Energy, Bio Fuels and Development : Comparing Brazil and the United States [1 ed.] 9780203833858, 9780415567206

Citation preview

Energy, Bio Fuels and Development Comparing Brazil and the United States

Edited by Edmund Amann, Werner Baer and Donald V. Coes

Routledge Studies in Development Economics

Energy, Bio Fuels and Development

This collection examines the important and topical issue of the economic, social and environmental implications of concerted attempts to diversify energy sources away from fossil fuels. The book expertly examines this issue by focusing on the contrasting experiences of two major economies: one developed, and the other a rapidly expanding, emerging market. Energy, Bio Fuels and Development evaluates the experience of Brazil, with elements of that of the United States highlighted for the purpose of comparison. A key area of concern surrounds the causes and consequences of the contrasting routes to bio fuel production represented by sugar cane (in Brazil) and corn (in the United States). The book also places the recent bio fuels drive in perspective by discussing the broader energy policy context. The book shows the complexity and interdependence of the issues involved in moving a society reliant on non-­ renewable energy sources to one based on alternative sources of energy. The key conclusion to emerge is that Brazil, in pursuing a flexible mix of fossil fuels and bio fuels, has greatly diminished its exposure to exogenous energy shocks. The US experience – in particular its development of corn-­based ethanol – has been more problematic, though by no means without successes. It is argued that bio fuels should not be seen as a panacea. There are clear limits to the efficiency and cost effectiveness of current bio fuel production technologies while there remain concerns surrounding potentially adverse effects on food production and rural livelihoods. This book should be an excellent resource for students focusing on economic development, particularly in the areas of energy, bio fuels, rural development and food supply. Edmund Amann is Reader in Development Economics at the University of Manchester, UK. Werner Baer is Lemann Professor of Economics at the University of Illinois, USA. Donald V. Coes is Professor of Economics at the University of New Mexico, USA.

Routledge studies in development economics

  1 Economic Development in the Middle East Rodney Wilson   2 Monetary and Financial Policies in Developing Countries Growth and stabilization Akhtar Hossain and Anis Chowdhury   3 New Directions in Development Economics Growth, environmental concerns and government in the 1990s Edited by Mats Lundahl and Benno J. Ndulu   4 Financial Liberalization and Investment Kanhaya L. Gupta and Robert Lensink   5 Liberalization in the Developing World Institutional and economic changes in Latin America, Africa and Asia Edited by Alex E. Fernández Jilberto and André Mommen

  6 Financial Development and Economic Growth Theory and experiences from developing countries Edited by Niels Hermes and Robert Lensink   7 The South African Economy Macroeconomic prospects for the medium term Finn Tarp and Peter Brixen   8 Public Sector Pay and Adjustment Lessons from five countries Edited by Christopher Colclough   9 Europe and Economic Reform in Africa Structural adjustment and economic diplomacy Obed O. Mailafia 10 Post-­apartheid Southern Africa Economic challenges and policies for the future Edited by Lennart Petersson 11 Financial Integration and Development Liberalization and reform in sub-­Saharan Africa Ernest Aryeetey and Machiko Nissanke

12 Regionalization and Globalization in the Modern World Economy Perspectives on the Third World and transitional economies Edited by Alex E. Fernández Jilberto and André Mommen

19 Finance and Competitiveness in Developing Countries Edited by José María Fanelli and Rohinton Medhora

13 The African Economy Policy, institutions and the future Steve Kayizzi-­Mugerwa

21 Mexico beyond NAFTA Edited by Martín Puchet Anyul and Lionello F. Punzo

14 Recovery from Armed Conflict in Developing Countries Edited by Geoff Harris

22 Economies in Transition A guide to China, Cuba, Mongolia, North Korea and Vietnam at the turn of the twenty-­first century Ian Jeffries

15 Small Enterprises and Economic Development The dynamics of micro and small enterprises Carl Liedholm and Donald C. Mead 16 The World Bank New agendas in a changing world Michelle Miller-­Adams 17 Development Policy in the Twenty-­First Century Beyond the post-­Washington consensus Edited by Ben Fine, Costas Lapavitsas and Jonathan Pincus 18 State-­Owned Enterprises in the Middle East and North Africa Privatization, performance and reform Edited by Merih Celasun

20 Contemporary Issues in Development Economics Edited by B.N. Ghosh

23 Population, Economic Growth and Agriculture in Less Developed Countries Nadia Cuffaro 24 From Crisis to Growth in Africa? Edited by Mats Lundal 25 The Macroeconomics of Monetary Union An analysis of the CFA franc zone David Fielding 26 Endogenous Development Networking, innovation, institutions and cities Antonio Vasquez-­Barquero 27 Labour Relations in Development Edited by Alex E. Fernández Jilberto and Marieke Riethof

28 Globalization, Marginalization and Development Edited by S. Mansoob Murshed 29 Programme Aid and Development Beyond conditionality Howard White and Geske Dijkstra 30 Competitiveness Strategy in Developing Countries A manual for policy analysis Edited by Ganeshan Wignaraja 31 The African Manufacturing Firm An analysis based on firm surveys in sub-­Saharan Africa Dipak Mazumdar and Ata Mazaheri 32 Trade Policy, Growth and Poverty in Asian Developing Countries Edited by Kishor Sharma 33 International Competitiveness, Investment and Finance A case study of India Edited by A. Ganesh Kumar, Kunal Sen and Rajendra R. Vaidya 34 The Pattern of Aid Giving The impact of good governance on development assistance Eric Neumayer 35 New International Poverty Reduction Strategies Edited by Jean-­Pierre Cling, Mireille Razafindrakoto and François Roubaud

36 Targeting Development Critical perspectives on the millennium development goals Edited by Richard Black and Howard White 37 Essays on Balance of Payments Constrained Growth Theory and evidence Edited by J.S.L. McCombie and A.P. Thirlwall 38 The Private Sector After Communism New entrepreneurial firms in transition economies Jan Winiecki, Vladimir Benacek and Mihaly Laki 39 Information Technology and Development A new paradigm for delivering the internet to rural areas in developing countries Jeffrey James 40 The Economics of Palestine Economic policy and institutional reform for a viable Palestine state Edited by David Cobham and Nu’man Kanafani 41 Development Dilemmas The methods and political ethics of growth policy Melvin Ayogu and Don Ross 42 Rural Livelihoods and Poverty Reduction Policies Edited by Frank Ellis and H. Ade Freeman

43 Beyond Market-­Driven Development Drawing on the experience of Asia and Latin America Edited by Makoto Noguchi and Costas Lapavitsas 44 The Political Economy of Reform Failure Edited by Mats Lundahl and Michael L. Wyzan 45 Overcoming Inequality in Latin America Issues and challenges for the twenty-­first century Edited by Ricardo Gottschalk and Patricia Justino 46 Trade, Growth and Inequality in the Era of Globalization Edited by Kishor Sharma and Oliver Morrissey 47 Microfinance Perils and prospects Edited by Jude L. Fernando 48 The IMF, World Bank and Policy Reform Edited by Alberto Paloni and Maurizio Zanardi 49 Managing Development Globalization, economic restructuring and social policy Edited by Junji Nakagawa 50 Who Gains from Free Trade? Export-­led growth, inequality and poverty in Latin America Edited by Rob Vos, Enrique Ganuza, Samuel Morley, and Sherman Robinson

51 Evolution of Markets and Institutions A study of an emerging economy Murali Patibandla 52 The New Famines Why famines exist in an era of globalization Edited by Stephen Devereux 53 Development Ethics at work Explorations – 1960–2002 Denis Goulet 54 Law Reform in Developing and Transitional States Edited by Tim Lindsey 55 The Asymmetries of Globalization Edited by Pan A. Yotopoulos and Donato Romano 56 Ideas, Policies and Economic Development in the Americas Edited by Esteban Pérez-Caldentey and Matias Vernengo 57 European Union Trade Politics and Development Everything but arms unravelled Edited by Gerrit Faber and Jan Orbie 58 Membership Based Organizations of the Poor Edited by Martha Chen, Renana Jhabvala, Ravi Kanbur and Carol Richards 59 The Politics of Aid Selectivity Good governance criteria in World Bank, US and Dutch development assistance Wil Hout

60 Economic Development, Education and Transnational Corporations Mark Hanson 61 Achieving Economic Development in the Era of Globalization Shalendra Sharma 62 Sustainable Development and Free Trade Shawkat Alam 63 The Impact of International Debt Relief Geske Dijkstra 64 Europe’s Troubled Region Economic development, institutional reform and social welfare in the Western Balkans William Bartlett 65 Work, Female Empowerment and Economic Development Sara Horrell, Hazel Johnson and Paul Mosley 66 The Chronically Poor in Rural Bangladesh Livelihood constraints and capabilities Pk. Md. Motiur Rahman, Noriatsu Matsui and Yukio Ikemoto 67 Public–Private Partnerships in Health Care in India Lessons for developing countries A. Venkat Raman and James Warner Björkman

68 Rural Poverty and Income Dynamics in Asia and Africa Edited by Keijiro Otsuka, Jonna P. Estudillo and Yasuyuki Sawada 69 Microfinance: A Reader David Hulme and Thankom Arun 70 Aid and International NGOs Dirk-­Jan Koch 71 Development Macroeconomics Essays in memory of Anita Ghatak Edited by Subrata Ghatak and Paul Levine 72 Taxation in a Low Income Economy The case of Mozambique Channing Arndt and Finn Tarp 73 Labour Markets and Economic Development Edited by Ravi Kanbur and Jan Svejnar 74 Economic Transitions to Neoliberalism in Middle-­Income Countries Policy dilemmas, crises, mass resistance Edited by Alfedo Saad-­Filho and Galip L. Yalman 75 Latecomer Development Innovation and knowledge for economic growth Banji Oyelaran-­Oyeyinka and Padmashree Gehl Sampath

76 Trade Relations between the EU and Africa Development, challenges and options beyond the Cotonou Agreement Edited by Yenkong Ngangjoh-­Hodu and Francis A.S.T Matambalya 77 The Comparative Political Economy of Development Africa and South Asia Edited by Barbara Harriss-­White and Judith Heyer

82 Reform and Development in China What can China offer the developing world? Edited by Ho-­Mou Wu and Yang L. Yao 83 Towards New Developmentalism Market as means rather than master Edited by Shahrukh Rafi Khan and Jens Christiansen

78 Credit Cooperatives in India Past, present and future Biswa Swarup Misra

84 Culture, Institutions, and Development New insights into an old debate Edited by Jean-­Philippe Platteau and Robert Peccoud

79 Development Economics in Action, 2nd edition A study of economic policies in Ghana Tony Killick

85 Assessing Prospective Trade Policy Methods applied to EU-­ACP economic partnership agreements Edited by Oliver Morrissey

80 The Multinational Enterprise in Developing Countries Local versus global logic Edited by Rick Molz, Cătălin Ratiu and Ali Taleb

86 Social Protection for Africa’s Children Edited by Sudhanshu Handa, Stephen Devereux and Douglas Webb

81 Monetary and Financial Integration in West Africa Temitope W. Oshikoya

87 Energy, Bio Fuels and Development Comparing Brazil and the United States Edited by Edmund Amann, Werner Baer and Donald V. Coes

Energy, Bio Fuels and Development Comparing Brazil and the United States

Edited by Edmund Amann, Werner Baer and Donald V. Coes

First published 2011 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Simultaneously published in the USA and Canada by Routledge 270 Madison Avenue, New York, NY 10016 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2011 selection and editorial matter; Edmund Amann, Werner Baer and Donald V. Coes, individual chapters; the contributors Typeset in Times by Wearset Ltd, Boldon, Tyne and Wear Printed and bound in Great Britain by TJI Digital, Padstow, Cornwall All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data Energy, bio fuels and development: comparing Brazil and the United States/edited by Edmund Amann, Werner Baer and Donald V. Coes. p. cm. Includes bibliographical references and index. 1. Energy development–Brazil. 2. Energy development–United States. 3. Clean energy industries–Brazil. 4. Clean energy industries–United States. 5. Energy policy–Brazil. 6. Energy policy–United States. I. Amann, Edmund. II. Baer, Werner, 1931– III. Coes, Donald V., 1943– HD9502.B72E536 2010 333.790973–dc22 2010027713 ISBN13: 978-0-415-56720-6 (hbk) ISBN13: 978-0-203-83385-8 (ebk)

Contents



List of figures List of tables List of contributors

  1 Introduction

xv xvii xxi 1

E dmund A mann , W erner B aer and D onald V . C oes

Part I

Macroeconomic and distributional dimensions of energy shocks

3

  2 Oil price shocks and the macro economy: the United States versus Brazil

5

T iago C avalcanti and J o ã o T ovar J alles

  3 Energy and income distribution in Brazil’s development process

23

E dmund A mann and W erner B aer

  4 The earth is finite and other irrelevancies about the world’s ultimate oil supply

38

F red G ottheil

  5 Energy restrictions to growth: the past, present and future of energy supply in Brazil A dilson de O liveira , E duardo P ontual R ibeiro , R osemarie B r ö ker B one and L uciano L osekann

51

xii   Contents   6 Oil prices and inflation in Brazil: exchange rate versus inflation targeting

65

C laudio A . C . P aiva

  7 Brazilian energy independence: petroleum, trade and economic efficiency

72

D onald V . C oes

  8 The role played by the BNDES in funding electricity investments in Brazil

95

L ui z R icardo C avalcante and S imone U derman

Part II

Social, local and environmental impacts of changes in the energy market

111

  9 Climate change, energy use and long-­run growth in Brazil

113

C arlos A z z oni , E duardo H addad and F abio K anc z uk

10 Spatial interactions between energy and energy-­intensive sectors in the Brazilian economy: a field of influence approach

122

G erv á sio F erreira dos S antos , E duardo H addad , J oa q uim J os é M artins G uilhoto , G eoffrey H ewings and D enise I mori

11 Determinants of the income of workers in sugar cane plantations and in the sugar and ethanol industries in the North-­Northeast and Center-­South regions of Brazil

137

M á rcia A z anha F erraZ D ias de M oraes

12 A framework for examining the impact of bio fuels on the poor in Brazil

151

M ary A rends - ­K uenning

13 Bio fuels, food, and trade: a comparison of bio fuel development efforts in two communities in Illinois G ale S ummerfield , K eith T aylor and S tephen G asteyer

164

Contents   xiii 14 Oligopolistic behavior of Brazilian gas stations

178

I gn á cio T avares de A ra ú jo J ú nior , A lexandre R ands B arros , A ndr é M atos M agalh ã es and L uciano M ene z es B e z erra S ampaio

Part III

The impacts of bio and alternative fuels

195

15 The journey to the next-­generation of bioeconomy: the US perspective

197

H ans P . B laschek

16 Between sustainability and development: bioenergy, land use, food security and lifecycle analysis

203

J ü rgen S cheffran

17 Bio energy efficiency and a flex-­mill simulation in Mato Grosso

221

P eter G oldsmith , G uilherme S ignorini , J oao G omes M artines F ilho , R enato R asmussen and C arolina G uimar ã es

18 The impacts of agriculture-­based energy sources on land use in Brazil

236

C arlos J os é C aetano B acha

19 Fossil fuels, bio fuels, and food: ranking priorities

256

G uilherme L eite da S ilva D ias and J oa q uim J os é M artins G uilhoto

20 The expansion of ethanol and land use in Brazil’s Cerrado

268

C harles C . M ueller and G eraldo B ueno M artha J r .

21 The viability of the biodiesel program as an instrument of social inclusion M arcos C osta H olanda , B runo M oreira W ichmann and P aulo A ra ú jo P ontes

285

xiv   Contents 22 The expansion of sugarcane cultivation and its impact on municipal revenues: an application of dynamic spatial panels to municipalities in the state of São Paulo, Brazil

292

A ndr é C hagas , R udinei T oneto J r . and C arlos A z z oni

Part IV

Conclusions

315

23 Conclusions

317

E dmund A mann and W erner B aer



Index

322

Figures

2.1 2.2 2.3 2.4 6.1 6.2 6.3 6.4 6.5 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 8.1 9.1 9.2 9.3 9.4 9.5 10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8

Net oil import share: Brazil and the United States Impulse response to an oil price shock, 1975Q1–1984Q4 Impulse response to an oil price shock, 1985Q1–2008Q2 Cost of gasoline in the United States and in Brazil Wholesale fuel prices and oil prices (levels) Wholesale fuel prices and oil prices (percentage changes) Ratio of wholesale fuel prices to oil prices (moving average) Wholesale fuel price increases and IPCA inflation Wholesale fuel price increases and core inflation Petroleum production and apparent consumption, 1953–1980 Exports and imports (FOB), 1968–1988 Ratio of international petroleum price to domestic Brazilian price Petroleum production and consumption 1980–2007 Short-­run and long-­run petroleum market equilibrium Real price of petroleum Comparative energy intensity Brazilian energy production by source Evolution of the electric power sector in Brazil Evolution of total factor productivity Evolution of productivity net of schooling and distortions Evolution of total emissions of GHGs Sectors losing participation in GDP Sectors gaining participation in GDP Energy-­intensity in Brazil Electricity-­intensity in Brazil Industrial consumption of energy in Brazil Electricity consumption of industrial sectors in Brazil Sector energy field of influence Spatial energy field of influence Sector energy field of influence (national coefficients) Spatial energy field of influence (electricity-­intensity coefficients)

7 11 12 16 66 67 67 68 69 75 76 77 79 82 83 85 86 98 115 116 118 119 120 124 125 126 126 132 132 133 134

xvi   Figures 12.1 12.2 12.3 19.1

World commodity price indices, 2000–2009 World food price indices, 2000 to July 2009 São Paulo food price index, January 2003 to July 2009 Share of imports in total output of selected developing and developed countries 19.2 Share of food and energy in total imports of selected developing and developed countries 19.3 Share of imported food and energy in total disposable food and energy of selected developing and developed countries 19.4 Share of imported inputs in the total value of mining and quarrying used in the sectors of coke, refined petroleum products and nuclear fuels of selected developing and developed countries 19.5 Share of food and energy in intermediate consumption of selected developing and developed countries 19.6 Share of food and energy in final demand of selected developing and developed countries 19.7 Share of food and energy in total demand of selected developing and developed countries 19.8 Share of disposable domestic value added of food and energy in GDP of selected developing and developed countries 19.9 Relation of disposable domestic value added between food and energy of selected developing and developed countries 19.10 Impact of a 10 percent change in the price of food on the wholesale price index (WPI) and on the consumer price index of selected developing and developed countries 19.11 Impact of a 10 percent change in the price of energy on the wholesale price index (WPI) and on the consumer price index of selected developing and developed countries 21.1 Market price and production margins for biodiesel from castor oil

152 152 153 258 258 259 259 260 261 261 262 263 264 264 287

Tables

2.1 2.2 2.3 2.4

Growth in total factor productivity and the real price of oil Correlation coefficients among oil price proxies Results of unit root tests Variance decomposition for Brazil and the United States, oil shocks (∆oilt) to the volatility in output growth and inflation 2.5 Variance decomposition for Brazil and the United States, oil shocks (NOPI) to the volatility in output growth and inflation 2.6 Summary statistics for US variables 2.7 Covariance matrix, United States 2.8 Summary statistics for Brazilian variables 2.9 Covariance matrix, Brazil 2.10 SVAR lag-­length selection criteria 3.1 Brazil: yearly rate of growth of internal energy consumption 3.2 Brazil’s energy mix 3.3 Brazil: installed electric capacity and production 3.4 Brazil: origin of resources of the electric sector 3.5a Indexes of the real price of electric energy 3.5b Brazil: tariff changes 1995–2007 3.6 Yearly percentage price changes 3.7a Brazil: petroleum – domestic production, imports and consumption 3.7b Brazil: domestic oil production 3.8a Brazil: gasoline – domestic production, imports and consumption 3.8b Brazil: value of petroleum imports 3.8c Brazil: imports of crude oil and derivatives 3.9 Production of ethanol, 1997–2006 3.10 Land use in Brazil 3.11a Planted area coffee and sugar cane, 1997–2004 3.11b Area under cultivation by crop, Brazil 1947–2007 3.12 Yields of key crops, Brazil and São Paulo State 3.13 Land distribution: trends in Brazil, 1970–1996

5 8 9 13 14 18 18 18 19 19 25 25 25 26 27 28 28 29 30 30 30 31 32 32 32 33 33 34

xviii   Tables 4.1 Oil well drillings completed: total, United States and OPEC 1964–1973 4.2 Barrels of oil per foot of drilling: selected regions 4.3 Estimates of world recoverable oil: 1920–1975 4.4 Estimates of ultimate oil recovery 4.5 International rotary oil rig count: 1985–2005 4.6 OPEC reported reserves: selected years 5.1 Energy external dependency, Brazil 1970–2007 5.2 Error correction model for energy and GDP, Brazil 1970–2007 5.3 Structural change in energy demand elasticities 5.4 Oil and gas auction rounds summary 6.1 Wholesale fuel prices and oil prices – correlation 6.2 Cumulative pass-­through of a 10-percent increase in fuel prices to inflation 6.3 Cumulative impact of other variables on inflation 6.4 Inflationary impact of a 10-percent increase in fuel prices under alternative VAR specifications 7.1 Balance of payments – major components 7.2 Petroleum production and consumption, 1980–2008 7.3 Crude oil production (terrestrial and off-­shore) and LNG equivalent 7.4 Energy production – all sources 8.1 Power generation capacity in Brazil 8.2 Annual rates of growth of the installed generation capacity in Brazil 8.3 Installed generation capacity (public, private and self-­producers), 1952–1964 8.4 Annual increases in the generation installed capacity, 1952–1962 8.5 Financing structure of the Brazilian electricity sector, by source of resources, 1974–1979 8.6 BNDES disbursements to the “electricity and gas” sector (million BRL), 1995–2008 9.1 Macroeconomic scenarios 9.2 Energy intensity by source 10.1 Shares of the industry sectors in the industrial production value, 1970–1980 10.2 Sectors of interregional input–output table 11.1 Indicators for workers in sugar cane plantations – 2006 11.2 Sugar cane: evolution observed in the number of employees in Brazil by age bracket 11.3 Sugar cane: evolution observed in the number of employees in Brazil by years of schooling 11.4 Income equations estimated for workers in sugar cane plantations in 1993 and 2006

41 41 42 43 44 46 54 56 57 60 68 70 71 71 87 88 89 90–91 95 99 101 101 102 106 117 118 124 131 142 144 145 146

Tables   xix 11.5 Income equations estimated for workers in the sugar and ethanol industry, 2006 147 12.1 Basic needs foods basket used to define a consumption-­based poverty line, Brazil 2002–2003 156 12.2 Price indices for year ending October 2008 – IPCA 158 12.3 Prices for the basic needs food basket, São Paulo Brazil, 2003, 2007–2009 159 12.4 Cost of basic needs food basket, São Paulo Brazil, 2003, 2007, 2008 and 2009 160 12.5 Estimates of income and price elasticity for food, transportation, and various specific foods 160 13.1 Community capitals and farming community ethanol production 167 14.1 Relationship between margins and coefficient of variation for gasoline – panel data 184 14.2 Results of Equations (2) through (6) 187 14.3 Margin equations for the regions 188 14.4 Results for equations (7) and (8) 190 14.5 Relationship between margins and coefficient of variation for gasoline – cross section, 2003 191 17.1 Descriptive statistics: case study sugar cane mill in Mato Grosso 226 17.2 Flex mill throughput scenarios when maize is the feedstock 227 17.3 Simulation descriptive statistics: case study mill in Mato Grosso Brazil 231 18.1 Evolution of energy production, supply and consumption in Brazil 239 18.2 Annual geometric growth rates of energy production in Brazil 239 18.3 Structure of Brazil’s energy supply matrix (values in percentages) 240 18.4 Structure of Brazil, OECD and world energy supply 240 18.5 The main users of charcoal and firewood in Brazil 241 18.6 Shares of charcoal production (PROD) and consumption (CONS) among the Brazilian states 242 18.7 The main uses of ethanol and bagasse in Brazil 244 18.8 Shares of Brazilian states in domestic ethanol production 245 18.9 Production, land productivity, price and oil yield for various crops in Brazil 247 18.10 Number of biodiesel mills and their installed capacity 248 18.11 Biodiesel production in Brazilian states from 2005 to 2007 249 18.12 Farming land, cropland and pastures in Brazil and its states 250 18.13 Results of the land price regression 252–253 20.1 Cerrado dynamic regions: sugar and/or ethanol plants under operation, new and under construction plants, and plants under consideration (mid-­2008) 274

xx   Tables 20.2 Cerrado dynamic regions: area in crops, and in sugar cane, 1996 and 2006; proportion of the area in crops of sugar cane, 1996 and 2006; and proportion of geographical area in sugar cane, 2006 21.1 Average price of castor beans in Brazil, Northeast and Ceará (R$/kg) 21.2 Cost of production of biodiesel from different raw materials 21.3 Biodiesel production in the Northeast and Ceará 21.4 Estimated social cost of biodiesel from castor beans 21.5 Family income from biodiesel 21.6 Program efficiency 22.1 Average per capita fiscal performance and population by type of municipality, 1990 to 2006 22.2 Balancing of the panel: number of observations included in each period 22.3 Models for per capita own revenue 22.4 Models for per capita ICMS revenue share 22.5 Models for the per capita share of IPVA

275 286 287 288 289 289 290 298 302 304–305 306–307 310–311

Contributors

Edmund Amann is Reader in Development Economics at the University of Manchester. His research centers on the economics of Latin America. Mary Arends-­Kuenning is Associate Professor in the Department of Agricultural and Consumer Economics at the University of Illinois at Urbana-­ Champaign. Her research focuses on economic and demographic issues in developing countries. Ignácio Tavares de Araújo Júnior is Professor of Economics at the Federal University of Paraíba. His research centers on the economics of Brazil. Carlos Azzoni is Professor of Economics at the University of São Paulo. His research interests focus on the economics of regional development in the Brazilian context. Carlos José Caetano Bacha is a Professor at University of São Paulo. He is interested in development issues, in particular the role of the pulp, paper and forestry sectors. Werner Baer is Lemann Professor of Economics at the University of Illinois at Urbana-­Champaign. His research centers on industrialization and development in Latin America, especially Brazil. Alexandre Rands Barros is Professor of Economics at the Federal University of Pernambuco. His research focuses on regional development issues. Hans P. Blaschek is Professor of Food Microbiology and Assistant Dean at the University of Illinois at Urbana-­Champaign. His research interests include the  genetic and physiological manipulation of Clostridia and exploiting the potential of these microorganisms for biotechnology application in the fermentation industry. Rosemarie Bröker Bone is a Professor at the Federal University of Rio de Janeiro. Her research concentrates on economic development issues. Luiz Ricardo Cavalcante is a Researcher at the Brazilian Applied Economics Research Institute, IPEA. Among other themes, his research concerns itself with the BRIC economies and technological change.

xxii   Contributors Tiago Cavalcanti is a Lecturer in Economics University of Cambridge, UK. His research is focused on macroeconomic theory and growth. André Chagas is Professor of Economics and Research Coordinator at FECAP. His research focuses on the economics of development and industrial economics. Donald V. Coes is Professor of Economics at the University of New Mexico. His research interests center on the economic development of Latin America. Guilherme Leite da Silva Dias is Professor of Economics at the University of São Paulo. His research focuses on agriculture and agricultural development. Stephen Gasteyer is Assistant Professor in the Department of Sociology at Michigan State University. His research focuses on social networks and the management of water and other critical community resources. Peter Goldsmith is Associate Professor and Soybean Industry Endowed Associate Professor in Agricultural Strategy at the University of Illinois at Urbana-­ Champaign. His research interest is structural change in global agro-­industrial markets and its effect on agribusinesses and farmers Fred Gottheil is Professor of Economics at the University of Illinois at Urbana-­ Champaign. His research focuses on the economics of development, especially as it relates to the Middle East. Joaquim José Martins Guilhoto is Professor of Economics at the University of São Paulo. His research centers on the economics of regional development. Carolina Guimarães is a Researcher at the University of São Paulo focusing on sugar cane and the development of alternative fuels. Eduardo Haddad is Professor of Economics at the University of São Paulo. His research centers on the economics of regional development. Geoffrey Hewings is Professor and Departmental Head of Economics at the University of Illinois and Director of that university’s Regional Economics Applications Laboratory (REAL). His research is focused on study of the regional and spatial aspects of the process of economic development. Marcos Costa Holanda is Professor of Economics at the Federal University of Ceará. His research centers on the economics of development, especially as it plays out in a regional context Denise Imori is a Researcher in Economics at the University of São Paulo Her interests focus on development economics and the use of input–output analysis. João Tovar Jalles is a PhD student at the University of Cambridge. His research concerns itself with the macroeconomics and growth. Fabio Kanczuk is Professor of Economics at the University of São Paulo. He is particularly interested in the economics of inflation, finance and banking.

Contributors   xxiii Luciano Losekann is Associate Professor at Fluminense Federal University. His research has focused extensively on the Brazilian energy sector. André Matos Magalhães is a Researcher in Economics at the Federal University of Pernambuco specializing in development issues. Geraldo Bueno Martha Jr. is an economist working at EMBRAPA on agricultural and alternative fuel-­related issues. Joao Gomes Martines Filho is an Associate Professor at the University of São Paulo working on agricultural development issues. Márcia Azanha Ferraz Dias de Moraes is a Professor at the University of São Paulo. She has carried out extensive research in the fields of bio energy and sugar cane production. Charles C. Mueller is Professor of Economics at the University of Brasília. He specializes in the field of agricultural and rural development in the Brazilian context. Adilson de Oliveira is Professor of Economics at the Federal University of Rio de Janeiro. He has published extensively in the field of energy economics. Claudio A.C. Paiva is Associate Professor of Economics at California State University, Channel Islands. His research focuses on international macroeconomic issues and the economics of development, especially in the Latin American context. Paulo Araújo Pontes is a Researcher at the Federal University of Ceará specializing in development issues. Renato Rasmussen is a research student in the Department of Agricultural and Consumer Economics at the University of Illinois at Urbana-­Champaign. His research centers on energy project evaluation and energy efficiency. Eduardo Pontual Ribeiro is Professor of Economics at the Federal University of Rio de Janeiro. His research focuses on firm performance and industrial economics. Luciano Menezes Bezerra Sampaio is a Professor at the Federal University of Rio Grande do Norte specializing in rural development issues. Gervásio Ferreira dos Santos is a Researcher at the University of São Paulo. His research centers on spatial and regional development issues. Jürgen Scheffran is Adjunct Associate Professor in the Department of Political Science at the University of Illinois at Urbana-­Champaign. His research focuses on the policy implications of energy security and climate change. Guilherme Signorini is a Researcher at the University of São Paulo specializing in agricultural development issues.

xxiv   Contributors Gale Summerfield is Associate Professor, Human and Community Development at the University of Illinois at Urbana-­Champaign. Her research interests include gender, development, bio fuels and food security Keith Taylor is a Graduate Student in Human and Community Development at the University of Illinois at Urbana-­Champaign. His research interests center on the social aspects of rural development and the impacts of bio fuels. Rudinei Toneto Jr. is Dean of the Faculty of Economics, Administration and Accounting at the University of São Paulo, Ribeiro Preto. His recent research has extensively focused on the impacts and potential of bio fuels. Simone Uderman is Adjunct Professor at the State University of Bahia. Her main research interests concern regional development, industrialization and economic growth. Bruno Moreira Wichmann is a Researcher at the Federal University of Ceará specializing in regional and industrial development issues.

1 Introduction Edmund Amann, Werner Baer and Donald V. Coes

This book comes at a time of heightened concern about energy security and sustainability in the face of conflict in the Middle East, political uncertainty in Russia and the growing prominence of climate change/carbon agenda. Against this background, a number of developed and emerging economies are examining bio fuels. At the same time, more effort is being made to reduce reliance on imported hydrocarbons by increasing efforts to tap these conventional fuel sources domestically. Nowhere is this twin track approach being explored on such a large scale and with such wide-­ranging consequences as in Brazil and the United States. This book focuses on these two critical country cases, examining the impacts of changing energy sources in economic, distributional and environmental terms. By comparing and contrasting the Brazilian and US experiences, not only can lessons be drawn for the western hemisphere’s two largest economies, but they can also be usefully gleaned for other developed and emerging nations. Until the 1970s, when OPEC quadrupled the price of oil, surprisingly few economists paid much attention to the role of energy in the development process. This neglect ended in the late 1970s, a period when Brazil became one of the first countries to delve into alternative sources of energy with the development of its alcohol-­for-cars program, based on sugar cane. Though it was successful for a few years, the alcohol program was scaled back, if not abandoned, in the 1980s, when the price of oil declined again. Yet with the renewal of oil price rises in the second half of the 1990s, Brazil once more took the lead in the development of bio fuel energy based on sugar cane. The United States has followed hard on Brazil’s heels with an ambitious program to promote domestically-­ produced ethanol-­based fuels. Over the last decade the two countries have become the world’s largest producers of ethanol. In the case of Brazil, the ethanol boom has been spearheaded by the domestic development of flex-­fuel cars, a technology which the United States is keen to more widely adopt. Unlike the United States, the Brazilian quest for self-­reliance in energy has been boosted by recent discoveries of vast offshore oil fields. The chapters contained in this volume were prepared for a conference on “Energy and Economic Growth and Development” which was held in November 2008 in Ilha Bela, Brazil. The conference was co-­sponsored by the Lemann

2   E. Amann et al. Program of the Center for Latin American and Caribbean Studies of the University of Illinois,1 the Faculty of Economics of the University of São Paulo and Esalq of the University of São Paulo in Piracicaba. This collection examines the growth of the Brazilian energy sector – most especially its bio-­fuel industry – from various angles. These include its impact on the country’s general economic growth, on government finance and price stability; on world food prices; on the distribution of income; on the distribution of land; on employment; on the environment, including climate change; on the agricultural sector, including the trade-­off between bio fuels and food prices; and on the balance of payments. To provide comparative perspective, some of the chapters concentrate on the US experience, where the growth of ethanol was based on corn, which was much less efficient than ethanol based on sugar cane. The chapters that follow show the complexity and interdependence of the issues involved in moving a society reliant on non-­renewable energy sources to one based on alternative sources of energy. A particular lesson to emerge from this collection is that Brazil, in pursuing a flexible mix of fossil fuels and bio fuels, has greatly diminished its dependence on exogenous energy shocks, thus setting an example for both rich and developing societies. The US experience has been more problematic. Nevertheless, as the discussion reveals, recent years have seen unprecedented progress in trying to reduce US reliance on fossil fuels.

Note 1 This program became the Lemann Institute for Brazilian Studies in October 2009.

Part I

Macroeconomic and distributional dimensions of energy shocks

2 Oil price shocks and the macro economy The United States versus Brazil Tiago Cavalcanti and João Tovar Jalles1

Sharp increases in the price of oil and other energy products are referred to in the literature as classical examples of negative supply shocks (e.g. Brown and Yucel 2002; and Hamilton 2005). Increases in the price of oil lead to increases in the cost of production, which in general decrease the rhythm of economic activity and increase inflation. The response of nominal wages and monetary policies can amplify the shocks.2 In an important article, Hamilton (1983) argues that nine out of ten North American recessions after World War II until the mid-­ 1970s were preceded by sharp increases in oil prices.3 In addition, he shows that such a correlation between oil prices and output does not represent a statistical coincidence. In particular, he finds evidence of Granger causality between oil prices and output. Periods of low growth in real GDP and high inflation are preceded by high relative international oil prices. Price increases in oil have also been associated with the productivity slowdown in the 1970s. Table 2.1 relates the growth rates of total factor productivity (TFP) in the United States and in Brazil to the real price of oil for selected five-­ year sub-­periods. The overall relation is significantly influenced by a period of unusually low growth in TFP in 1975–1980 (for the Brazilian case one can observe a lagged effect of the early 1970s oil crisis, as TFP is only negatively Table 2.1  Growth in total factor productivity and the real price of oil

1960–1965 1965–1970 1970–1975 1975–1980 1980–1985 1985–1990 1990–1995 1995–2000 2000–2005

Real oil price averages

USA TFP growth averages

Brazil TFP growth averages

20.14 19.38 27.79 60.25 69.93 34.62 27.79 24.07 35.59

1.94 0.07 −0.034 −1.22 1.16 1.02 0.17 1.48 1.72

2.47 2.75 5.22 0.19 −1.65 1.69 −0.34 0.01 3.27

Source: IMF–IFS (oil price); TFP from Klenow and Rodriguez-Clare (2005).

6   T. Cavalcanti and J.T. Jalles affected at a later stage) which coincides with an unusually high real price of oil (see also Barsky and Kilian 2004). One way to see whether the relationship between the oil price and growth of output might not be just a coincidence is by performing a statistical regression of the real GDP growth rate on its lagged values and on lagged logarithmic changes in nominal oil prices, as suggested by Hamilton (2003). 4

4

j=1

j=1

growtht = α + ∑ ​      g​ ​     B​  ​  j oilt–j ​  j growtht–j + ∑

(1)

The OLS regression estimation of such relationship for the US economy, using quarterly data for the 1958–1980 period, shows that the parameters of the four lagged oil price variables are negative and statistically significant at 95 percent confidence level.4 An F-­test also rejects the null hypothesis of the joint estimate of the parameters of the lagged oil prices being all zero with a p-­value of 0.0058.5 Some studies (e.g. Blanchard and Gali 2008; and Killian 2005), however, have shown that while in the 1970s oil price shocks led to long periods of stagflation, recently the effects of such shocks on inflation and output have been mild in most economies. Blanchard and Gali (2008) posit that there are four sources for such a decline in the effects of oil shocks on the economy: (a) good luck (i.e. small concurrent adverse shocks); (b) decline in the dependence of oil in production; (c) more flexible labor markets; and (d ) improvements in monetary policy. A large body of empirical research (e.g. Barsky and Kilian 2004; and Gali and Gambetti 2008) has provided evidence of a remarkable decline in macroeconomic volatility experienced by the US economy since the mid-­1980s, often referred to as “the Great Moderation hypothesis”.6 This chapter studies the effects of oil price shocks focusing on the Brazilian and US economies. It is important to analyze different countries because they may react differently to oil price shocks due to the heterogeneous sectoral compositions of their main economic aggregates, different institutions and policies, as well as their dependence on oil imports. We make a historical analysis by investigating the effects of such shocks on the economy before and after the mid-­1980s for both Brazil and the United States.7 Do shocks have a different impact on output and inflation in Brazil than in the United States? How did such effects change over time? The Brazilian and the US economies are interesting polar cases, since they had completely different paths with respect to oil import dependence. While oil import dependence has increased sharply in the United States, it has decreased substantially in Brazil. The South American country has decreased oil import dependence not only by increasing domestic oil production8 but also by developing and increasing ethanol production.9 Figure 2.1 shows the net oil import share in Brazil and in the United States from 1980 to 2007.10 We observe that while oil net import share increased in the United States from about 37 percent in 1980 to roughly 60 percent in 2007, net oil import share in Brazil decreased from about 80 percent in 1980 to less than 5 percent in 2007.

Oil price shocks and the macro economy   7 90

Brazil USA

80

Percentage

70 60 50 40 30 20 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

10

Figure 2.1 Net oil import share: Brazil and the United States (source: Energy Information Administration office).

This chapter has three further sections. The first discusses the methodology to study how shocks affected business cycle fluctuations in Brazil and in the United States. The second presents the empirical results while the final section contains the concluding remarks.

Methodology Before analyzing the impacts of oil shocks on economic activity and inflation, we first investigate the stochastic properties of the series. In particular, Augmented Dickey–Fuller (ADF ), Phillips–Perron (PP) and Kwiatkowski–Phillips– Schmidt–Shin (KPSS) tests are performed to study the integration order of the series. We then use a multivariate structural vector autoregression (SVAR) model11 (see Hamilton 1983; Burbidge and Harrison 1984; Blanchard and Gali 2008; among others for a similar procedure) to study the impact, magnitude and reaction of inflation and output to oil price shocks.12 Variance decomposition analysis and cointegration tests are also conducted. The variables considered for the multivariate SVAR model are the following: average oil price (OIL), real Gross Domestic Product (Y) and the CPI-­based inflation rate (INF ).13 The Y and INF are included in the SVAR since our primary object of concern is the impact of oil prices on real output and inflation. We have taken the logarithm of all level variables in order to obtain rates of growth with the first difference. We have left the inflation rate in percentage terms. Due to the highly volatile behavior of oil prices (particularly in recent times), linear oil price specifications are not necessarily the best specification to study the true effects of oil price shocks on the economy. Hooker (1996) showed that, for the American economy, linear specifications of oil prices ceased to Granger-­cause most macroeconomic indicator variables, including the unemployment rate, real GDP,

8   T. Cavalcanti and J.T. Jalles aggregate employment, and industrial production. Based on Hooker (1996), we define three proxy variables for oil prices. The first is the evolution of the annual changes of world oil prices and is calculated as: ∆oilt = ln(oilt) − ln(oilt−a ),

(2)

where oilt is the oil price in period t (in quarters). This measure is used by many authors, including the recent article by Blanchard and Gali (2008). We use this measure in our exercises. Another measure is an asymmetrical one that captures only price increases: ∆oil+ = max(0; ∆oilt)

(3)

The rationale behind this specification relies on the observed asymmetry in the behavior of macroeconomic variables in reaction to oil price changes. It seems that only sharp price increases have a statistically significant effect on the economy (see Mork 1989; and Hooker 1996). Finally, we define the Net Oil Price Increase (NOPI). This variable takes into account an oil price change only if the percentage increase in price is above the observed values for the previous four quarters. Otherwise the variable takes the value zero. This specification eliminates any increase in prices that simply correct price volatility. With this measure it is possible to capture more effectively the surprise element, which may be at the origin of a change in spending decisions by firms and households. In our case, since growth rates are defined as quarterly growth rates, we shall calculate: NOPIt = max[0; ln(oilt) − ln(max(oilt−1, oilt−2, oilt−3, oilt−4))]

(4)

Table 2.2 contains the correlations among the three measures. Note the high correlation among them. For this reason, we present in the next section results using the traditional measure ∆oil and the NOPI measure.

Empirical results As a first step of the empirical analysis, unit root tests were carried out for all variables for the time period from the first quarter of 1975 to the second quarter Table 2.2  Correlation coefficients among oil price proxies

Oil ∆oil ∆oil+ NOPI

Oil

∆oil

∆oil+

NOPI

1 0.162 0.241 0.204

– 1 0.863 0.788

– – 1 0.927

– – – 1

Oil price shocks and the macro economy   9 Table 2.3  Results of unit root tests USA

Real GDP in first-log differences Inflation rate Oil price in first-log differences

Brazil

ADF

PP

ADF

PP

−10.64*** −19.37***   −8.98***

−10.68*** −19.63***   −8.98***

−4.68*** −3.76*** −8.98***

−17.52***   −3.67***   −8.98***

Note Sample is 1975Q1–2008Q2 (quarterly) for all variables. Data-driven lag selection procedures are used for the Augmented Dickey–Fuller tests, taking 9 as the maximum number of lags allowed in these procedures. The *, **, *** denote respectively the rejection of the null hypothesis at a 10 percent, 5 percent, 1 percent critical significance levels. KPSS p-values support the conclusions provided by both the ADF and PP unit root tests.

of 2008. Table 2.3 shows statistics for the ADF, PP (and KPSS) unit root tests.14 Results of these tests indicate that the first differences of all variables are stationary at usual significance levels. This hypothesis is corroborated by testing for both the ADF and P P tests for the cases with a constant and with both a constant and a trend.15 It is important to test whether the nature of the oil price-­macroeconomy relationship changed over time. Following Hamilton’s (1983) methodology, we perform the Chow Breakpoint Test and the Chow Forecast Test on the following equation: Yt = α + ∑βjyt–1 + ∑njoilt–j + ut

(5)

where y is the growth of real GDP and oil is the log-­difference of oil prices. We note that any arbitrary lag length choice can be subject to possible criticism. We chose not to test for breakpoints in the late 1970s due to the risk of obtaining results with little robustness, given that the first observation in our sample is 1975Q1. We tested for a breakpoint on 1985Q1 for two main reasons. First, there was a clear collapse of oil prices in 1985–1986 (see Brown and Yucel (2002) for some explanation of this breakdown in the relationship between oil and the economy). Second, several authors point to the mid-­1980s as the rupture point in the way economic agents react to oil prices.16 The Chow Breakpoint Test and the Chow Forecast Test provide evidence for the existence of a structural break in this point at the usual significance levels for both the US and Brazilian cases. This fact has two main implications for the remainder of this chapter. First, we found it more appropriate and insightful to estimate models for different time periods: for the entire sample, for a first sub-­sample (1975Q1–1984Q4) and for a second sub-­sample (1985Q1–2008Q2). Second, we chose to carry out the estimation both with the first log difference and also with the NOPI proxy to oil price shocks. This allows us to perform a comparative analysis and conclude whether or not the nature of the relationship has indeed changed in Brazil and in the United States. We now briefly describe our econometric methodology. We consider the following vector autoregression model of order p:

10   T. Cavalcanti and J.T. Jalles Yt = c + ∑φijyt–1 + Єt

(6)

Where yt corresponds to a (nx1) vector of endogenous variables, c = (c1, . . ., cn)' is the (nx1) intercept vector, Фi is the ith (nxn) matrix of autoregressive coefficients for I = 1, 2, . . . p, and Єt = (Є1t, . . ., Є)' is the (nx1) generalization of a white noise process. As we are working with quarterly data we use four lags of the endogenous variables. The usual lag length criteria provided support for this choice, so we estimated SVAR models of order 4.17 In order to identify oil price shocks, we follow the literature (e.g. Blanchard and Gali 2008) in assuming that unexpected variations in the price of oil are exogenous relative to the contemporaneous values of real output growth and inflation. We explore this system of equations by first analyzing impulse response functions (IRFs) to a one standard deviation shock in the oil prices variable (the precision of the estimation of the impulse responses can be gauged by looking at the confidence bands18). An impulse response function traces the effect of a one-­time residual shock to one of the innovations in current and future values of the endogenous variables. Next, we use variance decomposition to investigate by how much of the observed fluctuations in inflation and output in Brazil and in the United States can be accounted for by oil shocks. Figures 2.2 and 2.3 show the estimated impulse response functions (IRFs) for real GDP growth and inflation to an oil price shock for Brazil and for the United States.19 The magnitude of the shock corresponds to an increase of one standard deviation in the price of the oil, which is about a 10 percent increase. In Figure 2.2 we analyze the first period in the sample that goes from 1975Q1 to 1985Q1. The left hand panel shows the results for the United States. The response of inflation and output growth corroborates the conventional vision about the relationship between oil price and the macroeconomy. There is a jump in the inflation rate that remains positive for almost 12 quarters. For the first six quarters such a response seems to be statistically different from zero, since the range of the confidence interval is positive for these quarters. For later quarters, the results are not statistically different from zero. There is a lag in the response of output to a shock in the price of oil, but after four quarters output decreases and growth remains negative until the tenth quarter after the shock. In the case of Brazil (see right hand panel of figure), note that the magnitude of the response of inflation and output to an oil shock is much larger than it is in the United States. Neither the effect on inflation nor the one on output growth, however, seem statistically different from zero. The average effect on output growth is roughly zero. Recall that the first ten years in our sample correspond to a period of high volatility in inflation and output in Brazil, due not only to oil shocks and international increases in the interest rate, but also due to fiscal and monetary instability (see Baer 2008). Figure 2.3 presents results for the second period in our sample. The left hand panel again reports results for the United States. Note that compared to the previous period, the response of inflation and output growth to oil shocks is smaller in magnitude in the US economy. Inflation appears to increase only after the first

1

2

3

4

5

6

7

8

9

10

11

12

3

4

5

6

7

8

9

10

11

12

�0.03

�0.006

2

�0.02

�0.004

1

�0.01

�0.002

1

1

2

2

3

3

4

4

5

5

6

6

7

7

8

8

9

9

10

10

11

11

12

12

Figure 2.2 Impulse response to an oil price shock, 1975Q1–1984Q4. Left panel: United States; right panel: Brazil. First row: inflation; second row: output.

�0.008

0

0.01

0.02

0.03

0.04

�1.5

�1

�0.5

0

0.002

0.004

0.006

�0.002

�0.001

0

0.001

0

0.003

0.002

1 0.5

0.004

2 1.5

0.005

0.006

1

1

2

2

3

3

4

4

5

5

6

6

7

7

8

8

9

9

10

10

11

11

12

12

�0.03

�0.02

�0.01

0

0.01

0.02

0.03

�6

�4

�2

0

2

4

6

8

10

1

1

2

2

3

3

4

4

5

5

6

6

7

7

8

8

9

9

10

10

11

11

12

12

Figure 2.3 Impulse response to an oil price shock, 1985Q1–2008Q2. Left panel: United States; right panel: Brazil. First row: inflation; second row: output.

�0.002

�0.0015

�0.001

�0.0005

0

0.005

0.001

0.0015

0.0035 0.003 0.0025 0.002 0.0015 0.001 0.0005 0 �0.0005 �0.001 �0.0015

Oil price shocks and the macro economy   13 three quarters and by a smaller amount. In addition, while output remains negative for about 12 quarters, the effect is much smaller than was observed previously. With respect to the Brazilian economy (see right hand panel of figure), the average effect on output growth of an increase in oil price is roughly zero. Although inflation seems to increase in the first four quarters after the shock, such an effect does not seem to be statistically different from zero. In sum, oil price increases seem to have a negative effect on growth and a positive effect on inflation in the United States, but the magnitude of these effects is decreasing over time. In Brazil, changes in the price of oil have a positive effect on inflation, but this does not seem statistically different from zero. In addition, real output growth in Brazil seems to not respond to oil price shocks in either period of the sample. Table 2.4 provides the contribution of oil shocks to the volatility in output growth and in inflation in the two countries for the two sub-­samples. The first

Table 2.4 Variance decomposition for Brazil and the United States: the contributions of oil shocks (∆oilt) to the volatility in output growth and inflation United States

Brazil

Part I: first period: 1975Q1–1984Q4 Inflation

S.E. (%)

Contribution of ∆oil (%)

S.E. (%)

 1  4  8 12

0.51 0.76 0.98 1.00

  3.41 29.85 27.72 26.92

255.31 324.63 352.63 359.98

Output

S.E. (%)

Contribution of ∆oil (%)

S.E. (%)

 1  4  8 12

1.06 1.14 1.32 1.37

  3.34   4.01 15.87 15.46

4.99 5.49 6.69 7.25

Contribution of ∆oil (%) quarter:   3.02 11.36 15.23 16.09 Contribution of ∆oil (%) quarter: 3.64 5.19 6.48 6.64

Part II: second period: 1985Q1–2008Q2 Inflation

S.E. (%)

Contribution of ∆oil (%)

S.E. (%)

 1  4  8 12

0.40 0.45 0.49 0.50

42.33 50.86 47.87 46.62

1,802.27 2,757.91 2,907.85 2,928.34

Output

S.E. (%)

Contribution of ∆oil (%)

S.E. (%)

 1  4  8 12

0.45 0.51 0.53 0.54

0.09 4.79 8.42 9.04

6.43 7.50 7.77 7.72

Contribution of ∆oil (%) quarter: 1.72 3.49 4.35 4.33 Contribution of ∆oil (%) quarter: 0.54 5.41 6.77 6.76

14   T. Cavalcanti and J.T. Jalles part of this table contains the results for the first sub-­sample that goes from 1975Q1 to 1984Q4. Note that, in this period after 12 quarters, oil shocks contribute to about 25 percent of the volatility of inflation in the United States. The corresponding number for the Brazilian economy is 16 percent. With respect to output volatility, oil shocks accounted for about 15 percent of the volatility in output growth in the United States after 12 quarters, for the first period of the sample. For the Brazilian case in the same period oil shocks accounted for only six percent of the output growth volatility.20 The second part of Table 2.4 contains results for the second sub-­sample. Notice first that output volatility in the United States from 1985Q1 to 2008Q2 is about half of the observed output volatility in the previous period, which is consistent with “the Great Moderation” hypothesis. The inflation rate is also less volatile. Oil price shocks account for a larger fraction of the observed volatility in inflation than observed previously, but their contribution to output growth in

Table 2.5 Variance decomposition for Brazil and the United States: the contributions of oil shocks (NOPI) to the volatility in output growth and inflation United States

Brazil

Part I: first period: 1975Q1–1984Q4 Inflation

S.E. (%)

Contribution of NOPI (%) S.E. (%)

 1  4  8 12

3.41 4.32 4.55 4.57

  0.91 23.60 25.14 24.50

Output

S.E. (%)

Contribution of NOPI (%) S.E. (%)

 1  4  8 12

0.53 0.78 1.01 1.04

  3.24   4.78 13.71 13.41

255.31 324.63 352.63 359.98

4.99 5.49 6.69 7.25

Contribution of NOPI (%) quarter:   0.22 17.26 23.48 24.69 Contribution of NOPI (%) quarter: 3.29 3.94 4.97 4.98

Part II: second period: 1985Q1–2008Q2 Inflation

S.E. (%)

Contribution of NOPI (%) S.E. (%)

 1  4  8 12

2.55 2.84 2.88 2.89

32.81 32.98 32.02 31.88

Output

S.E. (%)

Contribution of NOPI (%) S.E. (%)

 1  4  8 12

0.43 0.47 0.50 0.51

0.80 3.39 6.58 6.71

1,802.27 2,757.91 2,907.85 2,928.34

6.43 7.50 7.77 7.72

Contribution of NOPI (%) quarter: 0.85 1.02 1.37 1.35 Contribution of NOPI (%) quarter: 0.11 5.38 5.29 5.32

Oil price shocks and the macro economy   15 the United States is about one-­third of that in the first sample period. In Brazil, oil price shocks have about one-­tenth of the contribution in inflation volatility when compared to the United States. In addition, such shocks contribute to only six percent of the real output growth volatility in 12 quarters. Table 2.5 contains results using the oil price shock NOPIt instead of the first log difference, ∆oilt . NOPIt identifies price shocks only as those relatively large changes in the oil price. In addition, it implicitly considers that the effects on the economy of an oil price change are not symmetric. Positive changes in the price of oil affect the economy, while negative changes do not affect output and inflation. Observe that the results with this new variable are qualitatively and quantitatively similar to those observed in Table 2.4. We can therefore conclude that output growth volatility in the United States has been decreasing over time, as has the contribution of oil price shocks to such volatility, despite the increase in oil import dependence in the United States. Inflation volatility has also been decreasing but oil price shocks appear to account for a larger fraction of this volatility in the United States. In Brazil, such shocks do not seem to have a clear impact on output growth and they account for only a small fraction of Brazilian inflation and output growth rate trends. In this section we investigate gasoline demand both in Brazil and in the United States using vector error correction (VEC) and cointegration techniques. Demand elasticities for gasoline are important in the transmission mechanism of oil price shocks. The 1980s were a period of high demand for gasoline in Brazil. During this period, the gasoline price index (measured in Brazilian R$) had a sharp decrease. Additionally, during the entire decade of the 1990s, the gasoline price index remained at a lower level (see the bottom graph of Figure 2.4). In the United States price rose as a reaction to the 1970s oil crisis and then dropped in the following decade. It is also clear from the graph, however, that there was a sharp increase in the early twenty-­first century, originating from the demand-­ supply interaction and scarcity of fossil fuels (see top graph of Figure 2.4). Data used in this section are annual, from 1978–2007. The variables used are: gasoline consumption per capita, real GDP per capita, and the gasoline price index (properly deflated) for both the United States and Brazil. All Brazilian data were taken from IPEA data and the US data are from the Energy Information Administration (EIA) office. Population statistics were taken from IMF-­IFS to transform aggregate variables into per capita equivalents. The gasoline consumption variable is measured in thousands of barrels; real GDP is measured in each country’s respective currency and is deflated. The methodology for a correct treatment of the three variables above follows Engle and Granger (1987). It is well known that a time-­series model can only be built once the included series in the model are stationary. This is, however, not the case for most series of practical interest. Working with transformed series, moreover, makes it difficult to interpret the results or impossible to use the model for forecasting. To overcome this dilemma, Engle and Granger (1987) show that if independent series are integrated of the same order d, denoted by I (d), and if the residuals of the linear regression among these series are integrated

2004 2004

2002 2002

2000 2000

1998 1998

1996 1996

1994 1994

1992 1992

1990 1990

1988 1988

1986 1986

1984 1984

1982 1982

1980 1980

1978 1978

USA: cost of motor gasoline in real 1982–1984 US$ (costs per gallon) 160 USA: cost of motor gasoline in real 1982–1984 US$ (costs per gallon) 160 140 140 120 120 100 100 80 80 60 60 40 40 20 20 0 0

2006 2006

16   T. Cavalcanti and J.T. Jalles

m3

2006 2006

2004 2004

2002 2002

2000 2000

1998 1998

1996 1996

1994 1994

1992 1992

1990 1990

1988 1988

1986 1986

1984 1984

1982 1982

1980 1980

1978 1978

6,000 6,000 5,000 5,000 4,000 4,000 3,000 3,000 2,000 2,000 1,000 1,000 0 0

Brazil: average gasoline price per 1,000m3 in R$ (deflator: IGP-DI) Brazil: average gasoline price per 1,000 in R$ (deflator: IGP-DI)

Figure 2.4 Cost of gasoline in the United States and in Brazil (source: Ipeadata and Energy Information Administration (EIA)).

of the order d − b, I (d − b), then the series are said to be co-­integrated of the order d, b, denoted as C I (d, b). There is a great advantage in finding long-­term co-­integration relationships, since the series need no longer be transformed and hence the forecasting power increases substantially. All three time series (gasoline consumption per capita, real GDP per capita and the gasoline price index) proved to be stationary in first log-­differences at usual significance levels. The common ADF and P P tests were performed to evaluate the existence of unit roots in these series. In this context, for the purpose of the V EC models, log-­differentiated variables are used for both countries. The Johansen cointegration test confirms the existence of one cointegrating relation for the three variables for both the United States and Brazil. We therefore incorporate the error correction terms into the adjusted model to study short­run relations. The cointegrating vector gives us information regarding long-­run elasticities. We found for Brazil a long-­run price elasticity for the demand of gasoline equal to −0.11 (p-­value = 0.004). For the US case the long-­run price elasticity of demand was estimated to be −0.077 (p-­value = 0.039). Therefore if the price of gasoline increases by 100 percent, gasoline demand decreases by 11

Oil price shocks and the macro economy   17 percent in Brazil and by only 7.7 percent in the United States.21 Indeed, one should expect the price elasticity of demand to be higher in absolute value in Brazil than in the United States, given the (partial) substitutability between ethanol-­fuel and gasoline.

Concluding remarks This chapter investigated the effects of oil price shocks in the last 30 years on the Brazilian and American inflation rate and on the rhythm of economic activity. The Brazilian and the US economies are interesting polar cases, since they had completely different paths with respect to oil import dependence. While oil import dependence has increased sharply in the United States, it has decreased substantially in Brazil. We applied standard econometric techniques, such as Structural Vector Autoregression Model, to study the response of inflation and output growth to changes in oil prices. We found that output growth volatility in the United States has been decreasing over time as has the contribution of oil price shocks to such volatility, despite the increase in oil import dependence in the United States. Inflation volatility has also been decreasing but oil price shocks now account for a larger fraction of this volatility in the United States. In Brazil, such shocks do not seem to have a clear impact on output growth and they account for a very small fraction of the volatility in the Brazilian inflation and output growth rate.

Appendix 2.1  Data sources and summary statistics The data used in this chapter are mainly obtained from International Monetary Fund (IMF ), International Financial Statistics and IPEA data. Data used to estimate VAR models have quarterly frequency and are from the first quarter of 1975 to the second quarter of 2008 for both the United States and Brazil. See below for the description of variables: GDP: Nominal GDP for the United States; real GDP for Brazil. Source: IMF and IPEA. GDP def: GDP deflator for the United States, used to compute the real GDP. Source: IMF, IFS. INF: CPI-­based Inflation rate. Source: IMF, IFS for United States and IPEA data for Brazil. OIL: international crude oil prices. Source: IMF, IFS. Table 2.6 reports summary statistics for the variables used in VAR regressions for the US economy. The corresponding covariance matrix is in Table 2.7. Table 2.8 reports summary statistics for the variables used in VAR regressions for the Brazilian economy. The corresponding covariance matrix is in Table 2.9. Data used to calculate gasoline demand elasticities are annual from 1978–2007 were:

18   T. Cavalcanti and J.T. Jalles • • •

Gasoline consumption per capita: Total gasoline consumption in thousands of barrels divided by population. Source: Brazil, Ipeadata; United States, Energy Information Administration (EIA) office. Real GDP per capita. Source: Brazil, Ipeadata; United States, IMF-­IFS. Gasoline price index. Source: Brazil, Ipeadata; United States, Energy Information Administration (EIA) office.

Table 2.6  Summary statistics for US variables

Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis Jarque–Bera Probability Sum Sum Sq. Dev. Observations

dOIL

INF

dGDPdef

Y

0.007686 0.003975 0.209994 −0.213776 0.058027 −0.190282 6.177329 57.17473 0.000000 1.029986 0.447829 134

0.010727 0.008956 0.038217 −0.008560 0.007740 1.139068 4.788135 46.82925 0.000000 1.437371 0.007968 134

4.354512 4.442458 4.803275 3.612255 0.320768 −0.719925 2.544958 12.73128 0.001720 583.5046 13.68468 134

0.007515 0.007339 0.038645 −0.020383 0.007659 −0.207406 5.969043 50.17899 0.000000 1.006992 0.007802 134

Table 2.7  Covariance matrix, United States

dOIL INF dGDPdef Y

dOIL

INF

dGDPdef

Y

0.0033 0.0001 0.0007 5.4122e−06

–   5.9462e−05 −0.0014 −3.0515e−06

– –   0.1021 −0.0001

– – – 5.82e−05

Table 2.8  Summary statistics for Brazilian variables

Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis Jarque–Bera Probability Sum Sum Sq. Dev. Observations

dOIL

INF

Y

0.007783 0.004052 0.209994 −0.213776 0.058236 −0.194586 6.137040 55.37496 0.000000 1.035138 0.447663 133

18.14313 6.372792 150.6757 −1.128204 27.51146 2.430627 9.413675 358.9172 0.000000 2,413.036 99,908.21 133

0.224276 0.092187 1.692416 −0.112731 0.286039 2.016963 8.383701 250.7981 0.000000 29.82877 10.79999 133

Oil price shocks and the macro economy   19 Table 2.9  Covariance matrix, Brazil

dOIL INF Y

dOIL

INF

Y

  0.0033 −0.1062 −0.0014

751.1895    7.2220

0.08120

Appendix 2.2  SVAR lag-­length selection criteria for United States and Brazil Table 2.10 contains the statistics for the usual test of length criteria in VAR models. Table 2.10  SVAR lag-length selection criteria United States: Endogenous Variables: dOIL, INF, Y Lag

LogL

LR

FPE

0 2 4 6

1,153.489 1,207.654 1,238.278 1,262.647

– 12.635 56.221 16.143

7.32e 4.27e−12 3.09e−12* 3.24e−12 −12

AIC

SC

HQ

−17.126 −17.666 −17.98923* −17.949

−16.996 −17.147 −17.275 −16.652

−17.073 −17.455 −17.699 −17.422

Brazil: Endogenous Variables: dOIL, INF, Y Lag

LogL

LR

FPE

AIC

SC

HQ

0 2 4 6

−313.444 −185.416 −155.180 −143.700

– 41.796 40.980*   8.861

0.030 0.005 0.004* 0.005

5.031 3.297 3.105* 3.207

5.164 3.835* 4.045 4.551

5.085 3.516 3.487* 3.753

Note * indicates lag order selected by the criterion; LR: sequential modified Likelihood Ratio test statistic (each test at 5 percent level); FPE: Final prediction error; AIC: Akaike information criterion; SC: Schwarz information criterion; HQ: Hannan-Quinn information criterion.

Notes   1 We have benefited from comments at Ilha Bela Conference on the Socio-­Economic Impacts of Energy in the Past, Present and Future: A Comparison of Brazil and the US. We thank Werner Baer and Alexandre Rands Barros for helpful conversations, and Carlos Azzoni and Joaquim José Martins Guilhoto for their hospitality. We are responsible for any remaining errors.   2 If wages increase after a negative shock, then inflation will increase further. A contractionary monetary policy would then increase unemployment further. On the other hand, a loose monetary policy would increase inflation. See Bernanke et al. (1997) for more on oil shocks and monetary policy.

20   T. Cavalcanti and J.T. Jalles   3 Early studies documented and tried to explain the inverse relationship between increases in the oil price and aggregate economic activity. Among those, see Pierce and Enzler (1974) and Darby (1982).   4 Data are from IMF-­IFS (US data and international oil price) and IPEA data (for Brazil). For the sake of space, we omit the estimated parameters and associated standard deviation.   5 We found similar results for the Brazilian economy, using annual data from 1954 to 1980. We use annual data since there is not quarterly data for the Brazilian economy for a period before the first quarter of 1975. Given the annually time frequency, we use only one lag variable in Equation (1) for the Brazilian case.   6 Gali and Gambetti (2008) show that the reduction in output volatility was followed by an increase in hours worked volatility. Such result provides some evidence of a different hypothesis associated with the “good luck” explanations of the Great Moderation.   7 The international price of oil decreased sharply in the mid-­1980s (see Figure 2.1) and this sharp drop is associated with a decrease in the ability of OPEC to keep a high and stable price of oil (see Barsky and Kilian 2004; and Rotemberg and Woodford 1996).   8 Brazil has made a great effort to increase its total energy production, particularly oil, over recent years. Recently, Petrobrás (the Brazilian oil company) announced that it had discovered an estimated 6–8 billion barrels of recoverable reserves (including both oil and natural gas) in the Tupi field, located in the Santos Basin. The reserves occur in a subsalt zone that is an average of 18,000 feet below the ocean surface. The Tupi find is the largest oil discovery in recent years (American Energy Information Administration 2008). Considerable challenges must still be overcome in order to extract oil from these reserves; but Petrobrás has a strong record in overcoming financial and technical challenges.   9 After the 1973 oil shock, the Brazilian military government launched the Programa Nacional do Álcool (National Alcohol Program) or ProAlcool. It was designed to increase the production of ethanol from sugar cane. During the early 1980s, Brazil’s ethanol program flourished with high oil prices, but its political support started to wane as oil prices declined in the late 1980s and early 1990s. However, the introduction of new technologies in the last years caused another boom in ethanol production in Brazil. Ford introduced flex-­fuel cars in 2002, followed by Volkswagen in 2003. Flex-­fuel cars can operate on ethanol, gasoline or any blend of the two fuels. The government created incentives for consumers to purchase flex-­fuel cars and they cost no more than single-­fuel models. Approximately 90 percent of all new cars produced in Brazil are now flex-­fuel (Associação Nacional dos Fabricantes de Veículos Automotores). See Schmitz et al. (2007) and Weidenmier et al. (2008) for a historical analysis of the Brazilian ethanol program. 10 The net oil import share corresponds to the ratio of oil domestic consumption minus oil total domestic production over oil domestic consumption. 11 A VAR model can be seen as a reduced form of a simultaneous equations model and therefore can be estimated by Ordinary Least Squares, equation by equation. These estimates will be both consistent and asymptotically efficient. 12 Sims (1980) advocates the use of VAR models as a theory-­free method to estimate economic relationships, thus being an alternative to the “incredible identification restrictions” in structural vector models. 13 Summary statistics and the covariance matrix are presented in Appendix 2.1. Blanchard and Gali (2008) use also GDP deflator and wage inflation as variables to proxy for price level changes together with the CPI-­based inflation. In their paper all proxies for inflation yield similar results. 14 Results for unit root tests for the variables in levels did not show stationarity. The first log-­differences were therefore computed and then tested again for the order of integration. 15 Recall that these tests are suspect when the sample period includes some major events, as the oil shocks we observe in the oil price variable. Failure to consider it properly

Oil price shocks and the macro economy   21 can lead to erroneous conclusions in the case when the null hypothesis is not rejected. In order to solve this problem, Zivot and Andrews (1992) introduced an alternative formulation to overcome pre-­testing problems. 16 Other possible breakpoints could be tested instead (see Hooker 1996; and Rotemberg and Woodford 1996). 17 To find the suitable lag length, different tests are considered, namely, the modified Likelihood Ratio test, as well as the Final Prediction Error, Akaike, Schwarz and Hannan-­Quinn Information Criteria. All tests supported the use of 4 as the proper lag length at a 5 percent level. For the statistics on these tests see Table 2.10 in Appendix 2.2. 18 The two standard error bands around the impulse responses are based on Lütkepohl (1990). 19 Results are presented using the first log difference of oil price as a proxy for oil price shocks. We do not present results using variable NOPI given that the IRFs were almost identical. 20 Note that the volatility in output growth and inflation is much larger in Brazil. 21 Demand price elasticities for the short run in Brazil and in the United States are −0.031 and −0.037, respectively (both significant at usual confidence levels). This confirms gasoline demand as very inelastic in the short run for both countries.

References American Energy Information Administration (2008) “Brazil”, Country Analysis Briefs, October. Baer, W. (2008) The Brazilian Economy: Growth and Development, Boulder, CO: Lynne Rienner Publishers, sixth Edition. Barsky, R. and L. Kilian (2004) “Oil and Macroeconomy since the 1970s”, Journal of Economic Perspectives, 18 (4): 115–134. Bernanke, B., M. Gertler and M. Watson (1997) “Systematic Monetary Policy and the Effects of Oil Price Shocks”, Brookings Papers on Economic Activity, 1: 91–148. Blanchard, O. and J. Gali (2008) “The Macroeconomic Effects of Oil Price Shocks: Why are the 2000s so Different from the 1970s?”, MIT Working Paper 07–21. Brown, S. and M. Yucel (2002) “Energy Prices and Aggregate Economic Activity: An Interpretative Survey”, Quarterly Review of Economics and Finance, 42 (2). Burbidge, J. and A. Harrison (1984) “Testing for the Effects of Oil-­Price Rises Using Vector Autoregression”, International Economic Review, 25: 459–484. Darby, M.R. (1982) “The Price of Oil and World Inflation and Recession”, American Economic Review, 72: 738–751. Engle, R.F. and C.W.J. Granger (1987) “Co-­integration and Error Correction: Representation, Estimation, and Testing”, Econometrica, 55 (2): 251–276. Gali, J. and L. Gambetti (2008) “On the Sources of the Great Moderation”, Mimeo, Universitat Pompeu Fabra. Gisser, M. and T.H. Goodwin (1986) “Crude Oil and the Macroeconomy: Tests of Some Popular Notions”, Journal of Money, Credit and Banking, 18 (1): 95–103. Hamilton, James D. (1983). “Oil and the Macroeconomy since World War II”, Journal of Political Economy, 91 (2): 228–248. Hamilton, James D. (2003) “What is an Oil Shock?”, Journal of Econometrics, 113 (2): 363–398. Hamilton, James D. (2005) “Oil and the Macroeconomy”, Mimeo, prepared for the Palgrave Dictionary of Economics.

22   T. Cavalcanti and J.T. Jalles Hooker, M.A. (1996) “What Happened to the Oil Price–Macroeconomy Relationship?”, Journal of Monetary Economics, 38: 195–213. Kilian, Lutz (2005) “Exogenous Oil Supply Shocks: How Big Are They and How Much Do They Matter for the US Economy?”, Review of Economic and Statistics, 90 (2): 216–240. Klenow, P. and A. Rodriguez-­Clare (2005) “Externalities and Growth”, in P. Aghion and S. Durlauf eds, Handbook of Economic Growth, Vol 1A, Amsterdam: Elsevier, 817–861 (chapter 11). Lütkepohl, H. (1990) “Asymptotic Distributions of Impulse Responses, Step Responses, and Variance Decompositions of Estimated Linear Dynamic Models”, Review of Economics and Statistics, 72: 777–793. Mork, K.A. (1989) “Oil and the Macroeconomy When Prices Go Up and Down: An Extension of Hamilton’s Results”, Journal of Political Economy, 91: 740–744. Olson, M. (1988) “The Productivity Slowdown, the Oil Shock, and the Real Cycle”, Journal of Economic Perspectives, 2 (4): 43–69. Pierce, J.L. and J.J. Enzler (1974) “The Effects of External Inflationary Shocks”, Brookings Papers on Economic Activity, 1: 13–61. Rotemberg, J.J. and Michael Woodford (1996) “Imperfect Competition and the Effects of Energy Price Increases”, Journal of Money, Credit, and Banking, 28 (part 1): 549–577. Schmitz, T., J. Seale and P. Buzzanell (2007) “Brazil’s Domination of the World’s Sugar Market”, Working Paper, Arizona State University. Sims, J. (1980) “Macroeconomics and Reality”, Econometrica, 48: 1–48. Weidenmaier, M., J.H. Davis and R. Aliaga-­Diaz (2008) “Is Sugar Sweeter at the Pump? The Macroeconomic Impact of Brazil’s Alternative Energy Program”, NBER Working Paper 14362. Zivot, E. and D. Andrews (1992) “Further Evidence on the Great Crash, the Oil-­price Shock, and the Unit-­root Hypothesis”, Journal of Business Economic Statistics, 10: 251–270.

3 Energy and income distribution in Brazil’s development process1 Edmund Amann and Werner Baer

Income and property concentration has been a constant in Brazil’s economic history from colonial times onwards.2 Much of the literature concerned with this theme has centered on issues such as distribution of land, industrialization strategy and inflation. One theme which has been overlooked concerns the impact of the type of energy used on income distribution. An investigation of this question now seems especially pertinent given the Lula government’s commitment to both poverty alleviation and to the development of new energy sources. The purpose of this chapter is to examine the links which may exist between changes in Brazil’s energy mix and alterations in the pattern of income and property distribution. A special focus of the chapter is its analysis of the distributional implications of bio-­ethanol fuels. The chapter adopts the following structure. In the first section we analyze the development of energy sources in Brazil, tracing events from the pre-­industrialization period through the ISI period, and then to the oil shock of the 1970s. Next, the crisis years of the 1980s are reviewed. Following this, the discussion is brought up to date by analyzing the development of the energy sector from the early 1990s onward. Next, we examine the possible links between changes in the pattern of energy supply on the one hand, and distribution on the other. In particular, the distributional implications of Brazil’s growing energy self-­sufficiency are investigated. Themes reviewed here include the possible distributional impacts of the pro-­ alcohol ethanol program and the increasing role of hydrocarbons in Brazil’s electricity generation. Finally, in the concluding section, we draw together the main arguments and comment on the possible future evolution of the energy-­ distribution relationship.

Historical review It is well known that Brazil’s economic history consisted of a number of export cycles: the Brazil wood cycle, the sugar cycle, the gold cycle and the long coffee cycle (Baer 2008). During those periods the source of energy consisted mainly of wood, animal power (mules), direct water power, and slave labor (Dias Leite 2007: 48).3 Those who profited most from these sources of energy were the landowning class. In the second half of the nineteenth century, as railroads became a

24   E. Amann and W. Baer major means of transportation, the source of energy was basically wood transformed into charcoal. Wood was also used as the main source of fuel in Brazil’s fast-­growing coffee sector. Charcoal, was used in Brazil’s nascent iron industry.4 Imported coal was used in port terminals and in factories located in port cities, while in the interior the emphasis was on wood-­stoked boilers.5 Petroleum in commercial quantities was not found until after World War II, and as Brazil’s fleet of motor vehicles rose in the 1920s, the country relied increasingly on petroleum imports. From the late nineteenth century onwards Brazil developed an industrial sector (mainly textiles and food processing) which relied on electricity. At first Brazil relied on thermal generation of electricity, which began in the 1880s. But due to the need to import coal, the country relied increasingly on hydro-­ generation, which, by 1900, surpassed thermal generation. Brazil’s total electrical capacity passed 1 MW in 1890, 10 MW in 1900, and 100 MW by 1908, reaching 152 MW in 1910.6 The 1920s witnessed the construction of large dams and improvements in the generation and transmission in the states of São Paulo and Rio de Janeiro, and by the end of the decade the capacity had grown to 779 MW (see Table 3.1). However, the continued growth of the small scale iron and steel industry was still based on charcoal. The industrial surge of the 1930s, induced by the world depression, continued to be based on the expansion of hydro-­electric sources and the use of charcoal. But it was hydro-­electric power which became predominant. This, according to Dean, Was as fortuitous a technological advance for southern and south-­eastern Brazil as coaking coal had been in England two centuries before. It is not possible to imagine the development of industry on the limited base charcoal, and the cost of importing coal and petroleum would have been . . . deleterious.7 (Dean 1986) The post World War II import substitution era and the simultaneously burgeoning urban population had a substantial influence on the energy mix in Brazil. There was a substantial expansion of hydro sources. However, the development of an integrated steel industry required a type of coal which was not available in Brazil (the local coal had too high an ash content to be useful in that industry). In addition, a key element of the accelerated import substitution which was taking place in Brazil was the automobile industry. To accommodate that industry, the government promoted the large scale construction of highways. This resulted in a dramatic decline in railroads while an increased proportion of goods were transported by road. Thus, the import substitution industrialization of Brazil in the middle of the twentieth century depended on a dramatic increase in reliance on petroleum products which were mainly imported at that time. The increasing energy dependence of the country led to a rising consciousness concerning the distributional implications of energy. Until 1930, Brazil’s

Energy and income distribution   25 Table 3.1  Brazil: yearly rate of growth of internal energy consumption (%)

Petroleum Coal Hydro Charcoal Sugar cane Other Total

1915–1930

1956–1964

1964–1974

1974–1984

8.8 4.3 7.6 – – 5.7 –

8.0 −0.3 8.3 1.9 5.9 – 3.6

9.2 4.5 11.3 −0.6 8.6 – 4.3

1.6 9.7 9.8 −0.1 13.3 15.1 3.5

Source: Adapted from Dias Leite (2007).

Table 3.2  Brazil’s energy mix (%)

Oil Coal Gas Nuclear Hydro Sugar cane Other Total

1973

1980

2004

  45.0   24.8   16.7    –    1.8   11.2    0.5 100.0

  50    5    1    –   10   34    – 100

  42    7    8    2   14   27    – 100

Source: Adapted from World Energy Outlook, 2006.

Table 3.3  Brazil: installed electric capacity and production* (MW)

1900 1914 1920 1930 1940 1945 1950 1960 1970 1980 1987 2004 2006

Total

Hydro

Thermal

10 303 367 779 1,244 1,342 1,883 4,800 11,233 31,147 50,329 86,500 96,634

5 253 301 630 1,009 1,080 1,536 3,642 8,828 27,014 43,244 71,795 73,434

5 50 66 149 235 262 347 1,158 2,405 4,133 7,085 14,705 23,200

Sources: IBGE, Estatisticas Historicas do Brasil, (1990), p. 493; Ipeadata. *Includes concessionaires and self-producers.

26   E. Amann and W. Baer government intervened very little in energy markets. In 1904 the federal government issued a decree (no. 5.407) which established rules for concessionary contracts for firms using hydroelectric sources. Among various rules contained in the decree was one requiring tariff revisions every five years. The general impact, however, was mild. A concession was given in São Paulo and Rio de Janeiro for the Light company to revise its tariff more frequently as its concession called for the application of the so-­called “gold clause”, which permitted the firm to automatically readjust tariffs when the exchange rate was devalued. Until 1930, governmental power in the electric energy industry resided to a large extent with municipal governments. In 1931 the federal government began to make drastic institutional changes. All previous arrangements were suspended, including the old “gold clause”. The precarious situation of the Brazilian economy due to the world depression and a nationalistic campaign against foreign investments led to a decree in July 1934, known as the Codigo das Aguas [Water Code], which became the basic legal instrument for the federal government to regulate the water and electric energy sectors.8 With the Codigo das Aguas, the government began to regulate the electric power industry in such a manner as to control the distributional aspects of this source of power through regulating tariffs. These tariffs were to be set by taking into account costs, depreciation and a reasonable rate of return on investments. The latter was based on the historical cost of capital, allowing a rate of return of 10 percent. This was to be the main bone of contention in the following years, as the foreign companies would constantly ask for a rate based on replacement costs.9 For political reasons these tariffs were not allowed to accompany general price increases for many years. This made the industry less and less attractive to its (mainly) foreign owners. The net result of this was that the traditional foreign investors in the electricity sector were unwilling to make the kind of investments necessary to accompany the rapidly expanding industrial sector. This resulted in the creation and expansion of state companies in that sector and by the 1970s most power generation, transmission and distribution was in the hands of state companies.10 With this there emerged the policy problems of tariff setting by the industry in which the welfare of the consumer was pitted against the rate of return needed to promote further investment. Table 3.4  Brazil: origin of resources of the electric sector (% distribution)

Internal Forced Loans State Resources Domestic Loans Foreign Financing Total

1973

1979

1984

44.9 9.4 20.3 6.6 18.8 100.0

24.2 7.6 6.1 30.1 32.0 100.0

17.9 3.9 6.0 9.4 62.8 100.0

Source: Adapted from Baer and McDonald (1998), p. 512.

Energy and income distribution   27 As Brazil faced many years of inflation during the period spanning from the 1950s to the early 1990s, it is not surprising that the government came to use its enterprises as tools of macroeconomic policy to try to control the general rise of prices. This was especially the case of the electricity sector. Allowing tariff increases below the rise of the general price level contributed not only to a general decline of internally generated funds, but often produced losses which forced the government to provide subsidies. This, in turn, increased government expenditures and worsened government budget deficits which were usually financed in an inflationary manner. Distributionally, this policy favored both consumers and industry, but the inflationary consequences would hurt the wage earners. In the first years of the military governments during the late 1960s and early 1970s, electricity tariffs were increased to make up for past price distortions, but after the oil price shock of 1973 tariffs were again allowed to lag behind general price rises. Taking the average real electricity tariff in 1964 as the base (= 100), this rose steadily until 1972 (reaching 175), but then were allowed to decline, reaching 72 in 1991.11 As the state came to dominate the provision of energy both through the state electricity companies and through Petrobrás, it faced a choice between financing investment through tariff increases or through drawing on taxpayer funds. The choices made here would obviously have important distributional implications. Had the choice been made to finance investment through tariff increases, this might have imperiled equity through its impacts on poorer consumers. On the other hand, financing investment through increased public borrowing was not without distributional consequences as rising public deficits contributed to inflationary conditions. Inflation, in turn, would fall most heavily on the poorer income groups. Tables 3.5a and 3.5b shows how electricity tariff policies affected the internal financing capacity of the sector. From 1964 to 1969 electricity tariffs increased by 62.4 percent per year, while the general price level rose by 39 percent and until 1973 tariffs continued to be readjusted slightly ahead of inflation. This policy enabled the sector to increase internally generated resources from 34 percent Table 3.5a Indexes of the real price of electric energy: 1979–1988 (1987 = 100) 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988

90.9 90.3 107.7 100.0 86.4 86.4 90.5 77.6 100.0 96.4

28   E. Amann and W. Baer Table 3.5b Brazil: tariff changes 1995–2007 (percentage change) Residential Industrial Commercial Rural Government Total average Consumer price index

287 411 238 255 269 333 139

Sources: ANEEL and Conjuntura Economica, and Pelin (1997), p. 99.

in  1967 to 44.9 percent in 1973 (see Table 3.4). The oil price shock of 1973–1974 and the foreign debt crisis of the 1980s contributed substantially to the inflationary outbursts which appeared in the second half of the 1970s and would last throughout the 1980s and into the first half of the 1990s.12 The government tried (unsuccessfully) to break the high rates of inflation through various types of price controls, including the energy sector. It has been noted in Table 3.5a that the real price of electricity was low throughout most of the 1980s and the capacity of the sector to generate internal funds for expansion was severely limited. It should also be noted that the 1980s was a period of re-­democratization and the government was not in a mood to dramatically raise tariffs on as basic a service as electricity. With the advent of the Real stabilization plan and the general adoption of neo-­liberal policies from the mid-­1990s on, the regulation of electric power in Brazil changed significantly. The regulatory agency adopted a tariff policy which favored firms in the electricity sector, many of which were being privatized.13 As Table 3.6  Yearly percentage price changes

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

General prices

Food

Transport

Communications

Fuels

22.41 9.56 5.22 1.65 8.94 5.97 7.67 12.53 9.30 7.60 5.69 3.14 5.16

8.42 1.72 1.20 1.95 8.12 3.20 9.63 19.46 7.48 3.87 1.99 1.23 0.94

17.34 18.10 14,47 0.88 20.34 12.08 8.00 9.96 7.28 11.00 8.07 3.02 0.28

– 69.21 89.60 2.00 9.20 12.89 7.60 11.27 18.69 13.91 6.45 −0.24 0.02

1.05 12.9 10.5 6.3 53.7 35.9 13.6 19.2 34.3 8.8 11.4 13.5 −1.1

Source: IBGE and IPEA.

Energy and income distribution   29 can be seen in Table 3.5b tariff adjustments in the period 1995 to 2007 were substantially higher than Brazil’s consumer price index. Vieira demonstrates that tariff adjustments fell more heavily on the smaller than on the larger consumers of electricity.14 In other words, the tariff policy that followed after 1995 had a regressive impact on the society. To protect some of the low income families from the burden of rising tariffs, attempts were made in 2002 to subsidize them, though their impact was relatively small.15

Oil, income distribution and the launch of the alcohol fuel program In 1974 OPEC succeeded in quadrupling the price of oil. At the time, 80 per cent of Brazil’s oil needs were supplied by imports. Brazil reacted to the situation throughout the 1970s by borrowing massively from international banks to finance the substantial oil-­driven trade deficit.16 At the same time Brazil began to formulate ways to reduce dependence on imported energy. This was to be achieved through a combination of policies: huge investments in petroleum substitutes, in the building of new hydro-­electric dams and transmission lines, in the expansion of ethanol and in the exploration and domestic production of petroleum.17 Domestic output of crude oil rose from an average of 165,000 barrels a day in 1975 to 638,018 in 2007. Proven reserves also grew as new offshore fields were discovered off the coast of the state of Rio de Janeiro in the 1990s and early 2000s. By 2007 the estimate of proven reserves stood at 9,138 billion barrels. This did not count massive oil reserves which were discovered in the ultra deep Tupi field in 2007 which were estimated at up to 8 billion barrels. The government’s oil price policy in reaction to the OPEC oil shock was to increase the domestic price accordingly. The price of gasoline was set high not only to reduce its consumption, but also to finance Petrobrás’ exploration and to subsidize other petroleum products. It was assumed in the late 1970s that the international price of oil would remain high. This policy was continued even Table 3.7a  Brazil: petroleum – domestic production, imports and consumption (1,000 m3)

1935 1940 1945 1950 1960 1970 1980 1987 2000 2006

Domestic production Imports

Exports

Domestic consumption

– – 12 52 4,708 9,534 10,563 32,829 71,844 100,241

– – – – 667 79 68 – 1,084 21,357

3 – 24 65 10,586 27,675 60,772 68,522 92,437 99,109

3 57 12 13 6,556 18,220 50,277 35,693 23,109 19,421

30   E. Amann and W. Baer Table 3.7b Brazil: domestic oil production (1,000s of barrels) Production 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

305,983 354,655 400,782 450,626 471,862 530,855 546,080 540,717 596.255 628,797 668,680

Sources: IBGE, Estatisticas Historicas do Brasil, (1990), p. 495. IpeaData.

Table 3.8a  Brazil: gasoline – domestic production, imports and consumption (m3)

1907 1914 1920 1930 1940 1950 1960 1970 1980 1987

Production

Imports

Exports

Consumption

– – – – 26,394 26,076 3,398,148 9,550,184 11,890,280 12,621,349

1,552 12,313 50,886 390,902 515,242 2,262,948 1,246,133 104,095 105,416 15,540

– – – – – – – – 296,223 5,152,182

1,552 12,313 50,886 390,902 541,636 2,289,024 4,644,281 9,654,279 11,389,473 7,484,707

Source: Adapted from IBGE, Estatisticas Historicas do Brasil, 1990, p. 496.

Table 3.8b  Brazil: value of petroleum imports (in US$1,000) 1935 1940 1953 1960 1970 1973 1974 1981 1987 2002 2006

10 134 978 112,635 243,273 853,383 2,902,043 11,289,109 4,424,280 6,281,000 15,201,000

Source: IBGE, Estatisticas Historicas do Brasil, 1990, p. 504; DIEESE.

Energy and income distribution   31 Table 3.8c Brazil: imports of crude oil and derivatives (% of total imports) 1960–1972 1975 1985 1992 2006

37.6 25.2 47.0 20.4 16.6

Source: Adapted from Baer (2008), p. 185.

after oil prices declined in the 1980s. However, the prices of diesel fuel and propane (used for cooking) were maintained artificially low, thus requiring subsidies. The low price of diesel was intended to keep transportation costs from rising and the propane subsidy was supposed to alleviate the situation of the low income groups. To promote the use of ethanol-­propelled cars (greater details in the next section) the price of ethanol was placed at 60 percent of that of gasoline. This subsidy was financed by selling gasoline with a 20 percent ethanol component. Since the unit cost of “pure gasoline” exceeded the unit cost of ethanol, resources were thereby generated to fund a subsidy for buyers of pure ethanol. The international oil price revolution was not fully reflected internally as subsidies meant that the price of petroleum products rose at a rate that was less than the world price. The establishment of subsidies represented another drain on public resources and resulted in more international borrowing (Baer 2008). Seeking to reduce the costs associated with dependence on imported oil, the government in the mid-­1970s began the Pro-­Alcohol program. This consisted of the following measures: (a) an incentive to increase sugar cane production; (b) the financing of distilleries; (c) an incentive for automobile producers to adapt the engines of cars to the use of alcohol. The net result of these measures was that by the late 1970s ethanol output and sales of alcohol-­compatible automobiles had risen dramatically. During the same period, efforts to increase self reliance in energy production were boosted by an expansion in the hydro-­electric program (Dias Leite 2007). After a second oil price shock in 1979 followed by a sharp rise in international interest rates as a result of the United States dramatically tightening monetary policy, most Latin American countries saw their debt skyrocket, forcing most of them to institute a debt moratorium by 1983.

Brazil enters a new era of bio fuels Although Brazil began to produce bio fuels in large quantities in the 1970s as a result of the first OPEC oil shock, reliance on this form of energy declined again in the 1980s as the price of oil fell. A substantial revival in the production of ethanol began again in the early 2000s as the international price of oil rose from a low of US$11 in 1999 to over US$100 in 2008. As can be seen in Table 3.9,

32   E. Amann and W. Baer Brazil’s ethanol production rose by around 70 percent between 2000 and 2006.18 Production of ethanol was also encouraged by a change in automotive technology. Whereas the first generation of ethanol fuelled vehicles in the 1970s had run solely on hydrous ethanol, by 2003 a new wave of vehicles had emerged which could run on a mixture of gasoline and anhydrous ethanol. These more powerful and reliable “flex-­fuel” vehicles achieved a 54 percent market share by 2005, bringing about a strong increase in demand for ethanol. The new flex-­fuel technology19 involved the development of special fuel injection systems, capable of adjusting the engine combustion and compression processes according to the relative concentration of gasoline and ethanol in the fuel. The increasing production of sugar cane for ethanol has had important implications for land use and, by extension, income distribution. It will be noted in Tables 3.10, 3.11a and 3.11b, that the area of agricultural land devoted to sugar cane increased sharply in Brazil between 1997 and 2004. This was especially so in the state of São Paulo. The tables also indicate the extent to which Table 3.9  Production of ethanol, 1997–2006 (1,000 m3) 1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Brazil 15,493 14,122 12,981 10,700 11,465 12,588 14,469 14,647 16,039 17,764 São Paulo State   9,525   9,008   8,482   6,472   7,037   7,734   8,744   8,861   9,853 10,958 Source: Data supplied by Ministry of Agriculture.

Table 3.10  Land use in Brazil (2006) (ha) Total land area Non agriculturable area Agriculturable area Area with perennial agriculture Area with annual agriculture (a) Area with sugar cane (b) % Area b/a Possible expansion area for sugar cane

851,404,680 497,793,441 353,611,239 7,541,626 34,252,829 6,252,023 18.30 51,359,289

Source: Data extracted from Censo Agropecuário (2006).

Table 3.11a  Planted area coffee and sugar cane, 1997–2004 (ha) 1997

Brazil São Paulo

2004

Coffee

Sugar cane

Coffee

Sugar cane

2,054,741 241,530

4,946,587 2,446,300

2,560,962 219,800

5,788,676 2,899,161

Source: Data extracted from IBGE.

Energy and income distribution   33 Table 3.11b  Area under cultivation by crop, Brazil 1947–2007 (ha)

Sugar cane Rice Beans Manioc Soybeans

1947

1978

1996

2007*

772,853 1,650,989 1,583,723 911,285 60,029**

2,391,455 5,623,615 4,617,259 2,148,707 7,782,187

4,184,598 2,968,125 4,069,613 1,215,495 9,240,301

6,152,929 2,974,596 4,016,797 1,901,561 22,006,677

Source: Data extracted from IBGE. Note ** Data are for 1952.

Table 3.12  Yields of key crops, Brazil and São Paulo State (kg/ha)

Brazil    Sugar cane    Rice    Beans    Manioc    Soybeans São Paulo    Sugar cane    Rice    Beans    Manioc

1947–1949

1978–1980

1995–1996

2007*

38,333 1,552 685 13,347

55,252 1,415 472 11,770 1,398

61,049 2,702 638 13,217 2,284

76,434 3,793 825 14,109 2,840

68,819 1,048 581 19,838

77,138 2,045 955 22,459

81,936 2,865 1,545 23,443

–

47,117 1,357 670 9,481

Source: Data extracted from IBGE. Note *Data for São Paulo are for 2006.

growth in the area of land allocated to sugar cane in São Paulo (the most important sugar cane producing state) has paralleled a decline in the land area given over to coffee, the crop which had dominated the state’s agricultural economy for over a century. The rise in sugar cane production has resulted not simply through more extensive use of available agricultural land but also by means of a sharp rise in average yields (see Table 3.12).

The distributional implications of the bio fuel economy As noted above, total land devoted to sugar production for ethanol has increased substantially, especially in the state of São Paulo. At the same time, the concentration of land ownership nationwide has continued to rise (see Table 3.13). According to statistics provided by the Brazilian Ministry of Agriculture, by

34   E. Amann and W. Baer Table 3.13  Land distribution: trends in Brazil, 1970–1996 Region: state

1970

1996

Brazil São Paulo state

Percentage of farms greater than 1,000 hectares   0.75   1.02   0.75   0.96

Brazil São Paulo state

Percentage of land accounted for by farms greater than 1,000 hectares 39.52 45.1 27.92 27.38

Source: Data extracted from IBGE.

2000 the Gini coefficient for landholding had reached 0.80 for Brazil as a whole, while in São Paulo state it stood at 0.75.20 The fact that the rise of the sugar agricultural economy has been associated with the persistence of patterns of concentrated landholding would tend to suggest, that in this respect at least, the bio fuel boom in Brazil may do little to alleviate income inequalities. Another key issue surrounding the land area given over to sugar cane production concerns its possible impacts on the growing of food crops and hence food prices. It might be supposed that as ethanol demand increases, growing demand for sugar cane would mean land is increasingly used for fuel rather than food production. The result would be higher food prices, a development which would obviously impact negatively on the poor. To what extent are we likely to witness such a displacement of land for food production? The evidence here (at least that provided by the IBGE agricultural census) is reassuring. The IBGE estimates that some 51,359,289 ha of suitable land is lying idle, allowing, in theory, the land area devoted to sugar to rise by 821 per cent without any encroachment into food crop production or cattle rearing. At the same time, as already mentioned, yields on existing land given over to sugar cane production have risen sharply (see Table 3.12), suggesting that the need to acquire new cultivable land may not be as acute as might be supposed. The sense that expansion in sugar cane cultivation may have resulted in a spike in food prices is undermined further when one examines long-­run price data. Between the base month of August 1994 and November 2007 the FGV general consumer price index stood at 299 while that for food prices had risen to only 240. The sense that the movement toward more intense ethanol fuel production may have income concentrating effects becomes more accentuated when one examines the ethanol distribution and processing (as opposed to the growing) components of the production chain. According to Brandão (2007) the largest five distributors control 73 percent of the market for anhydrous ethanol and 53 percent of the market for hydrous ethanol. The ethanol production industry demonstrates a more atomized profile: the 350 mills in operation nationwide are operated by approximately 120 groups. However, there is evidence of increasing consolidation in the sector: between 2000 and 2004, 20 mills – mainly in the state of São Paulo – were absorbed into the growing holdings of foreign groups.

Energy and income distribution   35 Among the key purchasers here were Louis Dreyfus (which acquired the Cresciumal mills in São Paulo), Béghin-Say (which acquired the Guaraní and Cruz Alta mills also in São Paulo) and Secden (which acquired five further mills). Over the past 20 years the number of mills operating in the key northeastern sugar producing state of Pernambuco fell from 43 to 22 (Mendonça 2005). Another potential impact of the burgeoning ethanol sector on the distribution of income stems from its labor intensity and employment generation. According to Goldemberg (2002) for every one job generated in the oil sector, 152 are generated in ethanol. This compares with four jobs in coal production and three in hydro-­electric power. However labor intensive the sector may be, it must be recognized that rates of pay in the sector are very low. Goldemberg reports that average rates of pay in the ethanol sector are just over 1 per cent of those in the oil industry. Moving from pay to other issues surrounding the quality of employment, the international press has made much of the supposedly poor conditions in the ethanol production sector, especially in its sugar cane growing component.21 It would seem that some of the poor labor conditions in the fields will be alleviated by growing mechanization (Smith and Caminada 2008).22 However, attempts to increase mechanization, while improving labor conditions, will by definition increase capital intensity and thus reduce employment, a factor which is likely to drive more intense concentration of income.

Conclusions Much of the literature concerning income distribution in Brazil has focused on issues of land distribution, wage policies, impacts of capital intensive technologies or educational opportunity. No attention appears to have been paid to the implications of energy availability for income distribution. We have shown that energy has played an important role in affecting the distribution of income. This role has been realized through a number of channels. In first place we demonstrated how, in colonial times, the energy mix adopted may have contributed toward a skewed distribution of income. Next, with the appearance of electricity, the initial providers were allowed to exploit their monopoly position with consequently unfavorable distributional implications. With the advent of regulation the adverse price impact of the monopoly was at first reversed through the setting of artificially low prices. Later on, however, the strategy of ISI saw the state using its regulatory powers to set tariffs in such a manner as to encourage industry at the cost of equity. We also have shown that industrialization was to large extent driven by the automobile industry and investments in road transportation rather than rail. This greatly increased dependence on consumption of petroleum and its derivatives. This favored the upper income groups since they were the beneficiaries of the automotive and related sectors. Finally, with sharp rises in the relative price of oil-­based fuels, Brazil’s notable development of the bio fuel industry, while effectively reducing the country’s dependence on imported oil, did little, if anything, to address underlying inequalities whether in terms of income or asset distribution.

36   E. Amann and W. Baer

Notes   1 We wish to thank Daniel Walsh for many helpful suggestions.   2 See Amann and Baer (2008).   3 Schwartz notes that sugar mills “were of two types: those driven by water-­wheels . . . and those powered by oxen or, more rarely, by horses” (1984: 432–433). With innovations animal-­powered mills gradually prevailed. Schwartz describes the use of slaves, both men and women, in the functioning of the mills.   4 Warren Dean noted that Coal deposits were low grade and exiguous; in 1929 output from the mines of Santa Catarina amounted to only 360,000 tons. For lack of this resource, wood and charcoal were burned, with further severe consequences for forest reserves, watersheds and topsoil. . . . By 1930 Brazil was clearing 330,000 ha of forest annually for this purpose alone. (1986: 711–712).   5 Dean (1986: 712).   6 Dias Leite (2007: 567).   7 Dean (1986: 712).   8 Baer and McDonald (1998: 507).   9 Tendler (1968: 48–49). 10 For details see Baer and McDonald (1998: 507–511). 11 Baer and McDonald (1998: 513). 12 Baer (2008: chapter 5). 13 For a discussion of Brazil’s privatization, see Baer (2008: 230–232); and Tankha (2009). 14 José Paulo Vieira (2007) offers a thorough analysis of electricity tariff policies in the second half of the 1990s and the early years of the twenty-­first century. 15 Vieira (2007: 158–163). 16 Much of the monies lent originated from re-­cycled “petro-­dollars”. 17 New oil fields were not discovered until the late 1970s. However investments in the energy sector increased from about 10 percent of total investment in the early 1970s to about 23.5 per cent in 1982–1983. As a proportion of GDP this was an increase from 2.8 percent o 5 percent. 18 Rises in ethanol output were especially concentrated in the state of São Paulo. 19 The technology underpinning flex fuel injection systems was developed in Brazil by a team based at the Brazilian subsidiary of the Italian autoparts firm, Magneti\Marelli. 20 For a detailed discussion of land distribution trends in Brazil – and in São Paulo state – please see Hoffmann (2007). 21 Some observers have claimed that the sugar cane cutters “are effectively slaves . . . and that Brazil’s ethanol industry is, in fact, a shadowy world of middle men and human rights abuses” (Philips 2007). The article also states that these cutters work on “12 hour shifts in scorching heat, earning just over 50p per tonne of sugar cane cut before returning to squalid, overcrowded guesthouses rented to them at extortionate prices by unscrupulous landlords, often ex-­sugar cutters themselves”. According to an article by Michael Smith and Carlos Caminada (Bloomberg 14 March 2008), from “2002 to 2005 . . . 312 sugar and ethanol workers died on the job, and 82,995 suffered accidents. . . . The number of accidents on the job increased to 23,787 in 2005 from 16,877 in 2002”. 22 According to the article, “two major distillery firms, Cosan and São Martinho had plans to replace cane cutters with combines. Cosan had plans in 2008 of investing US$96 million to mechanize”.

Energy and income distribution   37

References Amann, E. and W. Baer (2008). “Neo-­liberalism and Market Concentration in Brazil: The Emergence of a Contradiction?”. Quarterly Review of Economics and Finance, 48 (2): 252–263. Baer, Werner (2008). The Brazilian Economy: Growth and Development. Sixth Edition. (Boulder: Lynne Rienner Publishers). Baer, Werner, and Curt 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. Vol. 38, No. 3. Fall. Brandão, Antônio (2007). “The Sugar/Ethanol Complex in Brazil: Development and Future”. Presentation given at IFPRI Conference on Global Sugar Markets and Reform Options. Washington, DC. 1 June. Dean, Warren (1986). “The Brazilian Economy: 1870–1930”. In The Cambridge History of Latin America, Vol. V, “c.1870 to 1930”, edited by Leslie Bethell. (Cambridge: Cambridge University Press). Dias Leite, Antonio (2007). A Energia do Brasil. Segunda Edição, Revista e Atualizada. (São Paulo: Elsevier Editora Ltda.) Goldemberg, J. (2002). The Ethanol Program in Brazil. Online, available at: www.ccaa. org/pdf/JG%20ethanol%20program%20brazil.pdf Hoffmann, Rodolfo (2007). Distribuição da Renda e da Posse da Terra no Brasil. Brasilía: NEAD. Mendonça, Maria Luisa (2005). “The WTO and the Devastating Impacts of the Sugarcane Industry in Brazil”. Rede Social de Justiça e Direitos Humanos. Pelin, Eli Roberto (1997). “The Brazilian Energy Sector”. In The Brazilian Economy: Structure and Performance in Recent Decades, edited by Maria J.F. Willumsen and Eduardo Giannetti da Fonseca. (Miami: North–South Center Press, University of Miami). Philips, Tom (2007). “Brazil’s Ethanol Slaves: 200,000 Migrant Sugar Cutters who Prop up Renewable Energy Boom”. Guardian. 9 March. Schwartz, Stuart B. (1984). “Colonial Brazil, c.1580–c.1750: Plantations and Peripheries”. In The Cambridge History of Latin America, Vol. II, “Colonial Latin America”, edited by Leslie Bethell. (Cambridge: Cambridge University Press). Smith, Michael and Carlos Caminada (2008). “Brazil Ethanol Boom Belied by Diseased Lungs among Cane Workers”. Bloomberg. 14 March. Tankha, Sunil (2009). “Lost in Translation: Interpreting the Failure of Privatisation in the Brazilian Electric Power Industry”. Journal of Latin American Studies. Vol. 41: Part 1: February. Tendler, Judith (1968). Electric Power in Brazil: Entrepreneurship in the Public Sector. (Cambridge, MA: Harvard University Press). Vieira, José Paulo (2007). Antivalor. (São Paulo: Paz e Terra).

4 The earth is finite and other irrelevancies about the world’s ultimate oil supply Fred Gottheil

Introduction The conventional wisdom associated with the history of oil discovery and production and with estimating the world’s ultimate oil supply is anchored in a mindset that accepts the idea that there’s not much oil left and that finding it has become as difficult as finding that needle in a haystack. A substitute explanation of that history is anchored in the idea that the oil is there, in abundance, but that we are just not looking for it. Moreover, not looking makes perfect economic sense for those not looking. The idea that’s there not much oil left may be the more intuitively correct of the two proposed. Because planet Earth has finite dimensions, it stands to reason that there must also be some finite dimension to everything on or in it, including oil. Theoretically, then, we can run out of it. But does theory trump reality? After all, oil is no less buried away somewhere than is that needle in the haystack. There may be oceans of oil still undiscovered and untouched. There are 150 million m2 of land mass on planet Earth and even greater areas of ocean floor. That’s a mighty large haystack! The bedrock question then is: How much recoverable oil is left on planet Earth?1

Enter the idea of peak oil Much of the professional literature on ultimate recoverable oil has come to revolve around the concept of peak oil. The idea is simple and Malthusian, and very depressing. It asserts that while we once enjoyed an abundance of oil – drilling and producing to satisfy our most gluttonous consumption – we are running out of the stuff, pure and simple. As Exxon-­Mobil spokesperson William J. Cummings recently noted: “All the easy oil and gas in the world has pretty much been found. Now comes the harder work in finding and producing oil from more challenging environments and work areas.”2 What Cummings is telling us is that the law of diminishing returns has reared its ugly head to remind us that we live in a finite planet, housing obviously finite resources that are less and less accessible, and that’s all there is to that! This not-­too-difficult-­to understand idea was made the centerpiece of a model produced in the 1950s by one of the world’s foremost oil geophysicists, M. King

The earth is finite and other irrelevancies   39 Hubbert, then chief consultant for Shell Oil’s exploration and production research division. Hubbert framed the problem in terms of approaching and then ultimately having to live beyond a “tipping point” of world oil production, which he defines as peak oil: “Peak oil is that year or set of years somewhere in the future when the annual rate of global oil production will reach its maximum after which that rate enters an era of global terminal decline.” The rate of decline may be debatable, but not its inevitability. Much of the subsequent modeling associated with oil production and oil reserves is based on the Hubbert peak-­oil model template.3 The most optimistic scenario, according to the Wood–­Long–Morehouse three-­scenario forecasting model is the 3,896 billion barrel estimate of world recoverable oil that peaks in 2047. Their least optimistic forecast is the 2,248 billion barrels, peaking in 2026. The trajectory depicting rates of decline in world reserves beyond peak is calculated on assumptions of a fixed reserve–production ratio (R/P) of 10 and a 2 percent rate of growth in world GDP. The results are unmistakable. Anywhere in the model’s range of estimated recoverable oil, we peak well before the mid-­twenty-first century and we’re about flat out by the mid-­twenty-second century. Cast in shifting demand and supply curves, the consequences are dire. If you think oil prices are high now, just you wait! The Wood–­Long–Morehouse model reflects their master’s narrative: They write: All or very near all of the Earth’s prolific petroleum basins are believed identified and most are partially to near-­fully explored. All or nearly all of the largest oil fields in them have already been discovered and are being produced.4 Colin Campbell, exploration geologist and Trustee of the Oil Depletion Analysis Center concurs: “These are the main parameters for conventional oil. We have produced almost half what there is, and we have found about 90 percent.”5

It’s déjà vu, all over again What’s troubling about their observations? It’s déjà vu, all over again. The dire forecasts of very soon producing and consuming the last few ounces of commercially-­viable oil started almost as early as the discoveries of commercially abundant oil. The Bureau of Mines estimated in 1914 that there would be only enough oil for ten more years of consumption.6 A.J. Hazlett predicted in 1918, in Oil Trade Journal, that the commercial supply of crude oil would become exhausted in a matter of years. That was 90 years ago.7 In 1922, Sir Edward Mackay, a key player in the 1908 founding of the Anglo­Persian Oil Company was assured that: “Before long there will be a smash. An economic crisis is approaching . . . 150 million people are feverishly tearing from the Earth its irreplaceable wealth.”8 Some 17 years later, in 1939, the US Department of the Interior projected that oil would run out in 13 years.9 A half-­century

40   F. Gottheil later, the language, the analysis, and the predictions were pretty much the same. In fact, oil gurus in the 1970s were predicting that by the time we reached 2008, there would be no oil to speak of. They may have been dead wrong – as were their predecessors of the 1920s – but they were no less emphatic. James Akins, an industry-­valued private oil consultant, later head of the US State Department’s office of Fuels and Energy, and whose prestige as an oil expert allowed him to become US Ambassador to Saudi Arabia testified in Congress that conventional hydrocarbon energy sources would soon be exhausted, quite possibly before the end of the century. He didn’t mean this century. He meant the twentieth century. That doomsday view, of course, was by no means considered bizarre or even astonishing. It reflected the overwhelming opinion of industry, government, and academic people considered authorities on matters of oil availability and oil economics. Any challenge then to the view that the supply of global oil was simply running out was tantamount to reckless and irresponsible gibberish. The challenging evidence brought to bear on the issue was ignored with contempt and the challengers marginalized. But let’s revisit that evidence. And allow, for a moment, an analogy drawn between oil and back-­home fishing. Fisherman will tell you that if you dig for worms and find all you want in the first dig, you stop digging and go fishing. On the other hand, if you dig and find precious few worms, you are obliged to keep digging. What makes sense for these fishermen should make sense for oilmen. If you can’t find the oil you need, you keep digging.

The drilling record The drilling facts collide head on with the oil industry’s doomsday rhetoric. The 1964–1973 drilling record of Table 4.1 shows that for the nine-­year period preceding the 1973 oil crisis, the number of wells drilled both in the United States and elsewhere declined dramatically. Yet far from searching for more oil, the oil multinationals choose to decrease their drilling efforts. To MIT economist M.A. Adelman that made sense. He noted in 1972 that: For at least 15 years we can count on, and must learn to live with, an abundance of oil that can be brought forth from fields now operated in the Persian Gulf at something between 10 and 20 cents per barrel at 1968 prices.10 And while the United States and the Middle East are by no means unimportant regions of the world, what seems rather curious about Table 4.1 is the limited interest showed by the oil industry in the “other” region, that is, that vast world outside the United States and the Middle East. Drilling there accounted for less than 20 percent of total. Kirkby and Adams explained that lack of interest: “Outside the USSR and China, few prospective basins remain untouched by exploration, and most have been extensively explored.”11 In other words, it was

The earth is finite and other irrelevancies   41 Table 4.1  Oil well drillings completed: total, United States and OPEC 1964–1973

1964 1967 1971 1972 1973

Global

United States OPEC

Other

Other/global (%)

24,397 18,725 14,996 14,608 13,249

19,905 15,073 11,567 11,884   9,555

3,328 2,877 2,115 2,096 2,305

13.6 15.4 14.1 14.3 17.4

1,164 775 1,314 1,328 1,389

Source: Data derived from World Oil, Gulf Publishing Company, Houston, Texas, 15 August 1965; 15 August 1968; 15 August 1972; 15 August 1973; 15 August 1974.

not that the oil industry ignored the rest of the world so much as Mother Nature – who created oil – ignored the rest of the world. By their reckoning, there were just no needles in that haystack. A. Meyerhoff offered yet another explanation, “one of the principle reasons for not drilling in the LDCs is that the geology is unfavorable”.12 Wood, on the other hand, was less than persuaded: “Studies indicate that 145–155 major basin and sub-­basin areas are virtually undrilled at this time” and concluded that “although there is a great deal of petroleum to be found outside the conterminous United States, activity in such areas is woefully inadequate”.13 His critique of Kirkby et al. was echoed by B. Grossling who noted that while the less developed countries of the world accounted for as much as 50 percent of the world’s prospective area of oil reserves, it accounted for less than 5 percent of the world’s exploratory wells ever drilled.14 And to dispute the idea that the oil wells in the less developing countries were poor performers – which might explain the oil industry’s lack of interest – Grossling compared the barrels of oil generated per foot of drilling in the United States, Western Europe, Latin America, and Africa.15 The results are shown in Table 4.2. Whatever else may be garnered from Table 4.2, there is no discernible trend toward diminishing productivity in those oil wells of Europe, Latin America, or Table 4.2  Barrels of oil per foot of drilling: selected regions (1945–1974) Time interval

United States

Western Europe

Latin America

Africa

1970–1974 1965–1969 1960–1964 1955–1959 1950–1954 1945–1949

15.0 30.3 13.9 13.7 16.1 25.5

1,134.0 322.6 35.7 26.9 84.8 49.9

208.6 158.4 117.5 160.6 167.5 191.2

1,062.4 1,189.4 813.6 996.2 77.8 109.8

Source: Andrews, S. and Udall, R., “Oil Prophets: Looking at World Oil Studies Over Time,” ASPO Conference, 26–27 May, Paris France; “Peak Oil Theory: World Running Out of Oil Soon,” Cambridge Energy Research Associates, CERA Press Release, 14 November 2006.

42   F. Gottheil Table 4.3  Estimates of world recoverable oil: 1920–1975 (billions barrels)

1920 1942 1946 1946 1948 1949 1949 1953 1958 1959 1962

Estimate

Authority

43 600 400 555 610 1,010 1,500 1,000 1,500 2,000 1,250

Anon Pratt, Weeks Duce Pogue Weeks Weeks Levorsen MacNaughton Weeks Weeks King Hubbert

1965 1967 1968 1969 1970 1971 1972 1973 1974 1975

Estimate

Authority

2,480 2,090 1,800 2,100 1,800 1,200 1,952 4,000 2,000 2,000

Hendricks Ryman/Exxon Shell King Hubbert Moody Warman Jodry Odell Bonillas Moody

Source: Andrews, S. and Udall, R., “Oil Prophets: Looking at World Oil Studies Over Time,” ASPO Conference, 26–27 May, Paris France; “Peak Oil Theory: World Running Out of Oil Soon,” Cambridge Energy Research Associates, CERA Press Release, 14 November 2006.

Africa.16 Moreover, Mother Nature seems to be forever blessing us. Far from the world’s wells running low and dry, the estimates concerning ultimate recover­ able oil made over the period 1920–1975 grew to belie the peak-­oil advocates’ depressing view of the world’s oil potential. Look at Table 4.3. Two observations are noteworthy. First, the estimates increase over time – 40 billion barrels in the 1920s; 1,250 to 2,480 billion range in the 1960s; 1,200 to 4,000 billion range in the 1970s. Even Hubbert changes his own estimate by 60 percent, from 1,250 billion in 1962 to 2,100 billion in 1969 and Weeks’s 1948 estimate of 610 billion barrels is revised upward to 2,000 billion barrels by 1959. Second, Table 4.3’s impressive revisions overlap with the peak-­oil alarmists’ insistence that we’re about to reach the “tipping point” of oil discovery and production.

Thirty years later To sum up, what should appear as nothing short of amazing is that in the almost 100 years of discussion concerning estimates of ultimate oil recovery, the core argument hasn’t changed a whit. It was and still is about the imminent disappearance of oil. It was “imminent” to the leading authorities on oil in the 1920s, “imminent” again to leading authorities in the 1960s, and “imminent” yet again to leading authorities on oil in the 1990s. As well, this conventional wisdom expressed by the leading oil authorities in industry, government, and the academy – geophysicists in particular – continues to prevail despite the impressive contrary evidence.17 And that evidence builds. For example, look at Table 4.4, which extends the Table 4.3 data beyond 1979 to 2006. The estimates for 2000–2006 are all at or above 3,000 billion barrels. And with all exceptions admitted, note the steady upward movement in estimates

The earth is finite and other irrelevancies   43 Table 4.4  Estimates of ultimate oil recovery (billion barrels)

2006 2003 2000 2000 2000 1998 1998 1997 1997 1996 1995 1995 1995 1994 1993 1993

Author

Estimate

CERA Nehring EIA USGS Alhbrandt/USGS IEA Laherrere/Perrodon Edwards Campbell Ivanhoe Riva Laherrere Mabro Masters/USGS Townes Laherrere

3,750 3,000 3,000 3,000 3,000 2,300 2,750 2,850 1,800 2,000 2,300 1,750 1,800 2,300 3,000 1,750

1993 1992 1991 1991 1989 1987 1987 1984 1984 1983 1983 1982 1981 1981 1981 1980

Author

Estimate

OPEC Montadert/Alazard Masters/USGS Campbell Bookout/Shell Jenkins/BP Masters//USGS Ivanhoe Martin/BP Masters/Root Odell/Rosing Nehring/Rand Hubbert/Root Halbouty Colitti/AGIP W.E. Conference

2,150 2,200 2,200 1,650 2,000 1,700 1,750 1,700 1,700 1,700 3,000 2,350 2,000 2,250 2,100 2,100

Source: Andrews, S. and Udall, R., “Oil Prophets: Looking at World Oil Studies Over Time,” ASPO Conference, 26–27 May, Paris France; “Peak Oil Theory: World Running Out of Oil Soon,” Cambridge Energy Research Associates, CERA Press Release, 14 November 2006.

over this 26-year period. One again, it’s déjà vu. How can the constantly changing data shown in Table 4.4 be reconciled with the concept of peak-­oil? Two differing views are offered. Advocates of peak oil argue that the fundamental idea of a peak-­oil world is still alive and well but that its time frame must be adjusted to reflect the new data. That is to say, the newer data of recoverable oil shown in Table 4.4 only delays the timing of the inevitable. It’s not unlike Shakespeare’s soothsayer warning: “Beware the Ides of March” every Ides of March. The second view is the more severe. It is to discard the idea of peak oil altogether. After all, if the peak is repeatedly shifting over a century of revised estimates, perhaps there is no peak at all.18 Moreover, critics of the peak-­oil model look to the oil industry’s drilling behavior as evidence supporting their view. After all, given the dire forecasts offered by peak-­oil advocates, shouldn’t the oil industry be busy searching and drilling for more oil? But that’s not exactly what we see.

The non-­pursuit of recoverable oil, all over again Rather than ratcheting up oil drilling activity in response to expectations of “running out”, the number of active rigs fell rather dramatically in the 1980s, remained at relatively low levels during the 1990s, and increased only slightly over the 2000 to 2006 period.19 For the 1973–2006 period, data from Baker-­ Hughes, shows the oil industry’s continuing disinterest in world drilling. That disinterest appears to be universal. Look at Table 4.5.

44   F. Gottheil Table 4.5  International rotary oil rig count: 1985–2005 (monthly average)

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Europe

Middle East

Africa

Latin America

Far East

World

268 157 215 165 161 153 137 131 110 103 128 123 115 85 74 89 99 83 83 63 74

191 145 134 131 145 104 142 165 154 123 135 139 166 160 139 164 188 211 217 247 262

123 86 87 85 91 109 87 69 71 69 77 85 78 62 38 49 45 56 62 53 49

439 318 354 330 229 266 292 210 217 255 271 277 287 213 196 257 255 219 264 306 327

267 234 261 241 274 272 272 233 215 181 172 186 173 151 127 146 165 184 177 200 236

1,288 940 1,051 952 900 904 930 808 767 731 783 810 819 671 574 705 752 753 803 869 948

Source: Derived from “Industry at a Glance,” World Oil, Gulf Publishing Co., Houston, TX, years: 1987–2005.

Table 4.5 must raise once again the question: Just how hungry is the oil industry? Is it really interested in searching for more? The Brazil venture – or lack of venture – is instructive. Outlook 2006: International Worldwide Drilling reported that of the 1,134 block areas of potential oil put up for bids in 2006, only 251 blocks were actually picked by the oil industry.20 Disappointed in the industry’s lack of enthusiasm, Brazil’s own state-­run Petrobrás itself invested in its oil exploratory drilling. The nine potential oil fields they worked proved to be a bonanza, holding 50 to 80 billion barrels of light crude, or more than four times Brazil’s then proven reserves.21 The potential of Brazil’s Santos Basin could be one contiguous mega-­deposit of crude.22 Why did the oil industry choose to be a non-­player? And put more generally, what can possibly explain Table 4.5? Either the industry considered the quest for oil unpromising – which seems to fly in the face of continuing shifts in peak-­oil estimates – or they simply didn’t need the oil. If the oil reserves already in place were sufficient to satisfy anticipated demand, why drill for more? After all, drilling is not cost free.23 It may be more expedient then to keep whatever oil there may be – the proverbial needle in the haystack – unexposed.

The earth is finite and other irrelevancies   45 Keith Brown makes the point: In some areas (notably Saudi Arabia) large volumes of cheap-­to-find-­andproduce oil very probably exist, but, I believe, are not worth looking for because of the very large existing inventory of unutilized reserves. . . . Countries deliberately restricting production include not only OPEC members, but also such non-­member countries as Norway and Mexico. The production restrictions choked off the search for new reserves in many of the countries in which reserves could be expected to be found most cheaply.24 Randy Udall and Steve Andrews note that while peak-­oil advocates downplay the more promising estimates of Table 4.4, critics of the peak-­oil idea argue that oil producers, especially in the Middle East and other OPEC nations, have no incentive to explore for oil they don’t need immediately.25 Economist and oil industry analyst Leonardo Maugeri explains that new finds have never been high priority for the oil multinationals nor for the major oil producing nations because additions to supply “may create price-­sinking surplus”.26

The use (or abuse) of asymmetric information The lack of transparency – or the control of asymmetric information – allows the oil industry – companies and countries – to offer suspect data associated with “proven” reserves. While oil companies supposedly base their drilling activity and oil-­producing countries their production and price policies on “proven” reserves, it is more probable that their drilling, production, and pricing policies are what make reserves “proven”.27 Just a cursory glance at Table 4.6 shows the extent to which proven reserves estimates can be manufactured to serve any specific political or economic interest.28 Look at the strange year-­by-year entries for each of the OPEC members of Table 4.6. Kuwait in 1985 reported a 50 percent increase, 63.9 billion barrels of proven reserves to 90.0 billion, overnight. Are we looking at a “eureka” event? And note a “eureka” event in every column of the table. The boost in declared proven reserves had nothing to do with Kuwait’s real proven reserves. The declared increase was prompted by the OPEC decision to fix its members’ production quotas to reflect members’ proven reserves. To increase its production quota, Kuwait simply declared an arbitrary increase in its proven reserves. Kuwait was not alone. Iraq, in 1987 and for much the same reason, more than doubled its reported reserves as did Iran and Venezuela. For the first time, Venezuela included non-­conventional oil in its estimated reserves. Abu Dhabi and Dubai tripled their proven reserve counts. The OPEC giant, Saudi Arabia, increased its own in 1990 from 170 billion barrels to 257.5 billion barrels. Clearly, these reported proven reserves must be regarded as fictional. But for many peak-­oil advocates, these fictions parade as fact. Lack of transparency and data manufacturing have always characterized the industry’s reserve reporting.29 This problem of separating fiction from fact is an integral part of the more

46   F. Gottheil Table 4.6  OPEC reported reserves: selected years (billion barrels)

1980 1985 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

Dubai

Abu Dhabi

Iran

Iraq

Kuwait

Saudi Arabia Venezuela

1.4 1.4 4.0 4.0 4.0 4.0 4.3 4.3 4.0 4.0 4.0 4.0 4.0 4.0 4.0

28.0 30.5 92.9 92.9 92.9 92.9 92.9 92.9 92.9 92.9 92.9 92.9 92.9 92.9 92.9

58.0 48.5 92.9 92.9 92.9 92.9 92.9 92.9 92.9 92.9 92.9 92.9 92.9 92.9 92.9

  31.0   44.1 100.0 100.0 100.0 100.0 100.0 100.0 112.0 122.5 112.5 112.5 112.5 112.5 112.5

65.4 90.0 94.5 94.5 94.0 94.0 94.0 94.0 94.0 94.0 94.0 94.0 94.0 94.0 94.0

163.4 169.0 257.5 257.5 257.9 258.7 258.7 258.7 259.0 259.0 259.0 261.0 259.2 259.3 259.3

17.9 25.9 59.1 59.1 62.7 63.3 64.5 64.9 64.9 71.7 72.6 72.6 76.9 77.7 77.8

Source: Adapted from Kjell Aleklett and Colin Campbell, “The Peak and Decline of World Oil and Gas Production”.

general problem of separating fiction from fact on the bedrock issue of “are we running out of oil”.30

And then there’s unconventional oil . . . Relying solely on a literature of ultimate oil recovery that is based strictly on estimates of conventional oil – that is, oil extracted from wells – is not unlike ignoring the proverbial gorilla in the living room. The gorilla, in this case, is non-­conventional oil that can be extracted from tar sand, oil shale, as well as heavy oil, and bio fuel. Oil derived from these sources, if not already commercially competitive with conventional oil, wait only upon the right technology and the right price to be converted from their natural forms into useable oil. The quantities are enormous. According to Abdallah S. Jum’ah, unconventional world oil reserves, if fully exploited, would be more than twice as great as conventional proven reserves.31 These estimates make the analysis of peak oil virtually irrelevant. Consider the near 400 billion barrel estimate of North America’s non-­conventional oil; much of it tar sands and most of the sands located in Canada’s Alberta province. Other estimates of the Alberta sands are reported to run as high as one trillion barrels.32 These sands contain a semisolid form of oil known as bitumen. The distinction between crude bitumen and crude oil is quite arbitrary, depending upon the oil’s viscous grade. Oil analysts have estimates that production of this tar sand-­derived oil can become profitable with crude prices at US$30 to US$40 per barrel. Large deposits of tar sand are also found in the United States, Russia and Venezuela.

The earth is finite and other irrelevancies   47 The United States has the largest deposits of oil shale in the world, approximately 1,200 billion of the world’s estimated 1,662 billion barrels. Most of these deposits are in Colorado, Utah, and Wyoming. Like tar sands, oil shale requires processing stages – mining, crushing and heating – that not only generate negative byproducts but consume high levels of energy. Nevertheless, like tar sands, it can become commercially viable at the right price. This potential is so great that at US$40 per barrel – now a competitive price – there is enough shale oil to supply US energy needs for the next 250 years and, fully operational, is sufficient to cover those needs for the next 5,000 years.33 Heavy oil’s specific gravity is denser than the light crude of, say, Saudi Arabia – it is heavier than water – and flows less easily than the light crude. But there’s a lot of it and a lot of it can be made to flow. Consider the case of South and Central American non-­conventional oil. Much of it appears in the form of heavy oil located in Venezuela’s Orinoco Belt. Recoverable heavy oil reserves estimates in this belt are in excess of 1.3 trillion barrels, approximately equal to the world’s proven reserves of conventional oil.34 The belt contains an estimated 270 billion barrels of proven reserves, already matching the total reserves of Saudi Arabia. New technologies, specifically those dissolving the heavy oil by diluents which enable that oil to be transported to upgrading facilities have made substantial quantities of heavy oil already accessible. The 1.3 trillion barrel Venezuelan oil treasure is waiting upon the generations of more effective technologies and, of course, upon the heavy doses of international capital to exploit the resource. Is there a point here? Obviously: It is that the idea of peak oil, of the Malthusian world encroaching upon the twenty-­first century, of the idea that the world of cheap oil is over simply cannot stand the test of overwhelming counter evidence. M.A. Adelman sums it up: “It is commonly asked, when will the world’s supply of oil be exhausted? The best one-­word answer: Never.”35

Post-­script: oil drilling as a military expenditure If much oil – conventional and unconventional – is waiting upon discovery, how and where do we start? First, it is important to regard decision-­making concerning drilling activity and its location as properly belonging in the public domain. The externalities generated by oil are no less widespread or no less critical to the vitality of our society than are the externalities generated by military preparedness. We spend billions of dollars producing a missile defense system in the hope of never using it. No one would argue the fact that the system was never actually used proves it was a wasteful expenditure. Indeed, it is precisely when the missile defense system is not produced that situations develop which would have made its production necessary. The analogy to oil is proper. US policy with respect to oil should have been to explore, locate, drill, and create oil reserves in a multiplicity of locations regardless of prevailing oil prices. The argument isn’t that we actually need to discover or extract more oil. The counter evidence is impressive. But as with the

48   F. Gottheil missile defense system, our expectation should be never to use the oil. Its presence alone promotes our national agenda: to create greater oil supplies, and more important, greater number of oil suppliers. We don’t need oil as much as we need oil competition.36 The strategy must be to expand drilling in a blitzkrieg manner and to locate the drillings in as geographically dispersed a network as possible. The United States should engage in bilateral agreements with a host of countries assuming all the investment expenditure and requiring in return no commitment to us for any oil discovered. In a sense, these drillings would represent a form of economic aid. A good place to start might be in the African continent – look again at the oil rig count in Table 4.5 – where the prospects for oil have been virtually ignored by the oil multinationals. Since investments would be US funded, the recipient nations would have literally nothing to lose. US drilling activity would be viewed by them and by us as a unilateral grant, which is precisely the function it would serve. US benefits are not in what the United States. can ultimately get directly from these drillings, but from the competition it creates. Oil multinationals, for good reason, are not disposed to invest their private capitals in ventures that might be regarded as economically or politically high risk. After all, their concerns of confiscations are real. By contrast, confiscation issues are of no concern to our long-­run national interest. As well, political unrest – regional wars, civil wars, frequency of revolutions – might well discourage private investments but should not public investments. Moreover, and understandably, it is not in the interest of oil multinationals to flood the oil market with greater quantities or more diverse suppliers. But that strategy does serve our national interest.

Notes   1 Estimates of oil reserves are classified as: proved, probable, and possible. The distinguishing criterion is the degree of certainty associated with oil recovery. Proved reserves are those considered to have at least a 90 percent probability of recovery. Probable reserves are those located in either the same or adjacent fields of proved reserves and these have a 50 percent probability of recovery. The probable is highly dependent on known or expected changes in recovery technology. Possible reserves are those unproved reserves which analysis of geological and engineering data suggests have a probability of at least 10 percent recovery. The term ultimate recoverable oil – used in this chapter and in the oil industry literature – refers, in the main, to possible reserves.   2 Boston Globe (2005) 11 December.   3 See Wood, J., Long, G., Morehouse, D., “Long-­Term World Oil Supply Scenarios”. Online, available at: www.eia.doe.gov.   4 Ibid.: 2. My italics. In 1878, when Max Planck was a student at Heidelberg deciding on a career path, he was advised by his mentor, the celebrated physicist and mathematician, Johann von Jolly, not to go into physics because “in this field, almost everything is already discovered, and all that remains is to fill a few unimportant holes”. As you know, Planck ignored the advice and went on to open up the field of quantum physics.   5 Campbell, C.J. (no date) “Peak Oil: an Outlook on Crude Oil Depletion”. Online, available at: www.mbendi.co.za/indy/oilg/p0070.htm.

The earth is finite and other irrelevancies   49   6 Lomborg, B. (2001) “Running out of Resources”. Online, available at: http://images. guardian.co.uk/sys-­files/Guardian/documents/2001/08/14resources-biodiversity.pdf.   7 Deming, David (2000) “Oil: Are We Running Out?” Second Wallace E. Pratt Memorial Conference, San Diego, January.   8 Fanning, L.M. (1950) “A Case History of Oil-­Shortage Scares”, Our Oil Resources, ed., Fanning, L.M., New York: McGraw Hill.   9 Lomborg, B., op. cit. My italics. 10 Adelman, M.A. (1972) The World Petroleum Market, Baltimore: Johns Hopkins University Press: 77. My italics. 11 Kirkby, M.A., and Adams, T.D. (1974) “The Search for Oil Around the World up to 1999”, Petroleum Times, 1 November: 25. 12 Private correspondence, Meyerhoff to Gottheil. J.P. Riva of Congressional Research Services shares Meyerhoff ’s view. Private correspondence: Riva to Gottheil. 13 Wood, W.J. (1979) “There’s a Trillion Barrels of Oil Awaiting Discovery”, World Oil, June: 142. 14 Grossling, B. (1977) “A Critical Survey of World Petroleum Opportunities”, Project Independence: US and World Energy Outlook Through 1990, Congressional Research Service, Library of Congress, Committee Print 95–33, Washington, DC, November: 645. See also Broadman, H. (1984) “Incentives and Constraints on Exploratory Drilling for Petroleum in Developing Countries”, Paper given at the American Economic Association Meetings, Dallas, Texas, 28–30 December. 15 Grossling, B., op. cit.: 645. 16 Yet, in spite of the Table 4.3 evidence, Colin Campbell writes: “The larger fields are usually found first for obvious reasons, being too large to miss.” Op. cit., But if no or little exploratory drilling is undertaken elsewhere, how can anyone know that the drilling already undertaken is in the larger fields? 17 Even Paul Krugman, the 2008 Nobel Laureate in Economics, having no pretensions of being an oil expert, was confident in his New York Times article “Running Out of Planet to Exploit” to write: Meanwhile, resources are getting harder to find. Big oil discoveries, in particular, have become few and far between, and in the past few years oil production from new sources has been barely enough to offset declining production from established sources. (Krugman 21 April 2008) 18 Isaac Bashevis Singer, who won the 1978 Nobel Prize for Literature, was once asked why he continued to write in Yiddish, a dying language. Singer answered: “Yes, Yiddish may be a dying language, but in the Yiddish tradition, from dying to dead is a long, long way.” 19 Baker-­Hughes; Energy Information Administration, 2008. 20 Special Focus: OUTLOOK 2006: International Worldwide Drilling, Archive, February 2006, Vol. 227, No. 2. 21 “Oil Discovery Could Transform Brazil’s Economy” (2008) USA Today, 9 October. The Economist Intelligence Unit puts the find at 70 to 100 billion barrels. The International Herald Tribune reported the discovery at an estimated 33 billion, 5 November 2008. 22 Lyons, J., and Luhnow, D. (2008) “New Find Fuels Speculation Brazil Will Be a Power in Oil”, Wall Street Journal, 23 May. 23 The average cost of a floating rig varies from approximately $100,000 per day to over $220,000 per day depending on the rig’s depth capability. Khalid, Nazery (2006) “Kings of the Wild Frontier: Oil Rig Market’s Trends and Outlook”, Jack-­Up Asia Conference 2006, Singapore. Online, available at: www.google.com/search?hl=en&q =kings+of+the+wild+frontier+nazery+khalid&btnG=Search. 24 Brown, K. (1989) “Reserves and Reserve-­Production Ratios in Imperfect Markets”, Energy Journal, April. My italics.

50   F. Gottheil 25 Steve Andrews, Randy Udall (2003) “Oil Prophets: Looking at World Oil Studies Over Time”, ASPO Conference, 26–27 May, Paris, France. My italics. 26 Handwerk, Brian (2004) “Cheap Oil is Far From Over”, National Geographic News, 20 May. 27 M.A. Adelman is less kind on matters of transparency. His evaluation of some of the industry’s highly regarded oil peak-­oil advocates, among them R.W. Bentley – who uses data drawn from estimates by Campbell and Laheurre for Petroconsultants – allows him to comment not only on Bentley’s analysis but on Campbell’s and Laheurre’s as well. Because the data they use was never published and still remains private, they are to Adelman “untestable and unsupported”. That blatant lack of transparency allow him to dismiss Bentley’s (and Campbell’s and Laheurre’s) conclusions as being “without foundation; his source does not deserve even the most tentative and provisional acceptance”. Adelman, M.A. (2002) “Comment on: R.W. Bentley’s Global oil and Gas Depletion”, Energy Policy, vol. 30. 28 “To date, the question of the world’s true oil reserves, despite its huge importance for the future of global oil production, remains unresolved, as the detailed field data in many oil-­producing countries are considered as state secrets.” Tsoskounoglou, M., Ayerrides, G., and Tritopoulou (2008) “The End of Cheap Oil: Current Status and Prospects”, Energy Policy, vol. 36. 29 Adelman: “Except for the United States and a very few other countries, published reserves are not well defined and estimation methods are not revealed. Year to year changes usually do not mean much of anything” (Kovarik, B., Ibid.) 30 Lynch’s critical assessment of Campbell et al. shows how, time after time, these peak oil advocates have been proven wrong yet continue to generate the same kind of doomsday predictions, completely ignoring their previous and repeated miscalculations. Lynch, Michael (1998) “Crying Wolf: Warning about Oil Supply”, MIT, March. Michael Lynch is president of Strategic Energy and Economic Research, Inc. 31 Jum’ah, A. (2006) “The Impact of Upstream Technological Advances on Future Oil Supply”, Austria, 13 September. Online, available at: www.runet.edu~wkovarik/oil/. 32 “The Peak Oil Theory: Will Oil Reserves Run Dry?” Online, available at: www.cnbc. com/id/23728987. 33 Lomborg, B., op. cit. 34 Talwani, M. (no date) “The Orinoco Heavy Oil Belt in Venezuela (Or Heavy Oil to the Rescue?)”, Earth Science, Rice University. Online, available at: http://cohesion. rice.edu/naturalsciences/earthscience/research.cfm?doc_id=2819. 35 Adelman, M.A. (2004) “The Real Oil Problem”, Regulation, Spring. Adelman goes on to say: “US oil policies are based on fantasies not facts: gaps, shortages, and surpluses.” 36 Consider Adam Smith’s argument in The Wealth of Nations (New York: Modern Library) that it is the growing competition among employers for workers, not unionization that occasions increases in wages. When in any country the demand for those who live by wages . . . is continually increasing . . . the workman has no occasion to combine in order to raise wages. The scarcity of hands occasions a competition among masters who bid against one another, in order to get the workman, and thus voluntarily break through the natural combination of masters not to raise wages. (1937: 68) Substitute OPEC for “natural combination of masters” and “oil prices” for “wages” and you have Smith’s reading of effective oil policy, circa 2010.

5 Energy restrictions to growth The past, present and future of energy supply in Brazil Adilson de Oliveira, Eduardo Pontual Ribeiro, Rosemarie Bröker Bone and Luciano Losekann Introduction It is widely accepted that energy is a critical input for economic and social development (WEA 2000). Energy is used in every sector of human activity and energy consumption increases strongly in line with economic growth. Unsurprisingly, reliability of energy supply is a key objective of development policy. Shortages in the supply of energy, such as those experienced by the Brazilian economy in 1974 and 1978 (oil), 1990/1991 (ethanol) and 2001 (electricity), have a strongly attenuating impact on economic growth. Our chapter looks at the relationship between economic growth and the energy system, focusing on the Brazilian case in the post-­war years. We provide a historic overview of energy supply and evaluate the Brazilian growth-­energy elasticity for the first time using modern econometric techniques. We find that, paradoxically, there exists potential for the existence of energy supply shortages in a resource-­rich environment over the coming years. This is the key challenge that the current institutional and regulatory arrangements face and will have to adapt to. After World War II, Brazil experienced not only significant economic growth, but also a remarkable change in its energy sector. This occurred against a backdrop of industrialization and urbanization. In 1950, 63.8 percent of the population was living in rural areas and industry was 24.1 percent of GDP. Nowadays, the rural population share is only 19.8 percent and industry 37.5 percent of GDP (UN 2007; Baer 2008). The change in the energy system was also radical, as will be examined more closely in the following section. Wood, a non commercial energy source, was substituted by commercial fuels. A modern multi firm energy sector emerged, able to supply all urban areas of the country and most rural ones too. External dependency on oil was removed while a national electricity grid has integrated the entire country bar the Amazon region. Energy supply growth was remarkable over the period, and stemmed from different sources. While in the 1970s electricity supply was guaranteed from the large hydro plants, at least in the south and southeast regions, Brazil was dependent on foreign imports for its refined oil product needs. In the 1980s transportation-­related demand for such fuels was eased with the use of ethanol.1

52   A. de Oliveira et al. In the 1990s domestic oil production rose significantly. By the 2000s natural gas had become an important non-­renewable energy source. Some authors argue that productivity increases in ethanol have been such that there now exists excess supply (Sandalow 2006). Despite these accomplishments the energy system has come under stress at various points. In the 1970s the oil shock led to significant gasoline and diesel rationing, or at least restricted access to these fuels. In 1990, the changing relative prices of sugar and lack of internal support for ethanol production led to a shortage that shattered consumer confidence in this fuel. The same hydroelectric system and a more interconnected grid system that guaranteed power for larger and larger parts of the country over the 1980s came to a near halt in its supply expansion in the 1990s, leading to the 2001 power shortage in the face of below average rainfall (de Oliveira 2001). For the early twenty-­first century, the energy system in Brazil finds itself at a crossroads with significant resources at its disposal for the first time in its history. The available resources stem from large proven reserves of oil and natural gas, including the recently discovered “pre-­salt” deep water oil fields. There is also large energy output at competitive prices in the form of ethanol from sugar cane, a renewable energy resource. The electricity supply sector has become consolidated, based on a complementary system of thermal, nuclear and hydraulic plants, with the thermal plants fired by oil or natural gas. In short, while power shortages where commonplace in the 1950s, they represent news in the current era. But, the euphoria over availability of energy reserves and potential output has to be tempered with concern over the actual energy supply. This is the challenge for the twenty-­first century in Brazil, a challenge that the famous 2001 power shortage serves so well to expose. We turn to such challenges in the final part of the chapter.

The evolution of energy supply in Brazil, 1948–2008 The energy supply system in Brazil changed significantly after World War II. During the last seven decades, the Brazilian energy system has been struggling with the specter of energy shortages. In the 1950s, the main source of energy used in the country (firewood) was inadequate to supply the demand emerging from industrialization and urbanization. A consensus emerged between nationalists and liberals that state owned companies should control the energy sector. Petrobrás, a federal government monopoly, was created in 1954 to develop the incipient oil market while regional state-­owned power companies were created to develop the electricity market. In 1962, the federal government created Eletrobrás as a holding company to coordinate the regional power companies.2 These state-­owned enterprises operated with a mandate to expand the availability of modern energy sources, at prices that would foster industrialization (de Oliveira 2007). The rationale for a state enterprise was to guarantee control over resources, and to induce forward linkages and multiplicative effects in the economy.3

Energy restrictions to growth   53 These state-­owned companies used the large financial and regulatory incentives received from the state to remove the critical energy shortages in regions with strong industrialization and urbanization growth. Until the 1970s, Brazil’s rapid economic growth, regionally concentrated in the south and southeast, was pursued with no major energy constraint in these regions. Shortages were still a problem in the north and northeast where the hydro-­electric supply was lower and the interstate power grid was incomplete. Its infancy behind it, Brazil’s modern energy system faced new challenges with the 1970s oil crisis. The brutal increase in both oil prices and international interest rates badly strained the macroeconomic fundamentals of the country. The government introduced lackluster radical measures to curb rampant inflation (Baer 2007) and the economy shifted to low growth. However, urbanization kept its momentum, driving up energy consumption. Despite this, the financial situations of energy utilities were deteriorating as result of government control of energy prices. From the 1950s until the 1990s Petrobrás played an important role as a builder of infrastructure and major industrial player, both in oil refining and petrochemicals. By the 1970s, though, other energy sources had come to receive heavy government subsidies and incentives in an attempt to counteract the effects of the oil crisis. The biggest beneficiary here was ethanol. Anhydrous ethanol was used as an additive in gasoline, reducing pollution emissions from lead, while regular ethanol was used as a fuel in itself. In 1986 about 90 percent of new automobiles sold were ethanol fuelled by regular ethanol. Yet in 1990, ethanol shortages occurred which shattered consumer confidence in ethanol supply. The culprits for these shortages are usually considered to be changes in the relative prices of sugar and ethanol plus the existence of domestic price controls. Problems stemming from price controls were widespread in the sector. Energy companies were not allowed to increase their prices in line with their rapidly increasing costs.4 Moreover, government incentives to energy companies introduced during the 1950s were gradually removed. Inevitably, the financial situation of these enterprises deteriorated, reducing their ability to develop their supply networks. Investments, particularly in the power sector dwindled, and the specter of energy shortages reemerged in the 1990s (de Oliveira 2007). To move the energy system development forward, radical reforms were introduced in its regulatory and institutional arrangements. Privatization and competition were used to attract private investors to the energy system and to remove economic inefficiencies in the late 1990s. To highlight some of these trends, the following few paragraphs offer some statistical evidence. There have been three remarkable changes in energy supply over the past 70 years. The first change has involved the replacement of non-­commercial fuel (firewood and charcoal) by energy sources such as electricity and oil. This shift in the composition of energy was particularly acute up to the mid-­1970s, a period during which the thrust of the industrialization5 and urbanization process took place. Firewood replacement had positive impacts, given its poor efficiency and heavy environmental costs.

54   A. de Oliveira et al. The second change involved a strong, albeit irregular, move toward renew­ able resources over time, the key sources here being sugar cane products (ethanol) and hydraulic power. Hydropower was the chosen source for power generation in state energy planning from the 1950s, and took advantage of the favorable river basin conditions, particularly in the rich southeast. While in 1965 hydropower generated about 5 percent of all internal energy supply, it accounted for more than 15 percent in 2000. Ethanol reached 5 percent of total internal energy supply in 1995. It should be noted that oil use increased dramatically over the period, for different reasons. Oil based fuels replaced firewood and charcoal much more efficiently and they were irreplaceable in the transportation sector. Second, as the economy grew, manufacturing energy use rose remarkably. From 1970 to 1980, the manufacturing sector’s share of total energy consumption rose from 28 percent to 38 percent. Energy consumption in transportation followed a similar trend. The share of energy used among households diminished mainly as efficient commercial energy (electricity and gas) replaced firewood.6 The increase in energy use in the energy sector itself was remarkable. This is largely a result of the growing consumption of sugar cane bagasse in ethanol power plants. The third structural change in Brazil’s energy balance was a significant reduction in dependency on energy imports (Table 5.1). While in 1970, 27 percent of Brazilian energy needs were imported; in 2007 this figure was below 10 percent. This shift has resulted largely from a substantial increase in domestic oil production. Recently, Brazil has achieved oil net self sufficiency, although light oil imports are still necessary to complement the oil mix that feeds domestic refineries. The careful reader may have noticed the increase in imported electricity that occurred in 1990. These electricity imports are due to Itaipu, one of the largest dams in the world (14,000 MW capacity), built as a joint project between Paraguay (50 percent) and Brazil (50 percent) on the Paraná River. Brazil consumes most of Paraguay’s share of the plant’s output, this being registered as imports. Table 5.1  Energy external dependency, Brazil 1970–2007 Year

Energy

Oil and natural gas

Coal

Electricity

1970 1975 1980 1985 1990 1995 2000 2005 2007

27.1 39.9 42.6 20.1 25.2 30.2 22.2 10.2   8.0

67.6 79.8 83.0 43.1 43.4 49.0 27.1 −0.1   0.1

50.2 57.9 52.6 48.3 69.6 72.0 68.1 71.6 73.5

  0.0   0.1 −0.2   1.0 10.6 11.4 11.3   8.8   8.0

Source: Date extracted from BEN 2007. Note Dependency = (total demand − domestic supply), as percentage of total demand.

Energy restrictions to growth   55 From the 1950s to the 1980s, the supply of hydropower plants increased steadily. This period represents the heyday for state firms such as Eletrobrás, Furnas and Chesf. During the 1990s the construction of large dams came to a halt due to financing problems (mentioned in the previous section) and environmental concerns. The new institutional arrangements of the mid-­1990s led private investors to focus their attention on thermal power plants (a move encouraged by relatively cheap natural gas supply from Bolivia). Hydropower capacity growth averaged 13 percent in the 1970s but has fallen to below 5 percent since 1990. One major – and obvious – limitation of hydropower is its dependence on adequate precipitation rates. In case of below average rainfall, hydropower plants have limited generation capacity and this has a direct effect on supplies. As a consequence, hydropower systems have to operate with a certain level of thermal power installed capacity to avoid the risk of power shortages during dry periods. After the reforms of the 1990s, Brazil’s electric system was built basically around two economic players: state hydraulic energy generators and private thermal energy operators. The latter would provide energy based on the reservoir conditions of hydropower plants. When the energy available in the reservoir fell to given power shortage risk levels, thermal power plants would light up and produce the additional power needed to meet demand. However, thermal power capacity did not grow in the 1990s as expected, culminating into the 2001 power shortage. The difficulties faced here bring into sharp focus the question of energy supply security in the coming years. This is the topic of the next section.

Economic growth and energy consumption Many studies have measured the GDP elasticity of electricity consumption in Brazil, indicating that manufacturing demand is elastic, while household consumption is inelastic (e.g. Schmidt and Lima 2004). However, we propose a broader view, estimating income energy consumption elasticity. There are many energy sources, and there may be substitution effects over time due to technological changes or relative price shocks. Focusing on only one source of energy (namely, electricity), provides a limited view of the GDP elasticity. We consider that an “energy hurdle” to growth exists if the energy income demand elasticity is greater than 1. In this case, energy supply must increase faster than GDP, to avoid shortages. Our model recognizes the dual role of energy in GDP. Energy acts as intermediary input and final consumption good. This requires a system of equations, where GDP level and energy use are jointly determined in a Vector Auto-­ Regressive (VAR) representation, namely: Yt = b01 + b11Yt–1 + b12Et–1 + v1t and Et = b02 + b21 Yt–1 + b22 Et–1 + v2t where Yt and Et represent log GDP (in constant prices) and log energy use (in tep, or other energy units), respectively.7 Shocks to GDP and energy use are modeled as a jointly Normal vector of possibly correlated error terms (v1t, v2t).

56   A. de Oliveira et al. The above specification does not lend itself to the identification of the long-­ run energy consumption to GDP elasticity. Instead, we consider the long-­run relationship between the variables (the co-­integration vector between the two non-­stationary variables). Under co-­integration, the above VAR model can be written as a Vector Error Correction Model (VECM): Yt = a01 + a1ut–1 + v1t   and   Et = a02 + a2ut–1 + v2t where ut–1 = Et–1 – d0 – d1Yt–1. The long-­run elasticity of energy demand is given by d1 (see, e.g. Enders 2004).8 Last, but not least, energy demand differs across sectors (manufacturing, agriculture and services). If GDP growth over time is associated with structural changes we should expect changing demand elasticity and a structural break in the estimated relationship. While this can lead to non-­cointegration, we specifically test for changes in the co-­integration vector. There is co-­integration among the variables with the long-­run co-­integration vector, estimated from the vector error correction form for a VAR(1) model with unconstrained constant.9 The long-­run equation (co-­integration relationship), is given by (standard errors in parenthesis) Ln(Energy) = 11.38 + 1.6045 Ln(GDP) (0.0706) The hypothesis of a unit elasticity is rejected (p-­value: 0.0091), suggesting that energy demand is elastic. At the same time, there is clear endogeneity between the variables, as the estimated VECM has both adjustment coefficients significant, suggesting Granger causality from GDP to Energy and from Energy to GDP (Table 5.2). Structural change in GDP may be biasing the results. In the 1970s, while manufacturing GDP share was about 33 percent, and agriculture 18 percent, by 2007, their shares fell to 18 percent and 5 percent respectively. To overcome this, we estimate two VEC models, for the 1970–1985 and 1986–2007 Table 5.2  Error correction model for energy and GDP, Brazil 1970–2007

Cointegration vector Constant R-squared F-statistic

∆Ln(Energy)

∆Ln(GDP)

0.20708* (0.03992) 0.046087* (0.00536) 0.244535 11.32909

0.18944* (0.03870) 0.039366* (0.00520) 0.533764 40.06933

Source: authors calculations based on data from BEN and ipeadata. Note * indicates significant at 1 percent level.

Energy restrictions to growth   57 Table 5.3  Structural change in energy demand elasticities

∆Ln(GDP) Log-Lik. Unit elasticity test(p-value) Sample

∆Ln(Energy)

∆Ln(Energy)

1.289572* (0.03067) 65.67094 0.0002 1970–1985

0.930318* (0.19163) 114.0948 0.8116 1986–2007

Source: authors calculations based on data from BEN and ipeadata. Note * indicates significant at 1 percent level. Constant included.

periods. The break period was chosen as 1985 is the end of the II PND development plan and the start of the near hyperinflation crises. The cointegrating vectors indicate that GDP elasticity falls over time, as expected, given the fall in the manufacturing (an energy intensive sector) GDP share. While the elasticity for the period 1970–1985 is statistically greater than 1, we cannot reject a unity elasticity hypothesis in the second period. A Likelihood Ratio structural change test yields a 26.92 statistic (1 percent and 5 percent critical values 21.02 and 26.22, respectively). The test clearly rejects the coefficients stability over time. In sum, the recent Brazilian experience indicates that energy demand was elastic, but it has become unit elastic over time, due to the change in the GDP structure. Providing a consistent, sustainable energy supply may have direct and positive effects on economic growth in Brazil. While the energy restriction steaming from an elastic energy demand seems to be weakening, supply must grow in tandem with GDP to avoid shortages. Such supply growth hinges on tackling a number of issues in the current Energy Supply Model for Brazil. These include the role of natural gas; the role of Petrobrás in pre-­salt oil exploration and production; the extent of renewable fuels use by automobile and power generation and the availability of electricity supply at reasonable cost. All of these issues will require the discussion of the current regulatory framework for the main energy sources: electricity and oil.

Regulatory changes and challenges for the twenty-­first century From 1995 to 2000, there were significant changes in the regulatory framework of energy markets in Brazil, including constitutional amendments, and privatization of state enterprises, particularly in electricity distribution. These changes altered the roles of private and public players and one may argue that the impact of regulatory changes has not been completely absorbed. Before discussing ­possible changes to the current regime, we briefly present a portrait of it.

58   A. de Oliveira et al. The 1990s reform was meant to broaden private participation in the Brazilian electricity sector and to introduce efficiency incentives, mainly through electricity generation liberalization. Following the international experience, an independent regulatory agency (Aneel) was established as well as an independent system operator system (ONS) and a wholesale energy market (MAE). The privatization process started in 1995: 23 state owned companies were sold, generating US$22 billion in revenue (Losekann and de Oliveira 2008). The privatization process moved forward rapidly in distribution, but faced many challenges in generation. Only four generation companies (three state owned and one federal owned) were privatized. Adding up new plants, the private share in the generation market is now 20 percent, while in distribution, the private share is about 60 percent. Before the transition to a competitive model was completed, Brazil faced a major crisis in electricity supply. Since the late 1990s the storage level in the hydro-­electric reservoirs has progressively diminished, reaching critical levels in 2001. In May 2001, the government imposed a 20 percent cut on electricity consumption in the sub-­systems of the Southeast/Mid-­West and Northeast, with heavy fines for non-­compliance. Rationing lasted until May 2002. Electricity consumption was drastically reduced, with dire economic consequences. The estimated social cost of the rationing was close to 3 percent of the GDP (Sauer et al. 2003). The second reform of the energy sector aimed at avoiding a new supply crisis with a concurrent rise in electricity prices. In 2004, the new regulatory framework re-­established the planning role of the state and drastically altered the wholesale market. The energy research company (EPE) was created to assist the energy minister in sector planning, playing an important role at the sector expansion auctions. It was decreed that all energy trade must be carried out by long-­ term contracts. Two trade environments were created in the wholesale market: regulated contracting (ACR) and free contracting (ACL). At the ACR market, distribution companies can buy energy in public auctions. They submit demand projections in a five-­year horizon to EPE. Based on those projections, EPE sets the total market supply that will be offered in the auctions. The model distinguishes the energy generated by existing plants (“old energy”) from the energy coming onstream from the new ones (“new energy”); both being negotiated in the ACR in different ways. In 2004 and 2005 four auctions dealt with the major part of the “old energy” supply. As they took place at a moment of excess of supply (due to consumption decrease after the rationing), the resulting energy prices were around half of the long-­run marginal cost (Losekann and de Oliveira 2007). Since December 2005, nine energy auctions have been carried out to set the expansion of the generating system. Hydropower and thermal plants were treated differently. Whereas the hydropower plants compete with prices for the generated energy, the thermoelectric plants bid for the generating capacity. Auction winning thermoelectric plant operational costs are passed on to the final consumers. The lower prices attaching to “old energy” allow for a smooth transition to higher energy prices as new energy increases its share on ACR total sales.

Energy restrictions to growth   59 Nevertheless, any unanticipated need to run thermoelectric plants could alter this trend and may cause a sudden rise in energy prices. The second electricity reform did not address the main challenge of the Brazilian electricity system: coordination between the power and the natural gas sectors (Losekann and de Oliveira 2008). The dispatch rules of the power sector propose very low average operating rates for the thermal plants, but ask these plants to be fully dispatched when rain precipitation is short. This arrangement is financially costly for the infant natural gas industry. Gas that should be available for thermal power plants is often diverted to other consumers instead. Indeed, natural gas fuelled power plants are not participating in energy auctions, with oil fuelled plants seizing the opportunity and participating in the energy auctions. Thus, electricity regulation decisions are influencing the oil and gas markets and their regulatory structures. As mentioned in the introduction, by the mid-­1990s there was a belief that state enterprises, including Petrobrás, did not have the financial stature to lead the development of Brazil’s energy sector. The political mood for less state intervention induced regulatory reform across the oil sector. A constitutional amendment (EC no. 9, 9 November 1995) allowed a New Oil Law (9478/97) that regulated Article 177 of the Constitution. This constitutional amendment did not end the state monopoly over oil resources but allowed the state to franchise operations at any stage of the oil industry, under a variety of contract conditions.10 A regulatory entity (ANP) was created to oversee and regulate the oil and natural gas activities. The ANP is responsible for the concession auctions for oil and natural gas exploration rights.11 The ANP has so far carried out ten auctions (called “rounds”) with varying degrees of success. Table 5.4 presents a summary of the rounds. The sharp decrease in the auction success ratio (conceded/auctioned blocks) may be attributed to two factors: first, a change in the “block” definition, from a rectangular, large area form (about 200 × 800 km2) to an “exploratory cell” definition (used in the Gulf of Mexico) and, second; a large portion of land blocks have been auctioned lately, that are known to have less oil and lower success ratios than deep-­sea blocks. On the eve of the ninth round of oil exploration auctions in November 2007 the ANP withdrew from sale 41 blocks bordering the Tupi oil field. This was the field where pre-­salt layer oil was found, in extremely deep waters, but with enormous potential output of high quality oil (e.g. Petrobrás 2007). The official argument for the change in the auction format was that the new discoveries had opened up a new area for production, with little risk involved, as well as significant risk of different blocks sharing the same oil field. To avoid the “tragedy of the commons” of overexploitation of any block by aggressive companies, unitization of concessions was proposed as a solution. While many countries have experience with this sort of solution, little existed among the Brazilian authorities.12 Nevertheless, the potentially enormous pre-­salt reserves prompted federal government officials to question a concession framework that gives away the oil property, once extracted. In doing so they echoed nationalistic arguments. While

23 59,271 2,577 48,074 21 91.3

2000

1999

27 132,178 4,895 54,660 12 44.4

Round 2

Round 1

53 89,823 1,695 48,629 34 64.2

2001

Round 3

Source: Alvite et al. (2008) based on ANP (2008). Updated by the authors.

Auctioned blocks Auctioned area (km²) Avg. block size (km²) Conceded area Conceded blocks Conceded/auctioned    blocks (%)

Auction rounds

Table 5.4  Oil and gas auction rounds summary

54 144,106 2,669 25,289 21 38.9

2002

Round 4

908 162,392 179 21,951 101 11.1

2003

Round 5

913 202,739 222 39,657 154 16.9

2004

Round 6

1,134 397,600 351 171,007 240 21.2

2005

Round 7

271 73,079 270 45,329 108 39.9

2007

Round 9

130 70,371 541 48,154 54 41.5

2008

Round 10

Energy restrictions to growth   61 “previous contracts will not be broken” according to the Energy Minister E. Lobão (Zacconi 2008), a new regulatory framework is being drafted. As the new arrangements had not been set up by that stage, the tenth round, in 2008, included only onshore blocks. As of August 2009, the new regulatory regime is still unknown. While pre-­salt reserves are indeed significant, and may triple the Brazilian oil and natural gas reserves, their cost and technical difficulties are of equal stature. In addition, oil price uncertainty renders economic exploration a real challenge following the 2008 market slump. Regulatory uncertainty may push development of the pre salt fields many years into the future. Under current rules, Petrobrás estimates that the Tupi field will be operational only after 2012 or 2015. It must be noted that the Tupi field was auctioned in the second round, back in 1999. Although future time lags should decrease there was almost a decade of exploratory effort to reach the pre-­salt oil, and the financial resources required to actually exploit its reserves are substantial. Should the exploratory regime change, there is a good chance that the exploitation of pre-­ salt will be postponed.

Concluding remarks In the past 50 years Brazil has experienced radical changes in its energy sector. The energy matrix has changed radically with commercial energy sources, such as electricity and oil products, replacing firewood and charcoal. New energy sources have been introduced, namely, natural gas and ethanol. External dependency has been sharply reduced with the discovery of large oil fields in the continental shelf. Renewable resources have come to be the main source of energy, whether in the form of sugar cane-­derived ethanol or hydraulic electric power. These changes reflect the effort to ease energy restrictions to growth in the Brazilian economy. Such restrictions were acute in the twentieth century given the high income elasticity of demand for energy consumption. Energy needs up to the 1980s were growing faster than GDP, in the wake of the development process and associated industrialization and urbanization. Currently, these energy restrictions have eased to the extent that the income elasticity of energy consumption has lowered and is estimated to be either proportional to income or even inelastic. This places a lesser burden on the energy supply growth rate for the coming years. Nevertheless, the 2001 power rationing made clear that energy supply must be secure to play a positive role in the Brazilian economic growth. Electricity supply is still quite reliant on hydrology, with price shocks for consumers emerging from the almost continuous use of thermal power plants during dry periods. Recent international political changes have limited the supply expansion of natural gas. Indeed, supply restrictions for gas were reported in October 2007. These facts raise the issue of the current state of the regulatory regime for the energy sector. This chapter has argued that the regulatory regime has changed significantly over the past ten years with, arguably, varying degrees of success. The reforms,

62   A. de Oliveira et al. reviewed in the chapter, focused on two main energy sources: electricity and oil. The electricity sector still struggles with hydrology risks and it has not seen investment recover to previous levels. The oil sector appears to be on the cusp of a radically different scenario, both technologically and economically, given the “pre-­salt” reserves located off the coasts of Rio de Janeiro and São Paulo. Brazil is currently in a unique situation. On one hand it has large energy resources, both renewable and non-­renewable. Its hydraulic resources have not been fully exploited. Its ethanol sector is the most efficient in the world and has been producing with some excess capacity. Its oil reserves have risen significantly and should at least double in the next few years. Indeed, energy could be a trigger for growth, as Brazil starts to export energy (oil and ethanol) to other countries. On the other hand, resource availability itself does not guarantee energy supply. This requires a consistent regulatory framework for the whole energy sector. Whether this implies a renewed state participation as an energy producer and distributor, much like the 1960s and the 1970s, is an open issue. Past regulatory regimes must be evaluated, taking into account current and future resource availability, environmental and economic constraints and the need to develop an integrated energy system. In conclusion, Brazil is in a privileged position to engage in a transition from fossil fuels to renewable sources of energy. This represents a major comparative advantage in a world that has to cope with climate change. However, it is important to bear in mind that institutional and regulatory arrangements are critical components in the development of an energy system. If these arrangements are unable to attract investment, energy resources will remain locked in the ground. This would leave Brazil to cope with the specter of energy shortages whenever economic growth gained momentum.

Notes   1 In 1986, about 95 percent of all automobiles sold were ethanol powered (Anfavea 2008).   2 Besides the nuclear power plants and the Brazilian share of Itaipu, Eletrobrás had control of the four regional generation and transmission companies (Eletronorte, Eletrosul, Furnas and CHESF ), and minority shares in every other power company.   3 In fact, the cornerstone for the development of a modern energy system in Brazil was laid in the 1930s by the federal government. The Water Code and the Mining Code assigned to the federal government property rights of energy resources (Dias Leite 1997). However, by the end of World War II, amid a wave of nationalizations that discouraged foreign investors, Brazil was facing large difficulties in supplying a rapid increase in energy demand.   4 For a detailed description of the state enterprises in the 1980s, see Baer (2007).   5 Emerging manufacturing in Brazil used wood for energy generating, since the early stages of industrialization (Branstrom 2005).   6 It may be the case that the sector use shares are being unduly influenced by energy use imputation, particularly for firewood. This has been pointed out as a limitation to BEN statistics (De Oilveira and Gutierrez 1998).   7 To save space we present a VAR(1). The actual lag used is specified using usual information criteria.

Energy restrictions to growth   63   8 Variables non-­stationarity was tested using standard Dickey-­Fuller tests, available upon request. They are not presented to save space.   9 The variables used are GDP in constant prices, obtained from IPEA data, and energy production in Brazil, obtained from the Ministry of Energy’s energy balance (Balanço Energético Nacional BEN 2008). Energy production is measured in tep (oil equivalent tons), to allow aggregation of different energy types. We consider only commercial energy sources (oil derived, natural gas, coal, alcohol, electricity), as these are less prone to imputation errors in BEN and are the relevant energy supply sources for policy. Our sample covers from 1970 to 2007. 10 In the 1970s other firms could explore and produce oil according to the “contraltos de risco” (risk service contracts), but with limited success and interest from other firms (Campos 2005). 11 Other types of state-­firm contracts, with special emphasis for Latin America can be reviewed in Palacios (2002). 12 Actually, there is only one unitization agreement signed as of the writing of this chapter. This is between Aurizônia Petróleos and Petrobrás, on output individualization of the Lorena Field (BT-­POT-10) in Rio Grande do Norte. It is known that there are other unitization issues under analysis at the ANP. These include Petrobrás and Total at block BC-­2; Petrobrás and Shell at block BC-­10, and El Paso and Queiroz Galvão at BM-­CAL-4 (Alvite 2008 apud Schüffner 2008).

References Alvite, F. (2008) Análise Crítica das Ofertas das Rodadas de Licitações da ANP. Mimeo, Escola Politecnica, UFRJ. ANFAVEA (2008) Tabelas Estatísticas, São Paulo: ANFAVEA Araújo, J.L.R.H., Oliveira, A. de and Melo, P. (1994) 1954–1994: “Eletricidade No Segundo Governo Vargas e A Crise dos Anos 90”, in Angela, C.G. (ed.). Vargas E A Crise Dos Anos 50, Rio de Janeiro: Relume Dumara. Baer, W. (2008) Brazilian Economy: Growth and Development, sixth edition, Boulder: Lynne Rienner Publishers. BEN (2008) Balanço Energético Nacional, Brasília: EPE/MME. Brannstrom, C. (2005) “Was Brazilian Industrialization Fuelled by Wood? Evaluating the Wood Hypothesis, 1900–1960”, Environment and History, 11 (4): 395–430. Campos, A. (2005) “Transformações recentes no setor petrolífero brasileiro”, Perspectiva Econômica Online v34. Online, available at: www.perspectivaeconomica.unisinos.br/ pdfs/34.pdf. Dias Leite, A. (2005) A energia no Brasil, São Paulo: Elsevier. Enders, W. (2004) Econometric Analysis of Time Series, New York: Willey. Losekann, L. and de Oliveira, A. (2007) Technology Mix in the Brazilian Electricity Sector, 27th USAEE/IAEE North American Conference Proceedings. Losekann, L. and A. de Oliveira (2008) “Supply Security in the Brazilian Electricity Sector”, IAEE Energy Forum, Third Quarter: 21–24. Oliveira, A. de (2007) “The Political Economy of the Brazilian Power Industry”, in Heller, T. and Victor, D. Economy of the Power Sector Reform, Cambridge: Cambridge University Press. Palacios, L. (2002) “The Petroleum Sector in Latin America: Reforming the Crown Jewels”, Les Etudes du CERI, no. 88. Petrobrás (2007) “Another Oil-­bearing Well Discovered in the Santos Basin’s Pre-­Salt Layer”. Online, available at: www2.petrobras.com.br/ri/spic/bco_arq/DescobertaBM-­ S-21Ing.pdf.

64   A. de Oliveira et al. Sandalow, D. (2006) “Ethanol? Lessons from Brazil”, in Aspen Institute: A High Growth Strategy for Ethanol. Online, available at: www.aspeninstitute.org/eee/ethanol. Sauer, I., Rosa, L. and D’Áraujo, R. (2003) A Reconstituição do Setor Elétrico Brasileiro, São Paulo: Paz e Terra. Schmidt, C. and Lima, M. (2004) “A demanda por energia elétrica no Brasil”, Revista Brasileira de Economia, 58 (1). Schüffner, C. (2008) “Empresas com áreas no pré-sal devem negociar projetos”, Valor Econômico, 12 June. United Nations (2007) World Urbanization Prospects: The 2007 Revision, New York: United Nations. Vazquez, F., Silva, M.E. and Bone, R.B. (2008) “A regulação no processo de unitização na exploração de petróleo e gás natural no Brasil”, Rio Oil and Gas Expo and Conference 2008, IBP 2579_08. Weidenmier, M.D., Davis, J.H. and Aliaga-­Diaz, R. (2008) “Is Sugar Sweeter at the Pump? The Macroeconomic Impact of Brazil’s Alternative Energy Program”, NBER Working Paper, no. W14362. Zacconi, C. (2008) “Who Ends Up with Pre-­salt?”, Energia hoje. Online, available at: www.energiahoje.com/pops/materia.php?id=358761.

6 Oil prices and inflation in Brazil Exchange rate versus inflation targeting Claudio A.C. Paiva

Introduction This chapter examines the relationship between oil prices and inflation in Brazil during 1994–2008. The period is particularly rich from a research standpoint because it spanned two very distinct periods of monetary policy strategy: the exchange rate targeting (ERT) carried out in 1994–1998 and the inflation targeting framework (IT) adopted in 1999. Therefore, besides providing an estimate of the overall impact of oil prices on inflation in Brazil, this chapter also investigates whether this impact has changed after the adoption of IT. In addition, interesting byproducts of the empirical work include estimates of the degree of persistence of inflationary shocks and the impact of economic activity, the exchange rate, and the interest rate on inflation under the alternative monetary regimes. The empirical investigation relies mainly on estimates of VARs and their associated impulse response functions. Different model specifications generally include a measure of consumer prices, domestic wholesale fuel prices, the nominal exchange rate between the Brazilian real (BRL) and the US dollar (USD), a measure of economic activity, and a measure of interest rates. In addition, single equations provide estimates of the pass-­through from international oil prices and the exchange rate to domestic fuel prices. The data set comprised quarterly data spanning the period 1994:3–2008:2. Empirical estimates suggest that although the pass-­through from the cost of oil to domestic fuel prices has increased, the impact of fuel prices on inflation in Brazil has declined under IT. This result is in line with international evidence that an increase in the forward looking component of inflation expectations and in the credibility of monetary policy has reduced the inflationary impact of higher fuel prices in recent years.1 Anecdotal evidence of this change in price dynamics includes the relatively small response of inflation in Brazil and in industrial countries following the surge in oil prices (and other commodities) since 2003. This recent experience contrasts with the widespread acceleration of inflation following the oil shocks of the 1970s.

66   C.A.C. Paiva

Preliminary data analysis Wholesale fuel prices have largely accompanied the increase in the cost of oil measured in BRL, being adjusted to broadly reflect higher international oil prices and/or a weaker exchange rate (Figure 6.1). However, fuel prices have shown some downward stickiness: the main deviations between the two series occurred when wholesale fuel prices did not match temporary declines in oil prices expressed in domestic currency (Figure 6.2). First, during the ERT period, reflecting falling international oil prices; and later, during the IT period, reflecting periods of significant BRL appreciation, perhaps considered by the authorities as temporary. The ratio of wholesale fuel prices to the cost of oil expressed in domestic currency has remained relatively stable in the last eight years (Figure 6.3). This tendency contrasts with the wide fluctuation range observed during the ERT period and suggests that the degree of pass-­through from oil costs to domestic fuel prices has increased in recent years. This is particularly interesting given the escalation of international oil prices and greater exchange rate volatility observed in the period. Simple measures of correlation also suggest that the pass-­through from the cost of oil to wholesale fuel prices has increased since the floating of the exchange rate and implementation of IT (Table 6.1). The correlation coefficient for the two series in levels increased from 0.20 during 1994–1998 to about 0.95 after 1999. The correlation coefficient between the quarterly percentage changes

180 160 140 Wholesale fuel prices

120 100 80

Oil prices in BRL

60 40 20

2008.2

2007.2

2006.2

2005.2

2004.2

2003.2

2002.2

2001.2

2000.2

1999.2

1998.2

1997.2

1996.2

1995.2

0

Figure 6.1 Wholesale fuel prices and oil prices (levels) (source: data drawn from Banco Central do Brasil; Conjuntura Econômica).

35 Oil prices in BRL (moving average, quarterly percentage change)

30 25

Wholesale fuel prices (moving average, quarterly percentage change)

20 15 10 5 0 �5

2007.4

2006.4

2005.4

2004.4

2003.4

2002.4

2001.4

2000.4

1999.4

1998.4

1997.4

1996.4

�15

1995.4

�10

Figure 6.2 Wholesale fuel prices and oil prices (percentage changes) (source: Data drawn from Banco Central do Brasil; Conjuntura Econômica).

160 140 120 100 80 60 40 20

2008.2

2007.2

2006.2

2005.2

2004.2

2003.2

2002.2

2001.2

2000.2

1999.2

1998.2

1997.2

1996.2

1995.2

0

Figure 6.3 Ratio of wholesale fuel prices to oil prices (moving average) (source: author’s calculations).

68   C.A.C. Paiva Table 6.1  Wholesale fuel prices and oil prices – correlation Coefficient of correlation

Series in levels Percentage changes Moving average percentage changes

ERT

IT

0.20 −0.10 0.30

0.95 0.50 0.80

Source: Author’s calculations.

in the domestic price of oil and wholesale fuel prices went from −0.10 to 0.50 in the same period comparison. Using four-­quarter moving averages of the series as a way to smooth out short-­term price fluctuations does not alter the basic result: the correlation coefficient between changes in oil costs and fuel prices increases from 0.38 to 0.79 in the more recent periods. Although domestic fuel prices have increased sharply during the IT period, more closely reflecting oil costs, the impact on consumer price inflation has been limited (Figure 6.4). In fact, since the abandonment of the exchange rate peg in 1999, wholesale fuel prices have risen nearly 500 percent, whereas the accumulated inflation measured by the IPCA (the official target of the Central Bank) was about 90 percent. Episodes of particularly rapid fuel price increases seem to have had only a small impact on ensuing inflation, perhaps reflecting the credi-

20.0 16.0

Wholesale fuel prices (moving average, percentage change)

12.0 8.0

IPCA (moving average, percentage change)

4.0 0.0

2008.1

2007.1

2006.1

2005.1

2004.1

2003.1

2002.1

2001.1

2000.1

1999.1

1998.1

1997.1

1996.1

1995.1

�4.0

Figure 6.4 Wholesale fuel price increases and IPCA inflation (source: data drawn from Conjuntura Econômica and Banco Central do Brasil).

Oil prices and inflation in Brazil   69 18.0 15.0

Wholesale fuel prices (moving average, percentage change)

12.0 9.0 6.0 3.0 0.0 Core inflation (moving average)

�3.0

2008.2

2007.2

2006.2

2005.2

2004.2

2003.2

2002.2

2001.2

2000.2

1999.2

1998.2

1997.2

1996.2

1995.2

�6.0

Figure 6.5 Wholesale fuel price increases and core inflation (source: author’s calculations).

bility of the IT regime. In fact, the stronger response of inflation in early 2003 (to fuel prices and, more generally, to the earlier depreciation of the BRL) can be associated with speculation regarding the commitment of the new government to the IT framework. The behavior of core inflation – excluding administered prices – supports similar conclusions (Figure 6.5).

Empirical modeling and conclusions The data set comprises quarterly data for the period 1994:3–2008:2. It includes the following series: the consumer price index (IPCA); an index of wholesale fuel prices (IPAFUEL); the exchange rate expressed in BRL per USD (ER); international oil prices expressed in USD per barrel (POIL); the policy interest rate (SELIC); the real interest rate (RSELIC); the unemployment rate (UNEMP); the output gap (GAP); and a measure of core inflation that excludes energy and other administered prices (CORE). The VARs that support the main findings of this chapter were estimated using stationary series.2 When necessary, the original series were transformed to achieve stationarity. Augmented Dickey–­Fuller tests for unit roots and a 5 percent significance level underpinned the process. Lag lengths of the VARs were chosen to minimize the Schwarz information criteria. The preferred VAR specification had one lag and included the following variables: IPCA, IPAFUEL, ER, UNEMP, and RSELIC. The stationary unemployment variable was obtained by de-­trending the original series using an HP filter.

70   C.A.C. Paiva Impulse response functions and the associated pass-­through coefficients are used to assess the impact of fuel prices on inflation. Two alternative impulse response functions were estimated using Choleski decomposition (CHO) and generalized impulses (GI).3 Cumulative pass-­through coefficients were then calculated as the ratio of the cumulative response of inflation to the cumulative response of fuel prices after t+i quarters when there is a shock to fuel prices at t. The estimations suggest that a 10 percent increase in fuel prices raises inflation by almost 1.5 percentage point after one year (Table 6.2). This finding is broadly in line with Le Blanc and Chinn (2004) who estimate the impact of a similar increase in oil prices on inflation after a year to be between 0.1 percentage point and 2.5 percentage points for a group of industrialized countries over the period 1980:1–2001:4.4 The impact of fuel prices on inflation seems to have declined after the introduction of inflation targeting. The one-­year inflationary impact of a 10 percent increase in fuel prices is estimated at 1 percentage point when the sample is restricted to the period 1999:3–2008:2. Since the share of fuel prices in the consumer price index remained the same and the energy intensity of the economy is likely to have remained broadly the same, the lower inflationary impact of fuel prices may be attributed greater monetary policy credibility and lower persistence of inflationary shocks in general. Estimates of the impact of unemployment, the exchange rate, and the policy interest rate on inflation also suggest greater monetary policy credibility under IT. Impulse response functions and the associated cumulative pass-­through coefficients show that the impact of economic activity and of the exchange rate on inflation have declined whereas the impact of the policy interest rate on inflation has increased under IT (Table 6.3). This is consistent with the interpretation that the policy interest rate has acted more through the expectations channel than through economic activity to affect inflation and that, as expected, the pass-­ through from the exchange rate to inflation has declined after the abandonment of ERT. Table 6.2  Cumulative passthrough of a 10-percent increase in fuel prices to inflation Quarter

1 2 3 4 5 6 7 8

Full sample

Inflation targeting period

Choleski

GI

Choleski

GI

1.0 1.1 1.2 1.3 1.3 1.3 1.2 1.2

0.9 1.1 1.3 1.4 1.5 1.6 1.7 1.7

0.7 0.8 0.9 1.0 1.0 0.9 0.9 0.9

0.7 0.8 0.9 1.0 0.9 0.9 0.9 0.0

Source: Author’s calculations.

Oil prices and inflation in Brazil   71 Table 6.3 Cumulative impact of other variables on inflation (impact of a 1-percentage point increase after one year)

Full sample IT period

ER

UNEMP

RSELIC

0.21 0.12

−0.39 0.04

−0.04 −0.24

Source: Author’s calculations.

Table 6.4 Inflationary impact of a 10-percent increase in fuel prices under alternative VAR specifications Full sample

VAR with original unemployment VAR with output gap

IT period

Choleski

GI

Choleski

GI

1.2 1.1

1.5 1.5

0.6 1.0

0.6 1.0

Source: Author’s calculations.

Alternative VAR specifications corroborate the main finding of a lower pass-­ through from fuel prices to inflation. When the original unemployment series is included rather than its de-­trended series, the one-­year inflationary impact of a 10 percent increase in fuel prices is estimated at 1.2 percentage point (CHO) and 1.5 (GI) for the full sample and only 0.6 for the IT period (Table 6.4). When the output gap is used as the measure of economic activity, the one-­year inflationary impact of the same 10 percent fuel price shock is estimated at about 1.5 percentage point for the full sample and 1 percentage point for the IT period. Econometric estimates confirm that the pass-­through from oil costs to domestic fuel prices has increased after 1999. Simple OLS estimates show that the long-­term pass-­through coefficient from the cost of oil in BRL to fuel prices increased by 20 percent to 0.6 in the recent period. This makes it all the more important that IT has succeeded in reducing the persistence of inflationary shocks – including that of rising fuel prices.

Notes 1 Schmidt-­Hebbel and Mishkin (2006), “Does Inflation Targeting Make a Difference?”, presented at the eighth Annual Meeting of the Brazilian Central Bank. 2 The only exception is an alternative VAR model that includes the unemployment rate (I1) rather than the de-­trended series (I0). 3 The Choleski ordering was the following: measure of activity (UNEMP). 4 “Do High Oil Prices Presage Inflation? The Evidence from G-­5 Countries.” Le Blanc and Chinn use a different approach than mine, assuming inflation to be the only endogenous variable and thus estimating a single equation (for each country) through OLS.

7 Brazilian energy independence Petroleum, trade and economic efficiency Donald V. Coes

Introduction The big rise in international petroleum prices that began in October 1973, following the Arab–Israeli War, was the first of several negative “supply shocks” that hit the economies of major developing nations like Brazil. In the mainstream economic interpretation of this kind of event, such a supply shock lowers the level of real economic activity or slows its rate of growth and raises both the price level and the rate of inflation. The first shock, and the subsequent one of 1979–1980, clearly had such effects in Brazil. Real GDP growth rates, which had averaged more than 11 percent between 1968 and 1973, fell to a little more than 6 percent for the rest of the decade. Rates of inflation more than doubled after the first shock, and then more than doubled again after the second shock. In retrospect, while Brazil’s long period of macroeconomic instability in the two decades after 1974 may have been primarily home grown, it was clearly aggravated by these external oil shocks. It is therefore quite understandable that Brazil’s macroeconomic travails in the last quarter of the twentieth century continued to influence policy-­making in both the Cardoso and the Lula governments. Although the focus of the 1994 Plano Real was on domestic stabilization through a credible fiscal policy and monetary reform, the reduction in vulnerability to external shocks is an important element in making Brazil’s macroeconomic environment less volatile than it once was. Petroleum independence is only one part of a larger group of issues that might be grouped under the rubric of “energy independence”. Among the other sources of energy that are of potential importance in a Brazilian context are hydroelectric and nuclear power, as well as less technologically mature sources like solar energy, wind power, and renewable biomass sources such as alcohol. All of these potential saviors have attracted much attention, both in the Brazilian and foreign press. None of them, however, have had an impact on Brazil’s external payments comparable to that of petroleum. There are two basic reasons for this. First, several of these potential energy alternatives to petroleum, such as solar energy or nuclear power are basically “non-­tradable” activities. As such, their links to Brazil’s external payments are primarily indirect, through their potential

Brazilian energy independence   73 substitution for a highly tradable good like petroleum. Second, despite increasing interest in these alternatives and substantial technological progress in Brazil in developing them, their macroeconomic significance for Brazil has been much less than have been trends in the international petroleum market. For this reason, the focus of this chapter is on Brazilian economic independence in petroleum, with some attention given to the potential substitution effects of alternative energy sources. It begins with an examination of the major trends in petroleum production, consumption, and external dependence in three phases: (a) prior to the first oil shock, (b) from 1973 to the Plano Real and the liberalization of the petroleum market in the 1990s, and (c) since the 1994 reforms. It shows that most of the reduction in trade vulnerability came from the production side, rather than from a slowing of the growth in energy consumption. The section following then presents a simple model of foreign energy dependence, which distinguishes between short and longer run outcomes. It suggests that economic efficiency in energy use may be constrained both by past energy investment policies and by future trends in energy costs and technologies. The chapter concludes with a brief examination of some of the choices that may be faced in developing an economically efficient policy.

Patterns of Brazilian external energy dependence Energy dependence before 1973 Although there was interest in oil exploration and development in Brazil as early as the nineteenth century, government policies in the area really date from the first administration (1930–1945) of Getúlio Vargas. In 1933 the Departamento Nacional de Produção Mineral (DNPM) was established, in part as a reaction to political pressures for a national policy response to alleged threats to Brazil’s control of its mineral resources.1 In 1938 the Vargas government created the Conselho Nacional de Petróleo and created the legal basis for public control over oil refining and exploration. In the following year petroleum was discovered in Bahia, and in 1941 a commercially viable well was inaugurated in the Bahian municipality of Candeias. World War II had several important impacts on Brazilian energy policy. Access to imported petroleum supplies was sharply reduced, with rationing imposed and import quantities falling to less than half the level they had been even during the Depression era of the 1930s.2 The war highlighted the strategic importance of petroleum, a lesson taken very seriously by Brazil’s military. Brazilian participation in the war on the side of the democracies also increased the political pressures for a more open political system, helping lead to the demise of the authoritarian Vargas government in 1945. A nationalist petroleum policy was a key element in Vargas’ democratic return to power in 1950. Support came both from center-­left political parties and conservative nationalists, linked in a coalition skillfully managed by Vargas, under the slogan “O petróleo é nosso”. Law 2004 of October 1953 created the

74   D.V. Coes state enterprise Petróleo Brasileiro (Petrobrás), which was to become the central player in Brazilian energy policy. Although the new government company had a quasi-­monopoly in the exploration, extraction, refining, transport, and distribution of petroleum and its derivatives, private companies were permitted to operate in these areas. The establishment of Petrobrás nevertheless was a watershed in Brazilian economic and energy policy, and the underlying premise that petroleum production is of strategic importance has survived to today. The rapid growth of automobile production and extension of the road system in the 1950s led to a large increase in petroleum and petroleum derivatives demand. This market was almost completely supplied by imports, with domestic petroleum production in the 1950s accounting for less than 5 percent of consumption.3 The quantity of petroleum imports tripled between 1954 and 1964, before leveling off briefly following the 1964 stabilization program under the new military regime. Imports then rose five-­fold over the late 1960s and the 1970s, peaking in 1979.4 In its early years, much of the public support for Petrobrás’ mission was based on the widespread belief that there were major undiscovered petroleum reserves under the Brazilian landmass, but that earlier exploration had been discouraged by opposition from the major international oil companies. Petrobrás hired Walter L. Link, former chief exploration geologist of Standard Oil of New Jersey to head a team of geologists to assess Brazil’s petroleum potential. In 1960 the “Link Report” presented a fairly pessimistic estimate of the possibilities for major petroleum discoveries under Brazilian subsoil. The report appears to have been based fairly objectively on the geological information that was available at that time. But from today’s vantage point it vastly underestimated reserves, since it did not consider the off-­shore possibilities that were later to position Brazil for self-­sufficiency in petroleum. The report provoked an outcry among Brazilian nationalists, who certainly did not view it as objective, but as a deliberate attempt to limit Brazilian petroleum development.5 Despite the hopes that Brazil’s land area would contain large reserves, the course of petroleum production and development in the 1960s and 1970s appeared to justify the pessimism of the petroleum professionals. With the resumption of high real GDP growth in 1967, apparent consumption of petroleum steadily increased, while domestic production was virtually flat after 1969. As a result, Brazil depended on imports for nearly 80 percent of its consumption on the eve of the first petroleum shock in 1973. These trends are apparent in Figure 7.1, which shows production and apparent consumption from 1953, the year of Petrobrás’ founding, through 1980. The relatively flat trend in production was primarily due to the geographical focus of Petrobrás in its first two decades. Efforts to find petroleum in the Amazon basin were disappointing, and by the early 1960s, Brazilian production was heavily dependent on a few fields in Bahia. In 1967 the exploration focus began to change, as the proximity of the few economically viable fields to the coast suggested that off-­shore prospection and development might have more promise. New seismic techniques were applied, and in 1968 the first two off-­

Brazilian energy independence   75

Millions of cubic meters

70 60 50 40 30 20 10

Consumption

1979

1977

1975

1973

1971

1969

1967

1965

1963

1961

1959

1957

1955

1953

0

Production

Figure 7.1 Petroleum production and apparent consumption, 1953–1980 (sources: 1953–1969: IBGE, Estatísticas Históricas; 1970–1980, Ministério de Minas e Energia – Empresa de Pesquisa Energética, Balanço Energético Nacional 2007. Online, available at: www.ipeadata.gov.br (accessed 15 October 2008)).

shore wells were drilled on the continental shelf off the states of Sergipe and Espírito Santo. The oil shocks and the Brazilian response The relative stagnation of Brazilian domestic petroleum production that began in the late 1960s was macroeconomically manageable until the first oil shock in late 1973. Within a few months, however, it was apparent that if the growth of domestic petroleum consumption was not significantly reduced, there would be a sharp deterioration of Brazil’s current account. Policymakers were faced with a stark choice between borrowing to finance the trade deficit or accepting a devastating fall in real GDP growth. With the political pressures to maintain growth as the military government contemplated a return to a more open political system, the option for borrowing was obvious. This choice was facilitated, moreover, by several contemporaneous developments in international capital markets. Among them was the recent development of syndicated loans made by a consortium of banks to sovereign borrowers. The de facto indexing of these loans to world interest rates by making them short-­term, with nearly routine renewal at a new nominal interest rate, also facilitated the large increase in capital flows to borrowers like Brazil. The availability of “petrodollars” from the surpluses of the oil-­exporting countries made a huge increase in this form of international financing possible. The impact of the 1973–1974 oil price increase is apparent in Figure 7.2, which shows the steep rise in Brazil’s imports after the two oil shocks. Even during the rapid increase in GDP growth that began in 1967, Brazil had managed

76   D.V. Coes

Billions of US dollars

30 25 20 15 10 5

Total exports

Total imports

1987

1986

1985

1984

1983

1982

1981

1980

1979

1978

1977

1976

1975

1974

1973

1972

1971

1970

1969

1968

0

Energy imports

Figure 7.2 Exports and imports (FOB), 1968–1988 (sources: Banco Central do Brasil – time series online, available at: www.bancocentral.gov.br (accessed 25 October 2008); petroleum and coal imports from Ipeadata, online, available at: www.ipeadata.gov.br (accessed 25 October 2008). The primary source for the latter series, Fundação Centro de Estudos do Comércio Exterior (FUNCEX), with the series beginning only in 1974).

to maintain its trade balance in equilibrium over the six-­year period through 1973.6 The first oil shock destroyed this balance, with the value of total imports more than doubling between 1973 and 1974. Export growth after 1974 helped reduce the trade deficit in succeeding years, resulting in a small trade surplus in 1977. In 1979, in the wake of the Iranian revolution, Brazil was hit by the second oil shock. By year’s end, prices per barrel in international markets were nearly triple their level of a year earlier, and the value of Brazil’s imports rose steeply.7 Once again, Brazilian policy makers were faced with a choice between economic stagnation and increasing external borrowing to finance an expanding current account deficit. One of the key features of Petrobrás’ near total domination of the Brazilian petroleum market was that the domestic price of petroleum derivatives was an administered price, rather than one set by market participants within Brazil. Increases in international prices like those of the 1973 and 1979 oil shocks were therefore not automatically passed on to the domestic market until Petrobrás adjusted the local currency prices. The time lag between the increase in international prices and similar local increases meant that the state enterprise was in effect partially subsidizing domestic consumption until local prices were adjusted. This feature of the Brazilian petroleum market in the post-­1973 period can be seen quite clearly in Figure 7.3, which shows the ratio of international prices, converted into local prices by the nominal exchange rate, to an index of comparable local prices.8 If the pass-­through had been market-­based and approximately automatic, the two price series would be expected to track each other quite

Brazilian energy independence   77 6.00 5.00 4.00 3.00 2.00 1.00

Jul. 69 Feb. 71 Sept. 72 Apr. 74 Nov. 75 Jun. 77 Jan. 79 Aug. 80 Mar. 82 Oct. 83 May 85 Dec. 86 Jul. 88 Feb. 90 Sep. 91 Apr. 93 Nov. 94 Jun. 96 Jan. 98 Aug. 99 Mar. 01 Oct. 02 May 04 Dec. 05 Jul. 07

0

Figure 7.3 Ratio of international petroleum price to domestic Brazilian price (August 1994 = 1.00) (source: author’s calculation from indices of domestic price index of petroleum derivatives, nominal R$/US$ exchange rate, and international petroleum price per barrel. Primary sources: Fundação Getúlio Vargas, IGP-OG-combustiveis e lubricantes, Banco Central do Brasil, R$/US$ exchange rate, and IMF, International Financial Statistics, average monthly price/barrel, from Ipeadata, online available at: ipeadata.gov.br (accesed 19 October 2008)).

closely. The domestic price of petroleum derivatives, however, only slowly adjusted to the rise in international prices after both oil shocks. The divergence between international and local prices was in part due to the predominance of an administered price in local markets. It should be noted, however, that high levels of domestic inflation before 1994 also increased the likelihood of divergences between international and local prices, even when the domestic prices were largely market-­determined. Another key feature of Petrobrás’ dominant position in the Brazilian petroleum sector was its relative freedom to invest in exploration and development with little market pressure on it to deliver short-­term results. In view of the rather discouraging results of in-­shore oil exploration in Brazil in the first two decades of Petrobrás’ existence, it is questionable if the management of private corporations would have had such patience or whether private capital markets would have permitted such investment. In the mid-­1970s Petrobrás’ shift in focus toward offshore exploration and decades-­long development of a domestic petroleum technology began to pay off. The first significant off-­shore discovery was the Garoupa oil field in the Campos Basin, on the coast of northern Rio de Janeiro state, in late 1974. This success strengthened the argument for greater off-­shore exploration and accelerated the efforts to apply new seismic techniques to identify new fields.9

78   D.V. Coes In 1975 a large new field, the Namorado field, also in the Campos Basin, was discovered, and exploration extended to even deeper waters on the continental shelf. In the following year the first risk contracts were signed with a number of foreign petroleum companies, among them Shell, Esso, BP, and Elf. Over the following decade Brazil’s proven reserves more than doubled. In 1984 the discovery of the huge Albacora field, also in the Campos Basin, further increased reserves. By the end of that year over half of Brazil’s proved reserves were off-­ shore. The domination of petroleum exploration and development by a state enterprise in the 1980s and the early 1990s may have provided some protection from the macroeconomic instability that characterized Brazil until 1994. Petrobrás was not immune, however, from these pressures. Inflation in Brazil accelerated with the successive failures of a series of flawed stabilization programs, beginning with the 1986 Plano Cruzado during the Sarney government. Public investment was one of the victims of the loss of control over the price level, and Petrobrás investment expenditures for both exploration and for production and development both fell sharply after 1984.10 The move toward energy independence after the Plano Real Brazil benefited from its first successful stabilization, the Plano Real of August 1994, in a number of ways. Although most attention has been given to its effect on public sector finances and on inflation, the plan also had important indirect effects in the energy sector. Longer-­term investment planning – virtually impossible at triple-­digit rates of inflation – became much more viable for both state enterprises and private corporations. In 1997 the conditions for higher rates of investment in the energy sector were strengthened by the opening of a number of offshore exploration areas for prospecting and development by private firms, both domestic and foreign. This was one of the consequences of Law 9478 of August 1997, which profoundly modified the near-­monopoly of the petroleum sector that Petrobrás had held since 1953. The Constitution of 1988 had ended the concession of risk contracts, but only permitted contracts for proven discoveries judged to be commercially viable. A constitutional amendment in late 1995 in effect liberalized the rules, allowing private involvement at a number of stages of petroleum exploration, production, and distribution. The 1997 law carried this process a number of steps further. The most important feature was the elimination of a number of subsidies for petroleum derivatives and a phased deregulation of prices, to be completed by late 2000.11 As may be seen from Figure 7.3 above, the divergence between local and international prices was considerably lower after the mid-­1990s than it had been in the preceding decade. Although most of this closer linking of domestic prices to international ones was probably due to the success of the Plano Real in reducing inflation, the 1995 and 1997 reforms appear to have reinforced this linkage.

Brazilian energy independence   79

Production

Consumption

2006

2004

2002

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

3,000 2,500 2,000 1,500 1,000 500 0 �500 �1,000 �1,500

1980

The reforms in Brazilian petroleum policy and institutional arrangements had a positive effect on domestic production. Between 1985 and 1996 Brazilian petroleum production had grown at only about 2.4 percent annually, while petroleum consumption increased at about 5.3 percent annually. As a result, Brazil’s dependence on petroleum imports increased steadily during this period. After 1996 these trends were reversed: production growth averaged about 8.2 percent per year, while the growth rate of domestic consumption fell to 2.0 percent annually. These trends are evident in Figure 7.4, which shows production, consumption, and net exports between 1980 and 2007. The acceleration in petroleum production after 1996 was based largely on Brazil’s investments in offshore exploration and development, especially in the Campos Basin. In the early 1970s, offshore production of Brazilian petroleum had accounted for less than one-­tenth of total production. Over the following three decades, land-­based production grew only modestly, while offshore production increased nearly twenty-­fold.12 Brazil owed much of its production success to its acquisition and domination of deepwater oil exploration and drilling. By 2000, Petrobrás was routinely prospecting and drilling in waters several hundred kilometers from its coast, at depths of 150 to 1,500 meters. Exploration efforts were aided by three-­ dimensional seismic technology, which permitted more accurate mapping of potential fields, and by continued development of offshore platform technology. By 2005, the technologies used in Brazil’s off-­shore petroleum production were on a par with those of any other major producing areas. In November 2007 Petrobrás announced the discovery of the largest known field, the Tupi field in the Santos Basin, about 300 km from the coast. Preliminary estimates of the potential of the field ranged between five and eight billion barrels. If proved, the field would raise Brazil’s reserves by about 40 percent, placing the country among the world’s ten largest holders of reserves. The

Net exports

Figure 7.4 Petroleum production and consumption 1980–2007 (thousands of barrels per day) (source: US Department of Energy, Energy Information Administration, online, available at: www.eia.doe.gov (accessed 19 October 2008)).

80   D.V. Coes d­ iscovery was significant not only for its size, but also for the quality of the oil, which is considerably lighter than the heavy crudes extracted both on the Brazilian land and in some of the earlier off-­shore fields.13 The technological challenges in exploiting fields like the Tupi one are daunting. Fields discovered in the first phase of offshore development were much closer to the seabed, often at total depths of about 1,000 meters below sea level. Exploitation of the Tupi field will require drilling beginning from sea bottom about 2,000 meters down and then penetrating through about 5,000 meters of a salt layer before reaching the field. Although Brazil has the technological capacity to do this, the economic viability of drilling to such depths is still uncertain.

Efficiency issues in Brazilian energy development The significant increases in petroleum production, primarily offshore, that were examined in the preceding section have placed Brazil within reach of its long-­ desired goal of energy independence. Weighed against the vulnerability of the Brazilian economy to the first and second oil shocks, this is an impressive accomplishment. Brazil’s experience in the past two or three decades suggests that increased domestic production of energy was indeed part of its success in reducing that vulnerability. This path to energy independence, welcome as it is, raises a number of questions about both the costs of this path and other ways in which Brazil might have attained this degree of energy autonomy. In this section we turn to some of these questions. The production increase may be viewed as the longer run supply-­side response to the oil shocks of earlier decades. As the preceding section shows, it resulted much more from state enterprise-­directed investment in oil exploration and development than it did from market responses to the higher oil prices. Brazilian investment in the exploration and development of its domestic petroleum production was not closely related to real international oil prices – in retrospect, perhaps fortunately so. The real price of petroleum fell sharply from the early 1980s through the 1990s. Yet investment by Petrobrás was high, especially in the 1980s, declining significantly only in the following decade.14 An obvious question is how demand-­side policies might complement the effects of this increased production. One way is through the prompt transmission of price signals to final energy users. As is clear from Figure 7.3 above, this process appears to have worked better after the Plano Real than it had earlier. But there may be considerable room for non-­price containment or reduction of energy demand as well. Although our focus in this chapter is on petroleum, an obvious question is what economically viable substitutes might have done – or still do – to reduce Brazil’s external energy vulnerability. Substitution of petroleum on the production side, allied with consumption substitution on the demand side can both be encouraged with appropriate price incentives, but here too public policy may be as important as market forces in encouraging these kinds of substitution.

Brazilian energy independence   81 Finally, there are large international market uncertainties that overhang tradable energy sources, especially petroleum. Other critical uncertainties arise from both technological trends and from domestic exploration and production uncertainties. Few of these questions yet have very clear answers, but a simple economic model helps to focus the questions. Energy market adjustment in the short and in the long run One of the characteristics of the energy market that became dramatically evident in the months following the first and second oil shocks in Brazil, as in other important petroleum-­using economies, was the relatively limited ability of either domestic supply to respond to the high international prices, or of domestic consumption to fall. Both of these features of domestic supply and demand response are captured in panel A of Figure 7.5 below. This is a stylized representation of the short-­run impact on an economy like Brazil of a large rise in international oil prices, from P1 to P2 in Figure 7.5-A. The extremely low price elasticity of domestic Brazilian petroleum supply, modeled here by a nearly vertical domestic supply curve SB. Brazilian domestic demand is represented here by the curve DB which is somewhat more price elastic in the short run, but still relatively low.15 The quantity of petroleum imports at the lower, pre-­oil shock price P1, the differences between demand and domestic supply, or M1. With the rise in international prices, imports fall to the quantity M2. Given the very small increase in domestic petroleum supply, most of the fall in the quantity of imports is due to the price-­ induced retraction on the demand side. Since the percentage fall in the quantity of imports is much lower than the percentage rise in the price, total expenditure on oil imports in the short run rises sharply, from P1M1 to P2M2, resulting in a large negative balance of payments impact like that suffered by Brazil after both the first and the second oil shocks. In the long run, there is a significant domestic supply adjustment in response to higher prices. As was suggested in the preceding section, this supply response in the Brazilian case was as much the production result of a state-­enterprise directed increase in investment in the petroleum sector as it was any kind of private market supply response to a higher price, at least before the passage of Law 9478 in 1997. But in either case, the longer-­run response to high international petroleum prices may be represented as a shift in the domestic supply curve from SB in Figure 7.5-A (repeated as the dashed line in Figure 7.5-B) to SB' in the long run equilibrium in Figure 7.5-B. The long-­run increase in domestic supply reduces the quantity of imports from M2, the level immediately following the price shock, to M3. If world prices continue at the “oil shock” level P2, then long-­run expenditure for oil imports falls to P2M3 (