade, 131–32 Libya, 633, 691–92 life expectancy, 428f, 429, 595, 596t Light, 380–81 LIGHT group, 358 Linhares, José, 721 liquidity ratio index (LRI), 212–13, 213f, 217n43 Lispector, Clarice, 573n5 literacy, 44, 50, 55n13, 429, 490, 494, 507 literature review, 385–86 litigation, health care, 610–11, 611t Livelihood Empowerment Against Poverty (LEAP), 543–44 livestock, 272, 635, 635t, 675t LOAS (Lei Orgânica da Assistência Social), 517 logging, 590n4 London and Brazilian Bank, 202 long-term growth, 52–54, 147 simulations, 157–60, 158t, 159f, 160f Long-Term Interest Rate (Taxa de Juros de Longo Prazo, TJLP), 138, 191, 194n23 long-term solution, 235–37 Love, Joseph L., 6, 84n42 low and medium-low technology (LT-LMT) goods, 243, 250–51, 250f, 254–56, 253t lowess, 355n17 LRI (liquidity ratio index), 212–13, 213f, 217n43 LT-LMT (low and medium-low technology) goods, 243, 250–51, 250f, 254–56, 253t Lula da Silva, Luiz Inácio (Lula), 8–10, 80, 126, 134, 140–41, 224, 234, 622, 648–49, 711, 749 foreign policy, 535–39 industrial policy, 100–101, 714

index   795 neoliberalism, 127, 130 South-South cooperation, 538–39 Luz para Todos (Light for Everyone), 248 Macario, Santiago, 77–78 machinery and equipment Brazilian MNCs, 687–88, 687t exports, 635, 635t foreign direct investment in, 674 performance indicators, 251 production index, 257t, 260 tariffs, 404 transport equipment, 187, 251, 257t, 687–88, 687t Mackenzie University, 499 macroeconomic policy institutions, 221–40 macroeconomic volatility, 162–64 Maddison, Angus, 17 Maddison Project Database, 56n20, 58n40 Magnesitsa, 687t maize. See corn Malawi, 545 Malaysia, 338, 339t, 369–70, 370t, 642n47 Mali, 541 manioc, 277, 277t Mantega, Guido, 80 manufactured goods Brazilian exports, 96–97, 225, 226f, 231–32, 399, 399f, 407–10, 409t tariffs, 403 trade restrictiveness, 652–54, 653t world exports, 407, 407t manufacturing, Brazilian, 41–42, 51, 243–65, 338–57 comparison with retail sales, 225, 226f competitiveness, 254–58, 260–61, 346–51 contribution of services to, 346 demand for, 249–54, 249f, 250f early growth, 89 employment in, 93–94, 93t, 244f, 246, 566, 660 evolution of, 249–54 foreign direct investment in, 671, 674–75, 676t, 677, 678f long-term context, 244–46 performance, 251–54, 253t, 258–60 production, 249–54, 249f, 250f

productivity, 254–58, 257t, 343, 343f regional disparities, 426–29, 427t, 430 share of GDP, 102, 224, 225f, 244–46, 244f, 340–41, 341f SOEs, 712, 713f value added, 346 MAPA (Ministry of Agriculture, Livestock and Food Supply), 282 Maranhão, 202, 272, 283–84, 304 Maranhão, Tocantins, Piauí, and Bahia (MATOPIBA), 284, 286, 303 Marcondes, Renato, 11 Marcondes Meira, Severino, 756 Marconi, Nelson, 239n14 Marco Polo, 683 Marfrig, 687t Mariel Port (Cuba), 194n16 market reform, 125, 127, 129–30, 133–34 Martha, Geraldo, 10 mate herb, 271t, 272 mathematics, 493 Mato Grosso, 277, 303, 756 Mato Grosso do Sul, 277 MATOPIBA (Maranhão, Tocantins, Piauí, and Bahia), 284, 286, 303 MDA (Ministério do Desenvolvimento Agrário, Ministry of Agrarian Development), 282, 300, 305n12, 530n3, 540 MDS (Ministério do Desenvolvimento Social, Ministry of Social Development), 283, 512, 530n3, 540, 543 MDSA (Ministério do Desenvolvimento Social e Agrário, Ministry of Agrarian and Social Development), 283, 305n12, 540–41 meat production, 276, 276f mechanization, 51, 313 Medici, André, 612n1 Médici, Garrastazú, 105 medicine, 612 Mehran index, 476 MEI (micro-entrepreneur program), 560 Meirelles, Henrique, 188 Mello e Castro, Martinho de, 24 mensalão (big monthly stipend) scandal, 227, 749–50, 754–56 Mercadante, Otávio, 239n5

796   index MERCOSUL. See Common Market of the Southern Cone MERCOSUR. See Common Market of the Southern Cone Merger Control Agreement (ACC), 726 mergers, 724, 726, 728–31 metal industry, 51, 91, 257t, 635, 635t metalistas, 202, 214n11 metallurgy, 674 metal ore exports, 635, 635t methane emissions, 334n23 Methodists, 499 Methuen Treaty, 18 Mexico, 130 banking sector, 212 comparison with Brazil, 128 cooperation with, 537 debt, 639n9, 647t environmental issues, 589, 590n2 export share, 650–52, 651f, 651t foreign direct investment in, 1, 668, 671 IMF quota, 642n47 import-substitution industrialization, 77–78 internationalization, 671 minimum wage, 556 per capita GDP, 50 per capita income growth, 156t services, 346, 563t trade with, 417n7, 652–54, 653t MFI (More Food International), 546 MFN (most favored nation) status, 402 micro-entrepreneur program (MEI), 560 middle class, 254 Middle East, 43–44, 407–8, 407t, 409t. See also specific countries middle-income trap, 147, 151, 414 Midwest Brazil, 423–29, 424f, 427t, 434 migrant population, 654–56, 655t migration, 312, 654 milagre econômico (economic miracle) (1967–1974), 96, 147–48, 207, 401, 401t, 468, 647t milícias (local auxiliary troops), 23 milk imports, 276 mill owners (senhores de engenho), 200 milréis, 45, 58n41

Minas Gerais, 462n3 coffee production, 272, 275, 279, 279t, 280, 284 environmental issues, 578 gold excavation, 272 historical perspective on, 23–24, 42 hydroelectric stations, 358 import shares, 434 industrialization, 94 population and reported income, 476, 477t trade with other countries, 434 miners (mineradores), 200 Minerva, 687t minidesvalorações (crawling peg) policy, 647, 647t minimum price policy, 295 minimum wage (salário mínimo, SM), 111, 117, 247, 474, 477–78, 556, 572n4 mining Brazilian exports, 399f, 407 Brazilian MNCs, 687–88, 687t foreign direct investment in, 674, 675t productivity, 343, 343f, 346, 347f regional disparities, 426–29, 427t SOEs, 712, 713f world exports, 407, 407t MINUSTAH, 541 Mitsui, 133 MNCs. See multinational corporations modernization, 12, 546 agricultural, 292–93, 296–304, 309–37 Modernization Program for Public Finance Management (PMAT), 190 modern services, 102–3 Moldova, 640n24 monetary correction (correção monetária), 183 monetary overhang, 117 monetary reform, 118–19 money, 117–18, 202 Mongolia, 370, 371t Monte do Socorro, 201–2 Montessori schools, 499 Morceiro, Paulo César, 12 More Food Africa program, 545 More Food International (MFI), 546 mortgages (hipotecas), 207 most favored nation (MFN) status, 402

index   797 mothers and infants. See women Movement Account (conta movimento), 206–7 Movement of Landless Workers (MST), 300 Mozambique, 536, 541, 544–46, 692 MPF (Ministério Público Federal, Federal Public Prosecutors’ Office), 726, 746, 749, 756–57 MRE (Ministério das Relações Exteriores, Ministry of External Relations) (Itamaraty), 536, 539–42 MST (Movement of Landless Workers), 300 Mueller, Charles, 10 multinational corporations (MNCs) Brazilian, 12, 279, 666, 671, 682–97, 687t emerging market-based (EMNCs), 684–85, 694 ownership advantages, 682–85, 689–93 municipal councils (Senados das Câmaras), 22, 24 municipal hospitals, 598t, 599 municipalities (vilas), 23 Munksjö AB/Ahlstrom Corporation, 733 Murtinho, Joaquim, 203 My Home My Life Program (Programa Minha Casa Minha Vida, PMCMV), 248 Nabuco, Joaquim, 30 Nafinsa (National Financiera), 193n4 NAFTA (North American Free Trade Agreement), 410, 416, 657, 660–61, 668 Namibia, 545–46 National Agricultural Research System (SNPA), 315, 332n14 National Bank for Economic and Social Development. See Banco Nacional de Desenvolvimento Econômico e Social (BNDES) National Bank of Economic Development (Banco Nacional de Desenvolvimento Econômico, BNDE), 66–67, 70, 92, 132, 143n14, 216n37 National Bank of Housing (Banco Nacional da Habitação, BNH), 215n21 national champions, 138–39, 694–95, 702, 714 National Coffee Council (Brazil), 90 National Confederation of Industries (Confederação Nacional das Industrias, CNI), 66, 722

National Council of Justice (Conselho Nacional de Justiça, CNJ), 749, 756 National Council on Food Security (Conselho Nacional de Segurança Alimentar e Nutricional, CONSEA), 542, 545 National Energy Plan 2030 (Brazil), 372 National Financiera (Nafinsa), 193n4 National Health Agency (Agência Nacional de Saúde, ANS), 601–2, 610 National Household Sample Survey (Pesquisa Nacional por Amostra de Domicílios, PNAD), 470, 478–79, 484n1, 514, 604–7, 606t National Institute for Space Research (INPE), 581–83 National Institute of Social Security (Instituto Nacional de Previdência Social, INPS), 515 nationalism, 141 nationalization, 705 National Monetary Council (Conselho Monetário Nacional, CMN), 109, 206, 535, 583 National Privatization Program (Brazil). See Programa Nacional de Desestatização (PND) National Program for the Strengthening of Family Farming (Programa Nacional de Fortalecimento da Agricultura Familiar, PRONAF), 546 National Rural Credit System (NSRC or SNCR), 278, 295, 313 National School Nutrition Program (Programa Nacional de Alimentação Escolar, PNAE), 542 National Security Council (Brazil), 67 National Service for Industrial Apprenticeship (Serviço Nacional de Aprendizagem Industrial, SENAI), 498, 539, 541 National Social Protection Policy (Kenya), 544 National Steel Company (Companhia Siderurgica Nacional, CSN), 6, 91, 99, 116–17, 688 National Supply Superintendence (Superintedência Nacional de Abastecimento, SUNAB), 721 National System of Innovation (Brazil), 689–90

798   index National Transport Confederation (Brazil), 755 National Treasury (Brazil), 202 Readjustable Obligations (Obrigações Reajustáveis), 206 subsidies to BNDES, 186, 190–93, 191f National Treasury Bond (BTN), 115, 120 National Treasury Indexed Liabilities (ORTN), 109–11 National Treasury Liability (OTN), 111, 115 National Water Supply and Sanitation System (PLANASA), 384, 391n18 natural gas, 360f, 361, 361f, 369, 635, 635t natural resource–based (NRB) products, 1–2, 10, 12 NDB (New Development Bank), 2, 549, 627–29 neoliberalism, 98, 124–46, 185, 657 neoliberal pragmatism, 133–34 neostructuralism, 79–80 Nepal, 662n2 nepotism, 756 Nestlé-Garoto, 724, 730 Netherlands, 318, 506 Brazilian population, 655–56, 656t colonization of Brazil, 33 export share, 650–52, 651f, 651t services employment, 563t Netto, Antônio Delfim, 105, 124, 267 Neves, Tancredo, 469 New Brazil Plan (Collor Plan), 106, 116–21, 120f, 122n6, 223 New Christians (cristãos-novos, converted Jews), 19, 22, 268 new developmentalism, 141, 232 New Development Bank (NDB), 2, 549, 627–29 New Macroeconomic Matrix, 406 New Zealand, 53, 298, 563t, 626 Nigeria, 541, 545, 691–92, 754 Nóbrega, Mailson da, 115 Non-Aligned Movement, 537 non-tariff barriers (NTBs), 394, 396, 401–2, 647t, 652–53 Nordic countries, 562, 563t. See also specific countries Normano, João, 71 Normative Instruction 70, 398

Normative Instruction 113, 398 Noronha, Antônio de, 24 North, Douglass, 18, 310 North America, 407, 407t. See also specific countries North American Free Trade Agreement (NAFTA), 410, 416, 657, 660–61, 668 North Brazil agricultural development, 277–78 economic development, 426–29, 427t import shares, 434 income, 476, 477t, 480 population, 424, 425f, 476, 477t soybean production, 280, 280t states and regions, 423–24, 424f water and sanitation, 383–84, 383t Northeast Brazil, 67–68, 446–63 agriculture, 284, 301–2 climate, 448, 451 climate change, 438–39, 587–88 confisco cambial, 84n29 economic development, 426–29, 427t, 447–57, 448t, 462n1 economic policy, 434–35 ethnic groups, 448 European occupation, 450 Furtado on, 73–74 GDP, 446 geography, 447–49, 462n1 health resources, 599 historical background, 448–50 hospital beds, 597 import shares, 434 incomes, 320, 455, 455t, 476, 477t, 480 industrialization, 95 metropolitan areas, 446 per capita GDP, 447, 447f, 450 population, 424, 425f, 446, 476, 477t settlement of, 425–26 social indicators, 447–48, 448t soybean production, 280, 280t states and regions, 423–24, 424f SUDENE (Superintendência de Desenvolvimento do Nordeste, Superintendency for the Development of the Northeast), 67–68, 81, 458–59, 748 water and sanitation, 383–84, 383t

index   799 Norway, 171n17, 537, 563t Notes of Agribusiness Credit (LCA), 282 NRB (natural resource–based) products, 1–2, 10, 12 NSRC (National Rural Credit System), 278, 295, 313 NTBs (non-tariff barriers), 394, 396, 401–2, 647t, 652–53 nuclear energy, 360f, 362f, 365 Nurkse, Ragnar, 69 nutritional security, 542–46 Obama, Barack, 641n29 Obrigações Reajustáveis (Readjustable Obligations), 206 ODA (Official Development Assistance), 536 Odebrecht, 138–39, 194n18, 388–89, 695 OFDI (outward foreign direct investment), 683–89, 685f, 686t Official Development Assistance (ODA), 536 OGMA, 675, 691 oil and gas, 691–92 Brazilian MNCs, 687–88, 687t domestic energy supply, 360f, 361, 361f, 362f exports, 407, 407t, 635, 635t gasoline cartels, 737n7 imports, 359 income per capita and oil consumption, 369–70, 370t, 371t natural gas, 360f, 361, 361f, 369, 635, 635t offshore oil fields, 691 PAC program, 386–87 petroleum, 359, 360f, 361, 361f, 635, 635t petroleum prices, 97 SOEs, 712, 713f subsidies, 519 oil shocks, 97–98, 396, 397f, 400–401, 647t, 648 Oji, Funmi, 391n26 old age transfers, 516–18, 517f, 520–21, 520t, 529–30 OLI (Ownership-Location-Internalization) paradigm, 683–84 Olympic Games, 384–85, 577–78 O’Neill, Jim, 621, 641n32 onion imports, 276 Opel, 363–64 openness, 57n28, 650–56

Operation Anaconda, 748–49, 756 Operation Northeast (Furtado), 75 orange crops, 277, 277t orange juice exports, 297, 411 Ordenações Filipinas (Philippine Order), 200 Organization for Economic Co-operation and Development (OECD), 5, 47, 56n24, 137, 187, 298, 536, 727 Competition Committee, 732 Convention on Combating Bribery of Foreign Public Officials in International Business Transactions, 761 Latin American Competition Forum, 732 PSE/FGR, 283 Orphans’ Judge (Juízo dos Órfãos), 200 ORTN (National Treasury Indexed Liabilities), 109–11 Oswaldo Cruz Foundation (Fiocruz), 541 OTN (National Treasury Liability), 111, 115 Otto cycle motors, 363 Ouro Preto Protocol, 402 outward foreign direct investment (OFDI), 683–89, 685f, 686t overseas development assistance, 536 Owens Corning/Saint Gobain, 735 ownership land (see property rights) public (see state-owned enterprises (SOEs)) ownership advantages, 682–85, 689–94 Ownership-Location-Internalization (OLI) paradigm, 683–84 Oxfam, 543 PAA (Food Acquisition Program), 282, 542, 545 PAA-Africa, 542, 545 PABA (Plano de Ação de Buenos Aires, Buenos Aires Plan of Action), 537 PAC (Programa de Crescimento Acelerado, Growth Acceleration Program), 377–78, 386–89 PAC 1, 386, 391n21 PAC 2, 386, 391n21 PADCT (Programa de Apoio ao Desenvolvimento Científico e Tecnológico), 505 PAEG (Government Economic Action) Plan, 106–10, 111f, 121

800   index Pakistan, 654–56, 655t Panamerican Health Organization, 593 Pão de Açúcar, 735 papelistas, 202, 214n11 Pará, 202, 272, 305–6n17, 434 Paraguay, 647t, 654–56, 656t, 693 Paraguayan War, 202 Paraíba, 31 Paraná, 273–79, 279t, 302–3, 426, 434 Pardo, Caio, Jr., 56n23 Partido dos Trabalhadores (PT, Workers’ Party), 8–9, 80, 100–101, 126, 134, 223–24, 378, 498, 649–50 mensalão (big monthly stipend) scandal, 227, 749–50, 754–56 patents, 506 patrimonialism, 26 Pau Brasil (Brazilwood), 20, 71, 267–68, 440n3 paulista capital market, 56n17 Paz, Pedro, 70 PBF. See Bolsa Familia peacekeeping, 548 Peasant League, 67 Pedro I, 269 Pedro II, 269, 272, 358 peer reviews, 725, 732 pension, social, 516–18, 517f, 520–21, 520t, 529–30 pension funds, 137, 143n19, 713–14 PEP, 285 PEPRO (Equalizing Premium Paid to Growers program), 283, 285 Pernambuco, 20–22, 28–31, 268, 272, 425–26 peronismo, 76 Perroux, François, 64–65 personal services, 354n7, 562, 563t, 564–65t, 566–69, 567t Peru, 370, 371t, 536, 563t, 692–93 Pesquisa de Orçamentos Familiares (POF, Household Budget Survey), 471, 527, 530n6 Pesquisa Nacional de Saúde (PNS, National Health Survey), 597, 604–5, 606t Pesquisa Nacional por Amostra de Domicílios (PNAD, National Household Sample Survey), 470, 478–79, 484n1, 514, 604–7, 606t

PETI (Programa de Erradicação do Trabalho Infantil ), 512, 515f, 518, 521, 531n12 Petrobrás, 100, 126, 139, 227, 359, 363, 374, 388–89, 683, 688, 690–92, 705, 712 Car Wash scandal, 3, 14, 101, 387, 548, 686, 735, 744, 750–55, 758–59 CENPES laboratories, 368–69 challenges facing, 695–96 foreign assets, 687t petroleum, 359, 360f, 361, 361f petroleum exports, 635, 635t petroleum imports, 359 petroleum prices, 97 Petros, 133 pharmaceutical industry, 257t Philippine Order (Ordenações Filipinas), 200 Philippines, 754 PIA data, 355n15 Piaget schools, 499 Piauí, 283–84, 426–29, 427t, 508n11 Piesch coefficient, 476 Pigou, Arthur, 70 PinCADE, 733–34 Pinheiros river, 577–78, 580 Pinochet, Augusto, 127 Pinto, Anibal, 65, 70 PISA (Programme for International Student Assessment), 493–94 Pitágoras/Kroton, 503 PLANASA (National Water Supply and Sanitation System), 384, 391n18 Plano Brasil Sem Miséria, 522 Plano de Ação de Buenos Aires (PABA, Buenos Aires Plan of Action), 537 Plano de Metas (Targets Plan), 7, 70, 92, 244–45 Plano Real (Real Plan), 9–11, 99, 105–6, 121, 121n3, 125, 133–34, 209, 214, 223, 239n4, 296–97, 470, 647t, 710–11 planter class, 23 PMAT (Modernization Program for Public Finance Management), 190 PMCMV (Programa Minha Casa Minha Vida, My Home My Life Program), 248 PNAD (Pesquisa Nacional por Amostra de Domicílios, National Household Sample Survey), 470, 478–79, 484n1, 514, 604–7, 606t

index   801 PNAE (Programa Nacional de Alimentação Escolar, National School Nutrition Program), 542 PND. See Programa Nacional de Desestatização PNS (Pesquisa Nacional de Saúde, National Health Survey), 597, 604–5, 606t POF (Pesquisa de Orçamentos Familiares, Household Budget Survey), 471, 527, 530n6 polarization, 481, 482f police, 747–49, 756–57 political view, 181 Polity IV index, 631t, 632–33, 641n37 pollution, 577–78, 582, 588–89, 590n7 Pombal, Sebastião de Carvalho e Melo, Marquis of, 23–25 Pombaline Rule (Período Pombalino), 200 Pontifical Catholic University, 80 population, Brazilian, 639n7 characteristics of, 594–96, 594t, 596t foreign-born, 654 growth of, 55n4, 328, 594t migrant, 654–56, 655t outside Brazil, 655–56, 656t rural, 312, 594t urban, 312 populism, 234 Porcile, Gabriel, 84n41 Portal Brasil, 305n13 Porto Alegre, 385, 590n7 ports, 184, 383, 712, 713f Portugal Brazilian population, 19–20, 654–56, 656t colonization of Brazil, 17–40, 268–69 (see also imperial period) Portuguese language, 493 Postal Service, 749 post-ISI, 96 potato crops, 277, 277t poultry, 297, 301, 411 poverty, 164, 530n1 anti-poverty transfers, 511–34 reduction of, 525–29, 538 trends, 513–14, 513f working poor, 561, 566 PPCDAm (action plan for combating deforestation in the Amazon), 583, 585

PPP (public-private partnership) arrangements, 379, 387 Prado, Eleuterio, 81 Prado, Mariana, 14 pragmatism, 133–34, 548 Prebisch, Raúl, 17, 63–66, 70, 79–81, 95 The Economic Development of Latin America and Its Principal Problems, 68 Economic Survey of Latin America: 1949 (CEPAL), 82n15, 83n16 preferential trade agreements (PTAs), 622, 626, 640n19, 648–49 PRE-SAL deposits, 359, 368–69, 374 Pre-Salt investments, 404 Presbyterians, 499 Previ, 133, 137 Previdência Social Rural, 511–16, 515f, 522–26, 529–30 main components, 520–21, 520t transfers, 515–16, 516f, 518, 523 price controls, 111, 117, 139, 282–83, 285, 369, 737n7 minimum price policy, 295 Price of Reference Unit index, 115 Price Reference Unit (URP), 114 primary goods, 675, 678, 678f primary schools (ensino fundamental), 490–92, 508n2 printing, 36n9, 251 private banks, 202, 211, 211f private health, 599–602, 609–10 private hospitals, 598t, 599 Private Option Risk Premium program (PROP), 282–83 private schools, 498–99, 502–3 privatization, 116–17, 132–33, 137–38, 184–85, 214, 702–3, 708–12 National Privatization Program (Programa Nacional de Desestatização, PND), 132, 709–10 second National Privatization Program (PND II), 7–8, 97, 188, 229, 244–45 PROAGRO (Guarantee Program for Agricultural Activity), 279, 282 PROALCOHOL, 362–63 ProCADE (Procuradoria Geral), 725 PROCAP, 691

802   index Procuradoria Geral (ProCADE), 725 producer support estimate (PSE), 334n24 Productive Development Policy, 186, 194n12 productive services employment, 562, 563t, 564–65t, 566–70 low pay and wage inequality, 567, 567t productivity, 101 agricultural, 325, 326t, 327–28 labor, 189, 338–39, 339f, 339t, 343, 343f manufacturing, 254–58, 338–57 sectors of, 241–419 service, 340–50, 347f, 349f total factor productivity (TFP), 162–63, 163t, 166, 189, 338, 411, 666–67 productivity growth, 12, 354n6 PROER (Programa de Estímulo à Reestruturação e ao Fortalecimento do Sistema Financeiro Nacional, Program of Incentives to Strengthen the National Financial System), 209, 216n30 PROES (Programa de Incentivo à Redução do Setor Público Estadual na Atividade Bancária, Program of Incentives to Reduce States’ Participation in Banking Activity), 209–10, 216n30 professional services, 564–65t, 566 Programa de Apoio ao Desenvolvimento Científico e Tecnológico (PADCT), 505 Programa de Crescimento Acelerado (PAC, Growth Acceleration Program), 386–89 PAC 1, 386, 391n21 PAC 2, 386, 391n21 Programa de Erradicação do Trabalho Infantil (PETI), 512, 515f, 518, 521, 531n12 Programa de Estímulo à Reestruturação e ao Fortalecimento do Sistema Financeiro Nacional (PROER, Program of Incentives to Strengthen the National Financial System), 209, 216n30 Programa de Incentivo à Redução do Setor Público Estadual na Atividade Bancária (PROES, Program of Incentives to Reduce States’ Participation in Banking Activity), 209–10, 216n30 Programa de Metas (Targets Program), 70, 183 Programa Minha Casa Minha Vida (PMCMV, My Home My Life Program), 248

Programa Nacional de Acesso ao Ensino Técnico e Emprego (PRONATEC), 498–99 Programa Nacional de Alimentação Escolar (PNAE, National School Nutrition Program), 542 Programa Nacional de Desestatização (PND, National Privatization Program), 132, 709–10 second (PND II), 7–8, 97, 188, 229, 244–45 Programa Nacional de Fortalecimento da Agricultura Familiar (PRONAF, National Program for the Strengthening of Family Farming), 546 Programa Saúde da Familia (PSF, Family Health Program), 527, 597 Programa Universidade para Todos (PROUNI), 502–3 Programme for International Student Assessment (PISA), 493–94 Program of Incentives to Strengthen the National Financial System (Programa de Estímulo à Reestruturação e ao Fortalecimento do Sistema Financeiro Nacional, PROER), 209, 216n30 Projeto Crescer (Project Growth), 378, 389 ProManejo, 585 PRONAF (Programa Nacional de Fortalecimento da Agricultura Familiar, National Program for the Strengthening of Family Farming), 546 PRONATEC (Programa Nacional de Acesso ao Ensino Técnico e Emprego), 498–99 property rights, 581 land disputes, 303, 305–6n17 land distribution, 30–34, 44, 55n11, 288–308, 290t ProSavana, 546, 549 protectionism, 79, 137, 231, 401–4, 657, 659 protection rates, 646 Protestants, 499, 502 PROUNI (Programa Universidade para Todos), 502–3 Prova Brasil, 492–93 Provão, 504 PSE (producer support estimate), 334n24 PSE/FGR, 283

index   803 PSF (Programa Saúde da Familia, Family Health Program), 527, 597 PSI (investment maintenance program), 186, 190 PT. See Partido dos Trabalhadores PTAs (preferential trade agreements), 622, 626, 640n19, 648–49 publications, 506 public banks, 211, 211f. See also specific banks public education. See education public health care. See health care public illumination, 358. See also electricity public managers or gestores (EPPG), 733, 738nn19–20 public-private partnership (PPP) arrangements, 379, 387 public research. See research and development public safety, 386–87 public savings, 228–30, 228t public sector, 45–46, 570 public transportation, 382–83. See also transportation public universities, 500–502 public utilities, 184–85 publishing, 36n9, 251 Putin, Vladimir, 641n29 Putting Corruption Out of Business survey, 754 Quadros, Jânio, 67, 95 quasi-stagnation, 221–27 quinto, 27 Rada, Nicholas, 334n26 RAET (Regime de Administração Especial Temporária, Temporary Special Administration Regime), 209, 216n31 railways, 57n32, 269–70, 275, 379–83 PAC program, 386–87 privatization, 184 ranching, 272 Rands, Alexandre, 13 Readjustable Obligations (Obrigações Reajustáveis), 206 Reagan, Ronald, 127 real, 99, 236–37, 715n8 real estate bills of exchange (letras imobiliárias), 207 real estate services, 564–65t

Real Plan (Plano Real), 9–11, 99, 105–6, 121, 121n3, 125, 133–34, 209, 214, 223, 239n4, 296–97, 470, 647t, 710–11 recession, 101, 234–35, 258–60, 482–83 Recife, 74, 84n31, 446 REDD (Reducing Emissions from Deforestation and forest Degradation) projects, 584 Rediscount Portfolio (Carteira de Redesconto), 205 Regime de Administração Especial Temporária (RAET, Temporary Special Administration Regime), 209, 216n31 Regime Geral de Previdência Social, 516 regional development, 712, 713f regional development banks, 193n6 regional differences, 13, 73, 94–95, 408–10, 423–45, 484n9. See also specific regions agricultural, 426–29, 427t, 429–30 in education, 428f, 429, 460, 461f in GDP, 426–29, 427t Human Development Index, 428f, 429 in income, 428f, 429, 432–35, 455, 455t, 480–81 regional leadership, 548 regulation(s), 137–38, 185, 388, 390 Relatório sobre Cooperaçāo Brasileira para o Desenvolvimento International (COBRADI, Report on Brazilian Cooperation for International Development), 543 religious schools, 499, 502 Renato de Souza, Paulo, 492, 498 Renda Mensal Vitalícia, 517, 517f renewable energy domestic supply, 360f, 362f PAC program, 386–87 repression, financial, 179 research and development (R&D), 505–6 agricultural, 293–94, 312–24, 332n12, 333n17, 333n19, 406 focused research, 316–17 genetic research, 334n28 SOEs, 712, 713f research universities, 500–501, 690 retail sales comparison with manufacturing physical production, 225, 226f employment, 564–65t

804   index Revista Brasileira de Economia, 82n6 RFSSA (Federal Railroad Corporation), 380–81 rice areas of cultivated land, 270–72, 270t crop improvements, 322–23 crop production, 277, 277t domestic supply, 301 per capita production, 311 productivity gains, 325–26, 326t Ricúpero, Rubens, 722 Rio da Prata, 214n3 Rio de Janeiro, 424f agricultural development, 314 coffee production, 267, 269, 272, 280 credit before banks, 200 current financial crisis, 101 economic integration, 434 electricity services, 359, 379 environmental issues, 577–78, 580 first private banks, 202 foreign banks, 202 Galeão airport, 384–85 historical perspective on, 20, 23, 28–29, 42 human capital, 431 import shares, 434 industrialization, 94 manufacturing, 430 oil fields, 691 pollution, 590n7 population, 424, 476, 477t public illumination, 358 recession, 387 reported income, 476, 477t settlement, 426 trade, 434 Rio de Janeiro Botanical Garden, 269 Rio Grande do Norte, 31 Rio Grande do Sul, 28–29, 277, 302 risky behavior, 605–6, 606t roadways, 275, 284, 292–93, 303–4 Robinson, Joan, 67 Rockefeller Foundation, 333n18, 504–5 Rondônia, 304–6n17, 756 Rosenstein-Rodan, Paul, 69–70 Rousseff, Dilma, 8–10, 141, 224–27, 230, 234, 258, 522, 539, 580, 648–49, 711

competition policy, 725, 730 impeachment, 10, 101, 227, 389, 548, 751 industrial policy, 100–101, 227, 714 Royal Dutch/Shell, 692 rubber, 251, 257t rubber exports, 43t, 45, 58n37, 271t, 272, 311 rural credit, 278, 285, 295, 312–13, 332n5 rural extension, 312–14 Rural Labor and Land Statute, 278 rural population, 312, 594t Rural Product Note (Cédula de Produto Rural, CPR), 279 rural unemployment, 554 rural workers, 278, 514–16, 515f, 520–21, 520t, 529 Russia, 506. See also BRICS (Brazil, Russia, India, China, South Africa) countries demand for agricultural products, 328, 329t economy, 639n1 exports, 631t, 632, 635, 635t, 651f foreign direct investment (FDI), 636–37, 637t GDP, 631–32, 631t growth rate, 170n7 imports from Brazil, 408, 409t income growth, 156t migrant population, 654–56, 655t Polity IV rank, 631t, 632–33 poverty reduction, 530n5 trade, 633–35, 634t, 652–54, 653t trade agreements, 626, 640n24 Rwanda, 545 Sabó, 692–93, 695 Saint Gobain, 735 salário mínimo (SM). See minimum wage Salvador, 426, 446, 590n7 Samarco dam, 590n2 Samsung, 684–85 sanguessuga (bloodsucker) scandal, 756 sanitation. See water and sanitation Santa Catarina, 277, 302, 434 Santas Casas de Misericordia (Holy Houses of Mercy), 29, 200, 597 Santiago do Iguape, 32 Santos Neto, Nicolau dos (Lalau), 747–48 São Paulo, 424f agriculture, 270–73, 270t, 284, 292, 294

index   805 coffee production, 58n37, 267–75, 279, 279t, 280 economic development, 426–29, 427t economic integration, 434 electricity services, 359 environmental issues, 577–78, 580, 586 historical perspective on, 22–23, 30–32, 42, 55n10 human capital, 431 immigration, 55n10 industrialization, 94 land distribution, 31–32 manufacturing, 430 pollution, 590n7 population, 424, 654 population and reported income, 476, 477t port of Santos, 383 settlement, 426 slavery, 30 sugarcane industry, 31–32, 268, 586 trade, 434 São Paulo Guarulhos Airport, 384–85, 391n19 São Paulo Railway, Light and Power Company, 358 São Paulo Regional Labor Court of the Second Region (TRT), 747–48, 756 São Paulo State Bank (Banespa), 210 São Tomé and Príncipe, 28, 536, 662n2 São Vicente, 20, 22–23 Sarney, José, 68, 98, 223, 469, 648 Sarquis, Sarquis J. B., 11 Saudi Arabia, 156t, 654–56, 655t Saussure, Ferdinand de, 65 savings foreign, 233, 237 public, 228–30, 228t Savings Bank (Caixa Economica), 201–2 savings rate, 388 savings replacement, 239n11 SBDC (Brazilian Antitrust System), 722–23, 730, 734 scandals. See corruption; specific scandals school enrollment, 50, 490, 492, 496, 500–502. See also education School Stipend (Bolsa Escola), 474, 515f, 518–19, 527–29, 531n12 Schumpeterian theory, 135–36

Schwartzman, Simon, 506 science, 506, 508n12 science-based agriculture, 315–28 scientific cooperation, international, 318–19, 333n19 SDE (Secretaria de Direito Econômico, Secretary of Economic Law), 722–25, 734 SEAE (Secretaria de Acompanhamento Econômico), 722, 724–25, 733 secondary schools (ensino médio), 490–92, 495–96, 508n2 Secretaria de Acompanhamento Econômico (SEAE), 722, 724–25, 733 Secretaria de Direito Econômico (Secretary of Economic Law, SDE), 722–25, 734 Secretaria de Reforma do Judiciário (SRJ, Office for Judicial Reform), 749 Secretaria Especial de Controle de Empresas Estatais (SEST, Special Secretariat of State Enterprise Control), 704–5, 707 Secretariat of Internal Control, 746 security safeguard measures (SSM), 636 SELIC (Sistema Especial de Liquidação e Custodia, Special Custodial and Clearing System), 138, 191 Semiarido region, 302 semi-manufactured goods, 408, 409t Sen, Amartya, 545 Senados das Câmaras (municipal councils), 22, 24 SENAI (Serviço Nacional de Aprendizagem Industrial, National Service for Industrial Apprenticeship), 498, 539, 541 Sendas, 735 Senegal, 536, 541, 545–46, 662n4, 692, 754 Serbia, 370, 371t Sergipe, 359 Serra, José, 548 Serra do Mar, 49 Sertão, 448 service firms, 343–44 services sector nineteenth century, 42–43 early twentieth century, 58n44 average earnings, 566 average hours worked per week (AHWPW), 568

806   index services sector (cont.) average wages, 344 calculation, 355n16 competitiveness, 346–51 contribution to manufacturing, 346 contribution to productivity, 346, 347f, 348–51, 349f cost services, 348–51, 349f distributive, 562, 563t, 564–65t, 566–67, 567t, 568–70 employment, 561–70, 563t, 564–65t food and catering, 354n7 foreign direct investment in, 674–75, 675t, 676t, 677, 678f, 679 future directions, 53 informal workers, 557–61, 559f intermediate, 346 job turnover, 343 modern, 102–3 net wages/value added per worker ratio, 354n7 personal, 354n7, 562, 563t, 564–65t, 567–69, 567t productive, 562, 563t, 564–65t, 566–70, 567t productivity, 340–50, 343f, 347f, 349f regional disparities, 431–32 share of GDP, 93–94, 93t, 102, 340–42, 341f share of household budgets, 342–43 share of labor, 93–94, 93t, 102–3 social, 562, 563t, 564–65t, 566–70, 567t SOEs, 712, 713f tax burden, 344 union density, 568 value, 348–51, 349f wages, 354n8, 567, 567t workers’ characteristics, 567–68, 570 working poor, 566 servicification, 168 Serviço Nacional de Aprendizagem Industrial (SENAI, National Service for Industrial Apprenticeship), 498, 539, 541 sesmarias (land grants), 30–33, 44, 268 SEST (Secretaria Especial de Controle de Empresas Estatais, Special Secretariat of State Enterprise Control), 704–5, 707 Seven Essays on the Brazilian Economy (Barros de Castro), 80

Seventh-Day Adventists, 499 sewage collection. See water and sanitation SFH (Sistema Financeiro da Habitação, Financial System for Housing), 207, 215n21 Siemens, Werner von, 358 Silva, Peri, 10 silver extraction, 35n6 similarity test, 131–32, 143n13 Similars Law (Lei de Similaridade Nacional), 647t Simonsen, Mario Henrique, 108, 737n5 Simonsen, Roberto, 71 simulations, 157–60, 158t, 159f, 160f Singapore, 149 Singer, Hans W., 73 Single Registry system (Cadastro Único), 521, 542, 550n1 Sirohi, Rahul A., 103n8 Sistema de Avaliação da Educação Básica (SAEB), 492–93 Sistema de Prestação de Contas Eleitorais (SPCE, Accounting System of Campaign Financing), 750 Sistema Especial de Liquidação e Custodia (SELIC, Special Custodial and Cleaning System), 138, 191 Sistema Financeiro da Habitação (SFH, Financial System for Housing), 207, 215n21 Sistema S, 497 Sistema Único de Assistência Social (SUAS), 517 Sistema Único de Saúde (SUS, Unified Health System), 597–99, 612 “Sketch of a Program of Development for Brazil” (Furtado), 66 skin color, 478–80 slavery, 127 enslaved population, 55n4 historical perspectives, 19–21, 27–30, 28t, 35, 43–44, 268 Voyages: The Trans-Atlantic Slave Trade Database, 36n12 SM (salário mínimo). See minimum wage small agricultural producers, 298–300 Smith, Adam, 35n3

index   807 SNCR (National Rural Credit System), 278, 295, 313 SNPA (National Agricultural Research System), 315, 332n14 Sobral, 508n5 soccer, 384–85 social assistance, 511–12, 520–25, 530n2 classical, 529 conventional, 516–18 emergence of, 514–22, 515f for informal rural workers, 514–16, 515f LOAS (Lei Orgânica da Assistência Social), 517 main programs, 520–21, 520t pensions, 516–18, 517f, 522–23, 523f regional differences, 484n9 strategies for inclusion, 529 Social Assistance Reference Centers (CRAS), 544 social contract, 141 social debt, 514–15 social development in Africa, 542–46 South-South cooperation for, 535–51 Social Investment Fund (Finsocial), 193n7 “Social Policies for Development” seminars, 544 social security, 109, 712, 713f social services, 562, 563t, 564–65t, 566–67, 567t, 568–70 Sociedades de Crédito, Financiamento e Investimento (Societies of Credit, Funding, and Investment) (financeiras), 215n20 socio-demography, 556, 594–95, 594t SOEs. See state-owned enterprises Sousa, Tomé de, 22 South Africa, 642n46. See also BRICS (Brazil, Russia, India, China, South Africa) countries demand for agricultural products, 328, 329t export share, 631t, 632, 650–52, 651f, 651t foreign direct investment (FDI), 636–37, 637t GDP, 631–32, 631t IMF quota, 642n47 income per capita, 156t, 370, 371t oil consumption, 370, 371t

Polity IV rank, 631t, 632–33 poverty reduction, 530n5 services employment, 563t trade, 626, 633–35, 634t, 640n24, 652–54, 653t South America. See also specific countries share of world exports, 407, 407t water and sanitation, 383 South Brazil agriculture, 291–92, 302 confisco cambial, 84n29 economic development, 426–29, 427t health resources, 599 hospital beds, 597 individual incomes, 455, 455t, 476, 477t population, 424, 425f, 476, 477t soybean production, 280, 280t, 284 states and regions, 423–24, 424f water and sanitation, 383–84, 383t Southeast Brazil agricultural frontier, 291–92 economic development, 426–29, 427t, 434, 452 health resources, 599 hospital beds, 597 hydroelectric stations, 358 import shares, 434 individual incomes, 455, 455t industrialization, 94–95 population, 424, 425f, 426 soybean production, 280, 280t states and regions, 423–24, 424f water and sanitation, 383–84, 383t Southern African Customs Union, 404, 417n7, 640n24 South Korea, 5, 149, 188–89 export share, 650–52, 651f, 651t foreign direct investment in, 671 growth rate, 155 income per capita, 156t, 341–42, 341f internationalization, 671 labor productivity, 338 multinational corporations from, 684–85 share of services in GDP, 341–42, 341f trade/GDP ratio, 132 trade restrictiveness, 652–54, 653t South-South cooperation, 535–51 soybean exports, 297, 332n10, 411

808   index soybeans crop improvements, 322–23 Green Soybean contract (Contrato Soja Verde), 279 production, 277–80, 277t, 280t, 284, 311 productivity gains, 325–26, 326t Spain Brazilian population, 655–56, 656t cooperation with, 318 export share, 650–52, 651f, 651t migrant population, 654–56, 655t SPCE (Sistema de Prestação de Contas Eleitorais, Accounting System of Campaign Financing), 750 specialization, 52–54, 673 SPIAC-B, 547 SRJ (Secretaria de Reforma do Judiciário, Office for Judicial Reform), 749 SSM (security safeguard measures), 636 stabilization (1960–1990), 105–23 Brady Plan, 223 Bresser Plan, 106, 113–14, 114f, 121 Collor Plan (New Brazil Plan), 106, 116–21, 120f, 122n6, 223 Collor Plan II, 120 Cruzado Plan, 106, 111–13, 113f, 121 Cruzado II, 113 PAEG (Government Economic Action) Plan, 106–10, 111f, 121 Real Plan (Plano Real), 9–11, 99, 105–6, 121, 121n3, 125, 133–34, 209, 214, 223, 239n4, 296–97, 470, 647t, 710–11 Summer Plan, 106, 115, 116f, 121 Stabilization Office (Caixa de Estabilização), 52, 58n41 stagnation, 221–27 Standard Nutritional Units, 331n2 standards of living, 151–52 state bourgeoisie, 128 state capitalism, 141, 143n14 state enterprises. See state-owned enterprises (SOEs) state hospitals, 598t, 599 state-led development, 135, 142n12 state-owned enterprises (SOEs), 6, 143n19, 386, 701–17 current or former, 688–89

distribution by sector, 712, 713f number, 712, 713f number of employees, 712, 713f ownership, 715n7 pension funds, 137, 143n19, 713–14 privatization, 116–17, 132–33, 702–3, 708–11 Secretaria Especial de Controle de Empresas Estatais (SEST, Special Secretariat of State Enterprise Control), 704–5, 707 subsidiaries, 715n7 steel industry Brazilian MNCs, 687–88, 687t (see also Companhia Siderurgica Nacional (CSN)) privatization, 132–33 STF (Supremo Tribunal Federal, Supreme Court), 750–52 strangulation, 76 strangulation points (bottlenecks), 70 structuralism (developmentalism), 6–7, 63–88, 451–52 desenvolvimentismo, 6–7, 63–88 new or neostructuralism, 79–80, 141, 232 structural reforms, 127, 561–70 SUAS (Sistema Único de Assistência Social), 517 Sub-Saharan Africa, 10, 535–36 subsidies, 138, 178, 189–90, 283, 364–65, 519 Subsidy for Rural Insurance, 282 subsistence farming, 327–28 SUDAM (Superintendência para o Desenvolvimento da Amazônia, Amazonian Development Superintendency), 748 Sudan, 662n2, 662n4 SUDENE (Superintendência de Desenvolvimento do Nordeste, Northeast Development Superintendency), 67–68, 81, 458–59, 748 sugarcane plantations, 18, 20–21, 30–32, 36nn15–17 sugarcane planters (lavradores), 31, 36n16 sugarcane production, 268, 362–65, 590n10 areas of cultivated land, 270–72, 270t credit, 199–200 crop production, 276, 276f, 277, 277t domestic energy supply from, 360f, 362 environmental issues, 586 ethanol production from, 365, 374

index   809 historical perspective on, 20–21, 47–48 modernization, 297 productivity gains, 325–26, 326t slave labor, 28, 28t, 29 sugar exports, 43t, 271t, 272, 297, 411 sugar mills (engenhos), 21–22, 25, 31–32, 36n15 Summer Plan, 106, 115, 116f, 121 SUMOC (Superintendência de Moeda e Crédito), 92, 103n2, 214–15n13, 398 SUNAB (Superintedência Nacional de Abastecimento, National Supply Superintendence), 721 Sunkel, Osvaldo, 70, 74–75, 79 Superintedência Nacional de Abastecimento (SUNAB, National Supply Superintendence), 721 Superintedência para o Desenvolvimento da Amazônia (SUDAM, Amazonian Development Superintendency), 748 Superintendência de Moeda e Crédito (SUMOC), 92, 103n2, 214–15n13, 398 Superintendência para o Desenvolvimento da Nordeste (SUDENE, Northeastern Development Superintendency), 67–68, 81, 458–59, 748 Super Simples, 560 Support for National and Sub-regional Strategies for Food Security and Poverty Reduction in Latin America and the Caribbean, 545 Supremo Tribunal Federal (STF, Supreme Court), 750–52 Suprinyak, Carlos Eduardo, 81 surplus, primary, 134 SUS (Sistema Único de Saúde, Unified Health System), 597–99, 612 Suzano, 284 Sweden, 238, 563t Switzerland, 506, 563t, 655–56, 656t Synverse Holdings, Inc./WP Roaming III S.À. R.L. (MACH), 733 tablitas, 112, 115, 117 Tailândia, 590n4 Taiwan, 149 Tajikistan, 640n24 Tanzania, 541, 545–46, 662n4

Tanzi effect, 109 Tapajós, 584–85 Targets Plan (Plano de Metas), 7, 70, 92, 244–45 Targets Program (Programa de Metas), 70, 183 Tariff Law, 92, 402 tariff modicity, 367 tariffs, 98, 132, 400–404, 416, 647t, 648 tariff water, 634t, 636, 642n42 Tavares, Maria da Conceição, 77–78, 81 Taxa de Juros de Longo Prazo (TJLP, LongTerm Interest Rate), 138, 191, 194n23 Taxa Referencial (TR, Benchmark Interest Rate), 120 tax burden, 229–30 taxes, 248, 408, 741–44, 743t colonial period, 27 discriminatory, 137 inflation, 106 progressive, 238 rebates, 404, 578 regional, 435 value-added, 109 TCU (Tribunal de Contas da União, Federal Accounting Tribunal), 746–49, 757 teacher training, 496–97 teaching systems (sistemas de ensino), 500. See also education technical education, 497–99, 508n11 technology, 261, 682–83 agricultural, 12, 272 capability building, 689–93 HT-MHT (high and medium-high technology) goods, 243, 250–57, 250f, 253t information and communications (ICT), 348 LT-LMT (low and medium-low technology) goods, 243, 250–51, 250f, 254–56, 253t made in Brazil, 506 technology institutes, 690 Telebrás, 239n10, 690 telecommunications, 184, 690 Telecurso 2000, 500 Temer, Michel, 378, 387–90, 539 Temporary Special Administration Regime (Regime de Administração Especial Temporária, RAET), 209, 216n31 Terra, Osmar, 540–41

810   index tertiarization, 102 textile industry, 24, 89–91, 251, 257t TFP. See total factor productivity Thailand, 339t, 654–56, 655t Thatcher, Margaret, 127, 711–12 Theil’s L index, 471–72, 472f Theil’s T index, 470–72, 471f, 472f thermoelectric power, 367, 368f, 374, 375t Thorp, Rosemary, 83n25 Tietê river, 577–78, 580 Tigre, 692–93 TJLP (Taxa de Juros de Longo Prazo, LongTerm Interest Rate), 138, 191, 194n23 TNCs (transnational corporations), 253, 664, 666, 671. See also multinational corporations (MNCs) tobacco exports, 271t, 272 tobacco industry, 91, 251, 257t, 270–72, 270t Tocantins, 283–84, 305–6n17, 756 total factor productivity (TFP), 162–63, 163t, 166, 189, 338 agricultural, 327–28, 411 foreign direct investment and, 666–67 trade reform and, 402–3 total primary energy supply (TPES), 369, 370f tourism and travel agencies, 354n7 town councils (senados das câmaras), 24 TPP (Trans-Pacific Partnership), 657 TR (Taxa Referencial, Benchmark Interest Rate), 120 tractors, 313 trade, 1–2 agricultural, 411–12 Brazilian, 132, 405–12, 645–63 employment in, 564–65t external, 96 free trade areas (FTAs), 435, 626, 640n22, 640n24 general trends, 623, 624t global, 408, 623, 624t intra-BRICS, 624–26, 625t, 633–34, 634t non-tariff barriers (NTBs), 394, 396, 401–2, 647t, 652–53 preferential trade agreements (PTAs), 622, 626, 640n19, 648–49 restrictiveness, 652–54, 653t slave, 19–21, 27–28, 28t, 36n12, 268

terms of trade, 396, 397f unfair, 400, 403 trade policy, 98, 137, 394–419 historical perspective on, 646–50, 647t liberalization, 131–32, 402–3 traditional agriculture, 288–308 training Sistema S, 497 teacher, 496–97 technical, 497–99 tramways, 380 Trans Amazonian Highway, 382 “A Transformative Agenda for the 21st Century: Examining the Case for Basic Social Protection in Africa” conference, 543 transnational corporations (TNCs), 253, 664, 666, 671. See also multinational corporations (MNCs) Trans-Pacific Partnership (TPP), 657 transparency, 731 Transparency International, 753–56, 753f, 755t transportation, 57n31–32, 382, 386–87, 436 employment in, 564–65t SOEs, 712, 713f transport equipment, 187, 251 Brazilian MNCs, 687–88, 687t production index, 257t treasury financial bills (LFTs), 112 Treaty of Asunción, 647t Treaty of Tordesillas, 22, 268 Trebat, Thomas, 13 triangular cooperation, 536–37, 543, 547–48 Tribunal de Contas da União (TCU, Federal Accounting Tribunal), 746–49, 757 Tribunal Superior Eleitoral (TSE, Superior Tribunal for Electoral Law), 750 Triennial Plan, 108 trilemma, 138 Trinidad and Tobago, 691–92 tropas de linha (professionally trained regular troops), 23 tropical agriculture, 293–94. See also agriculture TRT (São Paulo Regional Labor Court of the Second Region), 747–48, 756 truck transportation, 382. See also transportation

index   811 Trump, Donald, 385, 657, 659 TSE (Tribunal Superior Eleitoral, Superior Tribunal for Electoral Law), 750 Tunisia, 417n7 Tupy, 687t Turkey, 156t, 339t, 651f, 652–54, 653t, 692, 754 Ugá, Maria Alicia, 612n1 Ukraine, 640n24, 654–56, 655t Ultramarine domain, 200 UNASUR (Union of South American Nations, União de Nações Sul-Americanas), 535–36 UNCTAD (United Nations Conference on Trade and Development), 671, 732 UNDP (United Nations Development Program), 536–37, 541, 543–44 unemployment, 164, 234, 247, 258, 554, 555t, 571, 571t UNESCO (United Nations Educational, Scientific, and Cultural Organization), 536–37 unfair trade practices, 403 UNFPA (United Nations Population Fund), 543 UNICAMP (Universidade Estadual de Campinas), 80–82 UNICEF (United Nations Children’s Fund), 544 UNIDO (United Nations Industrial Development Organization), 78 Unified Health System (Sistema Único de Saúde, SUS), 597–99, 612 Unified Social Assistance System, 512 union density, 568, 573n11 Union of South American Nations (UNASUR, União de Nações Sul-Americanas), 535–36 United Arab Emirates, 654–56, 655t United Kingdom, 212–13 anti-corruption institutions, 760 Brazilian population, 655–56, 656t Brexit, 657, 659 cooperation with, 318, 536–37 Department for International Development (DFID), 543–45 export share, 650–52, 651f, 651t income inequality, 577 income per capita, 156t, 369–70, 370t industrial development, 180 migrant population, 654–56, 655t

oil consumption, 369–70, 370t per capita GDP, 50 privatization, 711–12 services employment, 563t SOEs, 705 Thatcher government, 133 United Nations BRICS and, 627, 630 High Level UN Conference on South-South Cooperation, 537 Inter-Agency Cooperation Board, 547 United Nations Children’s Fund (UNICEF), 544 United Nations Conference on Trade and Development (UNCTAD), 671, 732 United Nations Development Program (UNDP), 536–37, 541, 543–44 United Nations Educational, Scientific, and Cultural Organization (UNESCO), 536–37 United Nations Industrial Development Organization (UNIDO), 78 United Nations Population Fund (UNFPA), 543 United States, 238, 691–92 anti-corruption institutions, 760 Brazilian MNCs, 693 Brazilian population, 655–56, 656t cooperation with, 318, 333n18, 536–37 demand for agricultural products, 328, 329t environmental issues, 589–90n2 export share, 650–52, 651f GDP per capita, 46–47, 50, 339 general trends, 623, 624t Glass-Steagall Act, 215n25 imports from Brazil, 408–10, 409t incentives for farmers, 334n24 income inequality, 577 industrial development, 180 interest rates, 229 job loss, 660 labor productivity, 101, 338–39, 339f, 339t manufacturing, 245–46, 660 mergers, 738n11 migrant population, 654–56, 655t, 660 protectionism, 659 PSE/FGR, 283 railway network, 57n32, 382 requirements for automobile efficiency, 374

812   index United States (cont.) science-based agriculture, 333n18 secondary schools, 495 services employment, 563t support to agricultural producers, 298 trade, 132, 400, 408, 652–54, 653t, 658 United States Agency for International Development (USAID), 504–5, 543 United States Department of Agriculture– Economic Research Service (USDAERS), 328, 333n19 Universal Banks, 215n26 universal health care, 608. See also Unified Health System Universidade de São Paulo (USP), 64–65, 80, 690, 693 Universidade Estadual de Campinas (UNICAMP), 80–82 universities public, 500–502 research, 500–501, 690 University of Brazil, 64 University of Paris, 68, 70 urbanization, 50, 312, 386–87 urban population, 312 urban unemployment, 554 URP (Price Reference Unit), 114 Uruguay Brazilian population, 655–56, 656t economic growth before WWI, 53–54 health expenditures, 609 immigration from, 654 literacy, 490 per capita GDP, 56n20 service employment, 563t, 569 Treaty of Asunción, 647t USAID (United States Agency for International Development), 504–5, 543 USDA-ERS (United States Department of Agriculture–Economic Research Service), 328, 333n19 US Eximbank, 137 Usiminas, 132–33 US Steel, 692 usury, 200 Usury Law (Lei da Usura), 198, 205 utilities, 184–85

Vale, 687t, 688, 690, 692, 695, 705, 712 Vale do Paraíba, 267, 269 Vale do Rio Doce, 133, 143n19 Valepar, 133 value-added taxes, 109 value services, 348–51, 349f Vargas, Getúlio, 6, 66, 81, 214–15n13, 223, 703 Venezuela, 187, 626, 647t, 649, 651f, 652–54, 653t Vergolino, José O., 29 Veyance, 733 Vianna, Marcus, 143n14 Viegas Andrade, Mônica, 612n1 vilas (municipalities), 23 Villa, Juan, 13 Villareal, Rene, 83n19 Villela, André, 6 violence, 305–6n17 virtual cooperation, 547 “Vision 2014–2034” (Embrapa), 320–22 Vitória, 590n7 volatility, economic, 164–68 macroeconomic, 162–64 simulations, 157–60, 158t, 159f, 160f Volkswagen, 363–64 Volta Redonda plant, 91 von Hayek, Friedrich, 126–27, 142n4 Votorantin, 138–39 Voyages: The Trans-Atlantic Slave Trade Database, 36n12 WA (Agricultural Warrant), 282 wages, 343, 354n8. See also minimum wage Waldorf schools, 499 Washington Consensus, 8, 128, 142n6, 223, 435, 686–87 water, tariff, 634t, 636, 642n42 water and sanitation Codigo das Aguas (Water Code), 380 contemporary challenges, 383–84, 383t industrial water use, 588–89 PAC program, 386–87 regional disparities, 429 water pollution, 577–78, 588–89 Weber, Max, 25 WEF (World Economic Forum), 340 West Germany, 180. See also Germany

index   813 Westinghouse, 365 wheat production, 277, 277t, 311, 325–26, 326t White Martins, 735 whites, 555t, 556 Williamson, Jeffrey, 53, 142n6 Williamson, John, 128 wind power, 361, 362f, 374, 375t women earning differences, 478–80 health characteristics, 595, 596t labor force participation, 555t, 556, 567–70 social assistance for expectant mothers, 519 worker rights, 278 Workers’ Party. See Partido dos Trabalhadores (PT) Worker Support Fund (Fundo de Amparo ao Trabalhador, FAT), 138 Working Group for the Development of the Northeast (GTDN), 462n2 working hours, 568 working poor, 561, 566 World Bank, 8, 381, 499, 505, 544, 547, 658 Centro Regional de Competencia para America Latina, 732 Control of Corruption (CC) index, 753, 753f Enterprise Survey, 340

World Cup, 384–85 World Development Indicators (WDI), 148 World Economic Forum (WEF), 340 world economy, 1, 262n1, 619–97 world exports, 406–7, 407t, 411 World Food Program (WFP), 536–37, 545 World Health Organization (WHO), 543 world trade, 408, 623, 624t World Trade Organization (WTO), 137, 535–36, 622, 640n22, 641–42n40 BRICS and, 627, 635–36 G20, 638, 642n49 World War I, 51, 89, 204 World War II, 83n26, 91, 205, 395–98 World Without Poverty (WWP), 547 WP Roaming III S.À. R.L. (MACH), 733 Yale University, 68 youth employment, 554, 555t, 564–65t, 571, 571t Zero Hunger (Fome Zero), 542 Zimbabwe, 545–46 Zona Franca de Manaus, 434 Z-Score (Distance to Default), 217n42 Zuleta, Gustavo, 79