The Oxford Handbook of the Italian Economy Since Unification 9780199936694, 0199936692

Citation preview

The Oxford Handbook of

THE ITALIAN ECONOMY SINCE UNIFICATION

This page intentionally left blank

The Oxford Handbook of

THE ITALIAN ECONOMY SINCE UNIFICATION Edited by

GIANNI TONIOLO

1

3 Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trademark of Oxford University Press in the UK and in certain other countries Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016,

© Banca d’Italia 2013 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence, or under terms agreed with the appropriate reproduction rights organization. Inquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Library of Congress Cataloging-in-Publication Data The Oxford handbook of the Italian economy since unification / edited by Gianni Toniolo. p. cm. Includes bibliographical references and index. ISBN 978–0–19–993669–4 (cloth : alk. paper) 1. Italy—Economic conditions. 2. Italy—Economic policy. I. Toniolo, Gianni, 1942– II. Title: Handbook of the Italian economy since unification. HC305.O94 2013 330.945c09—dc23 2012017352 ISBN 978–0–19–993669–4

1 3 5 7 9 8 6 4 2 Printed in the United States of America on acid-free paper

Contents

Contributors Foreword: Mario Draghi and Ignazio Visco Map: The Italian Peninsula before Unification (1848) Map: Italy, present day

PART I

ix xi xiii xiv

AGGREGATE GROWTH AND POLICY

1. An Overview of Italy’s Economic Growth Gianni Toniolo

3

2. Italy and the First Age of Globalization, 1861–1940 Harold James and Kevin H. O’Rourke

37

3. The Golden Age and the Second Globalization in Italy Nicholas Crafts and Marco Magnani

69

4. Italy, Germany, Japan: From Economic Miracles to Virtual Stagnation Andrea Boltho

108

5. The Italian Economy Seen from Abroad Marcello de Cecco

134

PART II

SOURCES OF GROWTH AND WELFARE

6. National Accounts, 1861–2011 Alberto Baffigi

157

7. Productivity Stephen N. Broadberry, Claire Giordano, and Francesco Zollino

187

8. Standards of Living Andrea Brandolini and Giovanni Vecchi

227

vi

contents

9. Human Capital Giuseppe Bertola and Paolo Sestito

249

10. Migrations Matteo Gomellini and Cormac Ó Gráda

271

11. Democratization and Civic Capital Luigi Guiso and Paolo Pinotti

303

PART III

INTERNATIONAL COMPETITIVENESS

12. A Long-Run Perspective on Comparative Advantage Giovanni Federico and Nikolaus Wolf

327

13. Real Exchange Rates, Trade, and Growth Virginia Di Nino, Barry Eichengreen, and Massimo Sbracia

351

14. Innovation and Foreign Technology Federico Barbiellini Amidei, John Cantwell, and Anna Spadavecchia

378

15. Old and New Italian Manufacturing Multinational Firms Fabrizio Onida, Giuseppe Berta, and Mario Perugini

417

PART IV

FIRMS, BANKS, AND THE STATE

16. Technology, Firm Size, and Entrepreneurship Franco Amatori, Matteo Bugamelli, and Andrea Colli

455

17. Resource Allocation by the Banking System Stefano Battilossi, Alfredo Gigliobianco, and Giuseppe Marinelli

485

18. Public Debt and Economic Growth: Italy's First 150 Years Fabrizio Balassone, Maura Francese, and Angelo Pace

516

19. Why the Italian Administrative System Is a Source of Competitive Disadvantage Magda Bianco and Giulio Napolitano

533

contents

vii

PART V

THE REGIONAL DIVIDE

20. Regional Convergence Giovanni Iuzzolino, Guido Pellegrini, and Gianfranco Viesti

571

21. Regional Disparities: Internal Geography and External Trade Brian A’Hearn and Anthony J. Venables

599

Data Appendix—Italy’s National Accounts (1861–2010) Alberto Baffigi, Stephen N. Broadberry, Claire Giordano, Francesco Zollino

631

References Index

713 773

This page intentionally left blank

Contributors

Brian A’Hearn: Pembroke College, Oxford Franco Amatori: Università Commerciale Luigi Bocconi Federico Barbiellini Amidei: Banca d’Italia Alberto Baffigi: Banca d’Italia Fabrizio Balassone: Banca d’Italia Stefano Battilossi: Universidad Carlos III de Madrid Giuseppe Berta: Università Commerciale Luigi Bocconi Giuseppe Bertola: EDHEC Business School Magda Bianco: Banca d’Italia Andrea Boltho: Oxford University Andrea Brandolini: Banca d’Italia Stephen N. Broadberry: London School of Economics Matteo Bugamelli: Banca d’Italia John Cantwell: Rutgers Business School Andrea Colli: Università Commerciale Luigi Bocconi Nicholas Crafts: University of Warwick Marcello de Cecco: Scuola Normale Superiore di Pisa Virginia Di Nino: Banca d’Italia Barry Eichengreen: University of California, Berkeley Giovanni Federico: European University Institute Maura Francese: Banca d’Italia Alfredo Gigliobianco: Banca d’Italia Claire Giordano: Banca d’Italia Matteo Gomellini: Banca d’Italia Luigi Guiso: Einaudi Institute for Economics and Finance

x Giovanni Iuzzolino: Banca d’Italia Harold James: Princeton University Marco Magnani: Banca d’Italia Giuseppe Marinelli: Banca d’Italia Giulio Napolitano: Università degli Studi Roma Tre Cormac Ó Gráda: University College Dublin Kevin H. O’Rourke: University of Oxford Fabrizio Onida: Università Commerciale Luigi Bocconi Angelo Pace: Banca d’Italia Guido Pellegrini: Sapienza Università di Roma Mario Perugini: Università Commerciale Luigi Bocconi Paolo Pinotti: Università Commerciale Luigi Bocconi Massimo Sbracia: Banca d’Italia Paolo Sestito: Banca d’Italia Anna Spadavecchia: University of Reading Gianni Toniolo: LUISS–Roma and Duke University Giovanni Vecchi: Università degli Studi di Roma Tor Vergata Anthony J. Venables: University of Oxford Gianfranco Viesti: Università degli Studi di Bari Aldo Moro Nikolaus Wolf: Humboldt University Berlin Francesco Zollino: Banca d’Italia

contributors

Foreword

In 2011 Italy celebrated the 150th anniversary of the unification of the peninsula, previously composed of seven states, into a single Kingdom of Italy. In 1893, after three decades marked by the coexistence of six issuing banks, the Bank of Italy was established as the country’s central bank. From its creation onwards, the Bank played a key role in promoting Italy’s economic development and, at times, its modernization. It is therefore not surprising that it chose to join in the celebrations by organizing an exhibition on the monetary unification of the country and by promoting an international research project on “Italy and the World Economy, 1861–2011”. The aim of the project was to enhance the understanding of Italy’s long-term economic growth. Contributors to the research included internationally prominent economists and economic historians, Italian scholars, and members of the Bank’s research departments. Preliminary results were discussed at a workshop held in Perugia in December 2010; revised papers were presented at a Conference held in Rome in October 2011 and, with further revisions, they are now published in the present volume, a true handbook of the quantitative economic history of Italy. Since its unification in 1861, Italy has been one of the ten largest economies in the world. However, the features of its “modern economic growth” are less internationally known than those of other countries of similar or lesser size. We trust that this book will help to bridge the knowledge gap concerning Italy’s long-term economic development. A latecomer to modern industrialization, in 1861 Italy was a backward country at the periphery of Europe’s industrial revolution. In the first three decades after unification the new Kingdom had to win the trust and respect of its neighbours, while at the same time creating new institutions and a single currency, and slowly unifying the domestic market. From the late nineteenth century onwards, the Italian economy largely converged on those of the most advanced European countries, with which it became increasingly integrated. There are several reasons why this process of convergence should be of interest to anyone curious about both the general pattern and the specifically Italian features of modern economic growth. Among the latter, scores of scholars have emphasized the vitality of Italy’s small and medium-sized enterprises, capitalizing on a secular tradition of widespread craftsmanship, the peculiar institution of formal and informal industrial districts, and the important role played by state-owned enterprises in promoting technical transfer, innovation, and managerial culture. On the negative side, a much studied peculiarity of the Italian economy is the

xii

foreword

stubborn persistence of the most pronounced case of regional income inequality in Europe, the North-South divide. Perhaps the most important reason why Italy’s economic history since unification is of interest, both inside and outside the country, lies in the economic link between Italy and the rest of Europe, which became increasingly tight with the passing of time. Italy’s economic growth could not have been as successful as it was in the absence of the progressive integration of the country with the rest of the continent. When, in the 1930s, autarky tried to shield Italy from its natural markets and competitors, the economy suffered a serious setback. At the same time, the sheer size of the Italian economy, together with its rapid growth, had a considerable impact on the overall development of Western Europe, an impact that has often been overlooked by scholars and policy makers alike. Immediately after the end of the Second World War, in the most desperate political, social, and economic conditions experienced in the last 150 years, Italy chose to link its future to that of its European neighbours. It was not an easy choice. It required both courage and vision on the part of the country’s leaders. The European choice strengthened the re-established democracy, which in turn fostered both economic growth and further European integration. The participation of Italy, as a founding member, in the European Economic Community (later the European Union) had a lasting influence on the culture, the society, and the political evolution of the country, an influence that went far beyond the positive impact of economic integration on Italy’s economic development. For their part Italy’s European partners benefited not only from stronger interrelations with a large growing economy but also from the contribution of Italy’s centuries’ old civilization to the building of Europe. As in the past, the commitment to a Union of stability and growth is still the answer to the challenges of the difficult period we are currently facing.

Mario Draghi (President of the European Central Bank and former Governor of Banca d'Italia) and Ignazio Visco (Governor of Banca d'Italia). Frankfurt am Main and Rome, September 2012

MAP 1

Map: The Italian Peninsula before Unification (1848)

MAP 2

TRENTINO A.D. Trento  

V. D’AOSTA Aosta  

LOMBARDIA! Milano

FRIULI V.G.

VENETO

Trieste Venezia  

PIEMONTE

Torino  

Genova

LIGURIA

Bologna

EMILIA R. Firenze

TOSCANA!

Ancona

MARCHE   Perugia

UMBRIA   L’aquila

LAZIO

Roma

ABRUZZO MOLISE

Campobasso Bari

CAMPANIA Napoli

PUGLIA   BASILICATA Potenza

SARDEGNA CALABRIA Cagliari

Catanzaro

Palermo

SICILIA

Map: Italy, present day

the oxford handbook of

THE ITALIAN ECONOMY SINCE UNIFICATION

This page intentionally left blank

Part I

AGGREGATE GROWTH AND POLICY

This page intentionally left blank

CHAPTER 1

AN OVERVIEW OF ITALY’S ECONOMIC GROWTH gianni toniolo

On May 5, 1860 when Garibaldi’s one thousand red-shirt volunteers, boarded the two steamships that would ferry them from Quarto in Liguria to the Sicilian port of Marsala, the Italian peninsula was a poor and slow-growing “geographical expression” (in Metternich’s words) at the periphery of Europe. Less than a year later, on March 17, 1861, this “geographical expression” became the Kingdom of Italy. In the following decade, the unification process was largely accomplished with the acquisition of the Venice region (1866) and of Latium, the remaining territory under the rule of the Roman Pontiff (1870). A few numbers suffice to describe the new kingdom’s absolute poverty. Its gross domestic product (GDP) per person was roughly equal to the average of today’s “richest” 42 African countries (Maddison 2001). Life expectancy at birth was about thirty years, considerably lower than in any of today’s least developed countries. Almost one baby out of three did not reach its first birthday. Four Italians in five were illiterate; recruits were short, averaging only 163 centimeters; and income distribution was extremely unequal (the Gini coefficient for 1861 was 50.4; Vecchi 2011, 430; see chapter 8). Income inequality was mirrored in expenditure (Rossi, Toniolo, and Vecchi 2001) and caloric-intake inequality. Although around 1861 a daily availability of about 2,500 calories per person made Italians on average better nourished than citizens of other countries with similar GDP per person, for as many as 30 percent of Italians the caloric intake did not reach 2,000 calories per day, making them chronically undernourished

4

aggregate growth and policy Table 1.1 Welfare indicators: 1861 and 2011 1861

2011

2,190

25,668

30

82

First year death rate (per 1,000)

289

4.5

Income distribution (Gini coefficient)

0.50

0.33

People in absolute poverty (% population)

40

5

Literacy rate (% population)

22

98

GDP per person (euros at 2010 prices) Life expectancy at birth

Sources: Chapters 6 and 8.

(Vecchi 2011, 12–23; Vecchi and Coppola 2006). About 40 percent of the population was close to or below an absolute poverty line defined as “sufficient income to buy the essentials of life” (Vecchi 2011, 295), equal to 1.5 euros per day per person at 2010 purchasing power. In the 150 years since political unification, Italy’s GDP per person multiplied by about twelve (see chapter 6). Life expectancy at birth has grown to eighty-two years (one of the longest in the world); only 0.45 percent of children die in their first year (the comparable figure is 0.7 percent in the United States); income distribution is much more egalitarian (the Gini coefficient is about 0.33); and absolute poverty has been reduced to 4 percent of the population (see chapter 8). Striking as these figures are (Table 1.1), Italy’s progress in production and welfare is broadly similar to the progress made by other Western European countries from the beginning of their “modern economic growth” (Kuznets 1967a) to the present day. Even so, Italy’s case is worth studying for the following reasons: (1) despite progress made in the last couple of decades, Italy’s quantitative economic history has thus far been underresearched; (2) its long-term economic history is important to shed light on the country’s poor economic performance since the beginning of the twenty-first century; (3) Italy’s sheer size also makes it an important contributor to Europe’s long-term growth; and (4) its growth pattern and growth factors present similarities to those of other European countries, but also not entirely understood peculiarities. This book is the fruit of new research on Italy’s “modern economic growth” conducted on the occasion of the sesquicentennial of the country’s political unification. Chapters 2 and 3 provide an overview of Italy’s economic progress in the period before and after the Second World War, respectively, focusing on the major policy issues. The other chapters, mostly coauthored by Italian and foreign scholars, contain the outcomes of their research on specific issues in Italian economic history. The expert reader will note that the chapters contain a considerable amount of fresh quantitative findings that advance the knowledge of Italy’s growth experience in a number of fields, including new national

an overview of italy’s economic growth

5

accounting and productivity statistics; welfare (including income distribution) indicators; and measures of human and social capital, comparative advantages, real exchange rates, public finance, innovation (patents), firm size, migration flows, and efficiency of the banking system. Two chapters are devoted to the thorny issue of Italy’s persistent geographic disparities. The book also contains a comparison of the post–World War II growth in the three countries who “lost the war but won the peace,” a view of Italy from abroad, and case studies of Italian multinationals. This introductory chapter draws from these research findings to sketch a “tale of convergence and two tails.” Framed in a standard convergence narrative, Italy’s economic history from 1861 to 2011 is characterized by a central period of about a century (mid-1890s to early 1990s) of robust catch-up, framed between two “tails” of sluggish growth (1861–1896 and 1992–2011) when the country lost ground to the productivity leaders. In the following pages the main features of Italy’s economic growth since 1861 are outlined and explanations for the secular rapid growth and for the disappointing beginning and final periods are advanced.

A Tale of Convergence and Two Tails To assess the economic performance of united Italy, the country’s initial backwardness must be taken into account. Gerschenkron (1962a) argued that, if and when a “traditional economy” manages to overcome the “stumbling blocks” on the road of “modern economic growth,” then initial backwardness may prove to be an “advantage,” allowing for faster growth than in more advanced countries. Today’s economists talk about convergence of low-income countries on those with higher income. Since the early 1800s, catch up by lower-income countries is a key feature of the pattern of “modern economic growth” as described by Kuznets (1967a). Less developed countries grow faster than the productivity leaders by borrowing technology from them at a low cost and adapting it to their own comparative advantages, by cheaply transferring resources from low-productivity agriculture to high productivity manufacturing, by exporting labor-intensive products, and by the emigration of their unskilled workers. The long-run economic performance of a given country must, therefore, be assessed taking into account its initial relative level of development, as measured by GDP per person. How economically backward was the Italian Peninsula at the time of its political unification? From the thirteenth to the sixteenth century, Northern-Central Italy was perhaps the most prosperous European region. In 1500 its GDP per person was about 35 percent higher than Western Europe’s average. From about the middle of the seventeenth century onward, Italy began to lose economic ground relative

6

aggregate growth and policy

Table 1.2 Italy and the advanced countries: GDP per person, 1500–1870 Italy

United Kingdom

France

Germany

Japan

Western Europe

Italy as a share of Western Europe (%)

1500

1,100

714

727

676

500

796

138

1820

1,117

1,706

1,135

1,135

669

1,245

89

1870

1,499

3,031

1,876

1,876

737

2,088

72

Source: Maddison (2001). US dollars at 1990 purchasing power parity.

to the faster growing northwestern part of the continent. In the eighteenth century output in the Peninsula hardly kept up with population growth. According to Malanima (2005 and 2006), by the end of the eighteenth century the average citizen of Central and Northern Italy might have been about 20 percent worse off than their ancestors had been two or three centuries earlier. During the Napoleonic wars, Italy’s real GDP per person probably reached its lowest level since the fourteenth century. This view is somewhat more pessimistic than Maddison’s (2001), on which Table 1.2 is based. There is broad agreement that slow recovery started in the 1820s, possibly slightly accelerating in the 1850s. At the time of unification, Italy’s income per person, allowing for purchasing power parity was roughly half that of Great Britain, the productivity leader of the time, and about two-thirds that of France, a neighboring and in many ways similar country to Italy (Maddison 2001). More generally, Italy lagged behind the Northwestern corner of the continent. Indirect indicators, such as life expectancy, infant mortality, height of recruits, and literacy rates, are all consistent with Italy’s relative backwardness. The composition of employment was typical of a backward economy, with agriculture’s share of full-time equivalent labor accounting for as much as 63.2 percent, followed by the services sector (19.1 percent) and industry (17.7 percent). The Italian Peninsula also lagged behind the more advanced parts of the continent in terms of social overhead capital (infrastructure). In 1861 there were only 2,404 kilometers of rail in operation, almost entirely located in the northern Po Valley, against 14,603 in Great Britain (smaller in size than Italy) and 11,603 in the area of the future German Reich (however almost twice as large as Italy). At the beginning of 1859, the Italian Peninsula was divided into eight states (including the tiny Republic of San Marino, which did not become part of the new kingdom), with varying degrees of backwardness (see map on p. xiii). Almost two millennia of disunion had produced deeply rooted cultural and institutional differences across the Peninsula. In particular, around 1861, estimates of the number of persons capable of shedding their local languages or dialects and communicating in Italian vary from 2.5 percent (De Mauro 1991)

an overview of italy’s economic growth

7

to about 10 percent (Castellani 2009) of the population. French was the working language of the Sardinian parliament based in Turin. How deep were regional economic differences at the time of unification? The question is not trivial, given the persistence of the North-South divide throughout the 150 years after unification. The issue of “initial conditions” looms large on the never-ending debate about the “Southern Question.” Was the economy of the Kingdom of the Two Sicilies (by and large coinciding with today’s Mezzogiorno) substantially less developed than the country’s northwestern corner? If so, the new state can only be “blamed” for not closing an already existing gap. Or was Southern Italy almost as developed as (some say even more developed than) the Northwest? In this case, the new Kingdom of Italy would be guilty of having actually created a hitherto nonexisting North-South divide. Despite a new wave of quantitative research the issue unfortunately remains unresolved. Estimates of regional GDP for the early years of unified Italy are still tentative; they become more solid from the 1890s onward. Daniele and Malanima (2007), on the basis of their assessment of national GDP per head in the eighteenth and early nineteenth centuries, argue that when average national income is close to subsistence, regional disparities cannot be wide. According to Felice (2007), in 1871 GDP per head for the whole South was about 10 percent below the national average with a large within-South variance (Calabria being better off than most northern and central regions). For the same year, the authors of chapter 20 estimate manufacturing value added per person in the South to be about 30 percent below the Northwest’s, but only 10 percent below the nation’s average. Fenoaltea (2007) points to an East-West gap in addition to the North-South divide. The notion of an initial relatively small income gap between the northwestern and southern parts of the country is not entirely consistent with the large gaps revealed by some indicators of well-being, normally quite strongly correlated with GDP per person. In 1861, the average height of recruits was 163.7 centimeters in the Northwest (165.2 in the Northeast) versus 160.9 in the South (Vecchi 2011, 57). Only 15 percent of the population aged over fifteen years could read and write in the South versus 47.9 percent in the Northwest (Vecchi 2011, 425). These data are mirrored in a higher incidence of child labor in the South than in the rest of the country (Vecchi 2011, 147; Toniolo and Vecchi 2007) and in lower rates of school attendance. Less clear-cut are indications on the North-South gap coming from measures of life expectancy at birth and infant mortality. Although a wide regional dispersion in the life expectancy and infant mortality indicators existed at the time of unification, no precise latitudinal pattern emerges: citizens of Liguria enjoyed the longest life expectancy (thirty-six years), and those of Basilicata the shortest (28.5), but the second longest life expectancy was recorded in Apulia (for a map of Italian regions see p. xiv). Despite the enormous progress recently made in quantifying regional income and living standards at the time of unification, evidence remains sufficiently contradictory to settle the long-standing debate on whether or not the new state

8

aggregate growth and policy 2.5 FIN SPA

2.0

NOR SWE ITA

CANAUT DEN

USA

FRA SWI GER

NED

1.5

AUS UK

BEL

1.0

0.5

0.0 6.8

7.0

7.2

7.4

7.6

7.8

8.0

8.2

Log of per capita GDP

Figure 1.1 Unconditional convergence, 1870–2008 Sources: Adapted from Maddison (2001) and OCSE (http://stats.oecd.org/).

is to blame for creating a “Southern Question.” Scholars are still unable to adjudicate between Fortunato (1973 [1904])’s thesis, which emphasized the unkindness of nature in the South, the persistence of feudal power, the low level of physical and human capital, and the trade disadvantages created by geography (see also chapter 21), and Nitti’s (1958 [1900]) opinion, which instead stressed the accumulated gold and silver wealth of the Bourbon Kingdom, its low public debt, and the postunification diversion of Southern savings to the payment of social overhead capital in the North. On the basis of all available evidence, one can only conclude that a fairly wide dispersion of regional indicators of income and welfare existed at the time of unification (not confined to the traditional North-South representation), and that the North-South divide, although evident in several dimensions, is likely to have been less pronounced than it became at later times (see chapter 20). Starting from the conditions of economic backwardness described previously, between 1870 (the first year when internationally comparable data exist for a large number of countries) and 2008 Italy’s real GDP per person grew at an annual average growth rate of 1.9 percent, slightly higher than Western Europe’s rate and about the same as the United States’ rate. Figure 1.1 plots the results of a simple unconditional convergence exercise for a number of today’s “advanced” countries. Italy’s growth rate, together with that of Spain and Germany, seems to be somehow lower than expected given its GDP per person in 1870. Why, over the 150 years after unification, did the Italian economy slightly underperform with respect to its initial catch-up potential? In trying to frame an answer, we shall consider the time pattern of Italy’s growth characterized by a central period of robust catch-up to the more advanced countries with two “tails” of sluggish growth on each side.

9

an overview of italy’s economic growth Table 1.3 A tale of convergence and two tails 1870–1896

1896–1992

1992–2010

Italy

0.6

2.4

0.5

France

1.4

2.0

1.0

Germany

1.5

1.9

1.2

United Kingdom

1.1

1.4

1.8

United States

1.4

2.0

1.5

Japan

1.2

3.1

0.7

Sources: Maddison (2001), http://stats.oecd.org/ Chapter 6, and Data Appendix.

From the mid-1890s to the early 1990s, the narrative of Italy’s economic development can be framed in the catching-up paradigm: one of Europe’s “peripheral” countries converging to the “core” of the early industrializers. A backward country at the turn of the end of the nineteenth century, for a hundred years Italy grew faster than the main Western European countries and the United States (Table 1.3). Italy’s economic history, in its first and final tails, however, is also proof that convergence does not necessarily follow from initial backwardness. For more than a generation after unification, the new country failed to exploit its “latecomer’s advantages” in a rapidly expanding Atlantic economy to catch up with its most prosperous Northwestern neighbors. From 1861 to about 1896, rather than catching up, Italy fell behind the more advanced areas of Western Europe. The economic history of the fifteen-odd years before the Great Recession of 2008–2009 is also characterized by slower growth rates than the rest of Western Europe (itself an area not shining for rapid growth). This is, however, not so much a story of failing to catch up, since Italy’s GDP per head was then close to Western Europe’s average, as rather one portraying the more subtle malaise of a slow, progressive relative decline; malaise already suffered by the Peninsula in the eighteenth century.

The Years 1861–1896: Tenuous Growth and Unfulfilled Expectations? Italy’s political unification took place at a time when the so-called first globalization was unmistakably under way, driven by technical change in transportation and communication and by a steady movement toward free trade, with “modern economic growth” gaining speed and spreading from the northwestern

10

aggregate growth and policy

cradle of the industrial revolution to the European periphery. After the repeal of the Corn Laws, European average tariffs kept falling throughout the 1850s (Findaly and O’Rourke 2007,396). In January 1860, France and the United Kingdom signed the Cobden-Chevalier treaty, which by incorporating the “most favored nation” clause opened the way to further reductions of import duties throughout Europe. As a result, international trade increased. More generally, in the 1850s and 1860s, Western Europe’s GDP per head grew at a good pace, particularly in the lands of the future German Reich (1.6 percent per annum in 1850–1870 versus 0.4 percent per annum in the previous twenty years). Great Britain and France also grew at relatively good rates. The timing of political unification, therefore, was not “ill chosen” from the vantage point of economic growth. International conditions in the so-called Atlantic economy favored trade, capital inflows, labor mobility, and the transfer of technology from the more advanced industrial “pioneers.” These conditions notwithstanding, for about thirty-five years growth-acceleration eluded the Kingdom of Italy. Rather than forging ahead, the Peninsula lost ground to the productivity leaders. The secular relative decline continued, despite unification. If the economy was not the main ideologic and political driver of the Risorgimento, it was nonetheless present in the minds of its architects (Ciocca 2007, 70–77). A customs union modeled on the Zollverein was proposed in the 1840s, and advocated by France in the wake of the 1859 Villafranca armistice, which stopped in its tracks the eastward thrust of the French-Piedmontese armies. The liberals who forged unification saw protectionism, particularly in the Kingdom of Naples, as an unwise choice for small economies at times of rapid integration of the Western European economies. In the 1850s, Piedmont’s free trade coupled with a program of public investment in transport infrastructure showed that such policies could result in growth acceleration. However, the idea that unification per se would perform the miracle of raising the welfare of the Italian people was illusionary in the absence of sufficient “social capability for growth” (Abramovitz 1989). From unification to the second part of the 1890s, Italy’s GDP per person grew an annual average rate about half that of the United Kingdom, the acknowledged economic superpower of the time (Table 1.3). Population grew by 0.65 percent per annum and total GDP by 1.24 percent. The average growth rate (1861–1897) of value added by agriculture was 0.97 percent per annum, by the service sector 1.37 percent, and by industry 1.56 percent (all data at present-day borders; see chapter 6 and appendix). It is now possible to measure the sources of growth for this period (see chapter 7). In the first two decades after unification, agriculture led total factor productivity (TFP) growth (0.6 percent per annum), with industry and services lagging behind at close to zero productivity increase. Growth in GDP was largely driven by labor and capital inputs: the number of full-time equivalent workers grew faster than total population (on average 0.74 percent per annum). According to the 1881 population census, labor force participation reached

an overview of italy’s economic growth

11

55 percent, the highest level in the economic history of united Italy. After a sluggish growth in the 1860s, fixed-investments grew annually by 5.7 percent between 1869 and 1888, at the peak of a cyclical boom followed by a long and deep recession. Slow growth in the thirty-odd years after unification has been variously explained. The debate was particularly lively in the 1950s and 1960s, when the first output statistics (GDP and industrial value added) for the period were compiled (Istat 1957; Fuà 1975). Explanations ran from infrastructure being a prerequisite for growth (Romeo 1959), to the lack of an agrarian reform (Sereni 1971), to “agents of industrialization” (universal banks) being created only in the mid-1890s (Gerschenkron 1962a). The debate was framed in the language, fashionable at the time, of take-offs and big-spurts or of Marxian “original accumulation.” Objection to such categorizations led to lines of research focusing on slow but steady growth acceleration as in the British industrial revolution (Bonelli 1994), or on the international cycle induced by fluctuations in international capital movements (Fenoaltea 1988, 2011b). In the early 1990s Zamagni (1993, 81) argued that the existing quantitative evidence only “allowed to postpone” a final judgment on slow growth after unification: the old question about why, for over a generation, catching up was precluded to this latecomer to Europe’s modern economic growth remained an open and interesting one. To answer this question, a preliminary query must be addressed: in what ways was Italy’s growth potential enhanced by the political unification that was almost completed in 1870? Unification entailed two processes, each with its own economic implication: the fusion of seven states into a single sovereign entity, and the establishment of independence from foreign domination or influence. Independence also implied the belated end of absolute monarchies. From an economic viewpoint, the most obvious result of the establishment of the Kingdom of Italy in 1861–1870 was the emergence in the Italian Peninsula of the world’s eighth largest economy (Table 1.4). Poor as they were, the Italians were numerous. In 1870, the population of the new kingdom totaled about 26 million, whereas the Kingdom of Naples, by far the most populous among the preunification states, only counted 9 million subjects. The advantages of a large single market had not escaped the moderates, such as Vincenzo Gioberti, who had advocated a federation of independent states and a customs’ union rather than outright unification in a single sovereign state (Gioberti 1845). A priori, therefore, the most obvious economic benefit deriving from political unification would come from the creation of a large single market out of a number of small closed economies, with the attendant improvement in resource allocation and economies of scale, particularly relevant for the most dynamic sectors of the so-called second industrial revolution. Although the degree of openness (import plus exports/GDP) of the Kingdom of Sardinia was about 40 percent (Graziani 1960), that of Lombardo-Veneto, the two Emilian Duchies, and Tuscany was only 20 percent. The Papal States and the Kingdom

12

aggregate growth and policy Table 1.4 The G10 in 1870 and 1998: Shares in world GDP (1990 PPP dollars) and population (%) 1870

1998

GDP

POP

GDP

POP

China

17.2

28.2

11.0

21.0

India

12.2

19.2

5.0

16.5

United Kingdom

9.1

2.5

3.3

1.0

United States

8.6

3.2

21.9

4.6

Russia

7.6

7.0

3.4

4.2

Germany

6.5

3.1

4.2

1.4

France

6.5

3.0

3.4

1.0

Italy

3.8

2.2

3.0

1.0

Japan

2.3

2.7

7.7

2.1

Spain

2.0

2.0

1.7

0.7

24.2

26.9

25.3

46.5

Rest of the world

Source: Data from Maddison (2001).

of Naples lagged behind with ratios of about 10 percent (Zamagni 2007, 42–43; see chapter 6). Moreover, the Bourbon Kingdom of Naples (excluding Sicily) traded almost entirely overseas, with Italian trade accounting only for about 13 percent of its exports (Graziani 1960). Market unification would, therefore, also imply a considerable amount of trade diversion, particularly considering that the Austrian Kingdom of Lombardo-Veneto was part of the Habsburg customs’ union. In addition to market unification, to assess the potential economic impact of independence from foreign domination and constitutional government, institutions at large must be taken into account: the adoption of civil and commercial codes, the introduction of checks and balances between government branches, the diffusion of compulsory elementary education, and increased freedom of enterprise all spring immediately to mind as potentially growth-enhancing factors (see chapter 19). In addition, benefits would come from the necessary complements to market unification: single currency, homogeneity of weights and measures, unified regulation of financial intermediaries, uniformity of taxation, a single tariff, commercial treaties, and the like. The end of foreign dominance might also have had a direct impact on growth, in so far as some policy decisions were made more in the interest of foreign than domestic citizens (the port of Venice, for example, was left with little investment by the Austrians who instead privileged Trieste, their traditional sea outlet (Toniolo 1977a). Moreover, one might assume that the government of a large single country would be more

an overview of italy’s economic growth

13

successful in international economic negotiations than those of uncoordinated individual small states. If these were the exante potential benefits of unification, expost Italy’s growth in 1861–1896 showed that they were either too small or not adequately exploited. Why then did a larger market and new institutions not impact on growth for a long period of time after 1861? Part of the explanation is quite simple: political unification was only a precondition to market unification. Its actual implementation took time. The reduction of transportation costs depended on building railways, roads, and harbor infrastructures. Italy’s orography made the task expensive and time-consuming (Fenoaltea 1983, 1984). The new Kingdom was relatively efficient in connecting the main urban centers of the country by rail: if in 1861 trains could reach only as far south as Ancona on the Adriatic coast and no connection existed south of Genoa on the Tyrrhenian side, by mid-1864 one could already travel by train as far south as Bari. In 1870 it was possible to reach Naples from Milan and Turin and to cross the Peninsula from Rome to Ancona and from Naples to Bari (Ciocca 2007, 87). However, as was persuasively argued by Fenoaltea, a consistent reduction in transportation cost derived only from the construction of the so-called secondary lines, connecting the main to the lesser cities and spreading from the coast to the interior: a longer task that produced its fruits in reducing price differentials across the country not earlier than in the 1880s. The replacement of old monies with the new single currency (the Piedmontese lira renamed Italian lira) was slow: by 1870 only 57 percent of the pre-1861 money stock had been changed into lira, and the last large chunk of Neapolitan minted silver was only brought to the Treasury for change in 1894 (De Mattia 1959). The unification of weights and measures was equally slow. Financial market integration, as measured by bond price differentials across the regional stock exchanges, did not take place until the early 1880s (Toniolo, Conte, and Vecchi 2003). Only by that time can infrastructure and market-institution be seen as sufficiently advanced for the economy to take advantage of a single market that was therefore only “announced” in 1861, but took at least two full decades to materialize. Institution building was also slow (see chapter 19). A Civil Code of French imprint but containing numerous idiosyncratic norms was voted by Parliament in 1865 together with a Commercial Code, regulating business practices, which derived directly from the Code of the Kingdom of Sardinia with minor modifications. Only in 1882 did Italy introduce its own new Commercial Code based on the “revolutionary” premise that, in business matters, commercial law should prevail over the civil discipline. As argued in chapter 19, the creation of a national bureaucracy and of a set of administrative laws turned out to be not as growth-friendly as elsewhere in Europe. Human capital creation was even slower than the building of infrastructure and institutions (see chapter 9). Compulsory primary education (for three years) was introduced immediately after unification but its implementation, mandated

14

aggregate growth and policy

to local (city) governments, was poorly funded and supervised. Only in the first decade of the twentieth century did the national government take over the task of providing free and compulsory five-year elementary education. Despite these shortcomings, literacy in the population aged 15–19 rose from 27 percent in 1861 to 45 percent in 1881 and 62 percent in 1901(Vecchi 2011, 425), albeit with huge regional disparities. Slowness in the implementation of the single market, in human capital creation, and in the diffusion of modern legal institutions accounts for part of the failure of the new-born state to immediately catch up with its richest European neighbors. External shocks and policy mistakes also delayed convergence. In its first decade, the New Kingdom fought two wars (in 1866 and 1870), with the attendant suspension of convertibility of the lira and increase in government spending. Southern brigandage—a mix of social unrest and legitimist guerrilla—was met by ruthless action by a large military deployment. Throughout this period, many in Italy and abroad doubted the viability and even the future existence of the new Kingdom. Perceived instability was hardly conducive to business confidence. Neither was a fiscal policy aimed at balancing the state budget by 1876. A commercial policy characterized by moderate (1878) and then (1887) higher protection for manufacturing and agriculture figures prominently in the contemporary and historical literature as one of the main culprits for unsatisfactory growth. As argued in chapter 2, Italy’s shift to protection was hardly an Italian peculiarity, including the pactum scleeris between iron and wheat. The tariff of 1887 ignited a trade war with France that damaged silk and wine exports. The import duty on wheat, a response to falling transatlantic transport cost, was undeniably technically irrational, but it was seen by moderate free traders of the time as a useful compromise to ease the cost of globalization borne by a relatively small number of people. By the late 1870s, the single market began to take shape and modern institutions to impact on the economy. At the same time, equilibrium had been restored to public finances, and the 1874 banking law had somehow better organized the financial system. Italy looked more stable to investors and the expectation of a return to convertibility of the lira (which took place in 1883) made the country more attractive to foreign capital. Growth acceleration in GDP and investment took place from the late 1870s to the late 1880s. It was then frustrated by the fragility of a poorly regulated banking system, fraught with inexperience and greed. In an only too-well-known sequence of events—familiar to late twentieth- and early twenty-first-century readers—convertibility allowed banks to borrow abroad at relatively low rates. In Luzzatto’s imaginative metaphor, “the atmosphere was hyper-oxygenated by gold [ . . . ] For the first time since 1861, the public could nourish the illusion—if short lived—of cheap money: in 1884 the official discount rate was reduced to 4 percent, but commercial banks practiced a 3 percent to their best clients” (Luzzatto 1963, 208). Easy credit was largely directed to the construction industry, particularly in Turin, Rome, and Naples, soon fuelling a bubble. For a while interest rates remained relatively low

an overview of italy’s economic growth

15

but soon inexperienced (and dishonest) risky borrowers entered the market and systemic leverage increased. By the late 1880s the flow of foreign capital dried up and then reversed; Italy’s international reputation was not aided by the tariff war with France and the sudden shift of alliances by a treaty with Berlin and Vienna. Italian banks were forced to reduce their exposition to industry, particularly to the construction sector. The gold outflow made it difficult for the six banks of issue to maintain the mandated reserve ratio. Questionable accounting practices were used to conceal the lack of full metal coverage of notes. It would be later discovered that, to avoid collapse, one of the main banks of issue, Banca Romana, even resorted to the criminal practice of printing two sets of banknotes bearing the same serial numbers. The construction bubble burst, a number of companies were forced to receivership, and the smallest and least capitalized banks began to fail. Contagion spread to the larger commercial banks and to some of the six banks of issue. By 1892–1893 most Italian commercial banks were either illiquid or insolvent. Convertibility, de facto unavailable since 1887, was suspended in 1893. Weak political leadership, in the venomous climate of the struggle between Crispi and Giolitti, scandals, and turncoats (vividly described in Federico De Roberto’s novel Imperio), allowed the crisis to drag on until the twenty-third hour. Only in the summer of 1893 did a sweeping banking reform create the Bank of Italy out of the merger of three banks of issue. In 1894 the two largest commercial banks were liquidated and two better-managed and capitalized banks grew from their ashes. The appearance of a modern central bank and reorganization of the banking system on a sounder basis put an end to the panic, clearing the board for a sustained period of catch-up growth to take place. Before moving to Italy’s century-long catch-up growth, it is useful to note a peculiarity of the postunification Italian economy. Recent research (Vecchi 2011 and various chapters in this book) shows that the most relevant welfare indicators monotonically improved during the postunification low GDP growth years. Between 1861 and 1891, life expectancy at birth grew from twenty-nine to thirtyeight years (catching up with France); infant mortality declined from 223 to 189 per thousand (catching up with France and Great Britain); and the average height of recruits increased from 162.9 to 164.4. These numbers point not only to improved nutrition but also to better sanitation, caused by investment in aqueducts and sewerage systems. The already mentioned literacy improvements were mirrored in a reduction in child participation in the labor force (Toniolo and Vecchi 2007). If income inequality did not change much until about 1900, consumption expenditure became more egalitarian (Rossi, Toniolo, and Vecchi 2001) and, what probably matters most, absolute poverty declined from 44 to 35 percent of the population (Vecchi 2011, 297). Slow GDP growth not withstanding, there were considerable welfare gains accruing from unification to the first generation of Italians, particularly to its less privileged members. If not in GDP per person, Italians caught up with their more advanced neighbors in other important aspects of individual and collective welfare: those who showed

16

aggregate growth and policy

their disappointment for the economic outcomes of the Risorgimento were only partly justified by facts.

The Long Convergence, 1896–1992 Hooking into the “First Globalization,” 1896–1913 By the late 1890s Italy’s GDP per person had fallen to 38 percent of Great Britain’s (down from 45 percent in 1870); by 1913 it reached almost 54 percent. Sharp growth acceleration took place from the late 1890s onward: the cyclical component explains about one fifth of it, the rest is caused by trend change (Ciocca 2007,142–143). For the first time since the seventeenth century, rather than falling behind, the Peninsula forged ahead reducing the income gap relative to the most productive country of the time (Toniolo 1990; Zamagni 1993). Between 1897 and 1913 Italy’s total and per person real GDP grew on average by 2.4 and 1.6 percent per annum, respectively; industrial production grew on average by 3.8 percent; and agriculture by 1.7 percent. The pace of productivity growth across all sectors increased from the 1880s onward and further accelerated at the end of the century. Between 1901 and 1911, a period also called the “Giolittian era” after Giovanni Giolitti, the leading political figure of the time, labor productivity grew by 2.5 percent per annum in industry and by 2.2 percent in the services sector (see chapter 7). Some industries (metal-making, steam engines, electricity, and electromechanics) expanded at double digit rates. In this phase some “first mover” firms established an enduring advantage in key sectors of the second industrial revolution: Fiat in automobiles; Pirelli in cables and tires; Ansaldo in shipbuilding; and Falck, Piombino, and Terni in the mass production of steel (see chapter 16). Power generation (mainly hydroelectric) and the electromechanic sector, already quite sophisticated at the beginning of the 1890s, expanded rapidly from 1900 onward. The first foreign direct investments by Italian companies date back to this period. Foreign direct investments and imports of machinery were responsible for technology transfers from the most advanced producers. If French, Belgian, Swiss, and British foreign direct investments concentrated mainly in public utility and transportation, German and later American foreign direct investments favored manufacturing (see chapter 15). Modern industry, however, was mainly concentrated in the Northwestern part of the country (in 1911 the regions of Milan, Turin, and Genoa produced 55 percent of the country’s value added),thereby increasing regional economic disparities (see chapters 20 and 21). After 1896, once the light at the end of the financial-crisis tunnel was in sight and the economy again took a central place in the political arena, monetary,

an overview of italy’s economic growth

17

fiscal, and exchange rate policies supported growth in a time-consistent fashion (Toniolo 1990; see chapters 2 and 13). A sequence of balanced budgets revived the interest of foreign investors in Italy. The debt-to-GDP ratio rapidly declined (see chapter 18). In 1906, confidence was so high that the voluntary conversion of Italian Consols (Rendita Italiana) from 5 to 3.75 percent was subscribed by all but a handful of bondholders who were redeemed at par. Cleverly, the gold standard was never officially reinstated, but monetary authorities consistently mimicked it, holding the exchange rate within the “gold points,” at times even above the old parity (Cesarano, Cifarelli, and Toniolo 2012). Interest rates slowly converged on those of Paris and London. A virtuous cycle set in whereby the exchange rate stability was a cause and an effect of emigrant remittances, emigration being one of the most conspicuous results of Italy’s increasing participation in the “first globalization’s” Atlantic economy (see chapter 10). Despite substantial export growth, the trade deficit increased (see chapter 21) because of the import of raw materials and cheap overseas grain but also, to a large extent, because of the huge import of machinery embodying new technology (Warglien 1985). The relatively high, but far from prohibitive, tariff of 1887 remained in statute books but commercial treaties containing the “most-favored nation clause” progressively tempered it. Industrial growth was mainly concentrated in the Northwest of the country (Ciccarelli and Fenoaltea 2009, 2010; see chapter 20), which enjoyed an original natural advantage and easy access to the largest domestic markets for the new industries, particularly engineering (see chapter 21). The never-tobe-solved “Southern question” arose. It was intertwined with a “social question” that sometimes took violent forms but never again reached the pitch of 1898, when Milan had been the theater of the most violent repression ordered by the government. It is perhaps surprising, given the record of the industrial revolutions that took place in other countries during the first half of the century, that Italy’s late industrialization turned out to be unusually “benevolent” toward the lower classes (Toniolo 2003): average life expectancy lengthened to forty-five years, the incidence of child labor sharply declined, and Italian recruits grew taller (see chapter 8; Vecchi 2011). Certainly surprising is the fact that income and consumption distribution became less unequal and absolute poverty was reduced (Rossi, Toniolo, and Vecchi 2001; Vecchi 2011; see chapter 8). The egalitarian and poor-friendly features of Italy’s late industrialization are in need of a rigorous explanation (Toniolo 2003). For the time being a tentative hypothesis points to its taking place in an increasingly open economy during the so-called first globalization. It is likely that migrations reduced both the rent-to-wage and the skilled-to-unskilled wage ratios (see chapter 10), while at the same time international trade increased the worker’s real purchasing power through the import of cheaper foreign “wage goods.” The social and political environment is also likely to have played a role: the Socialist Party was created in 1892; trade union membership increased; socially motivated Catholics grew stronger (in 1891 Pope Leo

18

aggregate growth and policy

XIII published his social encyclical Rerum Novarum); and, what matters most, the governments of the Giolittian era had a more open view of liberal democracy and took an inclusive stance towards moderate socialists and Catholics, both advocates of social reforms (Gentile 1977).

Economic Survival and Political Collapse Index-number problems exacerbated by the enormous shift in relative prices that is typical of war economies make it hard to determine Italy’s progress in catching up with the more developed countries during the First World War. What can be said is that the steel, shipbuilding, engineering, and chemical industries—all heavily subsidized—underwent rapid expansion. A factor of particular importance for subsequent growth was high investment in the hydroelectric power industry, a sector in which Italy was on the technologic frontier at a time when electrification was the world’s new “general purpose technology.” After a fall in 1915, the year Italy entered the war, industrial production and GDP recovered and remained in 1916 and 1917 at the all-time level reached in 1913. After the war, part of the excess productive capacity was eliminated, not without pain, in a crisis that saw a drop in GDP by almost 9 percent between 1917 and 1921. Part of the wartime productive capacity, however, endemically outstripped demand for many years, typically in the shipbuilding sector. Other wartime high-technology investments were converted to peacetime use. The war tragically marked the watershed between the “first globalization” and a period known to historians as the “globalization backlash.” As for convergence, the story of the Italian economy in the interwar years divides into two periods. Up to 1929 Italy’s growth rate was somewhat above the Western European average, significantly higher than in the United Kingdom but significantly lower than in France. In the 1930s, however, Italy’s growth trailed behind that of the major European powers. Between the 1917 and 1929 (both cyclical peaks) Italy’s GDP grew by 2.2 percent per annum (versus 2.5 percent in 1898–1913). Between 1922 (the year when the highest prewar GDP level was roughly regained) and 1929, Italy’s annual growth rate was an impressive 4 percent. The rate was 6.1 percent in 1922–1925, the “liberal” years of the Fascist regime, with exports growing yearly at the tune of 18.7 percent. The first Mussolini government resumed the successful prewar macroeconomic policy stance: the budget deficit was eliminated by 1925, the GDP/debt ratio fell, and the Bank rate was kept at a relatively low level. At the same time, the wartime suspension of import duties on grains remained in effect (Toniolo 1980; see chapter 3). The year 1925 was the political and economic watershed of the Fascist period. Coincidentally, with Mussolini formalizing his dictatorship and with a stock market crash (caused by overtrading by banks), the wheat import

an overview of italy’s economic growth

19

duty was reintroduced together with an increase in overall protection of the so-called heavy engineering and metal-making industries (see chapter 2). Giuseppe Volpi, a high-profile representative of the heavy-industry compound, replaced Alberto De Stefani, an economics professor, as the Minister of Finance. The following year, Mussolini launched his campaign for the revaluation of the lira aimed at the reintroduction of the gold standard. The years 1925–1926 witnessed possibly the most sudden and complete 180-degree turnaround of economic policy in Italian history. From then until the end of the Second World War, the Italian economy was increasingly inward-looking. The Fascist ruling elites did not understand that a medium-sized economy, still relatively backward and prevalently agricultural, lacking mineral resources and dependent on imports for food, raw materials, and advanced technology, could only prosper by competing on world markets and would suffer more than others from the rigors of autarky.

Catch-up Intermission, 1929–1945 Growth continued at a slower pace until the onset of the Great Depression (2.5 percent per annum in 1925–1929). Among the results of our research project (see chapter 6)is an estimation of industrial production in the 1930s, based on hitherto unexplored output and employment sources (Giugliano 2012). It is now known that the Italian Great Depression lasted longer than previously believed: only in 1935 did GDP and industrial output reach the level of 1929, only to fall again in 1936. The crisis would have been more devastating without the measures taken in 1931–1933 to bail out and reform the big universal banks that over the years had turned themselves into something similar to industrial conglomerates holding majority stakes in the largest Italian firms. Swift, effective, and secret action by the state and the monetary authorities prevented the banking crisis from producing the crippling outcomes seen in Austria and Germany (Toniolo 1995). As an initially unintended consequence, the crisis produced one of the most important institutional innovations in the history of the Italian economy: the establishment of the Institute for Industrial Reconstruction (IRI). The state took over control of most of the largest companies from the banks, creating one of the major tools it would use to support reconstruction and growth after the war (Castronovo 2012). During the 1930s, Italy’s secular convergence to the most advanced countries came to a halt: between 1929 and 1939 growth averaged about 1 percent a year, against a Western European average of 1.1 percent. Over the same period the autarkic German economy, despite its collapse between 1929 and 1932, grew by more than 2 percent a year, thanks to the mobilization of the country’s abundant natural resources and to its advanced technology. In Italy, protectionist policies, progressively tightened until reaching risible forms of autarky, the

20

aggregate growth and policy

stubborn overvaluation of the real exchange rate (Toniolo 1980; see chapter 13) especially after the 1931 devaluation of sterling, participation in the “gold bloc,” and tight exchange rate controls go a long way to explain the decade-long halt in Italy’s convergence. Nor did an increasingly dirigiste approach to resource allocation yield results comparable with those of Nazi Germany: a command economy would have required a stronger and more efficient state apparatus and probably coercive measures like those used by the Nazi regime, which Mussolini was unable or unwilling to manage. In contrast with the experience of 1896–1913, income inequality rose in the 1920s (see chapter 8; there are no reliable data for the 1930s). The number of average hours worked per person employed rose in the early 1920s, and then sharply declined after 1925. Elementary school attendance rates declined between 1926 and 1936 and the growth of aggregate accumulation of human capital slowed down (Vecchi 2011; see chapter 9). In the 1930s, the progress in life expectancy somehow slowed and the number of poor increased, as did that of children at work. Despite his far from negligible approval rate by the Italian people, and his obvious claims to the contrary, Mussolini was not as effective in promoting the welfare of the ordinary Italian citizen as his propaganda claimed.

Reconstruction: Setting the Stage for Long-Term Growth Italy’s participation in the Second World War turned out to be an economic disaster. Whereas between 1915 and 1918 GDP had grown on average by 1.9 percent per annum, it decreased on average by almost 10 percent every year between 1940 and 1945, with most of the decline concentrated in 1943–1945, when the country was divided and war was fought on its territory. In 1945, Italy’s GDP stood at the level of 1906. Reconstruction, however, was fast: by 1949 Italian GDP was already higher by 10 percent than in 1939, the best prewar year. It is an accepted tenet in the literature that the “second postwar settlement,” unlike the 1919 Versailles Treaty, was a key factor in setting the stage for the subsequent rapid growth in Western Europe and Japan (see chapters 3 and 4 and references thereof). The Marshall Plan was a crucial ingredient of the settlement, not so much for its size (less than 2 percent of Italy’s GDP in the years 1948–1952) but for easing the “dollar gap” for import of raw materials and US technology, and for lessening the distributional costs of monetary stabilization. The Marshall Plan was also an institutional tool in that, through the Organization for European Economic Cooperation in which Italy eagerly participated, it opened the long way toward international competition and European integration. The policy of progressive, irrevocable opening up to the international economy represented a sharp break with the past. It was a fundamental political choice, upheld even in the face of resistance from powerful economic lobbies, the same that from 1925 onward had advocated autarky (Toniolo

an overview of italy’s economic growth

21

and Salsano 2011). International trade liberalization was set on course by slowly dismantling the autarkic apparatus and progressively moving from bilateral to multilateral payments (Guido Carli, future Governor of the Bank of Italy, was the first chair of the European Payments Union board). As for domestic policies, runaway inflation was stopped in its tracks in the summer of 1947; however, not before it was conveniently instrumental in wiping off most of the government debt. From then onward, the early postwar governments and the Bank of Italy took an “orthodox” macroeconomic policy stance, including the reconstruction of foreign exchange reserves, which in the minds of the main policy-makers had served Italy so well before 1914. This policy was met by fierce criticism not only from the left, but also from the ERP administration (see chapter 5). Then as in future years, domestic market liberalization proved more difficult to realize than opening up to international competition. A comprehensive political economy study of the long-standing preference of Italians for international competition matched by tightly regulated domestic markets has yet to be made. In the immediate postwar years, however, the international free trade stance and the European treaties themselves were largely the initiative of political authorities that were able to overcome the opposition of powerful vested interests. The same political authorities, however, nurtured a good dose of mistrust—resulting from ideology and the experience of the 1930s—of the ability of private enterprise to generate adequate capital accumulation, technology, and productivity growth (see chapter 3). The decision, already made once back in 1937, not to privatize IRI was confirmed. A non-trivial corollary was the decision to keep practically the entire system of financial intermediation in public hands. The financial system remained bank-oriented and strictly regulated (see chapter 17). The stock market stayed thin, oligopolistic, and vulnerable to the speculative forays of a handful of raiders. Overall, the postwar economic institutions corresponded to the idea of a “mixed economy”—characterized by a strong role for the state as producer and regulator—that had inspired the constitutional compromise among the various political parties, which sealed the creation of the Republic in 1947. Chapter 4 argues that “the reconstruction period opened an institutional gap between Italy on one hand and Germany and Japan on the other,” a gap that may help explain the long-run weaknesses of the Italian economy compared with that of her wartime allies. In particular, Germany was better able than Italy to create a Soziale Marktwirtschaft commanding a social consensus on the country’s economic goals that was unattainable in Italy’s more polarized social and ideologic environment. The stock of human capital that the Republic inherited from previous decades was lower than that of the main European countries. Primary education had made huge strides, and by 1951 the illiteracy rate among the population older than fifteen years was down to 15 percent. However, the average number of years of schooling (4.1 years, with large regional disparities)was the lowest

22

aggregate growth and policy

among the twelve countries of Western Europe, and despite great advances, Italy continued to lag behind in education attainments over the next sixty years.

The Golden Age A good part of Italy’s secular convergence toward more advanced countries occurred between 1950 and 1973. In the previously mentioned context of rapid opening-up to international competition coupled with protection and regulation of the domestic market, per capita GDP grew by an average of 5.3 percent per year between 1950 and 1973. Industrial production increased annually by 8.2 percent and labor productivity by 6.2 percent. Together with Germany and Japan, the other former Axis powers (see chapter 4), during this quarter of a century Italy converged strongly toward the high-income countries. In 1950– 1973, Italian GDP per person income rose from 38 to 64 and from 50 to 88 percent of, respectively, that of the United States and United Kingdom. In 1951, Italy’s product per hour worked was only 46.5 percent of the United Kingdom’s; it was 101.6 percent in 1973. Total factor productivity (TFP) grew at the fastest rate in Italian history and explained more than half of GDP growth (labor and capital inputs accounted, respectively, for about 17 and 29 percent of GDP growth; see chapter 7). Italy’s growth in the Golden Age cannot be explained only by catch-up. Many factors in Italy’s growth were common to all of Europe but some were idiosyncratic to the Italian Peninsula. The technologies of the day, largely adaptations of the Fordist model, were well suited to a country endowed with an abundant labor force, possessed of basic education, and a relatively small but sophisticated cadre of engineers. Up to the mid-1960s, the plentiful availability of labor ensured that wages did not outstrip productivity growth. A large domestic market allowed the exploitation of economies of scale, especially in the production of durable consumer goods, which in turn generated competitive advantages in export markets. During this period, public enterprises under excellent middle managers were an engine of investment and technical progress. IRI produced at internationally competitive prices intermediate goods, such as iron and steel, in which the private sector had historically been uncompetitive and in need of protection from 1887 onward. The impact of international technology transfers, including rapidly increasing foreign direct investment, was mainly caused by the ability of Italian firms to acquire and diffuse technical knowledge through imitation, reverse engineering, and adaptation (see chapter 14). This was true not only of the large firms but also of the medium-size enterprises organized in “districts,” one of the most peculiar and successful Italian institutions (see chapter 5). The rapid transformation of the Italian economy is witnessed by changes in the composition of export trade, by then largely based on medium-technology

an overview of italy’s economic growth

23

manufacturing (see chapter 12), and by the fact that most manufacturing employment (60 percent in 1970) was concentrated in capital-intensive industries, such as automotive, chemicals, heavy engineering, steel making, and shipbuilding (see chapter 16). In the time-span of less than a generation, the life of the average Italian was culturally, socially, and economically transformed. This was also as the result of mass migration from countryside to city in Italy and across the border, this time mostly to neighboring European countries (see chapter 10). If aggregate caloric intake was already adequate, the daily diet became more diversified and, what matters most, malnutrition was almost eradicated and absolute poverty radically reduced (Vecchi 2011). Houses became larger and healthier, with universal access to sanitation and the rapid diffusion of central heating. In collective memories, the Golden Age stands out as the triumph of mass consumption in consumer durables; the small Fiat 500 and 600 crowned the ordinary person’s dream of private transportation (Toniolo and Vecchi 2010). The progressive improvement in the public provision of health services and retirement pensions together with longer-term job opportunities lowered workers vulnerability to poverty. Income distribution became steadily more egalitarian (see chapter 8). The North-South income gap narrowed for the first and only time since unification (see chapters 20 and 21). Overall, Italy’s postwar economic arrangements, based on extreme openness to foreign trade, accompanied by ample safeguards for domestic producers naturally sheltered from international competition, and on an important role of the state as producer of goods and services, proved capable of generating fast growth in an initially still backward economy. Things began to change toward the middle of the 1960s, when the North approached full employment. It became harder to meet increased wage demands, while at the same time maintaining high levels of investment. Endogenous innovation was growing too slowly at a time when the advantages of backwardness, typically unlimited supply of labor and imported technology, began to fade. By the end of the 1960s it was already clear that Italy had to adapt institutions, financial markets, training and research, and the economic role of government to the characteristics of a no-longer backward economy (Rossi and Toniolo 1996). Little was done, however, and this omission weighed on subsequent growth.

The Silver Age, 1973–1990 In the early 1970s the pace of growth slowed abruptly in Europe and the United States. In Europe, the impetus from such key factors of postwar growth as reconstruction and the progressive European market integration petered out. A long phase during which cheap raw materials had maintained favorable terms of trade for the West came to an end. In the initially more backward countries,

24

aggregate growth and policy

the advantages of an abundant supply of labor faded. Convergence itself reduced the productivity gains that could be wrung from technology transfers (Crafts and Toniolo 1996a, 2010; Toniolo 1998). All this makes the decline in productivity growth comprehensible, although the reasons why it was so sharp and sudden and occurred simultaneously in every country west and east of the Iron Curtain are not entirely clear. The negative expectations caused by the end of the Bretton Woods cooperative monetary order and by the first oil shock are only part of the explanation. The Italian economy, now well integrated into the world economy, participated in the events that affected the latter, notably the negative productivity shock, although with an evolution that reflected its own peculiar features. Between 1973 and 1992, per capita GDP grew at the respectable rate of 2.5 percent per year. The convergence on US per capita GDP reached 76 percent(up from 65 percent in 1973). The convergence on Western Europe was virtually complete: in 1992, per capita GDP in Italy was equal to that in Germany and the United Kingdom. Convergence of product per hour worked was also almost complete. If the convergence on US GDP came to a halt at about three quarters of the way, the productivity story is different. The productivity convergence needs to be stressed, not least in view of the contrast with the following twenty years. Hourly output in the Italian economy rose from 66 percent of the US figure in 1973 to 86 percent in 1992. In other words, per capita GDP grew more slowly than labor productivity; the Italian labor force participation rate was lower than the US rate, and the gap widened over time. This feature was common to all the continental European economies, but it was especially marked in the case of Italy. It is not clear whether this was a preference for free time over work or the result of an inability on the part of the Italian economy to meet the demand for employment. The continuation, albeit at a slower rate, of convergence in the twenty years that followed the Golden Age was achieved with a succession of policies that in the short term stabilized cohesion, a key ingredient of the “social capability for growth,” and boosted aggregate demand. The needed structural reforms were half-baked, because weak political leadership was unable to effectively counteract the weight of distributional coalitions. The cost of Italy’s inability to change its institutions as the advantages of backwardness faded and international competition changed the conditions for growth sustainability would come to a head in the 1990s. From the “hot autumn” of 1969 and throughout the difficult 1970s, a series of off-the-cuff measures were adopted to expand welfare provisions and provide indiscriminate subsidies to firms. Political micromanagement in resource allocation grew in importance, particularly in public enterprises. In the process, the allocative effectiveness of the credit system decreased (see chapter 17). Double-digit inflation was higher and lasted longer in Italy than in the competitor countries. If an increasing number of Italians acquired secondary and

an overview of italy’s economic growth

25

tertiary education, the quality of the school system deteriorated. The time required to obtain justice in civil and administrative courts lengthened (see chapter 19). Growing signs of frailty in large private-sector firms accompanied the weakening of public enterprises. The labor market became more rigid. Italy remained locked into a product specialization concentrated in sectors marked by low or medium technology (see chapters12 and 16), although there was a continuous drive for quality in many branches of traditional Italian products. Employment came to be increasingly concentrated in small and medium-sized firms, more flexible and adaptable but less capable of research and development. It was probably thanks to this flexibility, another ingredient of the country’s “social capacity for growth,” that Italy succeeded in maintaining its share of the world export market. However, although the system’s ability to sustain growth in the long run weakened, and although domestic terrorism and social turmoil shackled the country, the overall performance of the Italian economy during the 1970s was surprisingly good. Despite 1975 being the first postwar recession year with real GDP falling by more than 2 percent, over the decade real GDP grew annually at the respectable rate of 3.4 percent (3.6 percent between 1973, the year of the first oil shock, and 1979). Income inequality decreased at the fastest pace than in any decade since unification (see chapter 8). The good performance of the real economy in the 1970s can be explained by the expansive fiscal and monetary stance adopted to soften the economic and social impact of the two oil shocks. The cost was two-digit inflation, and foreign exchange depreciation. Social tensions were oiled by increasing welfare provisions (see chapter 3). The resulting budget deficits were partly monetized, hence the relatively moderate growth in the ratio of debt to GDP. The jury is still out on the costs and benefits of Italy’s macroeconomic policy in the 1970s: it has been argued that it kept profits and demand high, thereby supporting employment and lowering the costs of the subsequent disinflation in the 1980s. The opposite opinion holds that the neglect of the fiscal constraint and the heavy-handed state management of resource allocation, including credit (see chapter 17), simply postponed and thus made more costly the necessary macroeconomic adjustment. At the end of the 1970s, however, consensus built up about bringing inflation under control. Three main decisions were made: (1) participation in the European Monetary System, (2) the introduction of a form of incomes policy, and (3) repeal of the agreement between the central bank and the Treasury whereby the latter bought unsubscribed government bonds at every auction (the so-called divorce between the Treasury and the Bank of Italy). It was hoped (in vain as it turned out) that the latter decision would also increase fiscal responsibility by making policy-makers come to terms with budget constraints. Disinflation came at the cost of sluggish growth (only 0.8 percent per annum) in 1980–1983. GDP growth then resumed at a quite sustained pace (3.1 percent annually between 1983 and 1990). By 1984, social tensions had receded

26

aggregate growth and policy

and governments could count on larger and more stable majorities. A window of opportunity opened to address Italy’s macroeconomic problems. Freed from previous obligations toward the Treasury, the Bank of Italy succeeded in taming, if not entirely defeating, inflationary expectations. The government, however, did not seize the opportunity afforded by growing GDP to stabilize the debt-to-GDP ratio. The overall budget deficit, which averaged 7.6 percent of GDP in the 1970s, rose to 10.7 percent as the average for the 1980s. Public spending expansion was amplified by interest payments, because lower inflation meant high real borrowing costs to the government. Rather than being stabilized, in the 1980s the ratio of the debt to GDP rose from 56 to 94 percent (see chapter 18). No similar pattern is to be found in such a short period of time in any other developed country after the Second World War. The roots of the debt crisis of twenty years later are to be found in the missed opportunity of the 1980s. In the medium term, however, everything looked pretty shiny. Productivity convergence with France and Germany was completed, the United Kingdom was “overtaken,” and the United States was closely approached. Distributional equality reached its all-time highest level in the mid-1980s. Social tensions were reduced to a “physiologic” level, and public opinion was engulfed in a wave of optimism.

Divergence in the “Second Globalization,” 1990–2011 From Convergence to Divergence The secular process of Italy’s catch up came to an end in the early 1990s. In the following two decades, the growth in GDP per person was roughly the same as that of the first decades after unification and in both periods Italy, rather than catching up, lost ground to the most advanced countries. In 1992, Italy’s GDP and labor productivity stood at 76 and 86 percent of the respective US values; by 2010 the ratio of Italian to US per capita GDP was back to its 1973 level (64 percent). To be sure, during the same period of time the growth rate slowed down in several other advanced countries, in an international economy seeing not only the rapid emergence of the leading Asian countries but also renewed US economic vitality (Rhode and Toniolo 2006). Japan and Germany, in particular, turned from rapid catch-up to a divergence: their gap with the United States widened by about a dozen percentage points. Italy’s performance, however, was particularly disappointing from about the year 2000 onward (Table 1.5). Between 1992 and 2000 Italian per capita GDP annual growth rate (1.7 percent) was slower than the average for the twelve Western European countries,

27

an overview of italy’s economic growth Table 1.5 Economic performance, Italy and advanced countries, 1992–2010 1992–2010

1992–2000

2000–2007

2007–2010

TFP 1990–2007

Italy

0.5

1.7

0.7

– 2.2

0.7*

France

1.0

1.7

1.1

– 1.0

n.a.

Germany

1.2

1.5

1.3

0.1

1.5

Japan

0.7

0.7

1.5

– 1.2

1.1

United Kingdom

1.8

3.1

2.1

– 1.9

0.7

United States

1.5

2.7

1.4

– 1.2

0.9

Sources: http://stats.oecd.org/ Chapter 7, and Data Appendix. Average annual growth rate in GDP per person and Total Factor Productivity. * 1993–2007.

but faster than Germany’s and Japan’s. It was only 0.8 percentage points below the secular catch-up trend (1896–1992) and above the growth rate realized during the so-called Giolittian era (1900–1913). Output per worker and TFP growth, however, progressively slowed down (see chapter 7). The years of 2000–2011 were technically a lost decade, with total GDP in 2011 a mere 1.1 percent higher than it had been ten years earlier, and still about 5 percent lower than it was in 2007 (in the previous major depression, the 1929 GDP level was reached again in 1935). For a country that was the second largest industrial producer in the euro area, especially serious was the weakness of the manufacturing sector, where output contracted by about 4 percent between 2000 and 2007. Italy’s weak performance since the early 1990s and particularly in the first decade of the century is largely caused by low productivity growth. In 1995, Italy’s output per hour worked reached 90 percent of the United States, the narrowest gap ever. From then onward, Italy’s productivity growth did not keep pace with that of the productivity leaders (Table 1.6). Between 1995 and 2000 the increase in hourly output was about half that of the euro area. TFP growth was also abnormally low. However, the number of hours worked continued to increase, up to the onset of the crisis in 2008. This was partly caused by the increase in employment deriving from important, albeit incomplete, labor market reforms. After peaking in 1995, in less than ten years unemployment returned to levels similar to those in the early 1980s. Investment also continued to grow in the first decade of the 21st century; “the demand for factors of production continued to expand satisfactorily” (De Novellis 2012, 26). An economy that was somewhat recovering between 2005 and 2007 was hit by a crisis that in two years resulted in the loss of more than 5 percentage points in GDP per person, a decline comparable with that of the Italian Great Depression of the early 1930s.

28

aggregate growth and policy

Table 1.6 Labor and Total Factor Productivity Labor Productivity *

Total Factor Productivity†

Agriculture

Industry

Services

Total economy

Total economy

1993–2010

3.0

0.8

0.6

0.9

0.4

1993–2000

5.9

1.8

1.0

1.6

1.3

2000–2007

1.4

0.3

0.2

0.3

0.2

2007–2010

1.0

−1.0

−0.1

−0.3

−1.0

Sources: Elaborations from Chapter 7 and Data Appendix. Average annual growth rates, %. * Estimates built using the FTE measure of labor. Valued added is the October 2011 Istat data release. It includes housing and government services. † TFP estimates built using the FTE measure of labor and the capital services measure for capital. Value added from October 2011 Istat data release has been netted of housing. Factor shares are variable, based on actual wage data.

Losing Growth Capability From the mid-1990s onward, Italy seems to have somehow lost its “social capacity for growth.” As in the first decades of its history as a united country, in the most recent years divergence rather than catch-up characterized the Italian economy. Why did Italy’s robust secular convergence stop at the turn of the twentieth century? This question inspired the launch of the research project that resulted in this book. Each individual chapter provides building blocks for those who try to provide a comprehensive answer to this complex question. What follows is my own attempt to do so. Needless to say, no parallels can be drawn between the 1990s and 2000s and the years of divergence that followed unification in 1861, besides the similarity in the aggregate growth rates. In the 1860s, Italy was a backward country in a rapidly integrating Atlantic economy, trying to engineer a typical catch-up process by importing technology and exporting labor-intensive goods and labor itself to the most advanced countries of the time, once the disadvantages of foreign domination and internal market fragmentation had been overcome. In 1992, Italy’s economy was possibly slightly larger than the United Kingdom’s and the seventh largest in the world (when measured at purchasing power parities; Maddison 2001). What matters most, it was very close to the productivity frontier with a product per hour worked at about 90 percent of the US level (1995). In the international economy, major changes had been slowly building up in the 1980s, overlooked by Italian producers, trade unionists, and policy-makers alike. Between 1989 and 1992, a bundle of major shocks suddenly changed the playing field: a technologic revolution; a rapid opening of the world markets for

an overview of italy’s economic growth

29

goods, services, and capital (known as the “second globalization”); an acceleration of European integration leading to the creation of the Monetary Union; and the irreversible establishment of two giant countries, China and India (together almost two-fifths of the world population), as major players in the international economy. Each of these shocks required rapid adjustment, first and foremost cultural, by entrepreneurs, workers, and political decision makers (i.e., the whole of society). Not many European countries were ready to carry through this adaptation. The Scandinavian countries succeeded quite well, others found the adaptation to the new environment harder. Italy was the large European country that was possibly the least successful in making the productive, social, and cultural changes needed to turn each of the shocks from fetters into growth opportunities. In the search for the reasons for Italy’s diminished “social capacity for growth” one might usefully try to distinguish between the weakening of growth factors that sustained growth until the 1980s, and long-standing weaknesses of Italy’s economy, institutions, and society that did not prevent (some even fostered) the previous convergence process but which, in today’s conditions, have become obstacles to growth.

Missing Growth Factors Seen in a long-term perspective, three major changes took place in the Italian economy from the late 1980s onward that stand out as potentially affecting its growth rate: (1) a decrease in size, scope, and productivity of the large firms, (2) an increase of the debt-to-GDP ratio over and above the supposedly “critical” threshold of 90 percent, and (3) a shift from under evaluation to overvaluation of the real exchange rate of the currency. Chapters 15 and 17 highlight the quantitative and qualitative progressive weakening of Italy’s large firms. More than other countries, Italy has always been the country of small firms: at the beginning of the 1960s Italy’s average firm size was forty-nine workers, compared with France’s seventy and Germany’s seventy-eight. As often emphasized, the resilience of Italy’s system of small companies is one of the idiosyncratic features of its economy: it gave and continues to give Italian industry a flexibility that is one of the country’s competitive advantages. Small companies, however, drew major advantages, not only as subcontractors, from the existence of a sufficient number of efficient large enterprises as the main producers of applied research, with fallouts on the rest of the economy, and vehicles for foreign direct investment, probably the main channel of technology transfer. In recent decades, the average firm size has declined in every European country, but the phenomenon was especially pronounced in Italy. In 2008 the average number of workers in Italian

30

aggregate growth and policy

firms was half the average of the five leading EU countries. The downsizing of large companies entailed a reduction in Italy’s research and development capability at a time when the rapid development and adoption of new techniques was more important than in the past. Moreover, the quality of Italy’s larger companies also declined: between 2000 and 2005 output per hour worked contracted by about 20 percent in firms with more than 500 workers, more than in the rest of the system. The short-lived recovery of 2005–2007 was driven by an improvement in the productivity of large firms, thus confirming their systemic role. High public expenditure and a large public debt growth emerged in the last two decades as another potential cause of lower growth. Although Italy was always a high debt country (the debt-to-GDP ratio exceeded 90 percent for seventy-five years since unification), Italy’s postwar growth took place in a low debt and public expenditure environment. In 1979, the debt-to-GDP ratio was about 60 percent (see chapter 18). By the end of the 1980s the ratio had risen to 90 percent and reached 105 percent in 1992. Although it is hard to set debt thresholds, valid for all times and countries, above which aggregate growth is negatively affected, a number of economists argue that a debt-to-GDP ratio above 90 percent is likely to reduce growth (this is also the view emerging from chapter 18). A large public debt is a drag on growth, drives up interest rates, requires heavier taxation, and often as in Italy over the past decade results in reduced public investment in research and infrastructure. In the long run it is a threat to the welfare state, potentially undermining social cohesion. Italy’s economic history shows that the country is more “debt tolerant” than most other countries, and has an excellent reputation for paying interest and principals. Nevertheless, a prolonged period of high debt, with no discernible downward trend, coupled with tax rates higher than 40 percent (high in historical and comparative perspective) was not seen in peacetime since the first decades after unification, another similarity between the two “tails.” In the late 1890s the debt-to-GDP ratio was nearly 120 percent. A credible and time-consistent commitment to debt reduction was one of the macroeconomic conditions for convergence growth that took place in the following years. A third “novelty” of the 1990s and 2000s was the overvaluation of the real exchange rate, in contrast to Italy’s previous economic history when it was almost always undervalued or close to parity. Chapter 13 argues that undervaluation fosters growth by shifting resources from protected sectors to high productivity growth export-oriented industries. The authors of chapter 13 maintain, however, that this effect became weaker as the economy grew. It was probably a fairly significant factor until the start of the 1970s, much less so thereafter. It is thus unlikely that the historical novelty of an overvalued real exchange rate slowed down growth after 1993. More than a cause of the disease, it was probably a symptom. As the nominal exchange rate rose, Italy did not respond with adequate product and labor market policies aimed at keeping unit costs in line with its euro area competitors.

an overview of italy’s economic growth

31

The Impact of Old Weaknesses Most of the chapters in this book highlight cultural, social, institutional, and economic weaknesses of Italy’s long-term growth since the unification of the country. Many of these weaknesses (North-South divide, low human and social capital, weak competition in the domestic product and labor markets, financial intermediation defined a “petrified forest,” narrow capital markets, inefficient public administration, and insufficient research and development)show an incredible secular persistence. Many of them also figure prominently in the literature as the causes of Italy’s unsatisfactory economic performance since the mid-late 1990s. To argue that ossified and never corrected weaknesses explain Italy’s divergence in the 1990s and 2000s one must explain why the same weaknesses were not an obstacle to rapid catch-up growth until the last decade of twentieth century. This explanation needs to account for the fact that the 1990s witnessed a good number of reforms trying to improve on at least some of the previously mentioned weaknesses (see chapters 3, 17, and 19). In 1990–1993 a comprehensive banking reform, the first since 1936, set in motion the privatization of the savings banks and the IRI-owned “big three” creating conditions for the emergence of large banking groups in an open and market-oriented environment, which improved the allocative efficiency of the financial system (see chapter 17). In 1997 and 2003, legislation was introduced that considerably increased labor market flexibility (at the cost, however, of increasing the gap between protected “insiders” and “outsiders”). Important changes were introduced in the university system. In 1998, Italy’s product markets were the fourth most tightly regulated among those of the Organization for Economic Cooperation and Development (OECD) countries (after Poland, Turkey, and Greece); they are now less regulated than those of France, Belgium, and Austria. Even more impressive have been improvements in administrative regulation and overall “barriers to entrepreneurship” (Wölfi et al. 2009). Some of the chronic weaknesses of the Italian economy were addressed and partly solved after 1992, even though the implementation of the reforms met with grassroots’ resistance that rendered them often incomplete, ill-applied, and distorted. Moreover, at the turn of the century the reform drive ran out of steam, and Italy has continued to slip downward in the “doing business” rankings of the World Bank (www.doingbusiness.org/). To understand why old deficiencies are now much more of a drag on the economy than they were before, despite non-marginal steps taken to correct some of them, two obvious considerations are in order. First, the international environment has been transformed by globalization, technology, and the European single market and currency. Second, the very success of long-run convergence has altered the conditions that make its continuation possible. Once a country (almost) closes the productivity gap with the most advanced countries, new ways to economic growth must be found. As we have seen, an

32

aggregate growth and policy

economy that is far from the technology frontier can exploit its backwardness to grow faster than the leader by adapting imported technology to its own factor endowment and by shifting labor from low- to high-productivity sectors, but in the proximity of the frontier, other growth factors become crucial: apt institutions (in the broad sense), research, human capital, and physical and intangible infrastructure. There is no need to repeat here the list of Italy’s long-standing and never-corrected economic weaknesses: they are measured and discussed at length in the various chapters of this book. Most of them proved to be relatively harmless in a backward country engaged in catch-up growth in a context of mild protection from international competition. They became more binding as Italy moved close to the productivity frontier, and turned into heavy fetters with the “second globalization,” the single European market, and the new “encompassing technology” (ICT) of the twenty-first century. Italy ranks twenty-fourth out of the twenty-six countries for which the OECD has produced an index of the capacity to withstand globalization based on regulations, education, labor market flexibility, employment programs, and environment for innovation. To a good extent, Italy missed the chance to exploit the ICT revolution, the present era’s “general purpose technology,” to increase productivity, particularly in the service sector, now by far the largest share of the economy. This is a technology that by its nature cannot thrive in an overregulated environment. Italy’s timid deregulation was not sufficient to encourage the spread of information and communication technology, nor did it help small firms to grow in size. Above all, the transfer and the diffusion of ICT require more human capital than Italy had built up over the decades. The case of education (see chapters 8, 9, and 11) is a good illustration of how a long-standing weakness became a serious drag on growth only from the 1980s or 1990s onward.

Education, a Case in Point In Italy’s 150 years as a unified nation, school attendance among the population aged fifteen to sixty-four has risen from an average of just over one year to 10.8 years, but this enormous progress has not brought convergence in this field with the other main European countries. Some of these countries, themselves not richer than Italy in 1861, had already then significantly higher literacy rates. Others, such as Finland, began with even lower rates of school attendance but overtook Italy in the second postwar years. Italy’s school attendance rate is still the lowest in the twelve countries of Western Europe. According to de Mauro (1991), a distinguished linguist and a former minister of education, the (relative) neglect of universal education was the gravest sin committed by Italy’s elites from unification to the present day.

an overview of italy’s economic growth

33

It is likely that the relative scarcity of human capital, as gauged by school attendance, did not negatively impact on the rate of catch-up from the initial backwardness to a productivity level close to the frontier. Some scholars maintain that it may even have favored growth, because the kind of skills required in the catch-up phase derived from practical and informal know-how passed on tacitly on the job, rather than acquired by formal education. To simplify a complex issue: to successfully import and adapt foreign technology Italy needed a good pool of highly trained engineers and a workforce with basic education and skills acquired in agriculture, construction, or craftsmanship. It was well endowed with both. Close to the frontier, in a globalized world, low human capital is a barrier to the diffusion of ICT, to endogenous innovation, and to the adoption of state-the-art production processes. Moreover, low human capital makes it more difficult to generally “cope with globalization,” including through the understanding of different cultures; it is also a barrier to the formation of the intangible social capital of trust and sense of belonging that is conducive to community cohesion (see chapter 11). Italy’s relatively poor human capital may go quite a long way in explaining the entrenched reluctance to change of a large portion of the Italian electorate. In strictly quantitative terms, it is encouraging that the fastest improvement in average years of schooling in the population ever achieved by Italy came in the first decade of the twenty-first century: from 8.3 in 2001 to 10.8 in 2010 (see chapter 9), equal to an increase by 2.9 percent per year (3.6 percent in the South), compared with 1.7 percent over the previous thirty years. More important still, the share of population aged twenty-five to sixty-four that completed tertiary (university) education increased from 9.4 percent in 2000 to 14.5 percent in 2009 (see chapter 9). The rate of university enrolment in the relevant age group was in 2009 close to the OECD average, as was the dropout rate. In this area enormous progress was made in the first decade of the twenty-first century, despite low GDP growth. Given the low starting level, however, it will take several years for the ratio of university graduates in the population of working age to come close to the OECD average and make an impact on growth. Moreover, the quality of Italian education has ample room for improvement: the OECD PISA mathematics score of Italy’s secondary school students is not statistically different from that of the United States (a notoriously poor performer) and much below the European, let alone the Asian, averages (www.pisa.oecd.org/). Education is not the only area where Italy was improving in the years of low GDP growth. Between 2000 and 2009 life expectancy at birth increased by two full years, from 79.8 to 81.8, and infant mortality decreased from 4.3 to 3.7 per thousand (OECD 2012). The number of deaths in car accidents decreased from 5,669 in 2006 to 4,090 in 2010 (http://dati.istat.it/). Poverty, both absolute and relative, remained constant over the years, including the 2008–2009 crisis. As in 1861–1896, low growth and divergence did not mean lack of improvements in some crucial quality-of-life indicators.

34

aggregate growth and policy

The Roots of Sluggish Growth The points made in this final section can be summarized as follows. In the 1980s, Italy’s productivity continued to converge to that of the leaders, if at a lower pace than in the previous “Golden Age.” Growth was sustained by a lax fiscal stance (see chapter 18) and, to a small extent, by the undervaluation of the real exchange rate (see chapter 13). Neither option was available in the 1990s and 2000s. Moreover, a weakening of the public and private large enterprise as an engine of growth had taken place from the 1970s onward (see chapters 14–16). Both the high debt-to-GDP ratio and a weaker ability to generate research and development and economies of scale are likely to have impacted negatively on growth in the post-1992 competitive and technologic environment, although it is difficult to quantify their impact. In addition, and despite the reforms of the 1990s, in the new international environment a number of long-standing weaknesses of the system came to weigh much more heavily than before on the performance of the Italian economy. According to the The Global Competitiveness Report by the World Economic Forum, “Italy continues to do well in more complex areas measured by the Global Competiveness Index, particularly the sophistication of its businesses ( . . . ) producing goods high on the value chain, with one of the best world’s clusters (2nd). Italy also benefits from its large market size”(Schwab 2011,27). However, the Report notes that such strong points are hampered by structural weaknesses, such as those briefly highlighted previously, and by high levels of corruption and organized crime. Overall, “Italy ranked 88th (out the 142 countries considered in the Report) for is institutional environment.” The synthetic quantification of institutions is a complex and controversial task: a good case in point is the different appreciation by compilers of national rankings of Italy’s labor market efficiency. Its assessment by the previously quoted OECD team (Wolfi et al. 2009) is strikingly more favorable to Italy than that by the World Economic Forum. Nevertheless, most of the chapters in this book point in the direction of widespread and long-standing institutional failures. The adoption of ICT is a good case in point. As argued by Crafts (2011b, 8–9):“The diffusion of this new technology was hindered by oppressive regulation and shortfalls in human capital by comparison with the European leaders in the take up of ICT(Conway et al., 2006).[ . . . ] In the retail sector, where the potential for ICT to raise productivity was very considerable, it is clear that productivity performance was impaired by regulation; barriers to entry and mark-ups in retailing remained high on average with adverse consequences for TFP (Daveri et al., 2010). However, in districts where competition was stimulated by the 1998 regulatory reform both ICT investment and labor productivity increased (Schivardi and Viviano, 2011).”If ICT diffusion is aided by complementary investments in intangible capital and in high-quality human capital, shortcomings in these areas are ultimately caused by institutional failures.

an overview of italy’s economic growth

35

Why has Italy been so stubbornly resistant to deep institutional change? Why have far-reaching reforms been rare in the country’s history? The authors of chapter 3 offer four explanations: (1) the severity of the division between groups and parties, paradoxically increased by the demise of ideologies and of the traditional political parties, which only strengthened Olson’s (1982) distributional coalitions; (2) the difficulty, on the firm’s side, to efficiently allocate entrepreneurial talents because of a closed corporate governance centered on family controlled and pyramidal business groups; (3) the lack of political leadership in reforming the public sector (education, justice, and security are major growth factors, and their weakness is particularly crippling to the southern economy); and (4) an overall institutional (and constitutional) design that renders substantial policy decisions difficult to make. These weaknesses have ancient roots. The institutional design of the new kingdom in the aftermath of the unification was hastily assembled and, some argue, focused primarily on market unification, thus partly missing the opportunity offered by the caesura of 1861 to build the institutional infrastructure of a modern state, which besides Piedmont hardly existed in the Peninsula (see chapter 19). The institutional and political weakness of liberal Italy was exposed immediately after the First World War, another major caesura, giving way to populism and dictatorship. The “modernizing” features of the latter came from agencies created outside of the official state bureaucracy (Melis 1996), which was inefficient, pervasive, and swollen in the Depression to enhance middle-class consensus for the regime. Finally, if the second-postwar reconstruction was the result of good leadership and national cohesion behind the surface of fierce political competition, it is possible that Italy’s treatment by the occupation power, less invasive than in Germany and Japan, might have been a curse in disguise (see chapters 3 and 4) as a missed opportunity for a more radical institutional change. Italy, however, is not unable to change, particularly when cornered by serious crises. In the mid-1890s Italy was in the throes of the potentially devastating epilogue to a banking and balance-of-payments crisis, permeated with episodes of corruption and wracked by fierce political struggle. The country survived thanks to fundamental institutional changes, including the creation of the Bank of Italy, and—in the wake of the shocking military defeat on the battlefield of Adowa—the recovery of national cohesion, which permitted the design of timeconsistent economic policies. In the early 1930s a group of technocrats saved the banking system from the fate that befell the Austrian and German banks. After the war, Italy dug itself out of the material, moral, and political rubble thanks to a political elite farsighted enough to forge a compromise that produced the constitutional pact for the Republic. In the 1970s and again in the early 1990s Italian society found the strength to overcome the crisis and introduce relevant reforms. Unfortunately, institutional reforms were almost ever imperfect and incomplete, and sometimes watered down after the immediate threat was overcome. The truce among the various distributional coalitions often did not last

36

aggregate growth and policy

long enough to allow institutional reforms to permeate society. This is likely to change in the medium term as international competition and conditionality create incentives for institutional reforms. Moreover, and more important, the impact of the recent acceleration in human capital formation on ever-larger strata of the population will not only weaken negative aspects of Italy’s culture but encode individually learned techniques and experiences into organizational routines that are possibly more important than formal institutions. More than in other countries Italy’s future crucially depends on the quality and diffusion of education.

Acknowledgment I am grateful to the researchers and staff of the Economic History Research Division at the Bank of Italy for their invaluable support and to Alfredo Gigliobianco, Claire Giordano, and Matteo Gomellini for useful comments.

Chapter 2

ITALY AND THE FIRST AGE OF GLOBALIZATION,  harold james and kevin h. o’rourke

Introduction: Italian Policies and Performance in a Comparative Perspective There is a natural tension in economic history between the tendency of economists to generalize, and the insistence of the historian on what is specific to particular times and places. A comparative perspective then becomes particularly helpful in seeing what features of a country’s economic development are specific to that country, and in placing this history within a broader international context. It is natural for economic historians to blame their country’s failings on country-specific institutional weaknesses or policy mistakes, or to explain their country’s successes in a similar manner. However, when one looks at the broader international context, one may see similar behavior elsewhere, suggesting that at least to some extent, these failures or successes may be part of a much bigger story.

38

aggregate growth and policy

This chapter demonstrates that such a comparative perspective is indispensible when considering Italian economic policy and performance in the years between unification and World War II. In particular, Italian trade policy, monetary policy, and industrial policy were not that unusual in the context of the time, and particularly when one remembers that Italy was a poor, capital-scarce economy trying to deal with the challenges that all European countries faced during this period: how to industrialize, how to become or remain militarily competitive, and how to grapple with the challenges of globalization. The fact that Italy was capital-scarce is crucial in explaining its trade policy, its monetary and financial policies, and its industrial policy, with all of these policies having ramifications for each other. During the period discussed here, Italian trade and monetary policies were by no means out of the ordinary, however beneficial or harmful they may have been. If an Italian specificity can be located, this is more in the field of industrial policy, particularly during the interwar period, and one can argue that these innovations helped to set the stage for the Italian postwar miracle, if at an eventual cost. However, this development can also be understood by appealing to the interaction between the international economic situation of the time, and banking and industrial systems that were themselves the product of Italian capital scarcity. To understand what follows, one needs to start at the beginning, and ask what were the challenges facing the newly unified Italian state. The Kingdom of Italy was established in 1861, just at the time when the world was beginning to grapple with a new and intensive phase of globalization. Declining transport costs led formerly sheltered sectors to become exposed to international competition for the first time, including grain, olive oil, and citrus fruits, all important in the Italian context. Market integration also made it more difficult for peripheral economies to industrialize, because they were faced with low-cost competition from the expanding industrial core. In those parts of the world where local populations were deprived of the ability to determine their own trade policies, the nineteenth century was a period of deindustrialization, a dangerous phenomenon at a time of increasingly industrialized warfare. Politically, the late nineteenth century saw the beginnings of a rudimentary system of international economic governance. The 1860 Cobden-Chevalier treaty set up the basic institutional mechanism in the form of most-favored-nation clauses that would dominate trade negotiations between states for more than a century. In 1863, an International Postal Congress in Paris began to work on the integration of local and national postal systems. The globalization of the 1860s was not just a question of economics and economic governance: entertainment, ideologies, and humanitarianism all became internationalized. In the middle of the nineteenth century, the world was galvanized by the transatlantic cable and the ubiquitous railroad; by the music of Verdi and Wagner; and by the reforms of Cavour, Bismarck, Gladstone, and Lincoln. The battle of Solferino in the war of 1859 produced not just the final impetus for Italian unification but also the International Committee of the Red Cross, as the outcome of the experience of

italy and the first age of globalization, 1861–1940

39

the Genevois businessman Henri Dunant in trying to care for the wounded on the battlefield. Some commentators see the nation-state of the mid-nineteenth century in Europe as a response to the challenges of globalization. There was an external and an internal challenge: how could political societies be strong enough to compete with existing and well-established nations, and how could they overcome internal divisions. The developments of the 1860s were the outcome of vast and bitter civil wars. One should place the unification of Italy and the nearly simultaneous creation of the German empire alongside the drama of the much bloodier but also highly divisive American Civil War. In each case, a more industrialized north defeated a rural, and perhaps more romantic, south. Italian and German businessmen, thinkers, and politicians saw a need to catch-up with a new model of preeminence and power offered by Great Britain. Italy had included territories with the highest per capita income in the world between 1300 and 1600, but since then had undergone a significant relative decline, associated with absolute stagnation (Maddison 2010). Britain, however, had some rather unique advantages: above all it had security at a relatively cheap price because of its island location. For the small and comparatively dynamic ancien régime states that lay on the old historic Rhineland and Alpine trade route from the North Sea to the Mediterranean, such as the Netherlands, Tuscany, Venezia, or Baden, there was no such easy answer. Defense was too expensive, and territorial expansion would produce additional costs that could not be covered from additional revenue. Unconventionally shaped states had a greater incentive to change the status quo: Prussia, an odd amalgam of low-quality agricultural land with some rich territories in the west of Germany, in the Rhine corridor; or Piedmont, a prosperous and heavily French influenced territorial state linked to a large and poor Mediterranean island. These two states already incorporated huge contradictions, between rich and poor, between manufacturers and farmers, and between Protestant and Catholic. An expansion might be a way of reducing or relativizing these differences by setting them in a larger context. When the unconventional Italian and German state had triumphed, when Cavour and Bismarck stood supreme, there was a clear but problematical legacy of the way in which unity had been thought and fought. As Massimo d’Azeglio famously put it, Italy was made: now it was time to make Italians (“L’Italia è fatta, ora restano da fare gli italiani ”). A developmental strategy was needed that fit with the prevailing sense of the need to create the sinews of a new state. There was also a legacy of the security dilemma that had produced unification: that changing the state structure depended on altering the age-old continental Austro-French balance that went back to the rivalry of Habsburg and Valois. It was likely that the developmental strategy should have a heavy military orientation. The German approach made the house of Krupp the iconic German enterprise of the Kaiserreich; the apparent success of the German model pushed the

40

aggregate growth and policy

Kingdom of Italy to a developmentalism that in the 1880s focused on building up the improbable Umbrian town of Terni into a Central-Italian Krupp, a replica of Essen. Making Italians almost inevitably carried with it the demand for a strategy of developmentalism, which had major implications for trade and banking policy.

Italian Trade Policy in a Comparative Context, 1861–1929 In 1851 and 1859, the Piedmontese government adopted tariff reforms that greatly liberalized trade policy in the kingdom.1 This was consistent with the liberal philosophy of Cavour, and with the free-trade philosophy of the time, which culminated with the signing of the Cobden-Chevalier Treaty between France and Britain in 1860. Toniolo (1990) notes that in contrast with other policy areas, unification of the Peninsula’s customs policies happened suddenly and brutally, without the transition periods commonly imposed during more recent moves toward free-trade areas or customs unions. The liberal Piedmontese tariff, and existing Piedmontese trade and navigation treaties, were extended to the new Kingdom with barely any modifications. This was followed in 1863 by the signing of a trade treaty with France, as a result of which Italy became a member of the nascent “European network of treaties” based on the most-favored-nation clause. Bairoch (1989) notes that in the mid-1870s, at which stage this network of trade treaties was at its apogee, Italy had the second lowest tariffs of any major Continental European economy, with average tariffs on manufactured goods of between 8 and 10 percent (Table 2.1). Initially local producers retained some of the protection offered by poor transportation facilities, but as railways were constructed across the country, this protection was gradually eroded. Such a liberal stance was entirely typical in the context of the period. It could also have been expected in a largely agricultural economy, with net exports of agricultural goods and a powerful landowning class.2 However, manufacturers in a variety of industries, especially in the South (which had been heavily protected before unification) lost as a result of this abrupt shift to what almost amounted to free trade, and it was predictable that they would seek to redress matters. In 1870, a commission of inquiry into Italian manufacturing was set up on the grounds that the 1863 trade treaty with France would soon be up for renegotiation. The inquiry was the opportunity for businessmen in textiles and heavy industry (the latter supported by the military) to demand protection, and the outcome was the adoption of a tariff in May 1878, which is often taken to mark the end of free trade in Italy. The textile industry was the most-favored sector, and modest tariffs on wheat were also introduced. Overall, however, this was not a dramatic increase in protection, and tariffs were reduced as a result of trade

41

italy and the first age of globalization, 1861–1940

treaties being signed, for example with France in 1881. This was not sufficient for industrialists; another commission led, finally, to the adoption of a far more protective tariff in 1887, which imposed high duties on textiles, and (especially) iron and steel products. The duty on yarn amounted to some 27 percent according to Toniolo (1990, 84), compared with a tariff of just 7 percent on cloth (implying little or no effective protection for weaving); similarly, engineering had to pay for the very high tariffs on iron and steel, without being compensated with high tariffs for its own output. Wheat tariffs were also dramatically increased, amounting to the equivalent of 25 percent ad valorem in 1885. This shift to protection has been widely criticized in the Italian literature, but it was hardly an Italian peculiarity. Just as the 1860s saw a move toward freer trade across Western Europe, and not only in Italy, so too the 1880s saw a reversion to protectionism. The Italian tariff increases are also exactly the reaction to the grain invasion that would be predicted by Rogowski’s (1989) HeckscherOhlin model of the politics of trade policy during this period. Because Italy was a capital-scarce economy, Italian industrialists were bound to be protectionist, but initially they had to deal with the opposition of many landlords, because Table 2.1 European manufacturing tariffs, 1875 and 1913 Country

1875

1913 (1)

1913 (2)

1914

1913 (3)

1913 (4)

(percent)

(percent)

(percent)

(percent)

(rank)

(rank)

Belgium

9–10

9

9

10

11

11

Denmark

15–20

14

n.a.

18

16

I4

France

12–15

20

21

22

12

12

4–6

13

13

17

6

3

8–10

18

20

18

15

17

8

8

14

13

Germany Italy Norway

2–4

n.a.

n.a.

12

Portugal

20–25

n.a.

n.a.

n.a.

Spain

15–20

41

34

42

18

18

Sweden

3–5

20

25

23

5

6

Switzerland

4–6

9

8

7

3

7

The Netherlands

3–5

4

n.a.

3

1

1

United Kingdom

0

0

0

n.a.

4

5

Source: Bairoch (1989, 76), O’Rourke and Williamson (1999, 98–99). Notes: 1875: average levels of duties on manufactured products in 1875, from Bairoch (1989, 42, table 5); 1913 (1): League of Nations estimate, as reported in Bairoch (1989, 76, table 9); 1913 (2): Liepmann (1938) estimate, as reported in Bairoch (1989, 76, table 9); 1914: average duties on British manufactures, Bairoch (1989, 76); 1913 (3): rank among 18 countries (1 = least protectionist, 18 = most protectionist), based on the adjusted trade intensity ratios in Estevadeordal (1997, 104, table 6); 1913 (4): rank among 18 countries (1 = least protectionist, 18 = most protectionist), based on the openness measures in Estevadeordal (1997, 105, table 6). n.a. = not available.

42

aggregate growth and policy Table 2.2 Agricultural protection, 1913 Country

Wheat 1913

1913 (1) (rank)

1913 (2) (rank)

(percent) Belgium

0

8

7

Denmark

0

1

1

France

38

10

12

Germany

36

6

6

Italy

40

12

16

Norway

4

16

13

Portugal

Prohibitive

18

18

Spain

43

14

17

Sweden

28

7

8

Switzerland

2

17

14

The Netherlands

0

3

3

United Kingdom

0

4

2

Source: O’Rourke and Williamson (1999, 98–99). Notes: Wheat 1913: levels of duties on wheat, calculated by Bairoch (1989, 76, table 9); 1913 (1): rank among 18 countries (1 = least protectionist, 18 = most protectionist), based on the adjusted trade intensity ratios in Estevadeordal (1997, 104, table 6); 1913 (2): rank among 18 countries (1 = least protectionist, 18 = most protectionist), based on the openness measures in Estevadeordal (1997, 105, table 6).

Italy was an agricultural exporter in the context of Europe. The worldwide reduction of transport costs, however, switched landowners to the protectionist camp, by exposing them to competition from the land-abundant frontier economies; a coalition of iron and wheat was a logical consequence, with socialists supporting the free trade that was in the interests of relatively abundant Italian labor (Bairoch 1989, 134). The increase in protection implied by the 1887 tariff was exacerbated by a trade war with France that started in the following year, and which severely damaged the Italian silk and wine industries. Between 1887 and 1897, Italian exports to France fell by 57 percent, compared with a 21 percent decline in bilateral trade in the opposite direction (Conybeare 1985, 161). The trade war ended in stages: Italy abandoned its penalty tariffs against France in 1890, and France reciprocated in 1892. Even so, it was only in 1898 that France accorded Italy minimum tariff status (160). Wheat tariffs were also increased again in 1888, and yet again in 1894. By 1913 wheat tariffs in Italy were equivalent to a roughly 40 percent ad valorem tariff (Table 2.2). However, a series of trade treaties subsequently lowered rates

italy and the first age of globalization, 1861–1940

43

of Italian protection, favoring agricultural exports at the expense of protection for industry. Tariffs were thus on a downward trajectory between the mid-1890s and World War I, and indeed the wheat tariff was suspended when war broke out. Tariff policy remained relatively liberal until 1925: a tariff of 1921 was again watered down by a series of treaties, which had been the intention because the 1921 tariff was a tarif de combat. In 1925, however, wheat and sugar duties were sharply raised, ostensibly for revenue reasons, and in 1926 tariffs on manufactured goods were increased again. The wisdom or otherwise of Italian protection during the late nineteenth century has been the subject of ferocious debate among economic historians. Some authors argue that grain tariffs helped smooth the disruption associated with the influx of cheap cereals from the frontiers, thus lowering adjustment costs, whereas tariffs on iron and steel helped establish Italian heavy industries, which otherwise would never have been established, and which were crucial for industrialization in the long run. (A contemporary might have added that industrialization was crucial for military security, whatever the deadweight economic losses this might have implied.) However, others have argued that the structure of protection was irrational. Gerschenkron (1962), who in principle accepted that government intervention could be beneficial in a backward economy, such as Italy, argued that the lack of Italian coal reserves implied that Italian policy should have promoted industries that were not coal-intensive, such as engineering. Instead, the tariff favored the iron and steel industries, which were heavy consumers of coal (and were thus uncompetitive), and placed engineering at a disadvantage. Moreover, the tariff failed to protect “the equally promising field of the chemical industry, pregnant with many innovational possibilities and well suited to the conditions of the country,” whereas tariffs promoted the cotton textile industry, which was old, “with a moderate rate of modern technological progress and accordingly relatively limited possibilities in a backward country on the European continent” (Gerschenkron 1962, 81, 83).3 Gerschenkron also objected to the grain tariffs on the grounds that they damaged industry, and Fenoaltea (1993) has provided a formal rationale for this in the form of a quasi-classical model in which grain tariffs increased nominal wages (because the real grain wage was exogenously given as a result of migration flows)4; these higher nominal wages reduced employment in manufacturing, with the surplus labor emigrating.5 It is beyond the scope of this chapter to adjudicate between the various arguments that have been made about the effectiveness or otherwise of Italian tariffs, but it can provide some guidance to the following questions: How high were Italian tariffs? How high were they in an international context? Were they sufficiently high as to have had a big impact on the Italian economy? The first question to address is: how high were Italian tariffs? Tables 2.1–2.3 provide one answer to this question for the pre-1914 period, based on the work of Bairoch (1989), Liepmann (1938), and Estevadeordal (1997). As noted, average manufacturing tariffs were relatively low in Italy in 1875, so the focus here is on average levels of protection on the eve of World War I. The first two

44

aggregate growth and policy Table 2.3 Overall protection, 1913 Country

1913 (1) (percent)

1913 (2) (percent)

1913 (3) (percent)

1913 (4) (rank)

1913 (5) (rank)

Belgium

15.8

6

14

10

10

Denmark

5.8

9

n.a.

2

4

France

8.7

18

24

14

14

Germany

7.9

12

17

8

8

Italy

9.7

17

25

16

17

Norway

11,4

n.a.

n.a.

11

12

Portugal

23.7

n.a.

n.a.

17

15

Spain

14.3

33

37

18

18

9

16

28

7

7

Switzerland

4.4

7

11

9

9

The Netherlands

0.4

3

n.a

1

1

United Kingdom

5.6

0

0

3

3

Sweden

Source: O’Rourke and Williamson (1999, 98–99). Notes: 1913 (1): import duties as percent of special total imports (1909–1913), calculated by Bairoch (1989, 76, table 9); 1913 (2): League of Nations estimate, as reported in Bairoch (1989, 76, table 9); 1913 (3): Liepmann (1938) estimate, as reported in Bairoch (1989, 76, table 9); 1913 (4): rank among 18 countries (1 = least protectionist, 18 = most protectionist), based on the adjusted trade intensity ratios in Estevadeordal (1997, 107, table 8); 1913 (5): rank among 18 countries (1 = least protectionist, 18 = most protectionist), based on the openness measures in Estevadeordal (1997, 107, table 8).

authors provide a range of percentage tariffs: the ad valorem tariff on wheat and average tariffs for industrial products, agricultural products, and imports as a whole. For example, column (1) of Table 2.3 gives Bairoch’s average tariffs, computed as the ratio of customs duties to the value of imports. As is well known (Anderson and Neary 2005) this measure is biased downward, because it amounts to an import-weighted average tariff, and as tariffs increase on particular products the weight on these products declines. However, it has the advantage of being easily measurable for many countries. Bairoch’s data suggest that in 1913, the Italian average tariff was 9.7 percent, somewhat higher than French or German average tariffs, but much lower than average tariffs in Portugal or Spain, and such industrialized countries as Belgium or the United States.6 The overall impression one gets from the average tariff data in Tables 2.1–2.3 is that Italian tariffs were roughly comparable with tariffs in France, with manufacturing tariffs being generally a little lower in the Peninsula, whereas tariffs were clearly lower in Italy than in Spain and Portugal. However, the Italian wheat tariff was quite high in an international context.

45

italy and the first age of globalization, 1861–1940 .14 .12 .10 .08 .06 .04 .02 .00 1876

1880

1884

1888

1892

1896

1900

1904

1908

1912

1916

Year France Sweden

Germany UK

Italy

Figure 2.1 Average agricultural tariffs, 1870–1913. Source: Lehmann and O’Rourke (2010).

Estevadeordal (1997) adopts a different approach, inspired by Leamer (1988). This approach assumes that trade is driven by Heckscher-Ohlin forces, and that any divergence from the level of trade that the Heckscher-Ohlin model would predict must be caused by protectionism. As can be seen from the tables, this method concludes that Italy was one of the most protectionist countries in Estevadeordal’s sample, but the method is obviously suspect, because it is only as reliable as the econometric and theoretical framework used to generate the predicted “correct” level of trade. The overall impression, then, is that Italian protection in 1913 or 1914, although nonnegligible, was not out of the ordinary in the context of the time. This is the major theme of Federico and Tena (1998). However, there is one problem with judging Italy’s protection by the yardstick of levels of protection Europe-wide in 1913: tariff levels were not constant over time, and indeed were falling in Italy from some point in the 1890s onward. Figures 2.1–2.3 remedy this by plotting average tariffs in Italy and four other major European economies (France, Germany, Sweden, and the United Kingdom) between the mid-1870s and World War I. The graphs plot three different average tariffs: (1) those on imports of agricultural goods; (2) those for manufactured goods; and (3) those for “exotic” goods (i.e., revenue tariffs on goods clearly not produced domestically, or substituting for domestically produced goods). These data suggest that, compared with other large European economies of the time, with whom Italy would have compared herself, Italian tariffs were toward the high end of the spectrum for most of this period. Her agricultural tariffs were similar to German ones: slightly higher for most of the

46

aggregate growth and policy .12 .10 .08 .06 .04 .02 .00 1876

1880

1884

1888

1892

1896

1900

1904

1908

1912

1916

Year France Sweden

Germany UK

Italy

Figure 2.2 Average manufacturing tariffs, 1870–1913. Source: Lehmann and O’Rourke (2010).

period, but lower toward the end. It is important to note, however, that these agricultural tariffs include tariffs on sugar, which were extremely high in Italy. Her manufacturing tariffs were lower than tariffs in Sweden, but usually higher than tariffs in France or Germany. Her revenue tariffs (i.e., tariffs levied on such goods as coffee, tea, and other products clearly not produced domestically)

1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1876

1880

1884

1888

1892

1896

1900

1904

1908

1912

1916

Year France Sweden

Germany UK

Italy

Figure 2.3 Average revenue (exotic) tariffs, 1870–1913. Source: Lehmann and O’Rourke (2010).

47

italy and the first age of globalization, 1861–1940 70

shillings per quarter

60

50

40

30

20 1870

1875

1880

1885 France Italy

1890

1895

1900

1905

Germany Sweden

Figure 2.4 Nominal wheat prices (shillings per imperial quarter), 1870–1906. Source: O’Rourke (1997) and data courtesy of Giovanni Federico.

were much higher than in other countries in the late 1880s and 1890s, if unremarkable at other times. However, Italian tariffs were not wildly out of line with tariffs in other European countries. They were surely lower than in such countries as Spain (not included in the Lehmann-O’Rourke sample) and the highly protectionist United States. It was Italian revenue tariffs that really stood out during the periods when Italian protection was at its highest in relative terms. Wheat tariffs have been particularly criticized in the Italian literature. Although the authors do not yet have comparable data on wheat tariffs in Italy and other countries, there are data on wheat prices. Figures 2.4 and 2.5 place Italian wheat prices in the late nineteenth century into a comparative perspective. Figure 2.4 plots nominal wheat prices in four protectionist European economies: (1) France, (2) Germany, (3) Italy, and (4) Sweden.7 Tariffs were not enough to prevent wheat prices from falling in any of these countries: they could counteract the impact of cheaper transport between Western Europe and the granaries of Russia and the New World, but they were not high enough to counteract the impact of technological change, which was lowering the supply price of wheat in these locations (O’Rourke 1997). The degree to which prices in the four countries moved together is noteworthy: tariffs and transport costs created wedges between prices in different countries, but grain markets remained fundamentally linked together. Italian prices were by no means out of line with prices in the other three countries, although they were slightly higher in the early years of the twentieth century. Figure 2.5 provides the same data, and those for the United Kingdom, but expressed now as a percentage premium over prices in Odessa. There is a clear divergence between trends in the protectionist

48

aggregate growth and policy

100 80

percent

60 40 20 0 –20 1870

1875

1880

1885

France Sweden

1890

1895

Germany UK

1900

1905

Italy

Figure 2.5 Wheat prices relative to Odessa, 1870–1906 (% price gaps). Source: O’Rourke (1997) and data courtesy of Giovanni Federico.

countries on the one hand, and the free-trading United Kingdom on the other; Italy does not stand out within the former group of countries as being particularly protectionist, according to this criterion. Federico and Tena (1998) show that what was true of 1913 was true also for the mid-1920s: Italy was very much an average country when it came to tariff levels. According to League of Nations data, Italy came either ninth out of twenty countries, or eighth out of sixteen countries, in the protectionist league table; according to the data presented in Liepmann (1938), it came ninth out of fifteen. Table 2.4 shows that, as in the late nineteenth century, it was Italian revenue tariffs that were particularly high, although the relatively high tariffs on semimanufactured goods also suggests an overall tariff policy that was not as consistently pro-industrial development as it might have been. However, as has been seen in earlier years, Italian protection had been reasonably high in the context of other major Western European economies. Was it high enough to have had a big impact on the Italian economy? Table 2.4 Nominal tariff rates, 1927 France

Germany

Italy

Fiscal goods

30.9

51

61.1

Foodstuffs

12.7

19.6

15.8

Semimanufactures

12

10.4

22.2

Manufactures

21.6

15.5

22.2

Source: Federico and Tena (1998, 89, table 7).

italy and the first age of globalization, 1861–1940

49

To answer this question one must first determine the structure of protection: what sectors were most protected, especially taking into account the fact that some of their inputs might have been made more expensive as a result of tariff policy. Federico and Tena (1998) provide data on effective rates of protection in thirty-five sectors in 1911. Their results show that the highest effective rates of protection were enjoyed by sugar beet processors (126 percent) and the coal and tar sector (136.3 percent). Within industry, steel-making was a big winner, with an effective rate of protection of 74.2 percent; textiles (26.9 percent) and clothing (15.3 percent) also benefitted. The effective rate of protection enjoyed by the promising engineering and chemical sectors was far more modest (8.2 percent and 17.9 percent), but at least it was not negative (unlike the case for foundries and shipbuilding). Overall, the big winners were wheat (30.7 percent) and sugar production, textiles, and iron and steel production, as the conventional wisdom has suggested. The choices that were made seem to be more easily explained by appealing to the political power of landowners, and the nascent military– industrial complex, than by positing a deeply thought-out strategy of industrialization behind tariff barriers. However, Italy was hardly alone in tailoring its trade policy to the interests of landowners, or in worrying about her military security. What was the impact of these tariffs? To answer this question, one inevitably has to commit to a model of the Italian economy, and different models give different answers. The answer here is based on a fairly standard static model, outlined in Federico and O’Rourke (2000); this will not satisfy those who believe that the important effects of trade policy are dynamic, and can only be understood within the context of models that involve nonconvexities of one sort or another. The finding here is that tariffs were probably not high enough to have a big impact one way or another, but it is important to recognize that there are alternative modeling frameworks in which small distortions can end up having big consequences in the long run. Federico and O’Rourke distinguish between nine sectors, four of them agricultural and five nonagricultural. The agricultural sectors are wheat and sugar, lumped together because they were so highly protected (WHEAT); nonwheat tillage crops (TILL); Mediterranean products, such as wine and oil (MED); and animal products (ANIMAL). The nonagricultural sectors are the military–industrial complex (steel-making, the production of rolling stock, and shipbuilding; MIC); other capital-intensive sectors, such as chemicals and other engineering activities (KII); textiles (TEXT); other industries (largely consumer industries, such as clothing and food-processing; OTHER); and nontraded goods including utilities and services (NT). There are four primary factors of production: unskilled or raw labor (L), skills (H), capital (K), and land (R). Land is only used in the agricultural sectors, and is sometimes sector-specific; skills are only used in the nonagricultural sectors, and are assumed to be sector-specific; capital and raw labor are mobile across sectors, although labor is imperfectly mobile between agriculture and the rest of the economy so as to account for the reality of rural-urban wage gaps.

50

aggregate growth and policy Table 2.5 Tariff levels, 1897–1911 (ad valorem equivalents, percent) Sector

1897

1911

Wheat and sugar

97.7

29.6

Other tillage

5.7

4

Mediterranean crops

17.8

8.1

Animal products

2.9

3.6

Military–industrial complex

26.1

15.7

Other capital-intensive industries

15.8

6.6

Textiles

6.8

4.1

Other industries

8

8.4

95.4

74.5

Exotic goods Source: Federico and O’Rourke (2000).

Having constructed the model, which is calibrated so as to replicate the state of the economy in 1911, one can then impose various tariff-related shocks on it and see how the economy adjusts. The most obvious shock is to impose free trade on the economy: this means setting all existing tariffs to zero. The impact of the shock depends on the sizes of the initial tariffs, which are given in the second column of Table 2.5. Because wheat and sugar production enjoyed the highest level of protection, their output would have changed the most, declining by 53.2 percent; protection led to the combined sugar and wheat sector being roughly twice the size that it would have had under free trade. This decline in the WHEAT sector would have led to land being reallocated to other agricultural sectors, with the result that the output of nonwheat tillage crops would have expanded by some 21.5 percent, and animal production would have increased by almost 10 percent. Within industry, the big loser from a switch to free trade would not surprisingly have been the military–industrial complex, whose output would have fallen by 14.2 percent. The big beneficiary would not have been engineering (whose output would have declined marginally), but rather textiles production, whose output would have expanded by 27 percent. This may seem surprising in light of the literature reviewed previously, but it makes sense given that the initial tariff on textiles production in 1911 was relatively low, and that as a labor-intensive industry in a labor-abundant country it should have benefitted from a move to free trade. Overall, agricultural output would have contracted by some 2 percent, whereas industry would have expanded by 6 percent: the overall structure of the economy would not have changed by very much. Nevertheless, protection did have a major impact on individual sectors, and would have had an even bigger impact in 1897, when (as can be seen from the table) the level of protection was higher.

italy and the first age of globalization, 1861–1940

51

Overall, the conclusion has to be that Italian tariff policy during this period was hardly out of the ordinary: it followed the ebbs and flows of international trends toward and away from freer international markets fairly faithfully. The pattern of production was not dissimilar from that in other European countries with a similar economic structure: as everywhere, Italy protected grain, whereas the protection afforded to heavy industry was similar to that in Germany, and reflected Italy’s capital-scarce status. That capital scarcity had even more direct implications for Italian monetary and banking policy, as shown next.

Monetary and Fiscal Policy, from Unification to the Resumption of the Gold Standard The exchange rate and its management constitute one of the most direct ways in which a national economy is connected to the world monetary system. In the course of the last third of the nineteenth century, the world’s major countries adopted the gold standard. There might even have been a more ambitious monetary unification. In 1867, at the instigation of Napoleon III, who launched an International Monetary Conference in Paris, it looked as if there would be a single world money, broken into divisions of pounds, worth five dollars, which in turn would be worth five lira or francs. Unified Italy for a short time looked as if it was on this path: a law of 1862 adopted a bimetallic gold and silver standard on the French pattern. In 1865, when Italy joined in the Latin Monetary Union, managed by France and Napoleon IIII, it looked as if the new currency bloc would move in the direction of a pure gold standard, which was the intention of Napoleon III’s advisers. Then war intervened, and Italy moved in another direction. In 1866, at the outbreak of the last of the wars of unification with Austria, the issue banks suspended convertibility in the so-called corso forzoso; by 1868, the currency had depreciated by some 12 percent. It was only in 1883 that convertibility was resumed. The banks of issue very quickly stopped free conversion, however, although the exchange rate was carefully managed to be close to the gold parity before 1893, and then again after 1899. Italy thus belonged to the periphery, rather than the core, of the prewar gold standard. The periods of formal adherence to all the rules of the gold standard, 1862–1866, 1883, and then again between 1927 and 1934 (when exchange controls were introduced), were thus rather brief. For peripheral, capital-scarce countries, such as Italy, the attraction of the gold standard lay mostly in opening up international capital markets, and allowing a lowered cost of borrowing. That was the irresistible attraction that drew them to the golden magnet. Fenoaltea (1988) has argued convincingly that

52

aggregate growth and policy

Italy’s success in attracting capital inflows, which was largely dependent on world capital market conditions, was a key determinant of the Italian construction cycle. However, because these capital movements were channeled through financial intermediaries, they also had consequences for Italian financial stability, of a sort that is still familiar today. Italy’s formal attachment to the gold standard in 1883, and the maintenance of parity even after the end of free conversion, did produce some of the benefits associated with the gold standard in other countries. It sparked a substantial surge in growth, financed in large part by inflows of foreign credits, and a real estate boom. There was a substantial current account deficit, reflecting the inflows of external money. Between 1882 and 1887, the lending of commercial banks increased by 15 percent a year (Toniolo 1990, 77). When the boom slowed down in 1887–1888, the Italian banks ran into great difficulties. The inflows now stopped, while the current account moved into surplus in 1891 and 1892. The sudden stop was fundamentally unsustainable in the political and economic circumstances of Italy, and a wave of scandals and crises erupted that required a complete reorganization of Italian banking, with a new central bank (the Banca d’Italia) taking the place of the previous banks of issue, and new commercial banks. The year 1893 saw a major financial crisis in which the two largest commercial banks, the Credito Mobiliare and the Banca Generale, failed, in addition to one of the note-issuing banks, the Banca Romana. For most of the period after the re-establishment of the gold standard, the lira-sterling exchange rate was very close to parity, but after June 1893 the exchange rate fell sharply. The depreciation of the lira reached as much as 13 percent. The subsequent bank reform led to the creation of a new major note-issuing bank, the Banca d’Italia, which absorbed the Banca Romana. Bank reform also led to the institution of new mixed or universal banks, explicitly copied from the German model, and financed at first largely by international consortia in which the German financial establishment was preponderant. The new banks, the Banca Commerciale Italiano, often called the Comit (1894), and the Credito Italiano (Credit, 1895), were intended to contribute to the financing of industrial development (discussed in the next section). In practice they set off a new wave of capital inflows that lasted until the next financial crisis. A major driver of exchange rate policy during this period was the problem of handling government debt. At the time of Italian unification, public debt amounted to about half of gross domestic product (GDP), but the ratio doubled over eight years, and the interest on debt amounted to almost 40 percent of government expenditure. In the 1870s, there was a serious attempt at fiscal consolidation in preparation for the readoption of the gold standard, but the largest bank of issue, the Banca Nazionale, continued to monetize a substantial part of the government debt. Government bonds (rendita) were mostly sold abroad, chiefly on the Paris market, and paid in foreign currency at the official parity. There was thus a strong incentive for Italian investors to buy renditas but to collect their coupon payments in the foreign currency. In 1873 an affidavit was introduced that

53

italy and the first age of globalization, 1861–1940

required a declaration under oath that the owner was not an Italian citizen, but the new regime was not vigorously enforced and was abolished in 1881 (Tattara 2003). Although in the 1870s some 40 percent of the Italian government debt was paid out abroad, by the early 1880s the ratio had reached 80 percent, and the readoption of a gold standard in consequence now looked as if it was a relatively costless operation, because the government debt was serviced in gold anyway. When the financial crisis of 1893 shook Italy, the affidavit was reintroduced. Prewar Italy was tested by two crises: while 1893 fundamentally shook every part of the Italian banking and financial system, and led to a sharp drop in capital inflows, 1907 showed the resilience of the new order. The financial crisis, associated with a stock market decline (28 percent from August 1906 to July 1908) and a radical contraction of bank lending in 1907, barely affected the exchange rate (Cesarano, Cifarelli, and Toniolo 2012). Italy had reached a new level of financial maturity. The Banca d’Italia under Bonaldo Stringher had managed a conversion of state debt in 1903 and then tried to encourage the purchase of government bonds through easy money and low interest rates (Confalonieri 1977, 137; Forsyth 1993, 43). However, borrowing in Italy remained expensive. As Figures 2.6 and 2.7 show, the difference between long-term bond yields in France and Italy (which had reached a height of 2.1 percent in January 1894) fell in the course of the 1900s, but it did not disappear (unlike in Switzerland, where the same premium had existed relative to the French market before 1907, but virtually vanished in the aftermath of the establishment of a new National Bank) (Bordo and James 2007). At the end of 1910, the difference was still 0.6 percent, but by 1914 it had faded to between 0.1 and 0.2 percent. Like other European countries, Italy suspended convertibility at the outset of World War I. The wartime era produced high levels of inflation, which continued in the atmosphere of social violence that immediately followed the peace, with a sharp deflation following in 1920 that produced a new wave of banking difficulties. As elsewhere, Italian leaders were determined to reintroduce the gold standard, and they eventually succeeded. 9 8 7 6 5 4 3 2 1 0

Italy

01/31/1871 10/31/1873 07/31/1876 04/30/1879 01/31/1882 10/31/1884 07/31/1887 04/30/1890 01/31/1893 10/31/1895 07/31/1898 04/30/1901 01/31/1904 01/31/1906 07/31/1909 04/30/1912 01/31/1915 10/31/1917 07/31/1920 04/30/1923 01/31/1926 10/31/1928 07/31/1931 04/30/1934 01/31/1937 10/31/1939

France

Figure 2.6 Yields on French and Italian 10-year government bonds. Source: Global Financial Data.

54

aggregate growth and policy 2.5 2 1.5 1 0.5

–0.5 –1

01/31/1871 02/28/1873 03/31/1875 04/30/1877 05/31/1879 06/30/1881 07/31/1883 08/31/1885 09/30/1887 10/31/1889 11/30/1891 12/31/1893 01/31/1896 02/28/1898 03/31/1900 04/30/1902 05/31/1904 06/30/1906 07/31/1908 08/31/1910 09/30/1912 10/31/1914 11/30/1916 13/31/1918 01/31/1921 02/28/1923 03/31/1925 04/30/1927 05/31/1929 06/30/1931 07/31/1933 08/31/1935 09/30/1937 10/31/1939

0

–1.5 –2 –2.5

Figure 2.7 Difference in yields of French and Italian 10-year government bonds. Source: Global Financial Data.

In August 1926, Mussolini announced his intention to revalue the lira, and Italy began a package of stabilization measures that included raising interest rates, budgetary stabilization and debt consolidation, and restriction of note issue by the Banca d’Italia. The idea behind the currency stabilization was to allow Italian business and the government to borrow on international markets. There seemed to be an international trend in favor of gold. In particular the major western central banks, the Bank of England and the Federal Reserve, and the British and US governments pressed continental Europeans to return to gold. The reward would be, as before World War I, a “good housekeeping seal of approval” and access to capital markets. The operation was a trade-off, in that the chosen exchange rate of the quota novanta represented a considerable overvaluation of the lira, and a blow to the competitiveness of Italian industry; but the major attraction lay not simply in prestige calculations (Cohen 1972) but above all, yet again, in a lowered cost of borrowing. However, the yield on long-term government bonds remained more or less constant, at around 5 percent. In contrast, France, with a stabilization that undervalued the French franc, achieved a much more dramatic reduction in the cost of government borrowing. Italy’s formal return to the gold standard occurred on December 21, 1927, with a 40 percent gold coverage requirement.

Industrial Policy from Unification to the Great Depression The history of Italian economic development through much of the nineteenth century is a story of a capital-scarce country’s response to economic

italy and the first age of globalization, 1861–1940

55

backwardness and a poor resource endowment for the crucial technologies of the European industrial revolution: in particular, the absence of large iron ore and coal fields. It is not surprising that Alexander Gerschenkron chose Italy as a case study not only for the economics of “backwardness,” but also for the question of how state action could overcome backwardness. The imbalance between state provision of financing and private control of policy was widely seen as the key to an Italian phenomenon of “capitalism without capital.” What preexisting Italian advantages there were lay in wide-reaching networks of relationships, with a strong family content, combined with an interest in technical progress. The great Milanese economist Carlo Cattaneo developed a theory according to which intelligence formed the major principle of public economy: “Control of new arts gave plentiful and reliable food to a number of families. They took their secrets from land to land; their wealth was the idea they had developed. They often remained divided from the crowd owing to their foreign origin and their different religion. Their knowledge became an inheritance, a perpetual privilege; they became a caste” (Cattaneo 2002 [1859], 72–73). Especially in Lombardy, the commitment to technical education became institutionalized at an early stage through the Milan Society for the Encouragement of Arts and Sciences established in 1838 (Lacaita 1990). That was never a solution, however, that could be hoped to apply to every region of Italy. As a result, politicians embarked on a search for a new way of galvanizing growth. Soon after unification the emphasis shifted dramatically to an insistence that state action was needed to make good the weaknesses of the private sector. Other countries, especially Germany, seemed to have evolved in a different way from the United Kingdom: should not Italy imitate the new model of success? In the 1880s, Italy adopted a spectacular policy of economic development. The origins of the new approach lay in debates of the 1860s and 1870s about how an Italian political entity could best be built. Emulation of a foreign model seemed a necessity for launching Italy on a path to greater power and greater economic dynamism (the two being connected). The construction of a steel industry, which seemed to be the foundation of British and German strength, looked like a national imperative that also might be the foundation for a native armaments industry. Already in 1867, in the immediate aftermath of the Austro-Prussian war, the Venetian businessman and politician Vincenzo Stefano Breda had proposed a “patriotic effort” to build cannon according to the latest Krupp method, using steel made by the brand-new Bessemer process. Breda proposed a mixed private-public form of enterprise. The proposal was rejected by the center-right governments, but after 1876 governments of the left were more sympathetic. Breda argued strongly for a location away from traditional iron centers in Turin or Lombardy, because they were too near the frontier and vulnerable to attack. By the 1870s he focused his attention on the Umbrian town of Terni, home of an established jute industry but with no obvious locational advantages for steel-making (except for a plentiful water supply that could also be used as an energy source). In 1879

56

aggregate growth and policy

he founded a company in Terni, Cassian Bon, which acted as a proxy for his own construction company, Società Veneta, and worked closely with the naval engineer and Minister of the Navy Admiral Benedetto Brin in securing a law of 1884 that created a large steel-making and armaments industry in Terni. It was heavily financed by banks (above all the Credito Mobiliare) but its major asset was always its political linkage. In particular, the personal commitment of the monarch seems to have been crucial. In the same way as Wilhelm I supported Alfred Krupp, Umberto I bubbled over with enthusiasm for the Terni project, and endorsed it with an official visit in 1887. The whole affair of the origins of Terni was surrounded by a miasma of charges of corruption and bribery. It was an inauspicious beginning to the new policy of developmentalism. Terni became the most obvious symbol of what liberal commentators, such as Amatori and Colli (1999), politely term “a precocious state capitalism,” not so much the direct control of the state through ownership but a preponderant role of the state as a result of procurements and tariff policy. Alternately, socialist critics detect the capture of the state by special business interests and see the steel sector as what Ernesto Rossi termed “la grande parassitaria” (Balconi 1991, 82). From the last years of the nineteenth century, the Italian story shifted to the politics of the search for institutional mechanisms for overcoming backwardness. The principal feature of Italian backwardness seemed to be the under-development of the capital market. At first, many business leaders believed that the best device for transcending Italy’s institutional limits lay in financial institutions that could mobilize domestic investment resources and also attract foreign investments. However, such institutions were themselves quite vulnerable. In 1893, there was a general banking crisis after a severe economic downturn, in which two of the largest Italian commercial banks, Credito Mobiliare and the Banca Generale, failed. After this episode, there was more or less continual political intervention in business structures, with the result that business and financial and political interests became closely interconnected. A new set of banks, closely linked to a German model of industrialization, became very powerful, but also immediately became the subject of political attention. Periodic crises destroyed financial values and demanded restructuring: after World War I, such restructuring usually occurred at the public initiative and with public funds. The financial story of twentieth century Italy can be summarized as the repeated destruction of capital, both of investors in the stock market, and of the government’s contributions. The first hope of liberal Italy at the beginning of the twentieth century was that Italy could be made financially and economically more robust through the operations of the newly established Banca Commerciale and the Credito Italiano. These Milanese banks8 rapidly became intertwined with factions in the political elite of liberal Italy. In addition to a German business model, these institutions initially operated with German capital. They derived most of their earnings from underwriting, and promoted quite speculative enterprises. After

italy and the first age of globalization, 1861–1940

57

a sharp setback in the stock market after the US crisis of 1907, the stock-buying public turned away from this kind of asset, and the banks had to retrench their position (see Forsyth 1993, 35–36). The Banca d’Italia and its energetic general manager Bonaldo Stringher played a vital part in the rescue of a major bank, the Società Bancaria Italiana, but also in the restructuring of particular industries, notably in the creation of the Ilva steel trust of 1911. During World War I, the state launched a new round of developmentalism, with a particular concentration on the ability of hydroelectric power to substitute for those resources needed for the development of heavy industry. A government decree of September 1916 (the Bonomi decree) gave favorable treatment to enterprises that would construct their own generators, networks, and distribution systems. The decree gave a boost to the already well-developed hydroelectric industry, and to new industries that used electricity rather than coal as the fuel for steel-making. In the 1920s and 1930s, many prominent managers of the steel industry again used politics as a way to gain support for their vision of a business model in which Italy would develop its own alternatives to a competitive and open international economy. They emphasized the necessity of the “ciclo integrale” in a large-scale path of developing a big steel and metallurgy sector, and that ignored the economics of resource scarcity (Sinigaglia 1946). These managers seemed to manage the shift of the political system from liberalism to fascism just as surely and confidently as they would manage the transition back in the opposite direction in the 1940s. The ideology of the state mattered less than a developmental philosophy. Oscar Sinigaglia, the brilliant technocrat who played a central role in the development of the state sector and then became the director of the state steel holding company Finsider, before he was ousted in the aftermath of the 1938 racial law, later referred to the slowness of the adoption of such a cycle as the characteristic “Italian problem.” By the 1930s, the large-scale state steel industry seemed to be competing on unfavorable terms with private producers. Sinigaglia designed schemes for autarky behind which the idea of the complete cycle could be realized. An adequate national production of steel could then provide the basis for a modern engineering industry. As it turned out, the interaction between the universal banking model of development, and the deflation that followed the resumption of the gold standard, and the outbreak of the Depression, would provide the conditions within which the State could become the dominant actor in Italian heavy industry. Ilva had in 1920 been rescued by the Banca Commerciale and the Credito Italiano in a highly costly operation, through a newly created Istituto Finanzaria, but its business continued to remain highly problematic, and it is not surprising that its bank-owners looked for a way out. The banks emerged as effective controllers of much of Italian heavy industry, and in particular of steel, but that made them highly vulnerable in the deflationary business climate of the 1920s, after the stabilization of the Italian lira in 1927 at the over-valued quota novanta. As seen in the next section, the result was that the industrial hegemony of the

58

aggregate growth and policy

universal banks was destroyed by new deflation during the world depression at the end of the 1920s and in the early 1930s, with the state definitively taking their place.

The Great Depression, Institute for Industrial Production, and Italy’s Switch to Autarky The deflation associated with the return to the gold standard is usually thought to be responsible for a falling off of growth in 1926 and in 1927, but the renewed expansion of 1928 looked at first like a vindication of the government’s strategy. Capital flowed in, and borrowing rates were lower. After 1929, however, deficit countries were pushed into adjustment by the freezing up of international capital markets. From 1929 to 1935 a contraction of the monetary base of about 25 percent was experienced. There was a sharp loss of reserves in 1931, in the midst of the Central European banking and financial crisis, but the cover ratio stayed above the 40 percent level. It was only when that cover ratio fell below 40 percent that Italy formally ended the commitment to the gold standard, in September 1936. The response of Italy to the 1931 banking crisis is often held to be particularly resourceful and even constructive (Fratianni and Spinelli 2004). Unlike Austria, Hungary, and Germany, where there was a very similar mixed or universal banking system, in which banks were vulnerable because of the large portfolio of industrial participations that they held in their portfolio, in Italy there was no open banking crisis, although this is what some outside observers, such as the well-informed Swiss financier Felix Somary, had predicted (Toniolo 1995). Bank deposits indeed only fell by 3.6 percent in 1931 (Mattesini and Quintieri 1997, 275). However, the major banks were insolvent, in large part because of the extensive holdings of industrial securities that they had been pushed to take over in the early 1920s as part of the rescue operation for Italian industry. The Credito Italiano, Italy’s second largest bank, was insolvent by the end of 1930 and the largest bank, the Banca Commerciale, by May 1931. There was also a substantial degree of political tension between the Banca Commerciale and the government, which saw the Commerciale as dominated by figures from the liberal era. A few days after the collapse of the German Darmstädter und Nationalbank, on July 19, 1931, the general manager of the Banca Commerciale, Giuseppe Toeplitz, appealed to the Bank of Italy for support. In September he was received by Mussolini. The next month a new holding company, SOFINDIT, was created to take over the industrial assets of the Commerciale. Together with parallel holding companies, SFI (which had taken over the assets of the

italy and the first age of globalization, 1861–1940

59

Credito Italian) and Elettrofinanziaria, the industrial assets were transferred to a new state holding company, the Institute for Industrial Reconstruction (IRI). By January 1934, IRI held 48.5 percent of the share capital of Italy. Then, in March 1934, IRI acquired the capital of the banks themselves. By 1937, it controlled all the capital of the military steel sector (Terni, Ansaldo, Cogne); 40% of nonmilitary steel; and 30% of the electrical industry. In the aftermath of the creation of a bad bank, a new banking law (1936) limited the industrial engagement of banks (see also Guarino and Toniolo 1993). In a dramatic series of speeches, Mussolini seemed to come back to the anticapitalist rhetoric of the early days of Fascism. In the immediate aftermath of the financial crisis, he explained that “if I tried to harness the capitalistic force up to now, it was because it was already active. But when the first resistance occurs, it will be replaced without hesitation by a new, entirely different and more worthy force” (Avagliano 1991, 35). He saw the major members of the Consorzio Siderurgico on March 13, 1933, and explained that the steel makers were not responding adequately to the Fascist initiatives for industrial reorganization: “Unfortunately it is a question of one of these circumstances which have recently become quite common, in which the business leaders say that the company can be completely cured if particular measures are taken. I make the appropriate orders, and things develop worse than they were before.”9 In May 1934, as the process of IRI’s takeover of bank assets was being worked out, Mussolini announced that: “Those who still speak of a liberal economy make me laugh – laugh or weep, both at the same time. But three quarters of the Italian industrial and agricultural economy is in the hands of the state. And if I dare to introduce to Italy state capitalism or state socialism, which is the reverse side of the medal, I will have the necessary subjective and objective conditions to do it” (Giordano 2007, 35). The financial and economic crisis of the early 1930s thus opened the opportunity to launch a new wave of developmentalism. Again, there was a strong emphasis on hydroelectric power as a way of solving the sensitive issue of energy dependence. From 1925 to 1937, the output of hydroelectric power more than doubled. In the geopolitical context of the 1930s, military action was another way in which the state influenced economic development. Indeed, according to Toniolo (1980, 272), it was only with the putting into practice of the plans for invading Ethiopia that the Italian economy received the stimulus necessary for economic recovery. The decision to activate the invasion plans was taken in December 1934, and the invasion itself took place in October 1935. The undertaking implied massive military expenditure. This may plausibly be taken to have had Keynesian effects on Italian GDP, because Italian deficits exceeded 10 percent of GDP in 1936 and 1937 (Almunia et al. 2010), despite increases in taxation of various sorts. Various types of compulsion were used to ensure that government bonds be taken up; expenditure was also financed by nationalizing foreign investment in 1935, and by the 1936 “day of the wedding ring,” when Italians were asked to hand over their gold to the regime (Zamagni 1993b,

60

aggregate growth and policy

253–254). Monetary financing of the deficit was also resorted to. There were large military orders for industrial output, which rose by almost 15 percent in 1937, by 3 percent in 1938, and by almost 9 percent in 1939. Overall, Italian GDP grew by 9.1 percent in 1937, by 2.1 percent in 1938, and by 5.3 percent in 1939.10 Increasing control by the state in industry and banking coincided with a shift toward much greater intervention in monetary and trade affairs. The currency decree of May 26, 1934, (Decree 804) introduced a wide-ranging exchange control and de facto ended the period of the gold standard, marking Italy’s turn away from the globalized world of convertibility and international capital flows. Subsequent decrees tightened the controls: on December 8, 1934, (Decree 1942) it was required that foreign securities owned by Italians should be registered with the Istituto nazionale per i cambi con l’estero; and in August 1936 (Decree 1631), such securities were compulsorily sold to the government in exchange for treasury bonds. Foreign bank accounts had to be declared. Even though at the beginning of August 1936 Italy joined France in devaluing relative to gold, the abandonment of the gold parity was not used as an opportunity to move back to convertibility or to end exchange controls. At the same time, trade policy was turned into a tool of diplomacy. The bilateralization of trade was a way of contributing to the political reordering of Central Europe. In May 1934, for instance, Italy concluded an agreement with Hungary to buy one million quintals of Hungarian wheat, with an option for a further million, at prices double those prevailing on the world market (Ránki 1983, 139). Within two years, Italy’s share of Hungarian trade grew from 7.8 to 13.6 percent. Italian trade was also reoriented toward her colonies: Italian colonies accounted for less than 3 percent of Italian exports in the late 1920s, but a quarter of total exports between 1936 and 1939. They were a particularly important market for industrial products, and especially “advanced” industries, such as chemicals and engineering (Chapter 12; Federico 1998). Again, however, it is important to stress that this was a general phenomenon during the 1930s, and not one limited to Italy, as Table 2.6 makes clear. Unlike in some countries, notably Britain and Germany in 1931 or the United States in 1933, the turn away from the principles of an open world economy was not directly associated with financial crisis. Openness to the world economy seemed a less attractive strategy after the state of capital markets ruled out the possibilities of access and renewed borrowing. The Great Depression killed off this major attraction of a globalized world for the Italian state, concerned as ever with the implications of Italian capital scarcity. However, it was the currency implications of managed trade that really made for the final shift of Italian politics against the liberal economic view, or against globalization, in May 1934. The much larger interventions of the 1930s left a troubled legacy for the postwar period. The establishment of the industrial holding company IRI made virtually all of Italian large-scale industry dependent on the state. It thus provided an immediate lever for quite spectacular growth at a time when the world

Table 2.6 The share of formal and informal empire trade, 1929–1938 (percent) In imports

In exports

Trade of

Share of

1929

1932

1938 1929 1932 1938

United Kingdom

British Commonwealth, colonies, protectorates

30.2

36.4

41.9

United States

Philippines

2.9

6.1

4.8

France

French colonies, protectorates, and mandated territories

20.9

25.8

Belgium

Belgian Congo

3.9

3.8

8.3

2.6

1.3

1.9

The Netherlands

Netherlands overseas territories

5.5

5

8.8

9.4

5.9

10.7

Italy

Italian colonies and Ethiopia

1.5

1.1

1.8

2.1

3.6

23.3

Portugal

Portuguese overseas territories

7.9

10.4

10.2

12.7

13.9

12.2

Japan

Korea and Formosa

12.3

26.2

30

16.8 21.6

32.9

Kwantung

6

4

1.6

4.8

6.8

13.7

Manchuria

1.9

2.7

9

2.5

1.5

8.1

Rest of China

5.8

4

4.4

10.9

7.3

8

Total Japanese sphere of influence Germany

12

1.6

49.9

2.8

2.8

18.8 31.5

27.5

36.9

45

35

37.2

62.7

4.5

5.5

12

5

3.9

13.2

Latin America

12.2

11.2

15.6

7.8

4.3

11.5

Total German sphere of influence

16.7

16.7

27.6

12.8

8.2

24.7

Bulgaria, Greece, Hungary, Romania, Turkey, Yugoslavia

Source: League of Nations (1939, 34–35).

26

44.4 45.4

62

aggregate growth and policy

economic environment was favorable. In this way, the interwar period’s policies can be seen as having created a springboard for the postwar miracle.11 But they also created and augmented a host of political economy problems, in particular the heavy engagement of the state in wage bargaining and wage determination, and a corporate governance in which managers were attached to ideas of gigantism or elephantism. This tension was not a uniquely Italian phenomenon. As Eichengreen (2007) has pointed out, institutional environments that were wellsuited to producing catch-up growth across Europe after 1950 were not necessarily appropriate once convergence on the technological frontier had been largely achieved.

Migration Policies A survey of Italian economic policies during this period is not complete without some mention of policies related to the migration flows discussed in Chapter 10. As a poor, capital-scarce country, Italy had an incentive to attract inflows of capital, but an alternative way of increasing capital-labor ratios (and land-labor ratios into the bargain) was to export people. A premise of liberal Italy was that emigration helped to solve national problems: unemployment and underemployment, and the overpopulation of the South. Liberal politicians also liked to emphasize the economic benefits to Italy that followed from remittances. The liberals resisted the idea that emigration represented some kind of loss of national force or vitality, and generally presented the phenomenon of a “greater Italy” that spanned the world as a trade, wealth, and power-enhancing exercise. The phenomenon was most graphically depicted by the young economist Luigi Einaudi in 1899, when he suggested that the phenomenon of emigration was the reason that “Made in Italy” was eventually destined to overtake “Made in Germany” (Einaudi 1900). However, by the end of the nineteenth century, these arguments became increasingly contested on economic and ideological grounds. The agrarians in particular worried that emigration was causing an excessive and debilitating rise in rural wages. Nationalists, such as the novelist Enrico Corradini, recast this case in ideological terms (Choate 2008). When it tried to regulate emigration, the initial concern of the Italian government was to end the abuses and excessive charges of the shipping companies, and thus by effectively reducing the price, to facilitate the phenomenon of emigration. A major reform came in 1901, with a law that swept away the previous system of agents and subagents, although in practice the people who had previously organized emigration remained under a new name (“vettori ”). The law established a General Commissariat of Emigration, with financial resources intended to sustain Italianness abroad (Italianità) through guidance in Italy and through the provision of educational resources abroad and the

italy and the first age of globalization, 1861–1940

63

creation of a “chain of assistance”; in addition, the Commissariat superintended the funneling of remittances through the Banco di Napoli. The rural post office network was also expanded to facilitate the relatives of migrants (Esteves and Khoudour-Castéras 2011). Another example of domestic reforms associated with emigrant welfare came with the 1904 Franco-Italian labor treaty. This was part of a wave of international treaties that were designed to protect nascent welfare states from the pressures to which they were being subjected by migration. Because the social insurance programs of the period were typically restricted to national citizens, immigrants provided a potential way for employers to circumvent the new rules and regulations. Immigration thus risked undermining the social compacts of the time, and indeed immigration from Italy and elsewhere caused strikes and even riots in France on occasion. One solution was for the host country to extend welfare benefits to immigrants, on condition that the source country implement similar reforms domestically; this also had the benefit of leveling the playing field for employers producing in different countries. This was exactly the bargain struck in 1904, and the Franco-Italian treaty became the model for similar treaties in the following years (Huberman 2012). Although the Commissariat was primarily targeted at overseas migration, the Franco-Italian treaty reminds one of the growing importance of Italian emigration to other European countries. Italian nationalists began to emphasize that flows of migration in Europe were more likely to be compatible with the maintenance of a strong national identity. In the immediate aftermath of World War I, and in part as a result of changing and more restrictive legislation in the United States, transoceanic migration fell more sharply than nonoceanic flows. That result pleased the nationalists. There was no immediate break after the establishment of Mussolini’s rule.12 Indeed, at the beginning, Mussolini seemed to echo the language of prewar liberal politicians. In a 1923 speech, for instance, he said: “For better or worse emigration is a physiological necessity of the Italian people. We are forty millions hemmed in in our adorable but narrow peninsula. So it follows that the problem of Italian expansion in the world is a life and death issue for the Italian race” (Primiceri 2010, 140). A few years later, however, the tone of his speeches changed and he began to refer to the “evil” of migration. In 1927, the Commissariat was abolished and replaced by a more state-controlled Directorate. There were some administrative restrictions put in place to try to make emigration more difficult (e.g., the need for would-be emigrants to show that they had either a work permit or relatives abroad), but these restrictions were promptly lifted in 1930 with the onset of the Depression. The new emphasis was more on controlling the politics of Italian emigrants and tying them in to the cause of an international propagation of Fascism, an Italian policy innovation that would later appeal to German Nazis. The Fascist philosopher and politician Giovanni Gentile told the Italian parliament in 1927 that solving the “demographic problem” had an internal and an external aspect, and that the

64

aggregate growth and policy

internal aspect required the vigorous pursuit of developmental policy, of “using to the maximum extent all the resources that our earth can offer” (Primiceri 2010, 198). By the 1930s, the government would also present emigration to the colonies as a key aspect of Italian development. Mussolini justified the invasion of Abyssinia as a search for an outlet for surplus Italians. The original vision of completely free migration did not work, or did not work fast enough, because Italy’s problems seemed to persist. There was in consequence a demand for more regulation, but even at the height of the Fascist era, Italian politicians saw an outflow of population as one part of a solution to the “demographic problem.” It just turned out to be an increasingly controlled and regimented solution, which turned into a justification of imperialism rather than a vision of a peaceful greater Italy spread over the world.

The Institutional Bases for Convergence in a Capital-Scarce Economy The previous sections have shown that Italian trade, financial, and industrial policies during the period in question were not all that unusual when viewed in an international context. They reflected the capital-scarce nature of the Italian economy, and a desire to achieve military security and economic convergence on the leading economies of the day. These preoccupations were hardly unique to Italy. Insofar as we have been able to identify Italian policy specificities, these largely involved interwar industrial policy: IRI was a uniquely Italian institution, even though it was clearly of its time. More generally, the Italian response to the Depression was relatively successful by the dismal standards of the time, and may even have helped provide some of the foundations for the postwar economic miracle discussed in Chapter 3. Migration policy was also considerably more developed in Italy than in Ireland, where it was left to such institutions as the Gaelic Athletic Association and the Catholic Church to provide emigrants with cultural links to their homeland. Italian policies may not have been that out of the ordinary, but how well did they succeed in achieving their aims? Italian military history is beyond the scope of this chapter, but how well did Italy succeed in achieving its economic objectives? Standard neoclassical economics suggests that economically backward countries should be able to benefit from globalization, and see their convergence on the technological frontier speeded up as a result of technological transfers, trade, and capital imports. Furthermore, emigration may help them to achieve convergence in per capita terms, if at a cost in terms of a diminishing population. However, a vast empirical literature emphasizes that this convergence is

65

italy and the first age of globalization, 1861–1940

not automatic: it requires that various social and institutional prerequisites be put in place for convergence to be achieved. The Italian economy saw major institutional reforms over the course of the period under review, which were successful in spurring growth in the short run, but which would eventually outlive their usefulness as the international economic environment changed. The first major wave of reforms highlighted here was the banking reforms instituted after the financial crisis of 1893. This crisis was intimately linked with the capital inflows that the new Italian state had actively sought, but with whose consequences its banks (no more than the Thai and Korean banks in the 1990s or Icelandic, Irish, or Spanish banks in the early twenty-first century) were ill-suited to cope. This crisis was a salutary one in many ways, with the new universal banks ushering in a period of faster economic growth (Figure 2.8). Between unification and 1899, economic growth in Italy was just 0.6 percent per annum, but it more than doubled to a respectable 1.3 percent between 1899 and 1915. Figure 2.9 shows that 1900 marks the dividing line between a period when Italian incomes were falling steadily further behind incomes in the leading economies of the day, and a period when, at least for a time, Italy managed some convergence vis-à-vis France, Britain, and Western Europe as a whole, and stopped losing ground relative to Germany and the United States.13 Between 1921 and 1929, the Italian economy experienced extremely strong growth, at a rate of 3.1 percent per annum.14 Unfortunately, the deflation of the 1920s (which was a worldwide rather than an Italian phenomenon) placed the universal banking model under severe strain, and the worldwide Depression that followed finished it off. The result was another crisis-driven response, which again can be seen as salutary, if attention is restricted to its economic dimension. As Figures 2.8 and 2.9 show, the Italian recovery from the Depression was quite strong in absolute and 4,500 4,000

2009 euros

3,500

3,000

2,500

2,000 1870

1880

1890

1900

1910

1920

1930

1940

Figure 2.8 Italian GDP per capita, 1861–1940 (2009 euros). Source: Banca d’Italia database.

66

aggregate growth and policy France

Germany

1.0

1.0

0.9

0.9

0.8

0.8

0.7

0.7

0.6

0.6

0.5

0.5 1870 1880 1890 1900 1910 1920 1930 1940

1870 1880 1890 1900 1910 1920 1930 1940

UK

USA

.56

.64 .60

.52

.56 .48

.52

.44

.48 .44

.40

.40

.36

.36 1870 1880 1890 1900 1910 1920 1930 1940

1870 1880 1890 1900 1910 1920 1930 1940

Spain

Western European 12

1.6

.75

1.5 .70

1.4 1.3

.65

1.2 1.1

.60

1.0 0.9

.55 1870 1880 1890 1900 1910 1920 1930 1940

1870 1880 1890 1900 1910 1920 1930 1940

Figure 2.9 Italian per capita GDP relative to those of other countries, 1861–1940. Source: Banca d’Italia database; Maddison (2010). 30,000

2009 euros

20,000

10,000 8,000 6,000 5,000 4,000 3,000 2,000

1,000 1875

1900

1925

1950

1975

2000

Figure 2.10 Italian GDP per capita, 1861–2008 (2009 euros). Source: Banca d’Italia database.

67

italy and the first age of globalization, 1861–1940

.035

Spain Portugal

Growth 1950−2008

Italy Austria

.03

Norway

Finland

Germany

.025

.02

Belgium France Netherlands Denmark Australia UK Canada Sweden

USA

Switzerland

.015

NewZealand

2000

4000 6000 8000 1950 income per capita (1990 international G−K dollars)

10000

Figure 2.11 Initial income and subsequent growth, 1950–2008. Source: Banca d’Italia database; Maddison (2010).

relative terms, especially in the second half of the 1930s, and it paved the way for the extraordinary success of the postwar period (Figure 2.10): between 1945 and 2008, per capita GDP increased more than tenfold, an astonishing performance, even taking into account the collapse in output during World War II (see Chapter 3). This was a period of supremely successful convergence on the technological frontier internationally, but Italy, in large part because it had a long way to catch up, was one of Europe’s star performers (Figure 2.11). Yet again, however, the new institutional framework became less of an advantage, and more of a handicap, in Italy as in other rapidly growing catch-up economies. By the end of the century, Italy faced a crisis of slow economic growth, in large part a result of its successful convergence experience, but until now this crisis has been insufficiently dramatic to force another round of major structural change in the economy. In the past (notably in the 1890s and the 1930s), deep crises were associated with a fundamental change of the policy paradigm. Crises were a learning experience. Will the intensification of the crisis in 2011, the 150th anniversary of unification, change this assessment? In Italy, as in the rest of Europe, we have not yet come to the end of (economic) history.

Acknowledgment This paper is a revised version of a paper originally presented at the conference on Italy and the World Economy, 1861–2011, held at the Banca d’Italia, Rome, 12–15 October 2011. The authors are extremely grateful for their comments and suggestions, support, and help with data to Gianni Toniolo; to the discussants

68

aggregate growth and policy

Stefano Fenoaltea and Vera Zamagni; to participants at the conference; and to Alberto Baffigi, Nick Crafts, Giovanni Federico, Alfredo Gigliobianco, Michael Huberman, and Cormac Ó Gráda. The research leading to these results has received funding from the European Research Council under the European Union’s Seventh Framework Programme (FP7/2007–2013) / ERC grant agreement no. 249546. The usual disclaimer applies.

Notes 1. Our description of the course of Italian trade policy draws heavily on such sources as Coppa (1970), Cohen and Federico (2001b), Federico (2006), and Toniolo (1990). 2. The policy would have suited Northern landowners at least, if not the owners of Southern latifundia. 3. Toniolo (1977b) confirms that the duties on iron and steel harmed the engineering sector, but doubts whether this had a significant impact on overall rates of Italian growth. 4. Quasi-classical, because classical models achieved this fixity of the real grain wage by appealing to Malthusian mechanisms, rather than migration. 5. This argument is of necessity a counterfactual one: wheat prices were in fact declining, because Italian tariffs did not rise enough to counteract the impact of continually falling transport costs and declining supply prices on the frontiers of the New World. One can certainly argue that in the absence of these tariffs, wheat prices would have declined by even more. 6. The latter country is not included in these tables, but had very high tariffs: its average tariff in 1913 was 21.4 percent, according to Bairoch (1989, 76). 7. The Italian prices are wholesale prices; the British prices are the Gazette averages; the other prices are market averages. 8. The Credito Italiano was founded in Genoa as the Banco di Genova, but in 1907 moved its headquarters to Milan. 9. Archivio Storico Banca Intesa, BCI, Sofindit. 10. Banca d’Italia database. 11. Similarly, Alexander Field (2011) has argued that the interwar period provided the springboard for postwar US growth. 12. For a useful survey, see Cannistraro and Rosoli (1979). See also the discussion in Ipsen (1996). 13. The exercise is a crude one: the new Banca d’Italia data are rebased using Maddison’s (2010) figure for 1990 (expressed in 1990 Geary-Khamis dollars) as a benchmark. 14. Growth rates here and subsequently are all based on regressions of the log of per capita income against a time trend.

Chapter 3

THE GOLDEN AGE AND THE SECOND GLOBALIZATION IN ITALY nicholas crafts and marco magnani

Introduction This chapter examines Italy’s growth performance in the period since World War II. We develop an interpretation of catch-up growth from an initial position where the level of per capita income was well below that of the leading economies, and then explore the experience of recent decades where the issue becomes one of further modernization to maintain a strong position among the elite. Economic policy decisions, in particular those made in response to challenges emerging from the external economic environment, have played a major role in determining productivity performance through their impact on the quantity and quality of investment and innovation. Evaluation of their part in growth outcomes is the central concern. An important starting point for analysis is the proposition that the process of catch-up growth typically entails a series of ongoing reforms with the danger that at some point the political economy of the next step in modernization becomes too difficult. As modern growth economics stresses (Aghion

70

aggregate growth and policy

and Howitt 2006), the institutions and policy choices that can galvanize a far-from-frontier economy differ in many ways from what is appropriate for a close-to-frontier economy. In particular, in the latter case stronger competition in product markets and high-quality education become more important. Similarly, as new technologies come along, institutions and policies need to be reformed. Yet, making the requisite adjustments may be problematic and achieved only slowly and incompletely such that catch-up growth falters. The constraints of the historical legacy are important in this context. The international economic environment affects growth performance in several important ways, although its impact depends on the details of the domestic policy response and interactions with domestic institutions. It follows that each country’s performance has a common component based on external circumstances and an individual component based on its own distinctive behavior shaped by history and politics, even though ideas as to what is “best-practice” policy are strongly correlated across the Organization for Economic Cooperation and Development (OECD) countries. Key aspects of developments in the world economy include high-profile shocks to aggregate supply (e.g., oil-price rises) and to aggregate demand (e.g., financial crises). Dealing with these macroeconomic disturbances can entail policy choices with long-term supply-side consequences. In turn, the extent of policy discretion depends on the international economic architecture of monetary arrangements, treaty obligations, and so forth, which influence the prospects for export-led growth and the extent of globalization. As globalization proceeds, changes in the international division of labor imply the need for structural change, and this places a premium on flexibility in domestic markets. Finally, technological change, which overwhelmingly comes from other countries’ research and development (R & D), provides opportunities for faster growth but exploiting them well may depend on policy reform. Across the world, the diffusion of and convergence in the use of technology has sped up and has been about three times as rapid post-1950 compared with pre-1925 (Comin, Hobijn, and Rovito 2006). The international economic environment with which the Italian economy interacts has been transformed in the last sixty years. The trajectory has been toward globalization (i.e., the greater economic integration of capital and product markets especially within Europe but also across the world as a whole), starting from a position of very low capital mobility and trade protectionism after the disruptions of the Great Depression and World War II. Tables 3.1 and 3.2 reflect these trends. Table 3.1 shows a big decline in trade costs during the postwar period; these trends are dominated by reductions in protectionist barriers rather than declines in transport costs. This is apparent from the variations for Italian trade with different partners, which reveal that European integration has been the most powerful influence on trade costs. The European Economic Community (EEC) stimulated both “juggernaut” and “domino” effects, which consolidated the

71

the golden age and the second globalization in italy Table 3.1 Trade costs: Italy with various partner countries Germany

UK

Spain

USA

India

1929

1.10

1.22

1.63

1.26

1.54

1950

1.27

1.36

2.40

1.40

1.97

1960

1.01

1.25

1.54

1.29

2.00

1970

0.79

1.21

1.42

1.22

2.30

1980

0.61

0.86

1.08

1.13

1.86

1990

0.56

0.84

0.87

1.13

1.85

2000

0.66

0.90

0.87

1.14

1.83

Notes: Trade costs are inferred from a gravity model of trade flows and comprise both policy and non-policy barriers to trade; world average trade costs normalized to 1950 = were 0.98 in 1970, 0.82 in 1980, and 0.84 in 2000. Source: Data underlying Jacks, Meissner, and Novy (2011) generously provided by Dennis Novy.

original trade agreement by increasing the weight of business favoring further liberalization and encouraging enlargement and multilateral tariff reductions (Baldwin 2006). Table 3.2 reflects the very low level of international capital mobility in the early postwar period when OECD countries, including Italy, generally controlled capital movements very tightly while making their macroeconomic trilemma choice for a combination of independent monetary policy and a fixed exchange rate under the auspices of the Bretton Woods international monetary system (Obstfeld and Taylor 2004). From the 1980s, the globalization of capital markets accelerated sharply and Italy moved to full capital-account liberalization by the 1990s, a very different policy stance compared with the early 1950s. After World War II, growth of world trade by volume was very strong. It outstripped world gross domestic product (GDP) growth and averaged 7.9 percent during 1950–1973 (World Trade Organization 2008), even though the world average decline in trade costs was only about 2 percent before 1970 in a period when trade liberalization mainly concerned OECD manufacturing. From 1973 to the eve of the current crisis, world trade growth slowed to 4.8 percent per year but was still much faster than GDP growth. A striking feature of these years was the emergence of a new international division of labor in which Asian countries became much more important as exporters of manufactures, whereas the shares of Europe and North America contracted sharply (Table 3.3). This reflected Asian success in putting in place policies and institutions that promoted rapid catchup growth first in Japan, then in the Asian Tigers, followed by China and India (Table 3.4). Adjustments to this new exporting prowess were required, especially of such countries as Italy where revealed comparative advantage was positively correlated with that of dynamic Asia (Rae and Sollie 2007).

72

aggregate growth and policy Table 3.2 Capital account openness (0–100) and foreign-owned assets/world GDP (percent) a) Capital account openness World Average 1890–1913

99.7

Italy 1950–1954

37.5

1955–1959

50

1949

42

1960–1964

67.5

1950s

53

1965–1969

75

1960s

55

1970–1974

75

1970s

47

1975–1979

75

1980s

50

1980–1984

75

1990s

68

1985–1989

85

2000–2005

78

1990–1994

92.5

1995–1999

100

b) Foreign-owned assets/world GDP 1870

7

1914

18

1945

5

1960

6

1980

25

2000

92

Sources: Capital account openness from Quinn (2003) and data appendix to Quinn and Toyoda (2008); foreign-owned assets/world GDP from Obstfeld and Taylor (2004).

Table 3.4 reports that the period from the early 1950s to the first OPEC oil shock, the so-called Golden Age of European growth, was one where the world economy as a whole grew fast. These were the years of the Bretton Woods system of pegged exchange rates and, quite unlike the interwar period, an era when OECD countries were largely free from financial crises and macroeconomic turbulence (Bordo 1993). It was an ideal opportunity for countries with undervalued exchange rates, faced with buoyant export demand, to industrialize and catch up (Boltho 1996). Conversely, the stagflation of the 1970s, driven by adverse aggregate supply shocks, was an environment that posed great challenges for macroeconomic policymakers who faced difficult trade-offs between objectives. For many European countries, including Italy, the policy framework that eventually emerged was a trajectory toward a single currency via the European monetary system of the 1980s and 1990s. This was an era of slower macroeconomic growth but, after the 1980s, a combination of good luck in terms

73

the golden age and the second globalization in italy Table 3.3 Shares of world output and world exports of manufactures (percent) a) Output 1953

1973

1990

2007

Western Europe

26.1

24.5

32.9

22.4

North America

46.9

35.1

23.4

25.7

Japan

2.9

8.8

16.8

16.5

China

2.3

3.9

2.7

11.2

India

1.7

2.1

1.5

1.6

Other Asia

1

3.1

4.9

8.6

19.1

22.5

17.8

14

1953

1973

1990

2007

Western Europe

51.9

55.9

54.2

40.8

North America

35.8

16.1

15.2

11.9

Japan

2.9

9.6

11.5

6.7

China

0.1

0.6

1.9

11.9

Rest of Asia

1.6

4.5

11.1

16.5

Rest of world

7.7

13.3

6.1

8.3

Rest of world b) Exports

Sources: Output from Bairoch (1982), United Nations (1965), and UNIDO (2002) (2009); exports from United Nations (1958; 1976) and World Trade Organization (2001; 2008).

Table 3.4 Levels and growth rates of real GDP/person at purchasing power parity 1950

1973

2007

Growth Rate 1950–1973 (percent p.a.)

Growth Rate 1973–2007 (percent p.a.)

Western Europe

4,569

11,392

21,589

4.05

1.91

United States

9,561

16,689

31,357

2.45

1.88

Japan

1,921

11,434

22,950

8.07

2.07

China

448

838

6,303

2.76

6.12

India

619

853

2,817

1.41

3.57

Asian Tigers

955

3,631

21,212

5.98

5.34

2,111

4,083

7,468

2.91

1.80

World average

Source: Maddison (2003) with updates from website. Levels refer to 1990 Geary-Khamis dollars.

74

aggregate growth and policy

of the absence of adverse shocks and the adoption of inflation-targeting policies produced macroeconomic stability, the so-called Great Moderation (Stock and Watson 2005). Fast European growth in the 1950s and 1960s was based on policies and institutions that facilitated high rates of investment and the diffusion of American technology in the era of Fordist manufacturing. Broadly speaking, this was a period where, to use the terminology of Hall and Soskice (2001), “coordinated market economies” prospered. This was generally a period of greater government intervention in terms of regulation, state ownership, and industrial policy. In later decades, growth opportunities centered more on the services sector and the diffusion of information and communications technologies (ICT). This seems to have been an era that placed a greater premium on flexible adjustment in labor and product markets and on intangible capital accumulation rather than on massive investment in physical capital (Nicoletti and Scarpetta 2005; Timmer et al. 2010). In this era, the “liberal market economies” were better placed. However, as ideas of “best practice” in microeconomic policy changed, there was a general change of emphasis among OECD countries toward competition and privatization policies. To a considerable extent, Italy has shared in these general European trends. At the same time, it is clear that in some respects the Italian experience differs. With regard to institutions, as Hall and Soskice (2001) note, Italy has represented an “ambiguous case,” neither a CME (German-style) nor an LME (American-style). Concerning policy responses, the Italian approach to coping with the stagflationary shocks of the 1970s resulted in a relatively high inflation economy and led to the build-up of a high ratio of public debt to GDP in the 1980s.1 The question of economic performance in the South marks Italy as a special case. In terms of growth outcomes, the key feature is that as a far-from-frontier economy in the age of Fordist manufacturing, Italy outperformed during the Golden Age; however, as a close-to-frontier economy in the ICT era, Italy has disappointed over the last twenty years. This highlights the general question of whether Italy was more in need of policy reform than its European peer group after the Golden Age or whether policy reforms were implemented more slowly or less successfully. This in turn raises a number of other issues that are explored in what follows. These include evaluating the pluses and minuses of the postwar settlement in the short and the long run; reviewing supply-side policies relevant to consideration of the implications of decisions Italy made with regard to the European integration agenda; and the possible reasons for the difficulties of implementing structural reforms, in particular after the Golden Age. The analysis ends in 2007 on the eve of the crisis, which is not explicitly considered. In so far as Italy’s difficulties during the present travails of the Eurozone have been exacerbated by inadequate public-finance and supply-side reforms, this discussion of Italy’s economic history provides some context for, although not an analysis of, these problems.

the golden age and the second globalization in italy

75

Table 3.5 Levels and rates of growth of real GDP/person in Western European countries, 1950–1973 (percent per year) 1950

1973

Switzerland

9,064

18,204

3.08

Denmark

6,943

13,945

3.08

United Kingdom

6,939

12,025

2.42

Sweden

6,739

12,494

3.06

Netherlands

5,971

13,081

3.45

Belgium

5,462

12,170

3.54

Norway

5,430

11,324

3.24

France

5,186

12,824

4.02

West Germany

4,281

13,153

5.02

Finland

4,253

11,085

4.25

Austria

3,706

11,235

4.94

Italy

3,502

10,634

4.95

Ireland

3,453

6,867

3.03

Spain

2,189

7,661

5.60

Portugal

2,086

7,063

5.45

1,915

7,655

6.21

Greece

Growth Rate 1950–1973

Source: Conference Board (2011). Levels refer to 1990 Geary-Khamis dollars.

Italy in the Golden Age It has become conventional in comparative studies of European economic growth to consider the period 1950–1973 as the Golden Age. This was a period of beta-convergence in which European countries generally experienced rapid catch-up growth based on technology transfer and structural change, and growth rates were inversely correlated with initial income levels. The basis of catch-up varied over time; initially, postwar reconstruction and reduction in the size of the agricultural sector played a part, but by the later 1960s reductions in technology gaps were the main source (Temin 2002). During the Golden Age, Italy enjoyed growth of real GDP per person at almost five percent per year, as seen in Table 3.5. This reflects the relatively large scope for catch-up reflected in the low initial income level in 1950. However, normalizing for initial income, Italy outperformed its European peer group and growth was about 0.65 percentage points per year faster than might have

76

aggregate growth and policy Table 3.6 Contribution of structural change to labor productivity growth, 1950–1973 (percent per year) Orthodox Measure

Broadberry Measure

Denmark

0.24

1.10

France

0.00

0.52

Italy

0.83

1.77

–0.31

0.29

Spain

0.80

1.77

Sweden

0.00

0.60

Netherlands

United Kingdom West Germany

–0.12

0.31

0.18

0.77

Source: Derived from van Ark (1996) based on a three-sector decomposition (agriculture, industry, services) where agriculture is deemed to be the declining sector. The orthodox approach considers that the contribution of structural change equals ΔAO/AO —ΣΔAi/Ai*Ai/AO*Si, where A is labor productivity, S is share of employment, and the subscripts O and i stand for the whole economy and sector i, respectively (Nordhaus 1972). Broadberry (1998) modified this so that labor productivity growth in declining sectors is measured using the overall rate of labor force growth not the sectoral rate.

been expected (Crafts and Toniolo 2008). Labor productivity growth averaged 6.8 percent per year in industry and 5.7 percent per year in agriculture during these years (Broadberry, Giordano, and Zollino 2012). The weight of agriculture in the economy fell rapidly; the sector accounted for 44 percent of employment in 1951 but only 18 percent in 1973. Italy enjoyed a remarkably rapid rate of total factor productivity (TFP) growth at this time; from 1945 to 1973 the average was 5.8 percent per year (Broadberry, Giordano, and Zollino 2012). As catch-up proceeded from a starting point of “backwardness,” TFP growth partly came from reductions in inefficiency, especially resulting from the movement of labor out of agriculture (Table 3.6), and from the realization of substantial economies of scale (Rossi and Toniolo 1996). In part, TFP growth reflected successful technology transfer together with creative adaptation and localized learning at which Italy excelled (Antonelli and Barbiellini-Amidei 2009), even though the country did not have a national innovation system that exhibited high levels of R & D or rates of patenting. In the 1960s, Italy spent only about 0.6 percent GDP on R & D (about one-third the West German level), and in 1973 obtained only 3.4 percent of the foreign patents in the United States compared with a West German share of 24.2 percent (Verspagen 1996; Pavitt and Soete 1982). Human capital was on the whole favorable. Although average years of schooling were low (4.8 years in 1950 rising to 6.8 in 1970; Morrisson and Murtin 2009), human capital was suited to incrementally adapt the imported technologies that

the golden age and the second globalization in italy

77

constituted the main channel of innovation. In comparative terms, Italy benefited from a high number of good engineers who played an essential role in designing the organization of productive processes in large firms. This expertise was inherited and further developed largely by public firms, which were the main promoters of convergence of the Italian industrial structure toward that prevailing in Europe (De Nardis and Trau 2005; Barca and Trento 1997). This evolution paralleled and complemented the earliest stages of the process driven by small-to-medium firms, often forming industrial districts specialized in traditional products, which accelerated dramatically later in the 1970s. The convergence process of the Golden Age was, however, incomplete: high-technology sectors were less well developed than in the other major European economies, as is reflected in the trade performance data reported in Federico and Wolf (2011). Moving to a trajectory of rapid catch-up growth was not automatic but required an appropriate institutional and policy framework, “social capability” in the well-known terminology of Abramovitz (1986). In this regard, a key ingredient in Western Europe was the postwar settlement, which for most countries was reached in the context of the Marshall Plan (Judt 2010). It is not entirely clear how much difference the Marshall Plan made to early postwar Italy. It is unlikely that growth was much affected by the direct economic effects of the $1.5 billion that Italy received from the United States in the years 1948–1951, amounting to 2 percent of GDP compared with an average of 2.5 percent for all countries (Eichengreen and Uzan 1992), but it may have been an important factor in deciding distribution of the costs of stabilization; in supporting modernization of industrial plants, especially in the electrical, steel, and metals sectors (Fauri 2010); and in heading off a possible “war of attrition” (Casella and Eichengreen 1994). It also added impetus to the move toward trade liberalization and the European Payments Union. The main features of the postwar settlement in Italy were early stabilization in 1947, a political constitution that embodied checks and balances, isolation of a strong Communist Party, progressive opening to international competition, and participation in the Bretton Woods system. Italy was at the forefront of the European integration process, moved relatively quickly to reduce barriers to international trade, and joined the Bretton Woods system at an undervalued parity.2 These decisions were strongly influenced by the positioning of Italy in the Western bloc led by the United States. At the same time, they were seen by the leading economic authorities (notably by Luigi Einaudi, the Governor of the Bank of Italy and, in 1947, Minister of the Budget) as useful steps to foster competition and to weaken the industrial oligopolistic groups. However, unlike what happened in the United Kingdom in 1948 and later in Germany where competition authorities were created, on the domestic front this aim was not accompanied by corresponding attempts to implement competition policies; corporate governance structures remained closed, and sectors sheltered from international competition (retail, professional orders, public utilities) remained highly regulated.

78

aggregate growth and policy

The domestic policy stance reflected the heterogeneous nature of the main political forces and interest groups. Free market supporters were in a minority—because of the prevailing ideological climate in Europe after the war and because of Italy’s historical legacy—and had to face not only the isolated but socially influential communist and socialist opposition but also the hostility of an important component of the ruling party (the Catholic Democrazia Cristiana), which favored an active role for the State in the economy. A similar confidence in the necessity of public intervention to promote accumulation and a corresponding mistrust of the capacity of private groups to accomplish this task shaped the ideas of leading managers of State-owned enterprises (Institute for Industrial Reconstruction [IRI]), among them Donato Menichella, general director of IRI in 1933–1944 and Governor of the Bank of Italy from 1948 to 1960. A distinctive feature of the postwar settlement was preservation of the unusually large role played by firms and banks controlled by the State, but operating as private profit-oriented autonomous entities concentrated in IRI, created in 1933 after the collapse of the big private financial and industrial groups. It was mainly the political acumen of the Prime Minister and leader of the Democrazia Cristiana, Alcide De Gasperi, which enabled a balance to be struck between these different forces, and between them and the interests of private industrial firms. After World War II, for the first time in Italian history the South—since 1861 the most prominent structural issue of the Italian economy (Iuzzolino, Pellegrini, and Viesti 2011)—became the object of specific systematic policies. At the outset of the Golden Age, the South benefited very little from European Recovery Program (ERP) funds (Fauri 2010; Del Monte and Giannola 1978). The promotion of growth in the South was not a priority of the policy agenda (Cafiero 1996). However, two factors rapidly changed this picture. First, acute social tensions in the huge Southern agricultural sector, largely based on large estates, greatly endangered political stability and the integration of Italy in the Western bloc. Second, the influence and quality of IRI management was a fundamental element in promoting a specific policy aimed at the reduction of the North-South gap. This policy, “l’intervento straordinario” (special development policy), was based on the principle of additionality with respect to “ordinary” policy (i.e., was targeted specifically at the backward areas and was therefore “extra” with respect to “ordinary” capital expenditure), and was conceived as a decisive part of Italian economic policy as a whole. It was animated by the belief that closure of the income gap, far from being the long period outcome of market forces (as mainstream economist thought had maintained), had to be achieved by policy means. The growth of the South and of Italy was seen as one thing: this was the central idea animating promoters of the special development policy and later of the structural-reform policy (programmazione) during the center-left governments at the beginning of the 1960s (La Malfa 1962). The main instrument of the special development policy, the “Cassa del Mezzogiorno,” was a body independent from the Public Administration.3

the golden age and the second globalization in italy

79

A significant degree of autonomy from ordinary Public Administration was a feature of Italian economic policy in the Golden Age, not just regional policy. The most successful period of the Cassa coincided with this period. Between 1951 and 1973, the North-South gap in terms of GDP per head experienced a strong reduction for the first and only time in history. In the Golden Age, the South had positive effects on the growth of the rest of Italy. Low-skilled and cheap labor, particularly suitable for the Fordist mode of production, poured from the Southern agricultural sector into the Northern plants. Especially before the creation of the European Common Market, the growing market in the South increased demand for Northern products. At the end of the 1950s, 70 percent of net exports of the Center-North were directed toward the South. The growth of the domestic market allowed economies of scale to be realized in Northern firms (Iuzzolino, Pellegrini, and Viesti 2012). Trade liberalization was an important factor in Western Europe’s Golden Age. Between the late 1950s and early 1970s, it raised the average growth rate by about one percent per year (Badinger 2005). Countries that were slow to liberalize, such as Spain and Ireland, paid a price in terms of slower growth. In the Italian case, the move to greater openness facilitated a process of export-led growth as Italy’s share of world manufactured exports rose from 3.7 percent in 1950 to 6.8 percent in 1973. The real exchange rate was kept low by the tendency for real wages to grow only a little faster than labor productivity in the context of an elastic supply of labor to manufacturing industries.4 The consequent expansion of the tradeables sector raised its productivity through internal and external economies of scale; according to some estimates, undervaluation of 20 percent on average during the 1950s and 1960s may have raised the growth rate by perhaps 1 percentage point per year at this time (Di Nino, Eichengreen, and Massimo 2012b). Eichengreen (1996) argued that catch-up growth in the European Golden Age was underpinned in many countries by a corporatist social contract that allowed a cooperative equilibrium in which wage restraint was rewarded by high investment, which benefited firms and their workers in the long run. A cross-country regression suggests that until the mid-1970s coordinated wage bargaining does seem to have raised investment and growth rates quite appreciably (Gilmore 2009) but this was only to a very limited extent the case for Italy. Wage moderation in Italy lacked these institutional underpinnings and should rather be seen as a result of weak union bargaining power where, as Lewis (1954) might have put it, labor supply was “unlimited” as labor was redeployed out of agriculture. However, the high degree of political and ideological conflict (Salvati 2000) accompanied by the ruthless attitude of entrepreneurs keen to exploit their bargaining power made this equilibrium potentially unstable. The first strike wave in the early 1960s and, to a much greater extent, the “hot autumn” of 1969 signaled the end of this phase. The subsequent instability of industrial relations confirms the absence of institutions of a coordinated market economy.

80

aggregate growth and policy

With regard to domestic supply-side policy, the postwar settlement was adequate for a process of rapid catch-up for a “backward” economy to be launched, but as Italy got closer to the frontier this was no longer the case. This is epitomized by a distinctive feature of Golden Age Italy, namely the state-owned enterprises IRI augmented by ENI in the energy sector from 1953. Initially, public enterprises were a route to speeding up investment and were given high managerial autonomy, which meant they could achieve economic efficiency. In particular, IRI helped development of the private machinery industry by providing cheap intermediate inputs; compensating for the weakness of private firms in this field was important. By the 1960s, however, public enterprises were increasingly hijacked for political purposes (Barca et al. 1998; Woods 1998). This degenerative process also involved banks owned by IRI. Whereas in the 1950s and 1960s they allocated financial resources efficiently, in subsequent years they were hampered in selecting innovative firms by the growing influence of political networks. The postwar investment boom was largely financed from a high share of profits in national income. Italy possessed neither a CME bank-based nor an LME equity-based financial system but rather one in which corporate governance was largely based on family-controlled pyramidal groups, State-owned large firms, and banks that were themselves very largely under State control (Pagano and Trento 2003). By the end of the Golden Age, as profits came under pressure, this resulted in major problems for the operation and financing of large firms (de Cecco and Ferri 2001). The peak of the Golden Age was reached in 1958–1963 with the culmination of the “economic miracle” that had turned Italy into an industrial economy, but there were emerging tensions that threatened instability, notably the potential conflict between sustaining growth on the basis of low wages and demands for higher consumption resulting from successful economic development. In 1963, at the end of the “economic miracle,” monetary policy suddenly turned restrictive in the face of mounting balance of payments disequilibria and inflation pressures. These latter were driven by the first significant wage push in fifty years, in the context of the abatement of unemployment and a resurgence of union militancy. The policy reaction was successful in stopping wage increases and in avoiding a much feared devaluation and the economy managed to return to an equilibrium (albeit slower) growth path until the beginning of the next decade. In retrospect, 1963 is also noteworthy because of new aspects of the design of economic policy similar in some ways to those emerging elsewhere in Europe. In many countries some sort of vaguely Keynesian-inspired income policies were attempted during the 1960s (e.g., France, Denmark, Germany, and United Kingdom). However, the most ambitious attempt was put forward in Italy, when the socialist party, previously under communist influence, entered the government in 1963 forming a center-left coalition. Because the economy had grown at incredibly high rates in previous years, the general assumption was that this

the golden age and the second globalization in italy

81

speed could easily be maintained and that the main task left for economic policy was to direct the resources produced by growth to favored objectives. To achieve this result a macroeconomic program (programmazione) was designed to tackle the main structural problems of the economy, namely the North-South gap, welfare, antitrust regulation, and company-law reform. Incomes policy would ensure the sustainability of growth by preserving competitiveness and by keeping inflationary pressures under control. With the programmazione there was also an attempt to render industrial relations less adversarial by involving unions in discussions about the compatibility of wage dynamics with macroeconomic objectives and to pave the way for a coordinated market economy. On the whole, the program was heavily influenced by a dirigiste approach not uncommon in European economic culture of the time. Somehow paradoxically, this was also caused by the intellectual contribution to the programmazione of the “liberal-socialisti,” a small but very active group in the second half of the 1950s, aimed at introducing competition on the product and service markets, in particular in public utilities, and at opening the structures of corporate governance of private firms. Both anticipated central themes of the policy debate decades later. At that time, the latter was successfully countered by the employers’ organization. The competition issue was tackled by the proposal to dismantle existing monopolies not by ensuring a proper competition-promoting regulation of the market but by nationalization of entire sectors. Whereas the case of failure of the market was acknowledged, the case of failure of the State was still to be taken into consideration. Policy reform proved unrealistic, however, for a number of reasons. On the macroeconomic side, the momentum of growth gradually decreased. On the whole, firms encountered significant difficulties in speeding up innovation to compensate for the waning of the catch-up model based on “unlimited” labor supply and on technological transfer. The growth of fixed investment, mainly aimed at implementing labor-saving technologies, slowed from 11 percent on a yearly average in 1958–1963 to 5 percent in 1964–1973. Current account surpluses (2 percent of GDP on average at this time) were matched by capital outflows surging to $1 billion annually (Biagioli 1995). The first, by changing relative factor intensity, reflected the aim of containing the squeeze of profits caused by wage increases of 1961–1963; the second was the result not only of the search for financial instruments still unavailable on the Italian market but also to a considerable extent of an emerging fear of left-oriented policy measures. The reform design dragged on during the entire decade and was never implemented. Incomes policy proposals ran up against the weakness of the reform culture of the communist opposition (Magnani 1997a), the targets set by the programmazione were not compatible with the inefficiency of public administration, and the antitrust regulation and company-law reform proposals were defeated by consolidated interests of the large industrial groups (Barca 1997a; Ciocca 2007). Moreover, the ambition of the State to guide the market in strategic directions turned into a radical reduction of the autonomy of

82

aggregate growth and policy

economic leaders, more and more subordinated to the greed of political lobbies and parties. The progressive contribution of State-controlled enterprises came gradually to an end, accentuating the 1970s crisis of large firms in the steel, energy, and chemicals sectors that had led convergence of the Italian industrial structure during the Golden Age. The hot autumn of 1969 marked a clear failure of the incomes-policy approach advanced by the center-left governments. Within a few years, the Italian economy went through a huge and persistent wage shock and through the oil shock. The macroeconomic context for reform worsened rapidly. Also, development policy suffered clear setbacks at the end of the 1960s. The abolition of regionally differentiated wage structures in national contracts raised relative Southern labor costs and reduced wage flexibility, although a partial compensation was found by charging employers’ social-security contributions to the budget. Moreover, construction of the first pillars of a modern national welfare state, in particular in terms of pensions and health service, inevitably tended to reduce the amount of public resources available for the special development policy. Was economic policy well-designed to exploit the favorable conditions of the Golden Age? In principle, two episodes seem particularly relevant in answering this question: the postwar settlement in 1946–1948 and the center-left turn in 1963. The former was mainly a product of political circumstances of the time. The growth model that resulted delivered extraordinary results up to the early 1960s, even by standards of the time. At the beginning of the 1960s political constraints were less severe and growth had produced an unprecedented volume of resources. Despite its contradictions, the turn of 1963 represents a missed opportunity for reforms that would have improved subsequent economic performance. The failure in 1963 to adapt institutions, regulations, and the scope of public intervention in the economy to new circumstances created by growth (end of the unlimited supply of labor, congestion costs, formation of a significant working class in the North, and resurgence of union militancy) may be thought of as a clear reflection of a long-standing feature of Italian economic policy (i.e., the weakness of the reform policy culture), squeezed between the “alternative to the system” at that time cautiously but nonetheless firmly put forward by the main opposition party on the one hand and the defense of diffused particular interests on the other.

From the 1970s to the 1990s After the early 1970s, growth slowed markedly across Europe (Table 3.7). The end of the Golden Age had a number of unavoidable aspects including exhaustion of

the golden age and the second globalization in italy

83

Table 3.7 Levels and rates of growth of real GDP per person, 1973–1995 (percent per year) Y/P 1973

Y/P 1995

Growth Rate 1973–1995

Switzerland

18,204

20,627

0.58

Denmark

13,945

20,350

1.74

Sweden

13,494

17,648

1.23

West Germany

13,153

19,849

1.92

Netherlands

13,081

18,700

1.65

France

12,824

18,206

1.61

Belgium

12,170

18,270

1.87

United Kingdom

12,025

17,586

1.75

Norway

11,324

21,578

2.96

Austria

11,235

17,959

2.16

Finland

11,085

15,970

1.88

Italy

10,634

17,216

2.21

Spain

7,661

13,132

2.48

Greece

7,655

10,321

1.37

Portugal

7,063

11,614

2.29

Ireland

6,867

12,734

2.85

Source: Conference Board (2011). Estimates for Ireland are based on GNP.

transitory components of fast growth, such as postwar reconstruction; reduced opportunities to redeploy labor out of agriculture; narrowing of the technology gap; and diminishing returns to investment. TFP growth slowed markedly. Moreover, the United States itself experienced a productivity growth slowdown. All-in-all, the scope for catch-up growth was much reduced, although by no means eliminated. In addition, in the face of internal labor militancy and external oil-price shocks together with the collapse of the discipline of the fixed exchange rate system macroeconomic policymaking became much more difficult. Indeed, from the end of the Golden Age, Western Europe’s catch-up of the United States in terms of real GDP per person stalled. The European level was 68.3 percent of the American level in 1973, 68.6 percent in 1995, and 68.2 percent in 2007. The picture in terms of labor productivity was different. Catch-up in terms of real GDP per hour worked continued until the mid-1990s and the European level rose from 63.3 percent of the American level in 1973 to 85.1 percent in 1995 but then retreated to 76.5 percent by 2007. From the 1970s to the

84

aggregate growth and policy

mid-1990s the discrepancy comes from a decrease in the amount of work done by Europeans compared with Americans, accounted for by a combination of rising unemployment, earlier retirement, and longer holidays. The implications for economic welfare depend on how far these outcomes result from differences in preferences or distortions to markets from regulations, taxes, and so forth. This remains unclear (Faggio and Nickell 2007). Italy has had a similar experience but with some distinctive features; real GDP per person was 63.7 percent of the level in the United States both in 1973 and in 2007 but had been higher at 69.9 percent in 1995. The gap in labor productivity between Italy and the United States narrowed quite steadily between the end of the Golden Age and the mid-1990s but then widened appreciably; the ratio was 64.9 percent in 1973, 81.8 percent in 1995, and 67.6 percent in 2007. The idiosyncratic Italian aspects that are a cause for concern are low female employment and high levels of inactivity among males over 55. These aspects have been a persistent feature of recent decades and do, at least partly, reflect distortions. Together, inflexibilities in labor markets that deny part-time employment opportunities and relatively low education attainments account for most of the difference in female employment levels compared with the European average (Del Boca, Pasqua, and Pronzato 2004), whereas early retirement for men has been heavily incentivized by the pension system and an astonishingly high implicit tax on working after the age of 60 (Duval 2003).5 Although catch-up of the United States in terms of labor productivity continued, the rate of convergence slowed down (Crafts 2007). European countries struggled to cope with the aftermath of macroeconomic turbulence of the 1970s, to embrace creative destruction in the context of the need to adjust to a changing world economy, and to achieve rapid productivity growth in the increasingly dominant services sector. The Eichengreen wage-moderation model broke down (Cameron and Wallace 2002; Gilmore 2009), whereas regulation, taxation, and expenditure on social transfers increased. Although these policy shifts were understandable in the context of the pressures of the time and the legacy of the social contracts of the earlier postwar period, they tended to slow growth. Kneller, Bleaney, and Gemmell (1999) estimated that an increase of 1 percentage point in the ratio of distortionary taxes to GDP slowed growth by 0.1 percentage point. Employment protection impeded growth in sectors intensive in the use of human capital and more innovative sectors (Conti and Sulis 2010). Product market regulation that inhibited competition slowed catch-up in TFP (Nicoletti and Scarpetta 2005). Tables 3.8, 3.9, and 3.10 indicate that there was cause for concern. This account has resonance for Italy where the Golden Age growth model had run out of steam. The service sector now had a much increased weight and its performance became increasingly important. From 1973 to 1993, the share of employment in services rose from 43.9 to 62.2 percent, whereas the share in industry fell from 38.4 to 31.3 percent but service sector labor productivity growth averaged only 0.5 percent per year during this period (Broadberry,

the golden age and the second globalization in italy

85

Table 3.8 Employment protection (0–4) 1960

1980

2000

2008

Austria

1.10

1.92

2.21

1.93

Belgium

1.38

3.21

2.18

2.18

Denmark

1.90

2.30

1.50

1.50

Finland

2.30

2.30

2.09

1.96

France

0.75

2.80

2.98

3.05

Germany

0.80

3.21

2.34

2.12

Greece

—

—

3.50

2.73

Ireland

0.00

0.90

0.93

1.11

Italy

3.45

3.60

2.51

1.89

Netherlands

2.70

2.70

2.12

1.95

Norway

2.91

2.91

2.56

2.69

Portugal

—

4.00

3.67

3.15

Spain

4.00

3.87

2.93

2.98

Sweden

0.00

3.50

2.24

1.87

Switzerland

1.00

1.00

1.14

1.14

United Kingdom

0.27

0.60

0.68

0.75

United States

0.21

0.21

0.21

0.21

Sources: Nickell (2006) and OECD (2010a).

Giordano, and Zollino 2012) and economy-wide TFP growth fell to 1.2 percent per year. Italy was notable in this period for a very high level of employment protection (Table 3.8); was slow to relax product-market regulation (Table 3.9); and was hampered by regulatory procedures and costs, and by relatively high barriers to entry (Klapper, Laeven, and Rajan 2006; Bianco, Giacomelli, and Rodano 2012). The estimates in Nicoletti and Scarpetta (2005) suggest that slowness to deregulate product markets cost Italy about 0.7 percentage points per year in the 1980s and 1990s compared with adopting a stance similar to the most liberal OECD country, whereas the estimates in Caballero et al. (2004) imply that employment protection slowed adjustment in Italy and had a labor productivity growth cost of about 0.3 percentage points per year in the period 1980–1998 compared with the economy with the least employment protection. At the beginning of the 1970s, several factors tended in Europe to reduce optimal firm size (minimum efficient scale): the crisis of the Fordist organization of labor caused by the antagonistic attitude of workers in large firms

86

aggregate growth and policy

Table 3.9 Price-cost margins (PCM) and product market regulation (PMR: 0–6) PCM

PCM

PMR

PMR

PMR

PMR

Manufactures

Services

1975

1990

1998

2008

Austria

1.15

1.28

5.2

4.5

2.33

1.45

Belgium

1.10

1.20

5.5

5.3

2.17

1.43

Denmark

1.11

1.25

5.5

4.7

1.59

1.06

Finland

1.18

1.27

5.5

4.6

2.08

1.19

France

1.12

1.26

6

5.2

2.52

1.45

Germany

1.13

1.25

5.2

4.6

2.06

1.33

Greece

—

—

5.7

5.7

2.99

2.37

Ireland

—

—

5.7

5

1.65

0.92

Italy

1.15

1.38

5.8

5.8

2.59

1.38

Netherlands

1.13

1.24

5.6

5.6

1.66

0.97

Norway

1.13

1.26

5.5

4.5

1.85

1.16

Portugal

—

—

5.9

5.3

2.25

1.43

Spain

1.14

—

5.1

4.7

2.55

1.03

Sweden

1.11

1.17

4.5

4.4

1.93

1.30

—

—

4.1

4.2

2.48

1.18

United Kingdom

1.11

1.16

4.8

3

1.07

0.84

United States

1.12

1.19

3.7

2.3

1.28

0.84

Switzerland

Sources: PMR indicator for 1975 and 1990 from Conway and Nicoletti (2006) and for 1998 and 2008 from Wolfl et al. (2009); the 1975 and 1990 numbers are not comparable with those for the later years. Pricecost margins from Hoj et al. (2007) refer to the mid-1990s.

and their demands for higher wages, the diversification and specialization of demand, technological innovations that made equipment capital more flexible, and reduced benefits of vertical integration of production. International competition was now based more on product innovations and less on standardized mass production, notably in automobiles (Bianchi 2002). These factors lay at the heart of the so-called flexible-specialization model (Sabel 1982). Italian industry shared these general tendencies with particular intensity for two reasons. First, industrial structure was already characterized in the 1950s and 1960s by a relatively large share of small- and medium-size firms and by the prominence of industrial districts (Brusco and Paba 1997). Second, after the “hot autumn,” large firms reacted by decentralizing production because of rapidly increasing rigidities emerging in the production process at the shop-floor level.

87

the golden age and the second globalization in italy Table 3.10 Distortionary tax revenues and social transfers (%GDP) a) Distortionary tax revenues 1965

1980

1995

2007

Austria

21.2

26.7

29.6

30.1

Belgium

19.5

30

32.4

32.3

Denmark

17.8

27

33.1

32.5

Finland

17.3

23.2

31.8

29.9

France

21.3

30

31.2

31.2

Germany

21.2

27.3

26.8

25.3

Greece

10

13.9

18.6

20.6

Ireland

11.8

18

19.6

22.2

Italy

15.4

21.8

29.2

30

Netherlands

22.4

31.3

31.3

26.2

Norway

17.4

27.5

25.2

31.4

Portugal

8.8

12.6

19.2

21.8

Spain

8.7

17.9

22.9

27.4

Sweden

24.1

35.7

34.8

34.6

Switzerland

11.5

19.5

21.7

22.3

United Kingdom

20.3

24.9

22.7

25.5

United States

19.1

21.7

22.9

23.2

1960

1980

1995

2007

Austria

15.9

22.6

26.6

26.4

Belgium

13.1

23.5

26.4

26.5

Denmark

12.3

25.2

28.9

26.1

Finland

8.8

18.4

27.4

24.6

France

13.4

20.8

28.3

28.4

Germany

18.1

23

26.6

25.2

Greece

10.4

11.5

19.3

21.3

Ireland

8.7

17.4

18.4

18.5

Italy

13.1

18

19.8

24.9

Netherlands

11.7

24.1

22.8

20.1

Norway

7.8

16.9

23.5

20.8

Portugal

—

10.8

18.1

22.5

b) Social transfers

88

aggregate growth and policy Table 3.10 (cont.) Spain Sweden Switzerland United Kingdom United States

—

15.5

21.5

21.6

10.8

28.6

32.5

27.3

4.9

13.9

17.5

18.5

10.2

16.6

20.4

20.5

7.3

13.3

15.4

16.2

Sources: Lindert (2004), OECD (2010b; 2011d).Distortionary taxes as defined in Kneller, Bleaney, and Gemmell (1999) and refer to direct taxes; Ireland in 1995 and 2007 as %GNP.

The emergence of new small- and medium-size enterprises was concentrated mostly in their historical locations in the Center and the North, shaping a sort of “regionalized model of capitalism” based on agglomeration economies, local networks, and nonmarket mechanisms of coordination (Trigilia and Burroni 2009), and on the accumulation of tacit knowledge (Antonelli and Barbiellini-Amidei 2009). By contrast, the weakness of indigenous entrepreneurship—the main failure of the special development policy—prevented the growth of flexible small-to-medium size firms in the South. These latter (twenty to fifty employees) were the main support of growth in industry during the 1970s, in the face of policy failures and growing macroeconomic imbalances. Their performance in terms of productivity, investment, production, and profits turned out to be significantly better than that of large firms (Barca and Magnani 1989). Italy did, therefore, find a partial solution to mitigate inflexibility. New social demands emerged. They were met by a progressive widening of the fiscal deficit. The level of Italian public debt had remained almost stationary until the end of the 1960s, well below 40 percent of GDP. By 1980 it reached 60 percent and in 1994, at its peak before the recent financial crisis, it amounted to 125 percent of the national product.6 Italy is the only major European country where, after the substantial reduction registered in the first years after World War II, public debt relative to GDP climbed back to values recorded in the second quarter of the century. This rise stemmed from strong increases of primary expenditures in the first half of the 1970s (from 31.2 percent in 1970 to 37.6 percent of GDP in 1975), which, however, still remained below the levels registered in France and Germany. Social transfers rose from 13.1 percent GDP in 1960 to 18 percent GDP by 1980 (Table 3.10). Italy looks similar to many other European countries regarding the expansion of welfare, even if the gap to be filled with regard to demands for social spending was wider, given the low initial conditions. What is distinctive in Italy is that the adjustment was made by keeping fiscal pressure

the golden age and the second globalization in italy

89

constant and thus by allowing a fiscal deficit to emerge that over time led to a rapid increase in the public debt-to-GDP ratio (Balassone, Francese, and Pace 2012). In the first half of the 1970s, intense social conflict discouraged counterbalancing revenue increases. It was politically more viable to let public debt, inflation, and the fiscal drag do the job of finding a new though unstable macroeconomic equilibrium. Moreover, the composition of revenues became less favorable to growth. Table 3.10 shows a substantial rise in distortionary taxation between the mid-1960s and the mid-1990s, which went along with higher social transfers but also reflects the increased outlays for interest payments on the public debt. Research by the OECD suggests that a growth-friendly tax system has low rates of corporate tax and collects a high proportion of its revenue from consumption taxes (Johansson et al. 2008). In both these respects, Italy was relatively badly placed. Against the general European trend, corporate income tax rates increased from 36 percent at the end of the 1970s to 52 percent in the mid-1990s, whereas Italy had the lowest C-efficiency score for value-added tax in Europe. The broad profile of monetary policies was similar in Italy and abroad; a tendency toward accommodation (actually more pronounced in Italy), followed at the end of the 1970s by a more restrictive stance in the context of the “Volcker turn” and participation in the European monetary system. According to the Governor of the Bank of Italy, Guido Carli, (Banca d’Italia 1975), monetary policy could not react effectively to the slackening fiscal discipline because this would have amounted to a “seditious act” against the State.7 Driven by enormous nominal wage increases (17 percent on a yearly basis between 1971 and 1975; real wages increased by almost 5 percent a year), the oil shock, and the devaluation of the lira, prices exploded. However, some authors believe that the degree of freedom in shaping macroeconomic policies was actually higher. Spinelli and Fratianni (1991), Andreatta and d’Adda (1985), and Onofri and Basevi (1997) underline that with a more market-oriented policy culture it would have been possible to implement more virtuous policies in the aftermath of the first oil shock, as occurred in other European countries. Their critique is severe; Italian economic policy in the 1970s was basically founded on the irrelevance of the public budget constraint, on the dependence of monetary policy on budget needs, and on a dirigiste approach of public interventions. Giavazzi and Spaventa (1989) argue instead that Italian macroeconomic policy in the second half of the 1970s reduced the costs of disinflation later, because it delivered high profit margins and financed new capital equipment such that pressure could be more easily exerted on industry to reduce costs. Boltho (1986), noting the still positive growth differential between Italy and the other advanced economies, supports this argument by pointing to the beneficial effects of expansionary policy during the whole decade on the development of small- and medium-size firms.

90

aggregate growth and policy

In the second half of the 1970s, in the middle of a dramatic crisis threatening the very survival of democracy, participation of the communist party in the coalition supporting the government (1976–1979: “governi di solidarietà nazionale”) allowed a deceleration of inflation and public debt growth; the willingness, albeit cautious, of trade unions to take into account the compatibility of wage dynamics with macroeconomic equilibrium induced wage moderation, increased profits, and indirectly supported technological restructuring processes. In 1979, the coalition broke up. Within a short time span, entrepreneurs regained control of production processes in large firms, strengthened by the substantial labor-saving process innovations undertaken in the preceding years (Barca and Magnani 1989). Around 1980, three events marked a turning point in economic policy: (1) the participation in the European Monetary System; (2) the Bank of Italy– Treasury “divorce”(by which the Bank was freed from the obligation of buying unsold public debt at Treasury auctions), which was aimed at inducing virtuous behavior by the public sector by increasing the cost of issuing new debt; and (3) the introduction of a form of incomes policy, through the predetermination of automatic inflation-related pay raises to guide price expectations. The first decision represented an attempt to “force” a reduction of costs in the sector exposed to foreign competition by a partial and delayed realignment of the nominal exchange rate in response to appreciation of the real exchange rate. The effectiveness of this policy with respect to that implemented in the first half of the 1970s is clearly shown by different price dynamics after the two oil shocks in 1973 and 1979. In the 1980s, conditions for addressing macroeconomic disequilibria looked more favorable than in the 1970s. Governments could rely on stable majorities. The drastic fall in oil prices reduced inflation and increased, ceteris paribus, available resources. After 1983, the major economies entered a long expansionary cycle. The “wage-push age” was definitely over and the concept of incomes policy (Tarantelli 1981) was, even if painfully, gaining some ground.8 Monetary policy, by pursuing flexibly an exchange rate target, was an important factor in reducing inflation from 1979 to 1986 (Gressani, Guiso, and Visco 1987). The adjustment was, however, incomplete. Inflation reached a low point in 1987 of slightly less than 5 percent after the fall of oil prices but rose back to 6 percent at the beginning of the 1990s, two percentage points more than the European average, mainly because of the existence of strong competition barriers in the nontraded sector (Barca and Visco 1993) and a still excessive wage growth. During 1985–1991, the nominal exchange rate changed little; competitiveness, measured by goods prices, worsened by 12 percent, in terms of unit labor costs by more than 15 percent. The current account, balanced in 1986, weakened gradually to a deficit of 2 percent of GDP in 1991. Firms were slow to react to competitive pressure by improving the quality of products and shifting toward more innovation-intensive production.

the golden age and the second globalization in italy

91

The major obstacle to complete adjustment was an inability to deal with the public debt problem. Programs to reduce deficits were regularly discarded. Although there was some progress in the final years of the decade, primary deficits generally remained substantial. Public debt rose further from 60 percent of GDP in 1980 to more than 100 percent in 1991. The foreign sector purchased significant amounts of State bonds attracted by the stability of the exchange rate and high interest rates. The negative net investment position of the country climbed to more than 10 percent of GDP. The long-term implications of allowing the debt ratio to increase so much are unfavorable for growth, as is highlighted in particular by growth models of the overlapping-generations variety. Adverse impacts can occur through a number of transmission mechanisms including reductions in market-sector capital formation, higher long-term interest rates, and higher tax rates. Empirical research on advanced economies has found negative effects; for example, Kumar and Woo (2010) estimate that a 10 percentage point increase in the initial debt-toGDP ratio is associated with a fall of about 0.2 percentage points in growth. It has also been suggested that the adverse effect on growth becomes substantial after the debt ratio reaches a critical level, which has recently been claimed to be around 90 percent of GDP (Checherita and Rother 2010). The risks inherent in the rising public debt ratio were clearly perceived, as shown by continuous attempts to reduce fiscal imbalances. There was no effective policy reaction because of the structural weakness of political power, more and more dependent on a multitude of powerful interest groups, unions often included.9 The failure to address the increase in public debt in a far better macroeconomic context than that of the 1970s must be considered as a major missed opportunity. In the period between the end of the Golden Age and the beginning of the 1990s the special development policy for the South encountered growing difficulties; reduction of the income gap came to a halt. The oil shock, changing technological and competitive framework, and increased rigidity of the labor market made this policy less and less effective (Iuzzolino, Pellegrini, and Viesti 2012). The progressive decline of IRI disproportionately reduced the efficiency of investments in the South, where State-controlled firms played a decisive role. A similar degeneration process undermined the Cassa per il Mezzogiorno, whose technical autonomy and authority were put in question. Finally, the attribution of relevant powers to the newly constituted Regions increased significantly the costs of coordination of regional policies and made local lobbies more powerful. Special development policy tended gradually to replace rather than add resources to the ordinary State policy. From the beginning of the 1970s up to 1992, total net current public transfers in the South became more negative by 3 percentage points of its GDP, from −3.5 to −6.5, but in the rest of the country they plunged from −13 to −20 percent of GDP of these regions. Including labor income in the South often reflecting a disguised social assistance policy, net

92

aggregate growth and policy

public resources flowing to the South increased slightly from 11 to almost 12 percent of GDP, with those pouring out from the Center-North from −3 to −10 percent of GDP (Magnani 1997b). The divergence in public net flows was not caused by expenditures but by taxes, these latter directly related to the level of incomes in the two parts of the country. However, what matters more is the fact that the composition of public expenditure flowing into the South started to shift more and more in favor of current transfers and labor income, with negative effects on growth prospects of the southern economy on the one hand and the national public budget on the other (Cannari and Chiri 2006; Staderini and Vadalà 2009). An implicit redistribution of resources therefore took place in the aftermath of the extension of the welfare state to the whole country. From being a factor supporting Italian growth, the relative stagnation of the South gradually became a drain on public resources. To sum up, supply-side adjustments in the 1970s were much more limited than elsewhere in Europe because of policy constraints resulting from the social and political emergency. In the subsequent decade, the long shadow of the 1970s mainly took the form of a lack of political leadership in subduing the pressure of interest groups, which was a precondition for ensuring a sustainable path for public finances. The crisis of 1992 exploded because of a combination of factors affecting the whole EMS (Eichengreen 2000) and the failure to correct the previously mentioned disequilibria. Italy was forced to exit the European Monetary system (EMS). This crisis finally induced some long overdue macroeconomic adjustments.

From the Mid-1990s to the Eve of the Crisis After the mid-1990s western European catch-up of the United States in terms of labor productivity ceased. Whereas between 1973 and 1995 real GDP per hour worked grew at 2.6 percent per year in Europe and 1.3 percent in the United States, between 1995 and 2007 these growth rates were 1.2 and 2.1 percent, respectively. The end of catch-up reflected both European slowing down and American speeding up. At about this point, employment started to grow faster than population in Europe but it no longer did so in the United States and the potential of ICT for productivity increased considerably, a potential that was more fully realized by the American than the European economy. ICT expenditure as a share of GDP averaged 2.2 percent in the late 1990s compared with 4.0 percent in the United States and in Sweden (Timmer and van Ark 2005) and productivity growth was especially weak in the service sector, averaging only 0.2 percent per year between 1993 and 2007 (Broadberry, Giordano, and Zollino

the golden age and the second globalization in italy

93

2012), where ICT had strong effects in the United States. These developments were important as proximate sources of the turnaround in relative productivity performance. To quite a large extent, productivity growth in Italy reflected these tendencies. The growth of real GDP per hour worked fell from 2.35 percent in 1973–1995 to 0.46 percent in 1995–2007, whereas the growth of hours worked per person went from −0.14 to +0.87 (Table 3.11). This latter can be expected to have reduced labor productivity growth in the short to medium term if investment failed to respond and the additional workers were lower quality (Dew-Becker and Gordon 2008). Nevertheless, productivity growth was very disappointing even by European standards, suggesting that there were further factors involved. This poor performance came even though Italy went through a period of major policy reforms during the 1990s including privatization, antitrust, banking, and company law reforms, which seemed appropriate as moves in the direction of reform required by a higher level of economic development. The 1992 crisis added momentum to reform and “the entire political system born after World War II came to an end” (Pagano and Trento 2003, 199). The big devaluation of the lira in 1992–1995 delivered a sharp improvement of competitiveness and confidence was for the time being restored. Significant progress was made in reducing the public debt-to-GDP ratio, which declined from 125 percent in 1994 to 103 percent in 2004. There is a need, therefore, to confront the puzzle of why growth performance deteriorated rather than improved. There seem to be two key reasons. First, Italy was relatively badly placed to exploit the opportunities of the ICT era. The diffusion of this new technology was hindered by the small size of firms, oppressive regulation, and shortfalls in human capital compared with the European leaders in the take up of ICT. Second, supply-side reform did not go far enough especially given the context of joining the Eurozone. In several respects, including the legal system, competition policy, regulation, and privatization, reforms were either incomplete or inadequately implemented. The international evidence is that diffusion of ICT has been significantly inhibited in countries that are heavily regulated. Employment protection has been shown to deter investment in ICT equipment (Gust and Marquez 2004) because reorganizing working practices and upgrading the labor force, which are central to realizing the productivity potential of ICT, are made more expensive. Restrictive product market regulation has deterred investment in ICT capital directly (Conway et al. 2006) and the indirect effect of regulation in raising costs has been relatively pronounced in sectors that use ICT intensively. There has been a strong correlation between product market regulation and the contribution of ICT-using services (notably distribution) to overall productivity growth (Nicoletti and Scarpetta 2005). The general story is not that regulation has become more stringent but rather that existing regulation became more costly in the context of a new technological era.

94

aggregate growth and policy

Table 3.11 Levels and rates of growth of real GDP per person and per hour worked, 1995–2007 ($GK1990 and percent per year) a) Real GDP per person Y/P, 1995

Y/P, 2007

Growth Rate 1995–2007

Norway

21,578

28,553

2.36

Switzerland

20,627

24,781

1.55

Denmark

20,350

25,060

1.76

Netherlands

18,700

24,405

2.24

Belgium

18,270

23,487

2.12

France

18,206

22,282

1.70

Austria

17,959

23,744

2.36

Germany

17,672

21,143

1.51

Sweden

17,648

25,381

3.07

United Kingdom

17,586

23,620

2.49

Italy

17,216

20,163

1.33

Finland

15,970

24,635

3.67

Spain

13,132

17,869

2.60

Ireland

12,734

23,338

5.18

Portugal

11,614

14,601

1.93

Greece

10,321

15,860

3.64

1995

2007

Growth Rate 1995–2007

Belgium

30.37

35.74

1.37

Norway

29.82

36.72

1.69

France

29.02

35.44

1.69

Netherlands

27.75

33.84

1.67

Denmark

26.98

30.52

1.03

Germany

25.10

30.78

1.72

United Kingdom

24.33

31.65

2.22

Italy

24.29

25.63

0.46

Austria

23.50

28.68

1.68

Sweden

23.13

31.32

2.56

Finland

22.36

30.42

2.60

b) Real GDP per hour worked

95

the golden age and the second globalization in italy Table 3.11 (cont.) Spain

22.21

23.43

0.64

Switzerland

21.92

25.82

1.38

Ireland

17.21

26.01

3.50

Portugal

13.60

15.62

1.16

Greece

11.63

16.78

3.10

Source: Conference Board (2011).Ireland is GNP.

This account clearly has resonance for Italy. Italy entered the ICT era with relatively high levels of regulation, albeit decreasing over time as reforms took place (Tables 3.8 and 3.9). The estimates in Tables 3.12 and 3.13 indicate that investment in intangible capital has been low and that the productivity growth contributions of ICT capital and intangible capital have been very modest. International comparisons confirm that ICT investment has been held back in Italy by regulation and a shortfall in human capital (Conway et al. 2006). Microeconomic studies of Italian manufacturing confirm this picture while adding further insights as to why diffusion of ICT had been relatively slow in Italy. The take-up of ICT has been strongly correlated with firm size and changes in organizational structure (Fabiani, Schivardi, and Trento 2005). In Table 3.12 Investment in intangibles in the market sector, 2006 (%GDP) Computerized

Innovative

Economic

Information

Property

Competencies

Austria

0.89

3.14

2.42

6.46

Czech Republic

0.71

2.80

2.93

6.45

Denmark

1.87

3.06

2.93

7.86

France

1.42

3.18

3.30

7.90

Germany

0.73

3.59

2.84

7.16

Greece

0.34

0.62

0.63

1.59

Italy

0.64

2.21

2.19

5.04

Slovakia

0.37

1.76

2.39

4.53

Spain

0.79

2.78

1.90

5.47

United Kingdom

1.55

3.16

5.84

10.54

United States

1.61

4.37

5.50

11.48

Source: van Ark et al. (2009).

Total

96

aggregate growth and policy

Table 3.13 Market sector growth accounting with intangibles, 1995–2006 (% per year) ICT

Non-ICT

Intangible

Labor

K/L

K/L

K/L

Quality

Austria

0.26

–0.02

0.55

0.22

1.35

2.36

Czech Republic

0.35

1.62

0.68

0.31

1.64

4.60

Denmark

0.44

0.24

0.72

0.17

0.53

2.11

France

0.12

0.31

0.48

0.40

0.69

2.00

Germany

0.20

0.48

0.38

–0.15

0.88

1.79

Greece

0.45

1.48

0.24

0.71

0.40

3.27

Italy

0.11

0.29

0.12

0.22

–0.45

0.29

—

2.72

0.21

0.46

2.78

6.17

Spain

0.19

0.49

0.12

0.64

-0.96

0.47

United Kingdom

0.63

0.28

0.69

0.22

1.23

3.06

United States

0.40

0.24

0.83

0.18

1.33

2.96

Slovakia

TFP

Labor Productivity

Source: van Ark et al. (2009). K= Capital; L= Labor. Intangibles investments are part of final output and ICT K included with non-ICT K in Slovakia.

this context, Bugamelli and Pagano (2004) found that many firms seem to be constrained in their ICT investment by the adjustment costs associated with reorganization, especially if their workforce has relatively low levels of human capital. These reflect regulatory burdens but, because they are fixed costs, they bear very heavily on the small- and medium-size firms that have been central to Italy’s distinctive variety of capitalism.10 Turning to supply-side policy reforms, there are a number of shortfalls that deserve to be highlighted. One major point is that their effectiveness in promoting competition has been insufficient. The retail sector is an important example. Labor productivity growth in retailing was 0.8 percent per year in 1995–2002 compared with 7.4 percent in the United States and 1.6 percent in the EU15 (McGuckin, Spiegelman, and van Ark 2005). It is clear that productivity performance was still impaired by regulation; barriers to entry and mark-ups in retailing remained high on average with adverse consequences for TFP (Daveri, Lecat, and Parisi 2011). However, in districts where competition was stimulated by the 1998 regulatory reform ICT investment and labor productivity increased (Schivardi and Viviano 2011). The competition-policy framework established in the 1990s has been rated below average relative to OECD countries using criteria relating to political independence, toughness, and investigative powers with the implication that productivity growth has been adversely affected (Buccirossi et al. 2009);

the golden age and the second globalization in italy

97

low-efficiency producers continued to survive in many sectors (Milana, Nascia, and Zeli 2008); and mark-ups in the service sector as a whole remained high (Table 3.9). Failure to deregulate professional services (in 2008 Italy still had the tightest product-market regulation in Europe after Luxembourg) has had significant adverse effects on the productivity of sectors for which they are important inputs (Barone and Cingano 2011). On the contrary, the reform process was successful in the banking sector where competition increased significantly. The volume of privatization in Italy was impressive; in terms of the proceeds, it was second only to the United Kingdom during the period 1979–1999. However, it was mainly aimed at reducing public debt. Only 30 percent of the value of transactions involved a transfer of control; opportunities to introduce competition in utilities, such as electricity and gas were missed; and regulation was inadequate. There were improvements in operating efficiency in sectors exposed to competition but not in sheltered sectors where it seems that privatization was simply conducive to increased quasi rents as market power was exploited more fully (Barucci and Pierobon 2007).11 Although on the surface reforms to corporate governance might have enabled an effective market for corporate control, in practice changes in ownership and control structures (pyramids have been mostly replaced by coalitions) have been quite limited, private benefits of control seem to have remained high, and the number of listed companies has increased only slightly (Bianchi and Bianco 2006). Bloom and van Reenen (2008) found that the quality of management in Italian companies in 2006 was below leading international standards, largely as a result of weak competition and family control, with adverse implications for productivity performance. The effectiveness of company law reforms has been limited by established interests and the deep-rooted Italian difficulty of linking economic and legal culture (Ciocca 2007). The supply-side reform agenda has been also undermined by the inadequacy of the legal system (Bianco et al. 2007; Carmignani and Giacomelli 2010) and the ineffectiveness of the government. The low rankings of Italy in the World Bank’s Governance Indicators reported in Table 3.14 underline this point. Similarly, Italy has a low “ease of doing business” ranking (eightieth in the most recent year) in the World Bank’s Doing Business report with enforcement of contracts seen as especially poor (157th).12 Other important factors for long run growth are less responsive to policy, at least in the short run. Despite progress, especially after World War II, the gap in human capital measured by the number of years of education remained significant compared with the other advanced countries. The percentage of persons aged twenty-five to sixty-four who attained at least an upper secondary education was around 32 percent in the second half of the 1990s, almost ten points below OECD average. With respect to tertiary education, the gap widened: 12 percent vis-a-vis 24 percent (OECD 2010a). Whereas in terms of the quantity of education Italy has slowly caught up, quality has lagged behind. The evidence provided by student competencies is

98

aggregate growth and policy

Table 3.14 Governance indicators (-2.5 to 2.5) a) 1996 Government Effectiveness

Regulatory Quality

Control of Corruption

Rule of Law

Austria

1.984

1.154

2.120

1.841

Belgium

1.935

0.971

1.350

1.406

Denmark

2.027

1.213

2.222

1.828

Finland

2.114

1.089

2.221

1.914

France

1.755

0.766

1.353

1.520

Germany

2.006

1.072

2.134

1.633

Greece

0.836

0.746

0.511

1.098

Ireland

1.732

1.224

1.486

1.589

Italy

0.919

0.649

0.505

1.050

Netherlands

2.056

1.285

2.178

1.734

Norway

2.064

1.063

2.216

1.928

Portugal

0.999

1.028

2.015

1.276

Spain

1.570

0.878

1.248

1.354

Sweden

2.010

1.075

2.203

1.763

Switzerland

2.146

1.082

2.166

1.946

United Kingdom

1.884

1.469

2.171

1.662

Government Effectiveness

Regulatory Quality

Control of Corruption

Rule of Law

Austria

1.776

1.649

2.049

1.928

Belgium

1.502

1.365

1.303

1.297

Denmark

2.235

1.855

2.426

1.964

Finland

1.907

1.511

2.408

1.861

France

1.433

1.255

1.433

1.378

Germany

1.606

1.549

1.698

1.698

Greece

0.661

0.850

0.319

0.798

Ireland

1.614

1.780

1.727

1.735

Italy

0.333

0.869

0.270

0.403

Netherlands

1.708

1.735

2.159

1.738

Norway

1.978

1.331

1.925

1.905

b) 2007

99

the golden age and the second globalization in italy Table 3.14 (cont.) Portugal

0.902

1.084

0.990

1.007

Spain

0.960

1.186

1.008

1.083

Sweden

1.942

1.569

2.206

1.855

Switzerland

1.966

1.620

2.134

1.816

United Kingdom

1.622

1.801

1.716

1.658

Source: World Bank. Governance matters indicators.Indicators are normalized such that the median = 0 in each year and so are not strictly comparable over time.

quite clear in this regard. The picture is especially worrying with regard to cognitive skills as measured by performance in international tests, such as the OECD PISA studies. Italy now ranks very low, especially in mathematics, and its average performance has been declining steadily since the mid-1970s in stark contrast with such countries as Finland and Sweden (Hanushek and Woessmann 2009). The econometric evidence suggests that this has exacted a serious penalty in terms of an adverse effect on growth. Among the possible reasons for the relatively low quality of education in Italy, one must exclude a shortfall in educational spending; on the contrary, Italy spends more per student than many countries that achieve much better results. Instead, a relaxation of educational standards in the 1970s and an inability to achieve coherent reforms seem to have played a significant role (Bertola and Sestito 2012). Moreover, international comparisons show that efficiency of schooling resources is enhanced by accountability, autonomy in hiring decisions, and effective competition within the school system, which all raise standards (Woessmann et al. 2007). The issue is the organization of the school system, which has been ineffective in addressing principal-agent problems in the delivery of education. Examined through this lens, Italy has lagged behind other countries especially in terms of lack of autonomy of schools in providing incentives and sanctions to teachers; in recent years, accountability has been weakened through not having external exit examinations (Boarini 2007). In the Golden Age, given the prevailing technology, the lack of formal education was not a serious constraint on growth because informal knowledge formation, such as on-the-job training, compensated. However, at more advanced stages of technological development, the gaps in formal education, in particular tertiary education, have mattered because it is essential in fostering innovation (Bertola and Sestito 2012). Another important structural factor potentially hindering growth is Italy’s “anomalous” position in international trade, which has its roots in the strength of small- and medium-sized firms producing in industrial districts (De Benedictis 2005). Revealed comparative advantage for Italy exhibits high persistence. Compared with other G7 economies Italy’s exports remain skewed toward low-technology and labor-intensive sectors, such as textiles and footwear,

100

aggregate growth and policy

and away from high-technology activities. This means that Italy is more exposed to competition from China and other dynamic Asian economies and is less well-placed to benefit from fast-growing sectors in world trade (Lissovolik 2008) while export performance, especially in distant markets, is held back by the small average size of Italian firms (Barba Navaretti et al. 2011). The implications of the historical legacy reflected in this trade configuration should not be exaggerated. The “market-crowding” impact on export growth has been much smaller than relatively slow growth in the EU15 (Italy’s main market) and trends in the real exchange rate (Breinlich and Tucci 2010). There has been an adverse trend in the terms of external trade but the effect only reduced real income growth by 0.1 percentage point over the ten years to 2006 (Bennett et al. 2008). Over time, however, the Italian economy is being subjected to a greater need to adjust to the changing international division of labor than such countries as France, Germany, or the United Kingdom. The big issue concerns the flexibility of the Italian economy; an index recently compiled by OECD economists places Italy twenty-fourth out of twenty-six countries in terms of ability to cope with globalization.13 European integration moved forward in the 1990s culminating in the establishment of the European Monetary Union (EMU) in 1999. Despite forced exit from EMS in 1992, Italy became a founder-member of the Eurozone. The decision to join the EMU sought to introduce competitive pressures into the system to stimulate efficiency improvements and thus to promote growth even in the short run by a general improvement of expectations, thereby counterbalancing at least partially the deflationary effects of the huge correction to the public deficit implemented to join the EMU. In conjunction with the decision to participate in the EMU, a radical reform of regional policy took place, benefiting from the strong European Union commitment to regional and social convergence. The basic aim was to spur growth of the whole country by refueling the catching-up process of the South, which had been interrupted in the 1970s, to offset the loss of the exchange-rate policy instrument and the severe fiscal clampdown to meet the Maastricht criteria. A new regional policy was launched, centered on a place-based approach that looked at innovation as the primary driver of development so as to enhance the potential comparative advantages of the southern regions. Participation in the EMU was a major change in the framework for macroeconomic policy: inflation and nominal interest rates converged rapidly toward the European average. For the first time in thirty years, price stabilization was finally achieved. It also had, however, supply-side implications through the intensification of competition. EMU membership can be thought of as a commitment technology; put differently, “Italian vices are overcome by importing European virtues,” as Guido Carli, Governor of the Bank of Italy (1960–1975) and Minister of the Treasury (1989–1992), emphasized retrospectively (Carli 1996). With better macroeconomic discipline, interest-rate convergence,

the golden age and the second globalization in italy

101

and improved prospects for fiscal sustainability, investment could respond positively. Overall, it might be hoped that joining the Euro could be good for growth. In any event, Italy’s participation in the EMU amounted to achieving the goal of the postwar generation of Italian leaders, anxious to anchor Italy within Europe. It was ultimately a political decision as was the creation of the EMU. At the same time, EMU membership was not a risk-free strategy, especially for a country with weak productivity growth and a high public debt-to-GDP ratio, given that the Eurozone is not an optimal currency area and the European Central Bank’s choice of interest rates might not be ideal from an Italian point of view. There clearly were dangers from a loss of international competitiveness and this has materialized. During 1999–2009, the real exchange rate measured by prices of production in the manufacturing industry appreciated by 7.5 percentage points, compared with an appreciation of 5 points in France and no change in Germany. The loss of competitiveness in the whole economy measured in terms of unit labor costs was higher at 13 percentage points, compared with a loss of 5 percentage points for France and a gain of 13 percentage points for Germany. Ex post, the experience of EMU membership has been much less comfortable than had been hoped ex ante. The benefit to fiscal sustainability anticipated from reduction of the gap between the interest rate and the growth rate is threatened by the ineffectiveness of supply-side reform, weak productivity performance, loss of competitiveness, and fears about the political will to maintain budget discipline. Investment and growth are potentially undermined rather than stimulated by a high real exchange rate and debt overhang. The growth rate up to the outbreak of the financial crisis (1999–2008) was half of that of the Euro area and of the average for the period 1980–1999. One question obviously arises. Did a superior policy alternative exist? Would growth prospects have been more favorable in the case of nonparticipation? Would it have been easier to implement the necessary structural adjustments? The analysis put forward in the previous sections suggests that the answer to all these questions is no. The experience of the 1970s is especially helpful in this regard. At that time, the structural weaknesses of the Italian economy, although less relevant because of the more favorable competitive and technological context, were already quite evident. The (much-used) freedom of maneuver granted by the floating exchange rate did not trigger any substantial policy interventions aimed at correcting them. On the contrary, the outcome was an unstable macroeconomic framework marked by inflation and fiscal laxity, two factors notably pernicious for growth in the medium-long run. Either inside or outside the EMU, supply-side reforms conducive to raising the rate of productivity growth were required; the pressure to make them would have been weaker without the prospect of joining the Euro. However, the consequence of not reforming well enough may eventually be more severe if a sovereign debt crisis is precipitated.

102

aggregate growth and policy

Conclusions After the Golden Age, Italy experienced increasing difficulties in adjusting its economy to the changing external context and to the requirements for sustaining catch-up growth at a higher level of economic development. The adjustment issue is common to advanced countries but the difficulties experienced in Italy look particularly severe. Italy had a greater need for reform after the success of the early postwar period and had more problems in making effective reforms. In the Golden Age it was different. The Italian performance in terms of growth and productivity was spectacular even taking account of the catching-up effect. The postwar settlement was heavily conditioned by the external policy environment but the main policy choices were not disrupted by harsh conflicts, reflecting somehow the sense of national responsibility of the political and economic leadership emerging from the war. The wage moderation determined by the unlimited labor supply in the agricultural sector, the features of human capital particularly suited to the existing technologies, the impulse to industrialization given directly and indirectly by the large State-owned firms, the effective allocation of resources provided by the banks controlled by the State, the capacity of private entrepreneurs in enhancing their performance especially in the traditional sectors, the liberalization of trade, and participation in the Common Market were all growth-promoting factors. At the beginning of the 1960s, after the “economic miracle,” this model was put under pressure and attempts to correct the destabilizing forces generated by the model itself failed. The wage and oil shock around 1970 and the decline of Fordism engendered everywhere in Europe—albeit in different forms and at varying speeds—macro and micro adjustment processes. Italy changed relatively little. Having arrived near the technology frontier, the economy proved too slow to adjust to the new context, despite success in mitigating inflexibilities by developing a regionalized model of capitalism based on small and medium firms. Cushioned by inflation and devaluation, growth remained relatively high in the 1970s. In the subsequent decade, despite improved conditions for addressing macroeconomic disequilibria (the long expansionary cycle of the major economies, the end of the “wage push era,” the deceleration of inflation, the new monetary policy after joining the EMS) structural adjustments were neglected, in particular with regard to fiscal discipline and competition policies in sheltered sectors. The catching up of the South came to a halt. Major supply-side reforms were eventually implemented in the aftermath of the 1992 crisis and also in connection with policies undertaken in the European Union. Nevertheless, in the second half of the decade growth fell below the European Union average. These reforms were an important step in the right direction but ultimately fell short of what was required. Although increasing the pressure for adjustments through price and exchange-rate stability and more

the golden age and the second globalization in italy

103

stringent constraints on fiscal policy, participation in the EMU did not help as far as the improvement of growth prospects was concerned. Clearly, this has made Italy’s position in terms of fiscal sustainability weaker and has made the country more vulnerable in the crisis that has erupted since 2007. Why did policy reform prove unable to stimulate growth? More specifically, what were the major difficulties in implementing sufficient structural adjustments after the Golden Age, despite the growing awareness of their necessity? These are the crucial questions we must try to answer. We can tentatively list a number of not unrelated reasons. The first refers to a long-standing feature of Italian society: the existence and severity, except for very limited periods, of conflicts between opposing groups and parties. In part, but by no means only, these were connected to the big ideological divide between the strong communist-oriented opposition and the ruling parties. Reforms in the history of the country are rare; much more commons are periods of violent conflicts or of policy stalemates because of the veto power of opposing actors. In the 1970s, this feature stands out most clearly and made reform almost impossible. In the 1980s, the “divided society” started to take other forms, less based on (declining) ideologies and more on the gradual proliferation of many consolidated interest groups, each one seeking to gain rents at the expense of others. This basically explains the inadequacy of the adjustment in public finances. Regarding the second possible reason, after the fall of the Berlin Wall the demise of the two main parties, the Communists and the Christian Democrats, worsened the picture. It may seem paradoxical in light of the end of the ideological divide, but these two forces were somehow able, each one within its area of influence, to represent the different interests of their constituencies in a policy vision that reached beyond them and looked at the perspectives of the country as a whole. Their disappearance strengthened instead the emergence of distributional coalitions at the expense of encompassing organizations (Olson, 1982) and stabilization (Alesina and Drazen 1991). The third reason concerns firms. Their capability to exploit favorable conditions varied through time. They were able to do so during the Golden Age thanks to low wages and the opening of the economy to external competition. In the 1970s, they overcame the emergency because of devaluation and generous fiscal policy; at the end of the decade large firms were able to renew the capital stock while small firms reacted by developing an original locally based model. Later, however, when the technological context started to change and required more product innovations, more ICT investments, and corresponding changes in the organization of the production process, Italian firms got into increasing difficulties that still continue. The difficulty of achieving efficient allocations of entrepreneurial resources resulting from closed corporate governance structures centered on family-controlled and pyramidal groups proved to be a major competitive handicap in this context. The gradual elimination of the exchange rate lever in the face of mounting competition from the emerging markets and

104

aggregate growth and policy

the increased stringency of fiscal policy could not be overcome. One can argue about the relative importance of firm endogenous and contextual factors but there seems to be little doubt that there is a historically rooted weakness of large private firms in Italy compared with other major European countries. The importance of State-owned enterprises (born from the ashes of the large industrial and financial private conglomerations) between the 1930s and the 1990s—a unique case in Western Europe—indirectly reflects this fact. In any case, whatever the historical reasons, this implies a greater need for supply-side policy reform compared with elsewhere. The fourth reason is related to this conclusion. There might be a structural problem in the implementation of change because of lack of political leadership in reforming the public administration to increase its effectiveness. The importance of and the reasons for this have been widely debated during the history of united Italy and invests all the domains of economic policy. We conjecture that this weakness is related to the increasing influence of interest groups that invade the public administration and the polity. In any event, because the quality of institutions and public services are important for growth, especially in the context of coping with globalization, lagging behind on this front entails serious risks. The disappointing performance of the South in the last thirty years shows this very clearly because it is in these regions that these lags are mostly concentrated. The funds used were remarkable, comparable with those of the golden years of special development policy. However, the new policy failed to meet its ambitious objectives mainly because of the inadequacy of national policies, whose target is the whole national territory rather than regional policy per se (Cannari, Magnani, and Pellegrini 2010). Education, justice, and security are growth and productivity promoting factors. Their weakness explains much of the unsatisfactory performance of the Italian economy. The Doing Business indicators reported in Table 3.15, which are strictly related to the quality of public services and administration, are in the South even weaker than in the rest of Italy, and far below the average of European countries. The factors hindering growth of the Southern economy in the last twenty years are of the same nature of those that curb the Italian economy as whole. In the South they are, however, more powerful. In this sense, the Southern question is today more than ever the magnifying mirror of Italy’s difficulties. The poor quality of public services brings us to the final point: changes in the political institutional design so as to increase the possibility of taking substantial policy decisions. The theme has quite a tradition in the political economy literature. The issue concerns the relationship between political institutions and the functioning of the economy. In particular, it has been often argued that the electoral rule is an important factor for the economic performance of a country (Persson and Tabellini 2005). In Italy, the crisis of 1992 marked the end of the so-called First Republic. Widespread criticism was addressed in particular to the strictly proportional electoral law for its alleged feature of producing

105

the golden age and the second globalization in italy Table 3.15 Doing business indicators: time needed to complete procedures 2008 (1) Country

Starting a Business

Dealing with Licenses

Registering Property

Enforcing Contracts

Time (days)

Time (days)

Time (days)

Time (days)

Austria

28

194

32

397

Belgium

4

169

132

505

Denmark

6

69

42

380

Finland

14

66

14

235

France

7

137

123

331

Germany

18

100

40

394

Greece

38

169

22

819

Ireland

13

209

38

515

Italy (Rome)

13

257

27

1,210

Netherlands

8

230

7

514

Norway

7

252

3

310

Portugal

7

327

42

577

Spain

47

233

18

515

Sweden

15

116

15

508

Switzerland

20

154

16

417

United Kingdom

13

144

21

404

DB Italy (2008) (2) Area

Starting a Business

Dealing with Licenses

Registering Property

Enforcing Contracts

Time (days)

Time (days)

Time (days)

Time (days)

Italy NW

13.2

414.8

10.6

1,826

Italy NE

13.3

442.3

10.4

1,866

Italy Center

12.1

355.5

11.2

2,095

Italy South and Islands

25.3

585

12.2

2,140

(1) Source: World Bank Doing Business. (2) Sources: Bianco and Bripi (2010). Weighted averages of regional values in the area (using regional GDP as weights).

106

aggregate growth and policy

a plethora of parties, thus often not allowing the government to rule with the necessary stability because of the power of even small players and the resulting tendency to look for compromises and even to change majorities to survive. The principle of “alternation” was invoked, so that clear policy options could be implemented by the winning coalitions during their mandate. A majority rule was introduced in 1993; it remained in force till 2005, when it was replaced by an adjusted proportional system. The results of the majority rule in terms of stability of the governments and the number of parties were, however, not encouraging (Bordignon and Monticini 2011). This suggests that easy solutions in terms of institutional engineering do not exist as far as the structure of the political system and, a fortiori, as far as the capacity to implement structural economic reforms are concerned.

Notes 1. Balassone, Francese, and Pace (2012), Fig. 18.2, p. 518, in this volume. 2. The parity of the lira was set at 625 per US dollar at the end of 1949. The US proposal of a more devalued parity was successfully resisted by the Italian government (Asso, Biagioli, and Picozza 1995). 3. In the first years it was subject to the supervision of the International Bank for Reconstruction and Development, which heavily financed the start-up of the Cassa. 4. This was very much the case in the 1950s when for the industrial sector real wage and labor productivity grew at 5.36 percent and 5.06 percent, respectively; it was less so in the 1960s when real wage and labor productivity growth averaged 5.61 percent and 4.45 percent, respectively (Broadberry et al. 2012). 5. Gordon (2008) points out that state-sponsored early retirement has very high costs in terms of reduced consumption, which are clearly much greater than any gain in terms of additional leisure time. 6. Balassone, Francese, and Pace (2012), Fig. 18.2, p. 518, in this volume. 7. With the support of the International Monetary Fund there was a serious attempt in 1974 to impose monetary restriction as had been successfully done in 1963; however, despite initial improvements in the balance of payments and in inflation in the subsequent cyclical recovery, the monetary stance soon turned permissive. 8. However, the degree of the difficulties in progressing along this road is indicated not least by the fact that the main instigator of these proposals, Ezio Tarantelli, was assassinated by left-wing terrorists in 1986. 9. This historical feature of Italian society is sharpened especially by the progressive loss of influence of traditional popular parties, in particular the Christian Democratic party that had ruled the country since 1945, which in the past had managed at least partially to reconcile the different demands by taking into account the conditions for sustainable growth. 10. In the mid-1990s, mean employment in Italian manufacturing firms was only 42 percent of the EU15 average (Pagano and Schivardi 2003).

the golden age and the second globalization in italy

107

11. Sheltered sectors include energy companies, banks, municipal utilities, airport management companies, Autostrade, and Telecom Italia. 12. It should be recognized that there are substantial regional variations in bureaucratic burdens on business within Italy; they are generally more onerous in the South and the Islands (Bianco and Bripi 2010). 13. The index reflects regulation, education and skills, labor market flexibility, the innovation framework, and active labor market programs (Rae and Sollie 2007).

Chapter 4

ITALY, GERMANY, JAPAN FROM ECONOMIC MIRACLES TO VIRTUAL STAGNATION

andrea boltho

Introduction Most chapters in this volume place Italy in an international context by looking at how it fared relative to other countries in a number of areas. This one is more openly comparative and directly contrasts Italy’s macroeconomic experience since World War II with those of Germany1 and Japan (often referred to as IGJ in what follows). The most obvious reason for this choice is the course of economic history since 1945.2 IGJ were defeated in war, enjoyed rapid growth in the reconstruction period, experienced “economic miracles” in the Golden Age, faced sharp decelerations after the first oil shock, and have recorded the slowest growth rates in the Organisation for Economic Cooperation and Development (OECD) area since the mid-1990s. Other countries have shared some of these trends; none have had quite so similar developments. Confirmation is provided by Table 4.1, which shows for per capita gross domestic product (GDP) an IGJ performance well above that of comparable economies in the first two periods and well below in the last one.3 There are, of course, many other similarities as well as glaring differences in the economic structures and histories of IGJ. Among the former, one could list high household savings; a bank-centered financial system; and ownership

109

italy, germany, japan Table 4.1 Italy, Germany, Japan: growth of GDP per capita (average annual percentage changes) 1946–1953 1953–1973 1973–1995 1995–2011 1946–2011 Italy Germany Japan Rest of OECD area

a

8.5

5.1

2.3

0.4

3.4

10.7

4.8

1.9

1.4

3.6

8.0

8.0

2.2

0.6

4.1

2.7

2.9

1.8

1.5

2.2

Sources: Baffigi (2011); Conference Board (2012); Maddison (2003). a The “Rest of the OECD area” (which excludes IGJ) is defined as Western Europe (excluding Turkey), North America (excluding Mexico), Australia, and New Zealand.

set-ups that, however much they may differ from each other, differ even more markedly from the Anglo-American paradigm (Carlin 1996; Barca et al. 2001). Among the latter are very different inflation histories, trade union behavior, and importance of regional problems. In what follows some of these issues are examined primarily from an Italian standpoint, with Germany and Japan being used as illustrations of how differently or similarly structures evolved or policies reacted. It should be noted that the approach is largely descriptive and many of the conclusions are based on subjective judgments. The Chapter follows (boring) chronologic lines, examining the reconstruction years (1945–1953); the Golden Age (1953–1973); the post–oil shock period (1973–1995); the recent years of semistagnation (1995–2011); and the issue of regional differentials.

Reconstruction In IGJ reconstruction turned out to be a much smoother process than had initially been feared given the post–World War I experience of a brief boom soon reversed by recession, inflation, and hyperinflation (Armstrong, Glyn, Harrison 1991). The amount of war-time destruction had been more limited in Italy than in the other two countries (Armstrong, Glyn, Harrison 1991). Hence, Italy returned to prewar levels of output by the late 1940s to early 1950s, somewhat earlier than the mid-1950s of Germany or Japan, but in all three countries recovery was remarkably speedy. Some of the reasons for why reconstruction turned out to be so much more successful than after 1918 are common to all the major belligerent OECD countries. They have to do, inter alia, with a different domestic policy stance in Western Europe and a different external policy stance in the United States (Boltho 2001a). Turning more specifically to IGJ, one striking similarity was the inflation and subsequent price stabilization experience. Inflation, rampant initially, was curbed by a sharp monetary squeeze in Italy in 1947, by a major

110

aggregate growth and policy

currency reform in Germany in 1948, and by a very restrictive fiscal stabilization plan in Japan in 1949. And in all three countries, but especially in Germany and Japan, the demand shortages arising from these policies were greatly relieved by the outbreak of the Korean War and the economic boom that ensued. A more important similarity was the early solution given to uncertainties about the postwar institutional set-up. Worries were widespread in the late 1940s, given strong worker militancy everywhere (Armstrong, Glyn, Harrison 1991) and the threat, particularly felt in Europe, of encroaching Soviet power. Although a shift to a command economy was not really on the cards, such countries as Britain or France were engaging at the time in large-scale nationalizations and indicative planning. The presence of US troops in IGJ made such policy shifts much less likely, but initially (and in Italy and Germany at least until the 1948 elections and currency reform, respectively) business felt unsure about its future role. In the end an open and liberal market system was chosen in all three countries, but there was no certainty of this at the outset. A much discussed issue is whether the period saw the shattering of so-called distributional coalitions (Olson 1982). Subsequent growth in IGJ, it has been argued, was greatly helped by the purge of prewar elites and by the destruction of cartels, lobbies, and other pressure groups whose rent-seeking activities sap the vitality of an economy. Yet, this factor does not seem to have been important. A recent survey of the evidence showed that there is little support for the Olsonian hypothesis in Germany or Japan (Heckelman 2007); similar studies of Italy are not available, probably because “Mancur Olson did not attract much interest” in the country (Da Empoli 1993, 81). In other words: “Although the war may have disturbed the earlier constellation of institutional arrangements and interest groups, it did not prevent them from being rapidly reconstituted” (Eichengreen 2007, 54). Thus, in Germany, the main distributional coalitions “re-emerged . . . with ideas, ideologies, practical purposes and personnel that were in full continuity with Weimar times” (Paqué 1996, 97). In Italy too, there is little evidence of a breakdown of prewar structures. Indeed, one of the main economic achievements of Fascist Italy, the 1933 creation of the large state-holding company IRI (Institute for Industrial Reconstruction), was preserved (despite initial US resistance) with virtually the same management (Barca and Trento 1997). It was only in Japan that some changes turned out to be growth promoting: zaibatsu dissolution raised competition and allowed junior managers to achieve power earlier than would otherwise have been possible (Kosai 1986), land reform stimulated agriculture, and labor reform raised incomes and demand (Nakamura 1981). These changes resulted from US reforms, however, not from a breakdown of the prewar system.4 Even if an Olsonian bonfire of distributional coalitions did not occur, the period provided a window of opportunity for radical changes in institutional set-ups. The opportunity was more fully seized in Germany and, especially, Japan than in Italy, in part because the US occupation authorities tightly supervised the former two countries, whereas Italy was left broadly in charge of its reconstruction.

italy, germany, japan

111

Thus, in Italy an old liberal tradition reasserted itself. The controls of the Fascist economy were dismantled relatively quickly (by the standards of what was happening elsewhere in Europe) and a return to free trade, in particular, was rapidly achieved (Graziani 1972), but there was none of the purging of top managers and cartel busting that occurred in Germany and Japan (Barca et al. 2001). Germany, just as Italy, also saw a return to a long-standing liberal tradition that contributed to monetary reform and rapid decontrol. Interestingly, however, Germany’s liberal instincts were tempered by the search for a “third way” (Abelshauser 2004), which began laying the ground for what became known as the Soziale Marktwirtschaft. No doubt encouraged by the New Deal–influenced US occupation authorities, some of the participants in the German discussion at the time advocated “liberal interventionism” and a “strong state” (Abelshauser 2004, 95), so as to moderate the social costs of unfettered markets.5 The outcome was the gradual introduction of welfare provisions and codetermination, measures that probably helped to foster social consensus. Nothing of the sort happened in Italy. On the contrary, polarization between a conservative business class and a radical trade union movement (polarization that in milder form had already existed before World War I) became entrenched in these years (Magnani 1997a). Japan’s prewar liberal tradition had been much weaker than that of Italy or Germany. Hence, there was little demand for rapid decontrol and “the system that emerged after reform was by no means a truly market-based system; instead it was characterized by a substantial degree of government intervention” (Teranishi and Kosai 1993, 15). Widespread land reform was also important. This not only contributed to subsequent agricultural growth, but also led to much greater income and wealth equality (Kosai 1986), an outcome that may have helped later economic success. Italy’s efforts in this area were much more timid and less successful.6 The reconstruction period may thus have opened an institutional gap between Italy on the one hand and Germany and Japan on the other. The latter two countries, whether of their own volition or under American pressure, adopted some fairly radical reforms that paved the way for a subsequent degree of consensus that had not existed before the war. Italy’s reforms were less farreaching and the country reverted to the nonconsensual pre-Fascist liberal order. Had Italian reformers been more audacious, or had the United States been more influential or assertive, subsequent developments in the country might well have been more favorable.

The Golden Age Developments in the period from the end of reconstruction to the first oil shock are well-known and broadly similar in IGJ. Output and employment

112

aggregate growth and policy

growth were exceptionally rapid (indeed, Germany and Japan enjoyed virtual full employment most of the time); export market shares rose sharply (if at different speeds); and welfare state provisions (if of different generosity) were made widely available. Not for nothing, the experiences of IGJ in these years were often referred to as “economic miracles”7 even if (blot on the picture) there was also a move from near price stability at the outset to high inflation by the early 1970s. All three countries faced, of course, a world economy that grew at unprecedented rates. In addition, trade was liberalized (with Italy and Germany also joining a customs union), and the terms of trade moved in IGJ’s favor.8 Demand management was little used by IGJ (indeed, in Japan, cycles may well have been policy generated) (Ackley and Ishi 1976). Yet, fiscal and monetary policies may still have helped investment by coordinating their stance in a growth-promoting way. In all three countries, fiscal policy was broadly orthodox, whereas monetary policy was broadly accommodating. Budgets were, through most of the period, kept in balance or even small surplus (thereby adding to the supply of savings). Italy was a partial exception, but even here deficits were contained to levels well below those of earlier or later eras. Money supply, on the other hand, grew rapidly (thus facilitating investment), without this leading to inflation in view of the concomitant buoyant growth in the demand for money. Even that paragon of anti-inflationary credibility, the Bundesbank, allowed broad money supply growth rates in the 1950s and 1960s of as much as fourteen percent per annum, as did Italy (while consumer price inflation remained negligible). Behind these various factors lies something all three countries shared: catch-up. High growth and investment were fostered by a backlog of opportunities vis-à-vis the United States and by the profits provided by elastic labor supplies (Kindleberger 1967a). Although full employment was eventually achieved in the urban sector of IGJ, underemployed labor was still present until the end of the period in the Italian and Japanese countryside, or in Southern Europe for Germany. The rapid productivity growth that buoyant investment generated, together with the low wage settlements that elastic labor supplies permitted, were further ingredients in the virtuous circle of high investment and high export growth. Catch-up provides a compelling explanation for the Golden Age in IGJ (as well as for the rest of the OECD area). Even a very simple specification of the hypothesis throws a good deal of light on the 1953–1973 differential growth experience (Figure 4.1); Italian and German growth can be almost fully “explained” by their relative position at the outset. Only Japan (and Ireland) stand out as partial exceptions. Backwardness provided the opportunity. Social capability then exploited it (Abramovitz 1986). Defining, let alone measuring, the concept is impossible. Yet selected socioeconomic aspects can, plausibly, be thought of as being important in facilitating the exploitation of a catch-up potential. Three are looked at next: (1) educational achievements, (2) the presence of cooperative institutions, and (3) the pressure of competitive forces. Educational achievement is easiest

113

italy, germany, japan Growth rate of GDP per capita Ch.Yi = (logYi 1973 – logYi 1953) % p.a. J 7 GR P

E

A

5

D I

SF

IR

3

F

B N

Ch.Yi = 27.7 – 2.76 log (Yi)1953 (8.2) – R2 = 0.77

1 2

NL

DK CN S UK AU NZ

CH US

3 4 5 6 7 8 9 10 Country’s GDP per capita in 1953 (Yi 1953) in 1990 ppp dollars (1000s.)

11

Figure 4.1 Convergence: 1953–1973. Source: Basic data from Maddison (2003).

to quantify. Table 4.2 presents estimates of the educational attainment of the total and adult populations. At the outset and through the period Germany and Japan stand out. This factor might well have facilitated, inter alia, the exploitation of foreign technologies.9 Turning to the possibility that institutions may foster cooperative behavior that, in turn, enhances growth, two not mutually exclusive approaches are worth noting: the idea that trade union structures are important in directly determining inflation/unemployment outcomes and in indirectly influencing macroeconomic ones (Calmfors and Driffill 1988); and the alternative view that particular institutions and policies encouraged cooperative solutions in some West European countries, greatly helping to secure moderate wage settlements and high investment rates (Eichengreen 2007). Calmfors and Driffill (1988) argue that centralized, all-encompassing national unions, or decentralized, enterprise unions, are preferable to sectoral-industrial unions, and place Germany and Japan in the first and second category, Table 4.2 Italy, Germany, Japan: educational attainments Population Aged 15–64 (equivalent Population Aged 25 and Over years of primary education) (average years of total schooling) 1950 Italy Germany Japan

1973

1950

1970

5.5

7.6

4.1

5.2

10.4

11.6

4.9

5.2

9.1

12.1

6.1

7.5

Sources: Barro and Lee, 2010; Maddison, 1995.

114

aggregate growth and policy

respectively, while relegating Italy to the third one. Further refinements have argued that in Japan centralized coordination between unions and employers’ federations dominates (Soskice 1990), but the ultimate conclusion has seldom been put in doubt. The evolution of unionism, largely dictated by history, could thus be one explanation for the better employment performance of Germany and Japan relative to Italy’s. Eichengreen’s (2007) view is that implicit work force–enterprise bargains (in which labor foregoes immediate wage claims in exchange for higher investment and, therefore, higher real incomes in the future), secured faster growth in some European countries than in others. Germany fits this hypothesis, because it had institutions and policies, some going back to Weimar days, others developed early in the period (e.g., codetermination), that could be seen as encouraging cooperation. Similar institutions and policies, however, can hardly be found in Italy. Japan is not covered by Eichengreen, but there is sufficient circumstantial evidence suggesting that the country’s labor force was ready to accept longer-term commitments (e.g., the presence of “life-time employment,” at least in large-scale firms, and the significant share of compensation accruing in the form of bonuses).10 All this suggests that for complex historical and institutional reasons, Germany and Japan were in a more favorable position than Italy to exploit relative backwardness with less social strife and conflict. A not dissimilar conclusion may be drawn from an alternative viewpoint that stresses not so much cooperation as the forces unleashed by competition. IGJ all moved away from the controlled prewar economies, but at different speeds. Of the three, Germany embraced competitive forces most vigorously. It opened itself fully to the rest of the world, even adopting unilateral tariff reductions in the 1950s. It also passed robust anticartel legislation. Italy largely matched Germany’s opening to external competition, but was much less ready to thwart domestic anticompetitive forces. Monopolies or oligopolies had been widespread in prewar Italy and little was done to dismantle them after the war (Rossi and Toniolo 1996; Barca 1997). This was particularly true for the State-owned corporations that dominated banking and large-scale industry (De Cecco and Giavazzi 1993). On the surface, Japan moved least in the direction of more competition. The liberalization of foreign trade was slow and patchy and antimonopoly legislation lacked teeth. Yet, interestingly, many observers argue that domestic competition was fierce, as the keiretsu conglomerates fought for market share (Tsuru 1976; Yamamura 1982). The foregoing would suggest that Germany’s policies were first best, whereas Italy and Japan were held back in establishing vibrant and competitive economies by the pressures of domestic vested interests in one case and by mercantilist attitudes in the other. Yet, arguably, Japan’s combination of competition at home and protection vis-à-vis the rest of the world may have been superior to Italy’s opposite set of policies. A comparative look at the two countries’ international performance shows large market share gains in both cases (as it

115

italy, germany, japan % 15

Germany

Japan

12

9 Italy

6

3

0 1950

1960

1970

1980

1990

2000

2010

Figure 4.2 Export performance. Share in world exports of manufactures (3-year moving averages). Sources: General Agreement on Tariffs and Trade, International Trade (various years); World Trade Organization database.

does for Germany) (Figure 4.2), but also points to a significant difference in specialization patterns. In the early 1950s, Italy and Japan shared comparative advantage in broadly similar lines of activity; over time, however, divergence set in, as shown by changes in their revealed comparative advantages (Boltho 2001b). Italy’s specialization changed only slowly and the country remained committed to, admittedly, increasingly sophisticated consumer goods; Japan, instead, moved into high-tech activities and this process was almost certainly helped by the protection its capital-intensive industries enjoyed. Active in a sheltered but still intensely competitive environment, Japanese firms invested massively, slid down their average cost curves as domestic demand grew, and eventually entered the world market, a process that has been aptly called “import protection for export promotion” (Krugman 1984). Overall, however, Italy broadly matched the growth rates of Germany and Japan in the Golden Age, and this despite some less favorable preconditions. If shortfalls there were, these lie not so much in macroeconomic performance, but in the labor market and social policy areas. Italy, unlike Germany and Japan, did not achieve full employment largely because of its pronounced regional differentials. It also faced a large wage shock at the end of the 1960s. So did Germany, but its wage increases were not accompanied by the wave of strikes that were an integral part of the Autunno caldo movement,11 nor did they bring about lasting changes in labor market regulation, as was the case in Italy. The Italian literature tends to lay the blame for this episode on the whole development process since the war: “In some sense the social

116

aggregate growth and policy

conflicts of those years were an almost natural response to the speed and complexity of Italy’s economic development. In a relatively short time-span, the country had moved from a mainly agricultural structure to one of widespread industrialization, with extraordinary migration flows . . . and with a huge increase in the rate of urbanization” (Signorini and Visco 1997, 54). More specifically, the intensity of strike activity in 1969–1970 has, at least in part, been attributed to the very poor working and living conditions that faced the huge number of migrants who arrived in the North of the country (Valli 1977; Castronovo 1995). There is a clear contrast here with how Germany and Japan coped with similar phenomena, since migration and urbanization were hardly uniquely Italian features. Just after the war, Germany saw the immigration of 10 million refugees from Eastern Europe and Japan saw the repatriation of 5 or 6 million nationals from the erstwhile empire. Their incorporation into the labor force was relatively smooth. Both countries then witnessed further internal migration (Japan’s almost certainly larger than Italy’s) as agricultural employment dwindled.12 Germany, in addition, received some 6½ million foreigners. The 3½ to 4 million Southern Italians who moved to the Center-North between 1955 and 1973 were not an impossible number to manage. Where Italy thus seems to have failed (at least in relative terms) was in its inability to accompany rapid economic change with parallel changes in infrastructure and social welfare provisions. This inability, in turn, led to an overreaction. The late 1960s and early 1970s saw the hasty adoption of generous welfare benefits and the beginning of indiscriminate handouts for business, measures that gave the politicians much greater scope for interfering with economic policy-making. Many of the fiscal problems encountered in the 1970s and 1980s stem from the belated and misguided responses to the late 1960s crisis. Migration and welfare reform were clearly managed more smoothly and successfully by the other two countries. The experience of the late 1960s thus points to a major Italian shortcoming relative to Germany and Japan. Administrative ability (by politicians and the civil service) seems to have been in short supply, leading to more pervasive rentseeking, slower judicial procedures, worse legislation, and so forth. Evidence in support of this judgment is not available for the late 1960s, but over the last thirty years indicators have emerged that put Italy into a much less favorable position than the other two countries, be this in the areas of an intrusive regulatory environment, the provision of law and order, the control of corruption, the presence of trust, and so forth (Mauro 1995; Knack and Keefer 1997). This is confirmed by more recent World Bank investigations for 1996–2010. In all but one of the six governance indicators covered by that institution, Italy is significantly below Germany and Japan. An even more damning picture is painted by survey evidence on infrastructure quality: in recent international comparisons of 125 countries, Germany was ranked sixth, Japan sixteenth, and Italy seventythird (World Economic Forum, various years).

117

italy, germany, japan

Slowdown A bird’s–eye view of developments in this period shows that trends were not that dissimilar in IGJ. All three countries slowed down relative to the Golden Age (Table 4.1), largely because the potential provided by catch-up had greatly diminished. Yet, on average, they still grew at or slightly above the OECD’s growth rate at the time. All three experienced recessions, two in the wake of the oil shocks and a third one in the early 1990s, and all three saw strong upswings in activity in the late 1970s and late 1980s. The parallels, however, stop here because in other respects IGJ went their separate ways. By far the most glaring contrast was in macroeconomic policies, with Italy behaving in ways that were almost opposite to those of Germany and Japan. This was particularly so for fiscal policy. Italy’s budget deficit averaged some 10 percent of GDP for twenty years and the public debt-to-GDP ratio rose to 120 percent by 1995. Public debt also rose in Germany and Japan, but this only happened late in the period, in response to the costs of unification in one case and of recession in the other. In Italy, the reasons for fiscal incontinence are to be found instead in the politicians’ aim to buy social peace. “Frenetic activism” (Arcelli and Micossi 1997, 280) was the defining characteristic of a policy that was used to raise pensions, improve health provisions, boost public payrolls, subsidize industry, and so forth. Political expedience rather than economic rationality ruled decision-making: “The struggle for income shares was shifted from the factory floor to the state’s budget” (Maier 1999, 281). Conversion to a more orthodox stance came only late in the period, as the deadline for entry into the European Monetary Union approached. Monetary policy was permissive in the second half of the 1970s, but was tightened in the early 1980s as suggested by the level of real short-term interest rates (roughly measured by using an ex post deflator). These were virtually identical to those of Germany and Japan (and significantly higher than Germany’s in the years 1981–1990). Two events were important in the monetary policy conversion to greater orthodoxy: entry into the European Monetary System in 1979 and a first step toward granting more independence to the Bank of Italy, by freeing it from the obligation to buy Treasury bonds, in 1981 (Basevi and Onofri 1997). Interestingly, however, real long-term rates, although positive,13 remained below those of the more prudent Germany and Japan. Clearly, having a large pool of domestic savings greatly facilitates public sector financing, as also shown by Japan’s experience in recent years. Yet, despite relative monetary tightness, Italian inflation remained in double digits until the mid-1980s, significantly outstripping that of Germany and Japan. One reason for this may have been the lack of competition in many service sectors (Ciocca 2007) (the same was probably true in Japan, but much less so in Germany). A second reason had to do with the attitude of the largest trade union confederation, which opposed any form of incomes policy until

118

aggregate growth and policy

8

Other OECD countries Germany

7

Japan Italy

6

5 1970 1975 1980 1985 1990 1995 2000 2005

Figure 4.3 The “Freedom” Index. Indicator of government interference with the economy. Note: Values go from zero to ten; higher values denote greater economic freedom. Source: Gwartney, Hall, and Lawson (2011).

as late as 1992 (Ciocca 2007). German and Japanese unions were much more conscious of macroeconomic constraints. A third reason is linked to the public sector’s continuing increases in employment, in wages, and in income maintenance transfers (Arcelli and Micossi 1997). On the other hand, the hope that European Monetary System membership would have had favorable expectational effects on wage bargaining, now that Italy’s monetary policy was, de facto, in the hands of the Bundesbank, did not materialize (Egebo and Englander 1992).14 Partly as a consequence, sharp nominal depreciation led to no real depreciation; indeed, the opposite was the case until 1993. In addition, the country was also plagued by excessive regulation. Labor market rigidities were almost certainly more pronounced than in Germany or Japan, be this in the area of employment protection or in that of collective bargaining, given the virtual absence of coordinated wage negotiations (Nickell, Nunziata, Ochel 2005). Foreign exchange and price controls were pervasive: at the end of the 1970s the government directly or indirectly may have controlled up to 30 percent of the items in the consumer price basket (Arcelli and Micossi 1997). More generally, the Fraser Index of “Economic Freedom” (Gwartney, Hall, Lawson 2011), a measure that tries to quantify how a government interferes with market forces, placed Italy significantly below Germany and Japan in the 1970s and 1980s, despite some progress in 1985–1995 (Figure 4.3). Italy thus suffered from both macroeconomic mismanagement and excessive microeconomic interference. One would expect this to have led to serious economic difficulties. Conversely, the lesser regulations and controls of Germany

italy, germany, japan

119

and Japan, and their more restrained fiscal and monetary policies, should have facilitated a better performance than Italy’s. Yet, real output, employment, and productivity rose as or more rapidly in Italy than they did in Germany or Japan. Nominal developments in the three countries may have been very different, but real ones were not. Two possible (and not mutually exclusive) explanations for such a surprising outcome can be advanced: the first one argues that the costs of fiscal profligacy and partial monetary accommodation are perhaps not as high as is often thought; the second one points to other forces that, at least partially, offset Italy’s “wrong” policies. For the earlier part of this period, there is some research that suggests that the policy mix chosen by the Italian authorities was not that suboptimal after all (Boltho 1986; Giavazzi and Spaventa 1989). The latter authors, in particular, argue that the initial monetary accommodation of the second oil shock allowed high profits, which sustained later investment demand. The inevitable disinflation costs were thus lower than they otherwise would have been. Gradual disinflation was more growth promoting (or less growth retarding) than the United Kingdom’s shock therapy of the time, or the unnecessarily strict policies followed by Germany through the period (Carlin 1996). The second explanation stresses something that arguably went “right” in Italy at the time: the employment shift to medium- and small-scale firms (SMEs), particularly in manufacturing. This shift was neither unique to Italy nor to the period. It was, however, more pronounced in Italy, partly in response to the rigidities introduced into large factories by the strike waves of the late 1960s. Table 4.3 shows a rough proxy for this phenomenon (the importance of firms or establishments with ten to ninety-nine workers). These firms’ rising employment share, while predating the Autunno caldo, was especially marked in the 1970s and 1980s, in contrast to smaller changes in Germany or Japan. The added flexibility that this generated probably helped Italy to hold on to its share of world manufacturing exports in these years (Figure 4.2), whereas Germany and Japan suffered pronounced losses (admittedly from much higher initial levels). Two interpretations of this period are possible. According to one, Italy suffered from prolonged macro and microeconomic mismanagement and grew only thanks to its proverbial flexibility. Had it embraced orthodox policies with greater conviction, and deregulation and reform with more vigor, its performance would have been much more favorable. An alternative view, although accepting the importance of flexibility, questions the policy judgment. Germany epitomized macroeconomic orthodoxy, yet its overall performance was below that of Italy.15 As for Japan, thanks partly to financial deregulation, it found itself in a bubble and bust whose consequences are still with the country today. Perhaps, after all, high inflation and large budget deficits are not as destructive of economic activity as the orthodox literature suggests. Italy eventually converged on orthodoxy, but its gradual shift in that direction may have been preferable to more immediately austerity, given in particular that the country

120

aggregate growth and policy Table 4.3 Italy, Germany, Japan: importance of small firms (share of manufacturing employment in establishment or firms with 10–99 employees) 1961

1971

1981

1991

2001

Italy

26.4a

28.3

33.6

38.9

41.8

Germanyb

18.3

17.4

22.2

20.5c

22.5

40.3

41.4

41.7

41.7

Japan

d

43.7

Note: The data may not be strictly comparable across countries. a Eleven to one hundred employees. b The data are approximate because definitions changed in the period and East Germany is included from 1991. A very rough attempt was made to try to make the data comparable. c 1990. d 1959. Sources: Ciocca, 2007; Japan Statistical Yearbook (various issues); Statistisches Jahrbuch (various issues).

did not benefit from the wage bargaining institutions and social consensus of Germany and Japan.

Stagnation In next decade and a half IJG shared one common feature: they all experienced near stagnation in per capita income growth and were the slowest growing of the twenty-two countries of the “old” OECD area.16 This apparent similarity masks, however, different initial causes: Italy had to follow very restrictive fiscal policies in the late 1990s to fulfil the European Monetary Union’s entry requirements, Germany had to digest the costs of unification, and Japan had to come to terms with the bursting of its bubble. Although Germany recovered strongly at the end of the “Great Recession” thanks to rising exports made possible by labor market reforms and a prolonged phase of wage moderation, Japan has continued to struggle: because of price deflation, nominal GDP in 2011 was below its value for 1995. As for Italy, it hardly grew between 2001 and 2011. One possible explanation for such trends could lie in the globalization process, which affected IGJ differently according to whether their industrial structures were complementary to or competitive with those of emerging economies. Germany and Japan were likely to benefit more from (and be hurt less by) this new competition than was Italy, given that their exports are more heavily weighted in favor of high-tech and investment goods.17 Germany, in addition, greatly benefited from the opening of Eastern Europe, an area with

121

italy, germany, japan ITALY - TRADE WITH EASTERN EUROPE

GERMANY - TRADE WITH EASTERN EUROPE

JAPAN - TRADE WITH CHINA

20

20

20

15

15

15

10

10

10

5

5

5

0

1990–91 1995–96 2000–01 2005–06 2009–10

0

1990–91 1995–96 2000–01 2005–06 2009–10

0

1990–91 1995–96 2000–01 2005–06 2009–10

Figure 4.4 Trade with Eastern Europe and China (percentage shares in total exports plus imports). Source: International Monetary Fund, Direction of Trade Statistics.

which it had had historically close relations and with which trade rose sharply (Figure 4.4).18 Something similar happened to Japan’s trade with China. Italy, for obvious geographic and historical reasons, was left behind. Moreover, as the importance of foreign trade rose, so did the role of the real exchange rate. Here Italy lost ground to Germany and Japan, reflecting its virtual absence of productivity growth (Figure 4.5). Italy’s loss of competitiveness was particularly severe inside the Eurozone, because unit labor costs rose by more than 30 percent relative to Germany between 1999 and 2011. A further tentative explanation for Italy’s relative failure may, paradoxically, come from the country’s earlier successes. Globalization may have weakened the SMEs that dominate the traditional industrial districts (Ciocca 2007). Not only do they have to compete with emerging countries that specialize in similar products but enjoy much lower labor costs, but they also find it more difficult and expensive, because of their size, to delocalize, something that larger German or Japanese firms did on a significant scale in Eastern Europe and China, respectively.19 A production model that seemed well suited to the European market at the end of the twentieth century may perhaps be less appropriate to the more globalized world of the early twenty-first century. A second set of explanations turns to institutional reasons for relative stagnation. It has been argued that the institutions that Europe (and, arguably, Japan) developed during the Golden Age became gradually less suited to the world’s changing conditions at the turn of the century (Eichengreen 2007). In an environment of rising technologic uncertainty, the need is for venture capital and flexible financial markets, not for bank-based systems (Eichengreen 2007). Similarly, new dynamic sectors require research and development and human capital, rather than fixed investment. Even within research and development, the emphasis has to change from the incremental innovations for which

122

aggregate growth and policy

150

Italy

130

110

90

Germany

70

Japan 50 1995 1997 1999 2001 2003 2005 2007 2009 2011

Figure 4.5 Real exchange rates (indices; 1995 = 100). Normalized unit labor costs in manufacturing. Source: International Monetary Fund, International Financial Statistics.

Germany, for instance, is well known, to the radical spurts more typical of US firms (Hall and Soskice 2001). In addition, Europe and Japan’s institutions are often accused of being sclerotic. Given the emergence of nimbler competitors, public sectors have to be more efficient, marginal tax rates lower, and welfare provisions less generous; even more importantly, competition in product and, especially, labor markets must be encouraged (OECD 1989). There is room to cast doubt on some of these propositions. The evidence linking, for instance, labor market reforms to lower unemployment outcomes is virtually nonexistent (Howell et al. 2007). Still, few would doubt that there are sclerotic institutions in IGJ. One piece of evidence that might indirectly support this is the relative presence or absence of foreign direct investment, on the plausible assumption that multinational corporations shun overregulated economies. It turns out that between 1995 and 2010 (or between 1980 and 2010, for that matter), foreign direct investment inflows in percent of GDP have been lower in IGJ than in any single other OECD country.20 A third and linked set of arguments suggest that IGJ were also reluctant to change their institutions, at least relative to what was happening in other advanced economies. Figure 4.3 had already shown how the reform efforts of IGJ were relatively modest, indeed nonexistent in Italy, since 2000. More detailed data on selected aspects of regulation present a mixed picture (Table 4.4). In Italy there seems to have been some progress in the ease with which business is conducted, but there was none in the legal area. Judicial procedures still take an inordinately long time and the presence of the rule of law seems to have sharply decreased relative to what was happening elsewhere. In addition,

123

italy, germany, japan Table 4.4 Italy, Germany, Japan: institutional reforms (selected indicators) Rule of Lawa Italy

Germany

Japan

Starting a Closing a Businessb Businessc

Enforcing Contractsd

2003

0.98e

16

39

1,390

2010

0.38

6

58

1,210

2003

1.62e

27

50

403

2010

1.63

12

53

394

2003

1.39e

21

93

242

2010

1.31

16

93

360

12

71

309

8

78

366

Rest of OECD area 2003

1.54

2010

1.61

e

Note: The data may not be strictly comparable across countries. a Deviations from world average. b Average of number of procedures and time spent (in days) to register a new firm. c Claimants’ recovery rate (in percentage) from an insolvent firm. d Duration of legal procedures (in days) needed to resolve disputes. e 1996. Sources: World Bank, “Worldwide Governance Indicators” and “Doing Business” Databases.

economic performance must also have suffered from the presence, through much of the period, of what must have been the worst Prime Minister Italy has had since World War II. Such slow progress in reform would not have come as a surprise to Olson, who argued that as democracies grow in peace time, so does the importance of their distributional coalitions. These, in turn, “slow down a society’s capacity to adopt new technologies and to reallocate resources in response to changing conditions, and thereby reduce the rate of economic growth” (Olson 1982, 74). IGJ may not have swept away their distributional coalitions after the war, but they may well have presided over a gradual build-up in their importance through time. The evidence in favor of this thesis, however, is not very strong. Trade unions and monopolies/oligopolies are the two major culprits of Olsonian institutional sclerosis. Yet, the importance of unions has clearly shrunk over the last ten to twenty years, after labor market deregulation, globalization, and structural changes reducing the importance of heavily unionized public enterprises or large-scale factories. Thus, most OECD countries, including IGJ, have seen significant declines in unionization and in strike activity (Table 4.5). Restrictive product market practices must also have been weakened by deregulation and, more importantly, by the much higher degree of competition brought about by freeing world trade through successive General Agreement on Tariffs and Trade/World Trade Organization liberalization rounds and, within Europe, by the 1992 Single Market Programme.

124

aggregate growth and policy Table 4.5 Italy, Germany, Japan: importance of trade unions and strikes Italy

Germany

Japan

1960

24.7

34.7

32.3

1970

37.0

32.0

35.1

1980

49.6

34.9

31.1

1990

38.8

31.2

25.4

2000

34.8

24.6

21.5

2010

35.1

18.6

18.4

Trade union densitya

Strike activityb 1950–1959

679

98

715

1960–1969

1,402

22

250

1970–1979

1,503

53

124

1980–1989

621

27

10

1990–1999

158

11

2

2000–2008

86

5

0

a

Union members in percent of total employees. Working days lost per thousand employees. Sources: Bean, 1989; Economic and Labour Market Review and Employment Gazette (various issues); Flora, Kraus, and Pfenning, 1987; International Labor Organization, Statistical Database; Labour Market Trends and Monthly Labor Review (various issues); OECD Database. b

It is true, however, that in IGJ (and in other OECD countries), resistance to change seems strong. Low social mobility could be one reason, particularly for Italy. A recent survey of intergenerational income persistence suggests, for instance, that in Western Europe this is lowest in Italy, together with Spain (Causa, Dantan, Johansson 2009), and according to earlier work certainly also lower than in Germany (D’Addio 2007) (no data were found for Japan). One would need, however, evidence on whether this has declined through time before arguing that it may have added to Italy’s lack of dynamism in recent years. A more plausible candidate for this could come from trends in the wealth of the population. Figure 4.6 shows trends in the ratio of net household wealth to income in IGJ and in four other major OECD countries. Abstracting from the fluctuations that came with the bursting of bubbles in Japan in the early 1990s and in Britain and the United States in the late 2000s, what the chart clearly shows is that IGJ, and Italy in particular, are among the wealthiest societies on Earth. It is not implausible to suggest that

125

italy, germany, japan % 9

Italy

Japan 7 Other countries*

5

Germany

3 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008

Figure 4.6 Wealth/income ratios (total household net assets in percent of disposable income). *Weighted average of Canada, France, United Kingdom, and United States. Source: OECD, Economic Outlook (various issues).

there could be a link between “higher private wealth, a reduced propensity to undertake effort and risk, and a deceleration in the pace of economic development” (Ciocca 2007, 346). Finally, one further reason for semistagnation could come from rapid ageing. In 2010 IGJ had the highest shares in the world of those over 65 in their populations (Figure 4.7). The links between demography and economics are complex. Yet, common sense suggests that there must be some truth in the simple saying that: “old men do not like change; old men do not take new initiatives.” The difficulties in reforming the welfare state can clearly be linked to ageing, as the median voter’s age rises and he or she is increasingly attached to his or her health and pension benefits. Similarly, decelerating productivity growth could be linked to lack of new initiatives. Survey evidence suggests, for instance, that older people are less entrepreneurial21 and that there is less entrepreneurial activity in ageing countries.22 If there is something in this hypothesis, then at least a common, if partial explanation for why IGJ performed relatively poorly in recent years would be forthcoming.

The Regional Problem Regional differentials in output, employment, and living standards exist in all countries. They are, however, particularly pronounced in Italy, at least compared

126

aggregate growth and policy POPULATION LEVELS (1990 = 100)

THE AGING POPULATION (share of people age 65 and up in total population) %

125

27

Other OECD countries

120

Japan

24

Italy

115

21

Germany

110

4.0 Italy

Germany

105

18

Other OECD countries 15

Japan 100 1990

1995

2000

2005

2010

2015

12 1990

1995

2000

2005

2010

2015

Figure 4.7 Demography. Source: UNPD, World Population Prospects Database (2010 Revision).

with those of Germany and Japan. The left-hand panel of Figure 4.8 presents a simple measure of geographic inequalities: the (weighted) coefficient of variation of regional GDP per capita for eleven regions in IGJ (and for sixteen regions in Germany after unification). The chart speaks for itself. Throughout the period, Italian differentials are by an order of magnitude larger than those of Germany and Japan. “Poor” region’s GDP per capita in % of “rich” region (current prices)

Weighted coefficient of variation for per capita regional GDP

90

0.30 Italy (11 regions)

80

0.25

“Poor” W. Germany/ “Rich” W. Germany Japan North & Islands/ Central Japan

70

0.20

60 Germany 4.0 (16 regions)

Japan (11 regions)

50 East Germany/ West Germany

0.15 West Germany (11 regions)

0.10 1970

1975

1980

Mezzogiorno/ Center-North

1985

1990

40

1995

2000

2005

30 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Figure 4.8 Regional income differentials. Sources: Italian, German, and Japanese national accounts.

italy, germany, japan

127

A feature of Italy’s differentials is that they correspond to a clear-cut division between the Center-North and the South of the country, much more so than do the (smaller) German and Japanese differentials. The income gap between the Mezzogiorno and the Center-North of Italy is plotted in the right-hand panel of Figure 4.8 together with tentative estimates of roughly similar gaps in Germany and Japan. For Germany, the choice of a comparable indicator is simple after 1991. Unification brought together two geographically separate regions at very different levels of development. Before 1991, however, West Germany’s richer and poorer regions were spread across the country. The indicator used shows the ratio of GDP per capita in four poorer Länder to that of the other more fortunate seven Länder.23 As for Japan, it was split into a Central and relatively rich area and a more peripheral part made up of the Northern region, together with the islands to the South.24 Here too, the Italian gap stands out for its size and persistence. Even East Germany, despite suffering from a much larger gap vis-à-vis the West at the time of unification, has managed to close a good deal of the difference. A question this raises is how much Italy’s performance was affected by such a major territorial divergence. One possible way to answer this question is to assume that, contrary to what happened, Italy would have conformed to a regional convergence rule that has been shown to apply to a number of advanced economies (Barro and Sala-i-Martin 1995). According to this rule, the gap between poor and rich regions has declined by some 2 percent per annum, be this in the United States, in Japan, or in Western Europe. Had Italy shared in this convergence process, the GDP per capita (constant price) differential between North and South would have shrunk over the period from nearly 40 to 10 percent, and the overall growth rate would have been higher.25 The results just obtained assume, however, that nothing else would have changed had the Mezzogiorno been more successful. This is clearly an untenable proposition. Thus: 1. More rapid income and employment growth in the South would have reduced migration flows to the Center-North; this, in turn, would have made for higher wages and lower international competitiveness, at least in the 1950s and 1960s; 2. However, a more successful Southern economy might itself have developed some competitive tradeable sectors; 3. Equally, a more successful Southern economy would have required a smaller public policy effort; tax rates might thus have been lower and/or other public expenditure higher. A first step in quantifying such a counterfactual is to assess the diminished push for migration in the presence of faster income convergence in the 1950s and 1960s (net migration fell drastically after the mid-1970s). Although the actual income gap in the years 1955–1973 (for which migration data exist) fell only slightly, it would have shrunk from nearly 40 to just over twenty percent of Northern per

128

aggregate growth and policy

capita GDP had the Mezzogiorno converged at 2 percent per annum. Assuming an elasticity of migration with respect to changes in this gap of −2,26 migration to the Center-North in the years 1955–1973 would have been reduced by perhaps 90,000 people per annum. Even assuming that migrants had a high participation rate (70 percent) and that half of them entered manufacturing (both rather extreme assumptions), the reduction in the Center-North’s industrial work force over the period would have been at most of some 7 to 8 percent. The evidence on the so-called wage curve shows that labor supply changes have only small effects on compensation levels. A consensus estimate is that the wage elasticity with respect to unemployment in developed countries is of roughly −0.1 percent (Blanchflower and Oswald 1992; Nijkamp and Poot 2005). Using such a figure, and assuming a 10 percent reduction in the Center-North’s labor force, would have resulted in only a 1 percent increase in the wage level and, perhaps, an even lower increase in unit labor costs (because lesser labor availability might have led to higher productivity). Given an export elasticity with respect to the real exchange rate of −1,27 annual export and GDP growth would have been reduced by perhaps 1 and 0.2 percent, respectively, thus not fully undoing the impact of faster Southern convergence. This very simplified calculation ignores, however, forces operating in the opposite direction. First, the North might have engaged in more capital-deepening investment. Second, in the presence of higher growth, the South might have developed some competitive tradeables, in line with a hypothesis that fast growth leads to increased exports (Krugman 1989). Third, and most importantly, if the strikes of the late 1960s, which led to significant increases in labor market rigidities, were in part a response to badly managed migration flows, in the presence of less migration Italy’s economic history could have been different. Quantifying these various effects is beyond the scope of this chapter, but an impressionistic judgment might be that they were at least as important, if not more so, than any negative first round effects of a somewhat lower labor supply elasticity to Northern industry in the years 1955–1973. In addition, a more successful South would have required less aid. Over the period as a whole, this may have averaged 1–1½ percent of GDP per annum (Boltho 2010). Assuming that as the income gap shrank help to the South would have diminished, Italian public expenditure and/or taxation would have been lower than it was. A conservative estimate might put the potential savings at ½ percent of GDP per annum. Had this been reflected in lower taxes, it might have had a (small) positive impact on entrepreneurship. Had it been used in more productive expenditure (e.g., infrastructure rather than income maintenance or subsidies), it could have had (somewhat larger) positive effects on growth. And to these benefits should be added the likelihood of less rent-seeking and crime (fostered by high levels of public spending). All this leads to an obvious conclusion: more rapid expansion in the South would have raised overall growth, some of the Autunno caldo’s costs might have been avoided, public funds would have used been more productively, and

129

italy, germany, japan

there might have been less crime. Even if a somewhat higher real exchange rate would have penalized exports in the Golden Age, it could still be argued that had Italy’s regional problems been more similar to those of the two other countries, its growth over the period might well have come somewhere in between those of Germany and Japan and well above what was recorded. A comparison with Germany’s experience is also instructive in showing why East Germany has been more successful in closing its initially much larger income gap vis-à-vis the West. East Germany benefited from massive public investment in infrastructure. More importantly, the West German institutions that it adopted were almost certainly superior to Italian institutions in terms of “a functioning legal system, control of crime and corruption, efficient administration of taxes” (Carlin 2010, 12). In addition, it was able to decouple itself from Western wage-setting mechanisms and labor costs declined rapidly in relative terms (Carlin 2010).28 It will be remembered that Italian unions had followed the exact opposite policy by enforcing wage equalization across the country in the late 1960s.

Conclusions IGJ have shared a number of similarities since World War II, most notably success. Growth of per capita income was rapid through much of the period; living standards are among the highest in the world; relatively generous welfare provisions (undreamt of in 1950) protect most of the population from sudden reversals of fortune; and educational and health standards are much higher than they were sixty years ago. In addition, IGJ are also very open to the rest of the world and are, if not welcoming, at least accepting the invigorating inflows of immigrants (Germany more so than Italy, and Italy more so than Japan). Yet, looked at from the standpoint of Germany and Japan, there are disappointing features in Italy’s performance. The relative lack of substantial reforms in the reconstruction years is one of them. Conversely, and surprisingly perhaps, a more permissive macroeconomic policy stance may not have been as pernicious as is commonly thought. In the 1970s and 1980s, when the monetary and fiscal policy divergences between Italy on the one hand and Germany and Japan on the other were at their peak, Italy’s performance was just as good if not better than that of the two other countries. Policy orthodoxy after all may not always be conducive to rapid growth. More important are three other differences: (1) administrative inefficiencies, (2) the continuing underdevelopment of the Mezzogiorno, and (3) the near-permanent state of conflict in industrial relations. German and Japanese administrative efficiency, functioning institutions, and infrastructure quality seem to be, on available evidence, greatly superior to Italy’s. Regional problems,

130

aggregate growth and policy

never serious in Japan, were severe in Germany after unification, but have so far been tackled rather successfully. Industrial relations were exceptionally peaceful in Germany and Japan compared with Italy (Table 4.5). Behind these differences lies a further one with much older roots: the comparative incapacity of the Italian state to efficiently run a modern economy. This is easily seen in the area of regional development where vast funds were wasted, often in rent-seeking and corruption. But it also permeates industrial relations. Seldom did the state play a pacifying role (as in various ways it did in Germany and Japan). The mismanagement of internal migration in the 1960s is one example of government shortcomings; the almost total failure of promoting reforms between 2000 and the very end of 2011 is another. While Germany (through tax policy) and Japan (through industrial policy) helped the formation of so-called patient capital, Italy had no such ambitions. Its corporate sector gave priority to short-termism, capital flight, and requests for public aid. Ultimately, Italy paid the price for remaining profoundly divided between a business class that looked only for immediate profits, and a trade union movement whose significant anticapitalist component was largely oblivious to economic constraints (Salvati 2000). Germany was spared such ideologic conflicts (Hennings 1982), whereas Japan never really had them. As rightly argued (Salvati 1984), Germany and Japan have shown through time a capacity to efficiently take and manage major decisions, a capacity that Italy lacks. Had the country had a more efficient state and more cooperative industrial relations, its regional split would probably have been less severe, and growth would almost certainly have been higher. Would such an alternative outcome have been possible? There are episodes in Italy’s postwar history that suggest that the potential for a better performance was there: rapid industrial growth in the 1950s, often led by state-run firms; partial North-South convergence in the 1960s; a successful incomes policy in the early 1990s; and a determined fiscal policy in the late 1990s. Unfortunately, they were sporadic. Yet, despite all this, the country’s performance was very respectable. Italy today enjoys living standards that are not much below those of Germany and Japan and grew nearly as rapidly as they did. The regret is that so much more could have been achieved. This is particularly true at present. The last fifteen years have seen very slow OECD growth, particularly so in IGJ. Some reasons for this are common (e.g. demography and, possibly, the dampening effects of rapid wealth accumulation). Japan, in addition, is still reeling from two decades of deflation. Germany, however, has been able through concerted company and union efforts to regain external competitiveness. Germany’s wage flexibility is particularly noteworthy. In the East it has promoted convergence; in the West it has stimulated exports. Italy cannot claim a Japanese-type excuse and has, contrary to Germany, lost competitiveness since the Euro came into being. More importantly, little progress is evident in the three areas singled out in this chapter as having hindered performance relative to Germany and Japan. Industrial relations have improved

131

italy, germany, japan

somewhat, but more because of the erosion of union power than because of any effort to build a social consensus, whereas the Mezzogiorno’s relative backwardness has hardly changed. Given the virtual absence of reforms, at least until 2012, it is difficult to see how the country can, despite the continuing vitality of its SME sector, return to significantly higher growth rates.

Acknowledgments Andrea Boltho is Emeritus Fellow, Magdalen College, University of Oxford. The author is grateful to Wendy Carlin, Marcello De Cecco, Michele Salvati, and Hans-Jürgen Wagener for helpful comments.

Notes 1. Germany stands for West Germany until 1991 and for the whole country thereafter. 2. Interesting parallels in the three countries’ histories occurred also in the nineteenth century, when Italy and Germany were unified and Japan saw a momentous internal transformation. 3. As with any periodization, issue could be taken with the time spans here shown. Although any choice of terminal years is arbitrary, it is believed that for present purposes it reflects a not unreasonable compromise given the histories of the three countries here considered. 4. America also imposed a Constitution under which Japan renounced war and curtailed its military spending. Avoidance of that burden may have contributed to the country’s growth (Kosai 1986), more so probably than in Germany or Italy, which also sheltered under the US security umbrella. Over the twenty years to 1973, defence expenditure amounted to just over 1 percent of GDP in Japan, as against 2.7 and 3.4 percent in Italy and Germany, respectively. 5. Actual implementation of these ideas took time. The 1949 Constitution vaguely referred to the principle, legislation on social housing came in 1950, on codetermination in 1951–1952, but a significant reform of the pension system had to wait until 1957. Still, even this came well before similar measures in either Italy or Japan. 6. Rough estimates suggest that Italy’s reform in the early 1950s transferred between 6 and 10 percent of total acreage from large to small farmers, of which less than half was expropriated (Graziani 1972; Ciocca 2007). In Japan, by contrast, close to 40 percent of total acreage was, de facto, confiscated in the late 1940s and transferred to the erstwhile tenants (Kawagoe 1993). 7. The words “miracolo economico” and “Wirtschaftswunder” regularly appear when the period is mentioned in the Italian and German literature. The Japanese, less given to hyperbolae, refer to it under the more neutral words of “kōdō seichō jidai,” or era of high economic growth.

132

aggregate growth and policy

8. Between 1953 and 1972 (just before the first oil shock), the ratio of export to import prices rose by 1½ and 2 percent per annum in Germany and Japan, respectively (Italy saw a more modest 0.3 percent annual improvement). 9. The data shown in the table cover only the formal educational system. This understates the achievements of Germany and Japan. In the former, a large share of the young obtains apprenticeships qualifications after leaving school; in the latter the same result is achieved by widespread provisions of on-the-job training. 10. Few, in any case, would doubt that Japan possessed at least some of the characteristics of “co-ordinated capitalism” that Eichengreen (2007, 3) considers crucial: “solidaristic trade unions, cohesive employers associations, and growth-minded governments.” 11. International Labor Organization data show that the number of working days lost in strikes in 1969–1970 averaged over 29 million in Italy, as against only 170,000 in Germany. 12. In Japan, gross internal migration into the richer areas averaged 1.5 percent of those regions’ population between the early 1950s and 1973; in Germany and Italy the figure was 0.8 percent (net migration from “poor” to “rich” was, however, smaller in Germany than in Italy; no comparable figures were found for Japan). 13. Real long-term bond rates (deflated by the ex post GDP deflator) averaged 4.0 and 3.3 percent for the period as a whole in Germany and Japan, respectively, but only 1.8 percent in Italy. They would probably have been even lower in Italy if a more appropriate “expected inflation” indicator had been used to deflate the nominal rates. 14. In the end, this was not so surprising. Earlier British and American experience had already shown that the effects on expectations of “regime changes” can be significant for financial markets but are hard to detect on labor markets. 15. It is true that it was burdened with the costs of reunification after 1991, but the growth shortfall was already more than one decade old. 16. On a per capita basis, France, Denmark, and Portugal are below Germany’s 1.4 percent annual growth, but are still way ahead of Japan’s and Italy’s dismal figures. 17. Thus, according to the World Bank, the share of high-tech goods in total exports in 1995–2010 was 22 percent in Japan and 14 percent in Germany, but only 7 percent in Italy. Investment goods (defined as SITC categories 71 to 77) represented 44 percent of Japanese sales abroad in 1995–2008, as opposed to 30 and 28 percent in Germany and Italy, respectively. 18. Imports are also shown in the chart because outsourcing figures strongly in Germany’s trade links with Eastern Europe. Indeed, outsourcing and the threat to jobs this represented are almost certainly a major explanation for why earnings in Germany have grown much more slowly than they have in Italy since the creation of monetary union. 19. There may well be some truth to the simple view that it is much easier for Siemens or Sumitomo to be in China than it is for Sassuolo (the capital of the Italian ceramic tile industry) given the multitude of firms of which it is composed. This being said, numerous firms in Italy’s clothing industrial districts, in particular, have outsourced some of their activities to Eastern Europe. 20. The only, small, exception is provided by Germany, which in the years 1995–2010 was somewhat more attractive than Greece, but still well below all the other OECD countries.

italy, germany, japan

133

21. The EU’s Eurobarometer surveys on entrepreneurship thus show that the percentage of people thinking of creating, or having created, a new enterprise is significantly higher among those aged twenty-five to thirty-nine years and, to a lesser extent, forty to fifty-four years, than it is among those over fifty-five. 22. International surveys of entrepreneurship, as those presented by the Global Entrepreneurship Monitor, show that entrepreneurial activity tends to be lower in countries with a high percentage of the elderly than it is in “younger” economies. 23. The four poorer ones (with GDP per capita below the West German average in both 1960 and 2008) are Niedersachsen, Rheinland-Pfalz, Saarland, and Schleswig-Holstein. 24. The Northern region is made up of Hokkaidō and Tōhoku; the Southern islands are Shikoku, Ky ūshū, and Okinawa. 25. Over the 1953–2008 years, for instance, annual per capita GDP growth would have been 3.7 rather than the recorded 3.4 percent, with the South itself growing at 4 instead of 3.1 percent. 26. Looking at the literature on the response of regional migration flows to changes in relative regional incomes (controlling for other variables, such as unemployment and distance), an earlier summary conclusion was that “most of the elasticities . . . have been greater than 1.0 but less than 3.0” (Gallaway, Rydén, Vedder 1975, 262). For present purposes, use was made of (a rounded-up) result obtained for Italy over the 1958–1974 years (1.8 for relative wages) (Salvatore 1977). 27. There are few reliable figures for international trade elasticities for the 1950s and 1960s. A well-known survey of the time could not find statistically significant results for Italy (Houthakker and Magee 1969). For eight other advanced countries for which such results were obtained, the average estimate was of the order of −1 (Houthakker and Magee 1969). 28. Wage moderation was one component of this decoupling. Rapid productivity growth in the tradeable sector was another, thanks in part to industrial policy and in part to a return to historical specialization patterns (Carlin 2010).

Chapter 5

THE ITALIAN ECONOMY SEEN FROM ABROAD marcello de cecco

Foreign economists, economic historians, and policy makers looked at Italy with keen eyes after it became a united country. They were interested to learn whether the economy of a country that had in the seventeenth and eighteenth centuries regressed to the role of agricultural raw material exporter after having been in the Middle Ages and the Renaissance the premier site of European industry, trade, and finance could redress itself and join the Industrial Revolution, making good use of its population and territory, which gave it the potential to be among the great powers of Europe. By what means this would happen and following which model was particularly interesting to foreign observers. Political unification was in itself regarded as a miracle, and other declining countries, or countries that had long smarted under foreign dominion, saw in Italy an example to emulate. Some adopted the Italian tricolor as their flag (e.g., Hungary and Ireland). Zionists were also keen students of the Italian experience, because the return of an ancient people after two thousand years of division as a potential modern great power augured well for an even older people who had for even longer been subjected to loss of land, dispersal, and persecution.

the italian economy seen from abroad

135

Natural and Unnatural Development Therefore, the greater part of the capital of every growing society is, first, directed to agriculture, afterwards to manufactures, and last of all to foreign commerce. This order of things is so very natural, that in every society that had any territory, it has always, I believe, been in some degree observed. Some of their lands must have been cultivated before any considerable towns could be established, and some sort of coarse industry of the manufacturing kind must have been carried on in those towns, before they could well think of employing themselves in foreign commerce. But though this natural order of things must have taken place in some degree in every such society, it has, in all the modern states of Europe, been, in many respects, entirely inverted. The foreign commerce of some of their cities has introduced all their finer manufactures, or such as were fit for distant trade; and manufactures and foreign commerce together, have given birth to the principal improvements of agriculture. The manners and customs which the nature of their original government introduced, and which remained after that government was greatly altered, necessarily forced them into this unnatural and retrograde order. (Smith 2009)

Adam Smith saw the Italian republics and states since the Middle Ages as archetypal examples of “this unnatural and retrograde mode of development.” Their example influenced all the countries of Europe. Smith was not surprised that the Italian city states had been exposed to volatility of fortune and fatal decline. He was at pains to show that their decline had not been brought about by the discovery of America. Italy’s decline had if anything preceded that event and Italy had somewhat recovered after the discovery of America, whereas Spain and Portugal declined. Smith believed that the small states of the Italian peninsula, which had experienced a meteoric rise in the early Middle Ages and fast decline afterward, had introduced an unnatural mode of development, avoiding what he considered the natural one. This natural mode did not depend on the vagaries and volatility of foreign commerce and plunder (the latter included in the category of primitive capital accumulation); rather, capital accumulation began in agriculture and proceeded to industry through reinvested savings and finally extended to foreign trade. Smith’s model of natural development had a profound impact on economists and politicians all over the world. Among them were foreign students of the Italian economy, writing after him, who having visited the peninsula or lived there for an extended period of time reflected on Italy’s decline. Smith also influenced the Italians who engineered Italy’s comeback as a political and economic power in the middle of the nineteenth century. The size of the country mattered to them, because it could permit a more stable and durable growth experiment, which ought to be based on modern agriculture affording a degree of capital accumulation sufficient to permit the installation of modern industry. Hence, their goal was to unite the entire country.

136

aggregate growth and policy

Thus, for the first seventeen years of the new Kingdom of Italy the governing elite (La Destra storica) applied Smith’s “natural development” model. For them, agricultural modernization came before industrialization.

Industry-Led Development or Balanced Growth? After the Destra storica lost power in 1876, Italian governing elites opted for industrialization at all costs as a necessary condition for increased international political power. Ricardo won the day against Adam Smith. However, in Italy perhaps a bit later than in other Continental European countries, Ricardian free trade gave way to protectionism as a necessary condition for industrialization. Not all foreign observers believed it was the right choice. British observers advised that Italy join free trade Europe. Richard Cobden had made, earlier on in the century, a very successful tour of Italy preaching just that. The Piedmontese adopted free trade before unification, especially to receive capital and direct investments from France and Britain, and then extended it to the heavily protected South after unification. Cavour, ever the great realist, did not declare unilateral free trade. Rather, he signed a large number of trade treaties with the principal powers with which Italy and before it, the preunification states, had or hoped to have important commercial relations. The Southern Kingdom, a couple of decades after removal of the Bourbon King, had chosen the model of American protectionism, preached by Alexander Hamilton and imitated in the German states by Friederich List’s Zollverein. This caused the enmity of the British. As the Florentine and Genoese had done in previous centuries, their merchants had thoroughly exploited all the potentialities the Southern Kingdom offered in the way of imports and exports. In particular, British merchants had virtually monopolized the export of sulphur, which was extracted by primitive methods and with much human suffering in the mines owned by Sicilian noblemen (Giura 1973; Davis 1982). Sulphuric acid was considered essential for industrialization and its price went through a veritable bubble in the early 1830s. After a few years, the bubble burst and the price declined as fast as it had risen, causing intense economic and social distress. Following its protectionist strategy, which had been introduced in 1823, the Neapolitan government in 1838 chose to stabilize the slumping price of sulphur by restricting exports, limiting them by one-third and auctioning the rights to the Sicilian sulphur mines, taking them away from de facto British control, to a newly formed monopoly run by a French company (in which the Duchess of Berry, sister of the Neapolitan King, was rumored to have an interest). The British government reacted, after representations were made by British merchants to their government, by a very open threat of naval intervention,

the italian economy seen from abroad

137

made by the foreign minister Palmerston. This led to the actual seizing of Neapolitan merchant vessels in the open seas. The vessels were taken to Malta until the dispute was resolved. Other ships under the Neapolitan flag were blockaded in Corfu and the commander of the British fleet was told to be ready to blockade the Gulf of Naples. This act of open imperialist bullying, an almost perfect case of gunboat diplomacy, was ordered in disregard of the advice of the British crown’s legal experts. It was, however, presented by Palmerston under the guise of a doctrinaire dispute between protectionists and free traders. The “objective principles of modern economic science” were wheeled out in defence of the status quo with the sulphur mines in the letter the British foreign secretary wrote to the Neapolitan King, after the latter had claimed his sovereign right to impose a monopoly on an economic activity conducted in his domains and of treating foreigners in a manner different from his own subjects. In countries where Government is arbitrary and despotic—wrote Palmerston— and subject to no responsibility or control, it may often happen that caprice, want of political knowledge, prejudice, private interest, or undue influence, may procure the promulgation of unjust and impolitic edicts, inflicting much injury upon the people of such state, interfering with the legitimate industry of individuals, deranging the natural transactions of commerce and causing great detriment to private interests and to national prosperity. (Giura 1973)

Palmerston then proceeded to draw an essential difference between a state monopoly and one granted by a sovereign to an individual. No doubt royal monopolies—he wrote—are . . . very objectionable ways of raising a revenue. They interfere injuriously with private enterprise, prevent the full development of the natural commercial resources of the Nation, and check the consequent increase of the public revenue; but in all the countries where the science of Political Economy has been imperfectly understood, such monopolies have constituted one of the sources of income for the Crown. (Giura 1973)

However scientifically and practically wrong they might be, he noted, state monopolies were not prohibited under the treaty of commerce signed between Naples and Great Britain in 1816. However, the case of sulphur was different, he affirmed, because the sulphur trade had been completely private from the start on both trading sides, and the imposition of a monopolistic restriction was a serious prejudice to the rights of British merchants and Her Majesty’s government had every right to come to their help. Palmerston made an additional didactic remark to his opponent: the Neapolitan Government “appears to imagine that sulphur is an article found only in Sicily, and that Providence has rendered all mankind dependent on that single island for a commodity which is extensively required for various uses” (Giura 1973). Sulphur, on the contrary, is to be found in many parts of the world and the persistence of the Sicilian monopoly, Palmerston affirmed, will bring other supplies into cultivation and the price will fall. When this result is

138

aggregate growth and policy accomplished—he concluded- the Government of Naples may perhaps regret that it allowed itself to listen to a scheme suggested by individual cupidity, which can only be carried into effect by sacrificing the interests of the many to the avarice of the few, which violates the national faith, and which must involve the Crown of Naples in a difference with a Power, whose fleet and armies have protected and preserved for that Crown the very island where the subjects of that Power are now about to be exposed to treatment oppressive and unjust. (Giura 1973)

Thus, Palmerston showed that the Neapolitan Crown was scientifically backward and incorrect, politically offensive, ungrateful, and would also be easily proved to be militarily weak compared with Britain. The Neapolitan response was fast in coming, in a letter personally dictated and signed by King Ferdinand. There, the king showed his ire at being defined as politically despotic and ignorant of economic theory. He wrote that his country did not take lessons in political economy from anyone, least of all England, because it had been the birthplace of such economists as Genovesi, Galiani, and Broggia, all earlier and just as great as the most eminent in the world. He stated flatly that Palmerston was also wrong in his arguments about monopoly, and that he had been contradicted even by his own legal advisers. The ebullient and arrogant foreign secretary, however, was not to be deterred by this display of theoretical and argumentative prowess. He also had a personal axe to grind with the Neapolitan Crown, because Carlo di Borbone, a cadet brother of the king, had married Penelope Smith, a beautiful Irish girl who was Palmerston’s niece but did not belong to the aristocracy. For marrying beneath his rank, Carlo had promptly been disowned by his brother and left without means of support, to fend for himself and his wife in London. Palmerston, fired by political, ideologic, and personal motivations, ordered the British fleet to cruise out of Malta and seize all Neapolitan merchant vessels they found; to seize them also in Mediterranean ports, such as Corfu; and finally to appear in all its persuasive strength in the Gulf of Naples. In private letters Palmerston also mentioned his determination to blockade the Kingdom of Naples altogether from the sea. The Neapolitan government appealed to other Italian princes, to the continental powers of the Congress of Vienna, and to France. It received general sympathy, and expressions of readiness to mediate, but not much else. In the end, mediation managed to bring the dispute to a lengthy resolution. As J. Goodwin, British consul in Naples, noted at the time, the British had made sure that the Bourbon king would take theoretical free trade lectures seriously: Reprisals were made upon Neapolitan commerce by the British fleet in the Mediterranean, the result of which was that the (monopoly) contract was dissolved by a Royal decree of 1st July, 1840. (Goodwin 1839)

After this ideologic-military spat, the Bourbon dynasty never recovered the special relationship it had enjoyed with London. The Kingdom of Naples

the italian economy seen from abroad

139

became, for the British, the scourge of Europe, as William Gladstone famously called it in 1850. The Neapolitans were utterly defeated in their attempt to behave as a European power of the first rank. In economic terms, the defeat was less brutal, but the political implications of this demonstration of British imperialism were profound, because no other Italian state after that tried to go the protectionist route. In Piedmont, Cavour embraced free trade. After his death in 1861, his political heirs extended the free trade regime to united Italy, including the southern provinces where protection had lingered. Effects were devastating, especially for the South, also induced by the new kingdom’s choice of a monetary standard that overvalued the lira in the hope of attracting foreign capital (see Chapter 2 by James and O’Rourke on Italy’s postunification monetary policies).

Werner Sombart on Italian Commercial Policy A very lucid analysis of Italy’s commercial policy since unification was provided in the early 1890s by Werner Sombart (1896). In his view, the new kingdom had chosen the wrong path to development by extending Piedmont’s free trade to the whole of the country. Sombart also believed that had Cavour lived to govern the country he had so cleverly put together, he would not have done that because he knew that what was a good policy choice for a small country, such as Piedmont, would not be appropriate to a large and uneven one, such as united Italy. Sombart provided statistics on the level of protection that existed in the various kingdoms before unification, which clearly showed that the Bourbon kingdom was the most protected one. He remarked that the small Italian states before unification could not adopt a custom’s union, such as the German Zollverein, because it would have meant giving economic leadership over the whole Italian peninsula to the Austro-Hungarian government, thus going against all the Risorgimento was about. He provided statistics to show how the economy of the South had wilted in the years after unification, being subjected to the harsh winds of free trade, and how the entire country’s imports had swollen in the same years, while exports remained stagnant. He noted that in 1860 Italy had some industrial plants but could not be said to have an industrial sector, even in its most developed parts. He added that the shift to protectionism in the 1870s was gradual and followed similar policy changes that were decided by neighboring countries, in particular by France. He very perceptively remarked how close had been the integration between the economies of the two “Latin sisters,” and how economically painful would be for Italy the commercial war declared against

140

aggregate growth and policy

France in 1887, how Italy was led into it by France’s decision to abandon the free trade adopted by Napoleon III, and how Italy’s switch was also influenced by foreign politics. Sombart made an equally perceptive set of remarks regarding Italy’s decision to industrialize under the Sinistra governments. He was convinced that the Italians had adopted industrialization as their main policy goal to affirm their national will to be a great power, in an age when industry and modernity seemed to go hand in hand and all leading countries had become industrialized. This prevalence of politics over economics seemed altogether reasonable to him. He believed that if a country stuck to “natural” economic activities dictated by geography and natural endowments, it would lose control over its own destiny, and be exposed to the vagaries of the international cycle, but that once that path was abandoned, protectionist policies required very careful planning and could easily misfire. Accordingly, he noted that the earliest duties had been imposed on foreign trade by the Italian government not to promote industry or agriculture, but mainly to boost fiscal revenues in a country where taxation was difficult to impose and manage. When Sombart wrote his article the great industrial spurt of the late 1890s and early twentieth century had not yet begun in Italy; he dealt with the industrialization of the 1880s. He did not have any objection to the Italian decision to provide heavy protection for the steel industry. It seemed clear to him that no industrialization effort could take place in a large country as Italy that was not based on national steel production, even if the country did not have coal. He expressed admiration for the steel works that were built by Stefano Breda in Terni. To him this was evidence that Italians could realize large-scale, state-of-the-art, projects in leading industrial sectors. However, Sombart was too good an economist not to know that the lira’s return to convertibility brought about by the Sinistra government in 1882 had negative consequences on the competitiveness of the Italian economy, particularly of the newly born industry. He noted that devaluations worked like protective measures and that their efficacy depended on domestic prices rising earlier and faster than domestic wages, shifting income distribution in favor of capital and against labor. Smith, Palmerston, and Sombart are excellent examples of the mix of ideology, high-level economic theory, and more down-to-earth considerations with which foreigners have traditionally approached the reality of modern Italy. Examples can be found in Gerschenkron’s crossing of swords with Rosario Romeo on primitive accumulation, in the role of German banks in Italian industrialization, and in the misconceived protection of basic industries; in Vera Lutz’s polemic about the Italian government’s Southern industrialization policy; in Andrew Shonfield’s and Posner and Woolf’s admiration for Italian publicly owned industry; and in the more recent debate about small-scale industry and industrial districts inspired by Brusco and Becattini but made popular among world economists by Piore and Sabel and even by a political science analysis of

the italian economy seen from abroad

141

North-South difference as that proposed by Robert Putnam. All are looking at different facets of the same subject, as discussed next.

Bonnefon Craponne: A Contemporary Witness of Italian Turn-ofthe-Century “Big Push” After an early belief shared by many foreigners and Italians that Southern soil hid fabulous riches and that enlightened and efficient administration would bring them to fruition, the realization of Southern backwardness and of the scarcity of resources hit foreign observers as it did Italians, and advice on how to resolve the arch problem posed by Italian unification would be forthcoming from abroad for the next 150 years. In the eyes of foreigners, the coincidence of a rapidly modernizing North and an increasingly backward South made the Italian growth experience a sonderweg, a special case. Many foreign observers thought it could be solved if the Southern regions specialized in agriculture and tourism. After a much bigger case of economic renaissance in Germany, foreigners began considering whether integration with the German Empire may be better for Northern Italy than special relations with France, which had been the experience of the previous two hundred years. Northern Italy was the region involved because it was clear that there the Italian elites would conduct their experiment with rapid industrialization. Still, there was no widespread contemporary awareness by foreigners of the success of the Italian industrial spurt in the two decades leading to World War I. After the war, keener observers took note of what had happened but the prewar and wartime growth of Italy’s new industrial structure was overshadowed by the more urgent problems posed by Italy’s huge foreign indebtedness and by the need to reconvert the new industrial structure to postwar world economic conditions. There were exceptions. Perhaps the most interesting among prewar observers was Louis Bonnefon Craponne, a French silk producer based in Piedmont, who was the first chairman of the Unione Industriale of Torino and later of the Confederation of Italian Industry. He published in Paris, in 1916, what can be considered perhaps the best contemporary work, L’Italie au Travail, a book written before the outbreak of World War I (Bonnefon Craponne 1916, 1991). He was highly competent; a graduate of one of the elite Paris grandes écoles, the Haute Ecole en Sciences Commerciales; a true lover of his adopted country; and a not uncritical admirer of what the Italians had managed to do especially in the ten years of what is now call the Giolittian economic miracle. His book is an extremely lively, well-documented, and statistically backed description of the

142

aggregate growth and policy

growth trajectory that some of the most successful sectors of Italian industry had experienced since unification. It includes a very thoughtful chapter on la Questione meridionale, clearly inspired by a man the author thought very highly of, Francesco Saverio Nitti. He was, however, able to distance himself from Nitti’s passionate approach to the subject. Bonnefon started as a silk producer, closely connected to the Lyon silk market, but then moved on to take part in the explosive boom of the Italian automobile industry in the first decade of the twentieth century. Naturally, his chapters on silk cocoon production, the silk industry, and the auto industry are particularly brilliant, but he is extremely well informed and informative on other sectors, such as the hydroelectric power industry. It is only through the brilliant pen of this French grand commis that one gets a pulsating feeling of what must have been the heady experience of fast growth in Northern Italy in the first decade of last century. He quotes production levels in the most dynamic sectors, giving a body to those figures through extremely competent descriptions of the production processes adopted, and of the technology levels at which Italian industry was able to rise in such a short time in such sectors as automobiles, tires and cables, hydroelectric power, and cotton and wool. In his descriptions one is able to fully appreciate the modernization of Northern Italy, its growing distance from Southern Italy, the essential role played by emigrant remittances in the Northern miracle, the problems with the young and militant workers’ unions, and the pitfalls to be avoided. Bonnefon also fully understood the dualism between large- and small-scale industry and vividly describes the self-made former workers, the well-educated engineers, or upper class former cavalry officers, such as Giovanni Agnelli, both types of industrialists sometimes even coexisting in the same sectors. He did not overlook the workers, whose recent provenance from the peasantry he stressed, because of the advantages and disadvantages that rapid social change entails, which helped explain why such hardworking people in the end were less productive than their French, British, and German counterparts. Bonnefon was a positivist living in an age of positivists, in a city where that persuasion was rife. His book, in the original French edition, is difficult to find, although an incomplete copy can be downloaded from the Internet, offered by a Canadian university (http://www.archive.org/details/litalieautravail00bonn). A new French anastatic reprint has recently come out and an excellent translation into Italian was made and published in 1991 by the Unione Industriale of Torino, but was not commercially distributed and is difficult to find. For these reasons, Bonnefon Craponne is hardly a household name in Italy today. He certainly was not mentioned by either author in the Gerschenkron–Romeo debate. He does not appear in Risorgimento e Capitalismo (Romeo 1959) or in Economic Backwardness in Historical Perspective (Gerschenkron 1962). Nor is he mentioned in the index appended to an important collection of essays put together by Caracciolo (1969). More recent works, such as Castronovo’s history of FIAT (2005), make full use of the book. Bonnefon’s book, however, is today known only to specialist

the italian economy seen from abroad

143

historians, although it ought to have been among those suggested to university students of history, economics, and engineering in the year of celebration of the 150th anniversary of the birth of modern Italy. His down-to-earth, yet vivid and scientifically cognizant prose has not aged at all in a century. Reading it calls back the enthusiasm of a decade essential to Italian economic and political history.

The Economic Recovery Administration’s Country Study on Italy and Its Consequences on Theory and Policy Making What should one make of the Italian mixed economy, with large-scale industry chiefly owned and run by the State? This question occupied foreign students of the Italian economy after World War II, especially during and after the Italian economic miracle. Before that, however, foreign advisers were involved in plans for the reconstruction of the Italian economy after the war and its adaptation to the new free trade environment created by the Bretton Woods Agreements. The launch of the Marshall Plan and its application to Italy gave rise to a famous dispute between the young Keynesian American administrators of the Marshall Plan and the Italian authorities, led by the central bank and its then governor and future president of the Republic, the distinguished economist Luigi Einaudi. He thought foreign financial aid should be destined to help stabilize the Italian currency and not to fund Keynesian public works to alleviate Northern unemployment, which according to him had structural causes. After the Marshall Plan administrators published a “country study” (Economic Cooperation Administration 1949) on Italy that was sharply critical of Italian poststabilization policy choices, Italian economic policy makers cleverly managed to mobilize conservative foreign economists, such as Friederich and Vera Lutz, who agreed with the Italian government’s set of priorities and policy sequencing. The report by the young American administrators of the Marshall Plan titled “Italy-Country Study,” which appeared in February 1949 but was written in the last months of 1948, was slender in size but it proved to be the first mover in the postwar debate about what models of economic development ought to be applied to the case of Italy, the ideal starting point of the Gerschenkron–Romeo debate. It has been condescendingly called a product of “rough Keynesianism.” It certainly was Keynesian to the core, but rough it was not. The appearance of this report was traumatic for the Italian governing elites who had, starting from the authoritative advice and policy making activities of Einaudi, governor of the central bank and budget minister in De Gasperi’s

144

aggregate growth and policy

cabinet, devised a very different policy sequence to follow the just achieved stabilization of the lira. De Gasperi and Einaudi had bet on the positive electoral consequences of their stopping in its tracks the inflation that had burned Italian monetary savings and the real value of the huge public debt inherited from Fascist war making. They won their bet. There were more people in favor of the deflation that followed than there were people who attributed their being unemployed to it. The fact that inflation had destroyed the real value of cash earned through black market sales of all sorts of goods did not inflame farmers and the stock exchange and commodity speculators into voting against the government, and certainly did not persuade them to vote for the Left, the true fear of the Americans. The authors of the Country Study reported that inflation had been initiated by the Italian authorities, when they had allowed banks to sell the holdings of government debt they had been forced to absorb during the war and to lend the proceeds to their commercial and industrial clientele. The authorities, especially Treasury minister Epicarmo Corbino, an extremely obdurate economist who believed in free competition whatever the cost, had repealed all rules that prevented banks from opening new branches and competing for clients, with the result that industrial production and imports grew because of pent up demand for goods but also because of speculative inventory accumulation. A wild stock exchange boom also ensued, another product of easy credit. When Corbino’s influence was finally neutralized, open inflation was rampant and it took Einaudi, then Governor of the Bank of Italy, and Menichella, former Director of the Institute for Industrial Production (IRI) and then Director General of the Central Bank, to put up a joint effort to fight it back. Stabilization of the lira was achieved at the cost of deflation, a deflation that was only mitigated by a large devaluation of the Italian currency and by special government intervention to rescue the engineering industry with subsidies. All these aspects were mentioned in two brilliant articles by Albert Hirschman (1948a, 1948b) written just after the stabilization had been successfully performed. Hirschman had earned an economics degree in Trieste just before the war and had deep knowledge of the Italian economy. He worked for the Board of Governors of the Federal Reserve, and as a result was careful with words, but neither did he mince them. The authors of the Country Study probably read one of the articles because it appeared in the American Economic Review. Hirschman wrote his papers after having spent time in Italy to collect information. There is a good chance that he had met his Marshall Plan colleagues. The Country Study was written in the same vein as Hirschman’s articles. He had underlined the importance of devaluation of the lira for the success of stabilization. He emphasized the importance of “real” variables, such as commodity prices, in dampening price rises in Italy after the autumn of 1947. He also underlined the partial freeing of exporters’ foreign exchange earnings, which had been used to import raw materials and other essential products, rather than being parked in foreign bank accounts.

the italian economy seen from abroad

145

The authors of the Country Study irked Italian policy makers and conservative economists when they began their report by quoting the very high number of unemployed people who lived in the cities of the industrial North of Italy and by affirming that increasing employment was the primary aim of economic policy. They thought that aim could not be achieved if the deflation after stabilization was allowed to persist. A vigorous reflationary policy was mandatory, also to allay permanently the threat of Communism. They added that a massive dose of public investment was the only way to bring this about, to substitute for weak consumption demand to be expected because of low wages and unemployment and especially for the extremely timid performance of private investment after the inflationary bubble had been successfully punctured by the authorities’ fast and determinate measures of mid-1947. Private entrepreneurs, wrote the Country Study, seemed not to have gained, after the resounding Christian Democrats electoral victory of April 18, 1948, enough confidence that Communist takeover was definitively out of the picture to take out their money and invest it in new productive capacity. In any case, private industry thought capacity utilization was still low because of the deflation after stabilization and that there were enough unused plants and machinery, not to mention workers, to supply almost any increase in demand in the immediate future. Public investment, the Country Study wrote in no uncertain tones, was the only way to decrease present unemployment in the North and to increase productive capacity in the less immediate future, and to avoid bottlenecks that would appear in crucial and strategically important industrial sectors. Its authors also wrote that public investment could be financed by the ample funds made available by Marshall Plan aid and accumulated in the special treasury counterpart account, where the proceeds of the sales of Marshall Plan aid goods were deposited. It was for this, they added, that the ECA program had been devised. That is where the postwar dispute on models of development and capital accumulation began, much earlier than it would be taken up by academic economists. The Keynesian economists of the ECA had in mind a progenitor of what would later come to be known as the Keynesian growth model, very similar to what Gerschenkron seemed to have in mind when he wrote his articles some years later. Public investment to increase capacity and employ workers would be at its core. In the immediate postwar years, however, no primitive accumulation was required to realize it, unless Marshall Plan aid could be defined as primitive accumulation. Foreign aid replaced credit creation by German-style banks that no longer existed in Italy or extraction of agricultural surplus through taxation of rich landowners or through relative price movements between industrial and agricultural goods, which penalized the latter. The authors of the Country Study also wrote that, to boost fixed investment in infrastructure and new industrial capacity, publicly owned industry (whose size had been greatly increased in the 1930s through the creation of IRI to rescue large-scale industry and banks that owned it) should be called to play

146

aggregate growth and policy

the protagonist’s role, because IRI owned most of the firms capable of planning and executing large-scale investment projects, to be financed with ECA funds. The Country Study remarked that IRI had not yet begun to plan a coordinated industrial effort, such as the one now necessary for Italy to move to a fast growth trajectory, but that they ought to be asked to do it now that resources to finance that giant effort were made available by the ECA. They thought no one in private industry had the resources and the size to replace IRI in this role. The Italian authorities, however, did not agree with the course traced by the ECA economists in the Country Study. They much preferred to use the counterpart funds to increase gold and foreign exchange reserve, to give once again a solid base to the Italian monetary system. Some of them also believed that Italian unemployment was not caused by a deflation-induced fall in effective demand. They thought it was structural and could not be reabsorbed by a rapidly outlived demand boost to be obtained by increasing industrial capacity through fixed investment and infrastructure building funded by the state (e.g., see Eichengreen 2007). They also feared that inflation was not permanently subdued and would be resurrected by a sudden increase in international commodity prices, leading to inventory speculation by Italian industrialists. In the following decade, the years of Italy’s economic miracle, events showed that the authors of the Country Study and the Italian economic authorities were both right. There had been Keynesian unemployment in Northern cities and disguised unemployment in agriculture all across Italy. The Korean War spurred another bout of speculative inventory accumulation by Italian firms and merchants, and the foreign reserves the Italian authorities had prudently piled up could be used to balance the country’s international accounts. The war also boosted foreign demand for Italian industrial goods, and that caused employment and capacity use to rise without unduly pressuring wages. The low level of the lira made Italian industry fully competitive with foreign producers. At the same time, IRI firmly placed itself at the core of the new Italian growth trajectory. It managed to close the historic gap that Italy showed in integral-cycle steel production by building, against the violent hostility of private steel makers who used scrap iron in electric steel mills, a new steel plant in Genoa with American machinery supplied by the Marshall Plan. American refinery equipment was also obtained to build new oil refineries and reequip existing ones, thus helping to close another historic trade gap, that caused by coal imports. Higher-value refined oil products were exported and low-value ones were burnt as fuel in the thermoelectric plants also imported from America. Moreover, IRI engaged in a huge program of motorway construction. ENI, the other large publicly owned firm, found natural gas in the Po Valley and built a complete pipeline network to make it available to Italian industry and reduce the fuel bill Italy paid to foreigners. Thus, often the same men who had played down the advice of the young authors of the Country Study put the content of that advice into practice. At

the italian economy seen from abroad

147

the end of the decade, Gardner Ackley coyly remarked that there must have been plenty of Keynesian unemployment in the late 1940s to allow such a large and rapid supply increase in the 1950s, while millions of Italian workers had at the same time been migrating to Europe, the Americas, and Australia (Ackley 1957). Full employment would finally occur in the North of the country in the early 1960s, rapidly pushing up wages and causing a balance of payments crisis that Italy, now on the Bretton Woods fixed exchange rate standard, could not meet by devaluation. Fodor (2000) and Cavalieri (2009) provide an analysis of the monetary assistance offered by the US administration to the Bank of Italy during the 1963–1964 balance of payments crisis.

Dualism in Economic Development In the 1970s, David Abulafia, a distinguished Cambridge economic historian, published his work on the Italian economy of the late Middle Ages, where he showed that the dualism of the Italian economy, with a developed Center-North and an underdeveloped South, could be traced back as far as the fourteenth century (Abulafia 1977). This was not as far-fetched as what other scholars had done before him, to show that dualism in the Italian economy dated back to classical times. Arnold Toynbee, the most radical among them, had written that the underdevelopment of Calabria was to be attributed to the revenge the Romans had taken on its autoctonous inhabitants and the devastation they had wrought on their homeland, after the Calabrians had sided with Hannibal (Toynbee 1965). Dualism, or uneven development, is thus a concept that permeates Italian economic and social history. It reemerged in the 1950s, when a vigorous growth process finally started in the Northern part of the country, whereas the South stagnated and even regressed. In this period the Italian authorities asked for foreign advice when Italy requested United Nations help to develop the South. Paul Rosenstein Rodan, Hollis Chenery, and Richard Eckaus were consulted (Eckaus 1961). That attracted foreign interest much more than what was happening to the industrialized North in the new trade liberalization environment. The Northern “miracle” took foreigners largely by surprise, hence their interest in the mixed economy model, which had prevailed in Italy since the rescue of banks and industry in the 1930s. This model had admirers, such as Andrew Shonfield and Michael Posner, and serious critics, such as Friederich and Vera Lutz. The debate began with an article published by Vera Smith, a conservative English economist married to the German economist Friederich Lutz (Lutz 1956). The couple started visiting Italy for extended periods, at the invitation of the Bank of Italy. The purpose pursued by the Italian authorities in inviting them

148

aggregate growth and policy

to work on the Italian economy was reported by Paolo Baffi in a well-known article (Baffi 1985). In the first article, in another published a little later (Lutz 1958), and in a subsequent book (Lutz 1962), Vera Lutz built a model of the Italian economy characterized by some pathologic features, which in her opinion caused the divergence between North and the South to come about and to persist in time. Her articles coincided with the preliminary work for and then the launching of the Cassa per il Mezzogiorno, an institution created by Donato Menichella and inspired by US President Roosevelt’s Tennessee Valley Authority, and supplied with large resources to build public works and infrastructures in the Southern regions with the declared aim of bridging the gulf that existed between economic conditions in the two parts of Italy. In the same years, the Vanoni Plan was also launched, with the aim of speeding up growth rates in the whole of Italy but in particular in the more backward regions. Foreign economists were recruited to help with both efforts. Vera Lutz’s articles had the aim of inserting doubts into the minds of the Italian governing elites, which were swinging in favor of both intervention plans, as to the soundness of their theoretical foundations. They started from classical and Keynesian premises. She, on the contrary, based her reasoning on strictly neoclassical principles. According to her analysis, the glaring dualism that persisted between North and South as far as levels and rates of growth were concerned was the result of one simple fact: large-scale industry, mostly located in the North, was dominated by powerful unions that imposed much higher wage rates to the entrepreneurs than the relative availability of factors of production in Italy justified. This artificial wage setting, dominated by a strong monopoly of labor resources exercised by the unions, compelled firms to use more capital-intensive techniques than the ones labor availability in the economy would have suggested, thus inducing the economy to warrant a less than full employment demand for labor. Large-scale industry, as a result, did not grow as fast as it might have, and labor supply was partly employed by small-scale firms, which were allowed, by the relative weakness or even absence of unions from their shop floors, to pay much lower wages and ensure much worse working conditions than those prevailing in large-scale industry. According to Vera Lutz, because of the presence of powerful unions the Italian economy was therefore suffering from an inadequate growth rate; a dualistic wage structure; and a highly dualistic economic structure where very capital-intensive big industrial plants were surrounded by a plethora of small firms, where big industry paid higher wages than they should have, where labor conditions were very different in the two sectors, and where small-scale firms were concentrated in the Southern regions and in nonindustrial sectors. In addition, greater industrial concentration than would have been justified by the labor supply of the country was another unfortunate result of the excess wage

the italian economy seen from abroad

149

rates secured by the unionized labor force of the big industry sector. This led to monopoly and excess profits. Monopolistic wage-setting in the large-scale industry sector thus led, according to Vera Lutz, to monopolistic prices and behavior on the part of the entrepreneurs in the same sector. Resources were siphoned off from the rest of the economy and starved it of necessary capital. How to get out of this dangerous impasse in which the Italian economy had been cornered by too strong unions? According to Mrs. Lutz, the answer was simple: wages in the capital-intensive sector had to be kept from rising long enough to allow the relative reduction of capital intensity in that sector and an increase of its growth rate so that the sector would start absorbing labor again and thus reduce the size of the backward small firm sector. On the contrary, if wages resumed their rise in the capital-intensive sector, the latter’s capital intensity would grow further and the growth rate of the whole economy would slow down, and advantages of capital intensity and of high wages would be enjoyed only by capitalists and workers in the high-wage, capital-intensive sector. Geographic and sectoral dualism, far from declining, would probably increase further. It is clear that Vera Lutz’s analysis had profound implications for the growth strategy the Italian leadership was launching, aided by very eminent foreign advisors (e.g., Paul Rosenstein Rodan, Hollis Chenery, and Harvey Leibenstein) and with the help of large soft loans from the World Bank. Had her line of reasoning been accepted, the first move on the part of Italian politicians should have been to try and break the hold the unions had on wages in the capital-intensive sector. Without this preliminary action, all attempts to increase employment levels by public works and direct investment by state industry were destined to have only a small and short-lived success. If one adds that, according to Mrs. Lutz, a dualistic economy would also work to increase agricultural prices and depress industrial prices, Southern excess supply of labor could only be relieved by mass emigration to the North, investment in employment-creating public works, and even directly in new industries in the South being largely useless. It is thus not surprising to see that Vera Lutz’s articles were subjected to extremely heavy criticism on the part of economists who did not accept neoclassical theory. The first line of criticism was taken by Luigi Spaventa, then a very young and brilliant Cambridge-trained economist, alone or writing with Luigi Pasinetti, who also rose to international eminence in later years and with the mentioned Gardner Ackley, an American Keynesian economist, who had good knowledge of the Italian economy and who would later serve as US ambassador in Rome (Spaventa and Pasinetti 1960; Spaventa and Ackley 1962). Spaventa and Ackley replaced Lutz’s analysis with one based on classical Keynesian theoretical premises, to which they also added the results of Sylos Labini’s work on oligopoly (Sylos Labini 1956). Their model yielded results opposite to those of Lutz: it was oligopoly in the modern part of the economy that reduced the growth rate of the whole economy and maintained divergence

150

aggregate growth and policy

between modern and backward parts. A reduction of oligopolistic power (i.e., an increase in competition in the modern part of the economy) was thus required to raise the aggregate growth rate and reduce dualism. Other foreign economists, such as Chenery and Leibenstein and especially Kindleberger, also contributed to the study of Italian dualism. Kindleberger (1966) used Arthur Lewis’s analysis of growth with unlimited labor supply. He attributed to the exhaustion of excess labor supplies the end of the so-called supergrowth the Italian economy had shown in the late 1950s and until 1963. Sudden and very large wage increases had stopped the virtuous circle of Italian postwar development. His work was criticized by Italian economists who noted, for instance, that in 1963 full employment had been achieved only in Northern Italy, whereas unemployment persisted in the South. Eminent foreign economists were thus working in the 1950s and early 1960s to explain the Italian economy’s persistent dualism. It was Mrs. Lutz who started the debate and drove the discussions. In retrospect her line of argument was the one that had the strongest influence on Italian public debate for years to come, despite the many theoretical weaknesses of her basic model. Double wage standards in different areas of the country and the break-up of union power continue to be suggested as solutions to the dualism by which Italy is still affected 150 years after unification. Finally, belonging to this season of foreign publications on Italy, one cannot but recall also the impressive study on Italian public enterprises by Posner and Woolf (1967).

Gerschenkron and Romeo on Primitive Accumulation and the “Big Push” Needless to say, many foreign economic historians and contemporary observers studied the “Italian case.” In particular, one of the most relevant contributions was made by Alexander Gerschenkron. While the “dualism in economic development” debate raged, an equally heated one involved Gerschenkron and another eminent economic historian, Rosario Romeo. They were soon joined by other Italian and foreign economic historians and economists. The debate focused on the role of capital accumulation in the crucial phase of economic development, which at the time was referred to as the “big push,” a strong acceleration in the rate of growth of an underdeveloped country that would finally allow it to break away from underdevelopment and take a fast growth path. More Stalinist than the Marxian economists whose views he challenged, though by no means a Marxist himself, Romeo advanced the opinion that in the first fifteen years after Italian unification, the Destra storica had adopted the correct approach in its attempt to bring Italy among developed countries

the italian economy seen from abroad

151

in the shortest possible time (Romeo 1959). In his view, the politicians who had replaced the Destra after 1876 had pursued the same strategy, at the core of which was the extraction of investible surplus from agriculture to shift it to infrastructure building and then to industry. This was a direct criticism of Antonio Gramsci’s view of the failed agricultural revolution as the key to understanding why united Italy had not experienced a bourgeois revolution (Gramsci 2007). By not distributing land to peasants, Gramsci noted, demand had not been created for industrial goods; hence, slow industrial development and the very slow and timid emergence of a modern bourgeoisie. Romeo’s view was that it had been exactly the failure to redistribute land that had allowed the government to extract the surplus from agriculture. Who could have kept poor peasants who had come into land ownership from eating up the product they were able to produce themselves? By keeping land in the hands of the landlord class, the Italian government had been able to tax landlords heavily, because agricultural wages had not been allowed to grow. With some justification, Romeo accused the Italian Marxists of not having read enough classical political economy, to have directly shifted loyalties from prewar idealism to Marxism for purely political reasons (hatred of the Fascist dictatorship that had plagued Italy for twenty years). In his opinion, “primitive accumulation,” a favorite expression of Italian Marxists, had to be interpreted as the extraction of existing resources from a stagnant sector to invest them in a dynamic sector, such as industry, or in indispensable public works, such as the building of a railway network. The latter course was the one chosen by the Destra storica and it was a first indispensable step leading to a strategy of industrialization. Romeo was clearly inspired by the literature on growth theory and growth models developed in those years in British and American universities by such people as Joan Robinson, Ragnar Nurkse, and Arthur Lewis. He referred in particular to the articles on Italy Gerschenkron had published in English in Italian journals (Gerschenkron 1962). Gerschenkron’s reaction to this peculiar use of his work was ambivalence. He praised Romeo and criticized his Marxist opponents, but he also reaffirmed what he considered his more important views of the Italian experience with economic development. He thought that railway building, much as it had been useful to unite the country, had been very weak as a factor of development promotion, because most of the track and rolling stock had been imported. He thus believed that a meaningful growth strategy was that initiated in the 1890s, after the great banking crisis, by replacing the failed French-type banks with German banks, which had exported to Italy their banking model and directly promoted industrial investment, as they had done in Germany. In Italy they had not managed, however, to dispose of their industrial investments, recovering their financial resources for another round of investment because of the shallowness of the Italian stock exchange. Gerschenkron also affirmed that the industrialization policy of the 1880s, which had directly promoted the steel

152

aggregate growth and policy

industry by giving it huge resources and heavy protection, much as it might have promoted Italy’s status as a great power, from the economic point of view had been mostly a waste of resources, because it had penalized the much more promising engineering industry, which had been ineffectively protected and was made to bear the costs of expensive homemade iron and steel products. Gerschenkron also criticized the choice made in the 1880s to protect the textile industry, which was hardly an innovative industrial sector. Romeo had used a partial and lopsided version of the Gerschenkron view, one that did not accept Gerschenkron’s criticism of the development strategy adopted by the Sinistra governments in the 1880s. Gerschenkron told him as much in a debate they had in Rome and in a review of Romeo’s book (Caracciolo 1969). Unlike Romeo, he had no interest in defending the pre-Fascist Italian governing class from Marxist attacks. He based his reasoning on his own statistics of Italian industrial value added, which he rated as better than the ones constructed by the Italian national statistical institute. Gerschenkron’s series were then revised and improved by his Italian pupil, Stefano Fenoaltea, and in turn used by Angus Maddison (Maddison 1991) in his construction of historical gross domestic product series, which have remained unchallenged until the present volume (see Chapter 6 by Alberto Baffigi), if not by the Spaniard Albert Carreras (1992). Finally, several other foreign economic historians produced memorable works on Italy: John Davis, who recently has been focusing on the social history of Italy; Jon Cohen (1972, 1988), who was particularly fascinated by the Italian Fascist regime; and Peter Hertner, who wrote abundantly on German capital in Italy.

Small Is Beautiful, Industrial Districts, and the “Third Italy” In the 1970s and 1980s, foreign observers were confirmed in the opinion that Italy was a sonderweg. Not only did the divergence between North and South persist, but there emerged even a “Third Italy,” Italy’s answer to the oil crises of 1973 and 1979, which seemed to defy and even negate world trends by promoting small-scale industry and launching its products on world markets. This new historical turn received outsize interest, as foreign economists echoed the research of some of their Italian colleagues who were studying what came to be known as “industrial districts,” singing the praises of “informal horizontal integration” among small firms that did not merge but remained self-standing (Becattini 2000). Foreign economists joined their enthusiastic Italian colleagues in extolling the virtues of the “Central Italian” model of balanced and nondestructive industrial growth, based on small-scale firms springing up in an environment

the italian economy seen from abroad

153

where agricultural sharecropping had prevailed for centuries and local social and political institutions that favored socially cohesive choices existed. Foreign sociologists and political scientists were eager to jump on the band wagon, some of them like Robert Putnam even going back to medieval Italy’s experience with the “Comuni liberi ” to explain why contemporary localized production methods based on social cohesion and cooperation had emerged in the Center and Northeast regions (Putnam 1994). They contrasted this with what prevailed in the South, where Edward Banfield had several decades before found inspiration for his theory of “Mediterranean familism” (Banfield 1958). The Third Italy model was so successfully propagandized abroad by foreign social scientists that even politicians, such as Bill Clinton when he was governor of Arkansas, came to inspect it directly, visiting Reggio Emilia and Modena to see their small firms and their welfare institutions, such as the famous infant schools of Reggio.

Conclusions Italy has been the object of attention and study by foreign scholars since ancient times. This chapter discusses the main phases of Italy’s development process, beginning in the early nineteenth century, seen through the eyes of international experts. This approach has the advantage of providing a more detached and disenchanted view of the country’s successes and failures than what would emerge from an assessment of solely Italian economists’ views. The interest Italy created in foreign observers did not evaporate soon after the country’s political unification. On the contrary, it increased and became more multifaceted as Italy evolved, moving from a peripheral agricultural backward country to a core industrialized international power. As is natural, in some periods when economic growth was strong, views were positive, and Italy was set out as an example for other countries; in other periods, generally of sluggish growth and increasing inequality and social conflict, positions were more critical and pessimistic. However, because of some peculiarities of the Italian economic structure, its institutions, and its development path, not all scholars were able to correctly comprehend the growth dynamics and the future prospects of the nation. The assessment provided in this chapter only focuses on those experts who better understood and gauged Italy’s problems. This perspective was often attained by spending long periods in Italy (recall, for example, Bonnefon Craponne’s biography); by holding key political positions that involved lengthy contacts with Italian authorities (e.g., in Gardner Ackley’s case); and by writing together with Italian scholars (e.g., Gardner Ackley wrote elbow-to-elbow with Luigi Spaventa).

154

aggregate growth and policy

This deeper knowledge of Italy’s workings often meant that the foreign observers mentioned stimulated lively and long-lasting national and international debates on the country. These discussions, furthermore, did not remain at a theoretical or abstract level. In some cases the views supported affected and molded Italian policy makers’ decisions (the debate which arose around the Country Study is, in this sense, emblematic).

part ii

SOURCES OF GROWTH AND WELFARE

This page intentionally left blank

Chapter 6

NATIONAL ACCOUNTS, 1861–2011 alberto baffigi

Introduction A great deal of new quantitative research has been produced over the last three decades that has considerably changed the received interpretation of Italian economic development as described by the old national accounts reconstructions. Against this backdrop, the Bank of Italy, the Italian National Institute of Statistics (Istat), and the University of Rome “Tor Vergata,” together with academics from other institutions, produced new estimates of national accounts time series, covering the 150-year period after political unification, and consistent with the most up-to-date results. We obtained a new gross domestic product (GDP) series from the supply and demand sides. For the supply side, value-added at factor cost for eleven sectors and GDP at market prices have been estimated; for the demand side, public and private consumption and three categories of investment goods are provided. All the time series are at current and constant prices. Hence, sectoral and aggregate deflators are available for the entire 150 years (see data appendix in this volume). This chapter presents the new dataset. Not all the methodologic details are reported fully in the following pages. They are presented in detail in the database NA150, available on the Bank of Italy’s Web site (see references), together with all time series reconstructed and the procedures used in this research. It is worth recalling some of the advantages of organizing the main results in Italy’s quantitative macroeconomic history in a single framework, built on modern national accounts concepts. First, it is a prerequisite for a thorough,

158

sources of growth and welfare

complete, and consistent interpretation of the whole of Italian economic history since unification; international comparisons also benefit from such a reconstruction. The quantitative sources used to reconstruct the time series are fragmentary by nature; their heterogeneous origins and quality could, in principle, prevent their use as sources for a consistent interpretation of the 150 years of Italy’s economic history. The study of a country’s economic evolution requires the availability of consistent statistical material over a long time span; many features of the economic system need to be cross-examined. In some ways an effort of “translation” is required to make all the available sources and results speak the national accounting language. Indeed, our work is an exercise in translation and, like all translators, we were faced with numerous options and had to make many choices. Second, ensuring transparency and accessibility to our data and calculations is a tenet of this project. A prerequisite for such a broad and complex exercise in connecting and homogenizing various and assorted statistical material is a systematic inventory of all primary and secondary sources, and this is an important by-product of our efforts. Our methodologic choices together with primary and secondary sources are documented in this chapter. The aim is to make all data and documentation easily accessible to scholars and students. We regard our work as an “open source” enterprise: replicability and detailed documentation are important features of the new data. We believe the quality of the data used in our national accounts reconstruction to be generally high according to international standards. We followed a two-layer approach. The first layer contains the input of our work, and the second is national accounts reconstruction in action. The first layer involved collecting and extending the results on different sectors obtained with different methodologies in different moments by different authors. More research was needed, however, before we could proceed to make national account estimates. Some of the results forming the overall picture are part of this project, namely the new estimates of the tertiary sector’s value-added from 1861 to 1951 by Battilani, Felice, and Zamagni (2011); the bank sector value-added (1861–2010) estimated by De Bonis et al. (2011); industrial value-added between 1928 and 1938 by Giugliano (2011); and the new benchmark for 1871 by Baffigi et al. (2011). In the second section, a brief survey of the first layer is sketched. Appendix 6.1 provides short abstracts of the original “first layer” research included in the project. This chapter is mainly about the second layer of the research, which aimed to reconcile and connect all the available results within national accounts concepts and definitions.1 All the inputs had to square to give shape to the national accounts time series. Methodologic details are provided only for the second layer of the project (Appendices 6.2–6.5, which should not be read before the second section). The third section is an overview of the new time series over the 150 years of Italian national life. The fourth and fifth sections are case studies: they show some historiographic implications of our data, respectively, for

national accounts, 1861–2011

159

the fifty years after unification (the industrialization years) and for the interwar years. Some reflections are set out in the concluding section.

Earlier Studies The first complete system of national accounts conceived as a support for Italian economic history research was developed by Istat and published in 1957, early by international standards (Istat 1957). In the following decade, Istat’s estimates formed the base for the work by the Ancona group led by Giorgio Fuà, with a central contribution by the statistician Ornello Vitali. Although an important pioneering effort, these attempts unfortunately failed to fully meet some of the basic requirement for scientific work, such as transparency and replicability: a common feature of such early studies was their scanty documentation. Despite these reservations, however, the data were widely used by Italian and foreign scholars. A new edition of Fuà’s book, containing only minor changes, was published at the end of the 1970s.2 In the meantime, a new empirical research strategy had been pursued for some years by Stefano Fenoaltea, who used a large array of new historical sources with acumen and originality. This work led to new estimates of disaggregate industrial value-added at constant prices for 1861–1913 (Fenoaltea 2006, 2011b). Carreras (1983) produced a new index of Italian industrial production from 1861 to 1980. Meanwhile, Federico’s research had produced more solid knowledge of agriculture during the first five decades after unification (see Federico 2003c and references therein). In 1990, Roberto Golinelli and Milena Monterastelli reconstructed national accounts from 1951 to 1990 to obtain data consistent with the most recent international accounting system (European System of Accounts [ESA], 1979). In 1991, Angus Maddison reconstructed the GDP series at constant prices from 1861 to 1989, using Fenoaltea’s data (Maddison 1991a). In 1992, a team coordinated by Guido Rey and sponsored by the Bank of Italy on the occasion of its centenary provided detailed national accounts for 1911: the year of the first Italian census of industry and commerce was taken as a benchmark. The project involved a later estimation of other benchmark years to be used as a constraint in a general reconstruction of annual national accounts time series from 1890: a 25 x 25 input-output matrix was estimated; supply and uses accounts at the same disaggregation level were calculated. This rather ambitious task was still far from accomplished when Rossi, Sorgato, and Toniolo (1993), searching for a measure of aggregate productivity, provided complete national accounts time series stretching from 1890 to 1990. Fenoaltea was still working on his project, building industrial production series for yet unavailable sectors. Moreover, between 2000 and 2002, three new benchmarks were produced by the Rey–Bank of Italy group for 1891, 1938, and

160

sources of growth and welfare 150 anni - NA time series Banca d’Italia – Istat – Tor Vergata – Bologna Istat (1970–2010) Battilani-Felice-Zamagni (1861–1950) De Bonis, Farabullini, Rocchelli, Salvio (1861–2010) Istat (1951–1970) Federico (1861–1913) Carreras-Felice (1911–1938) Fenoaltea (1861–1913) Giugliano (1928–1938) Ercolani-Fuà (1861–1951) BaffigiBattilaniFelice-TrigliaZamagni

1861 1871

Rey – Banca d’Italia

1891

1911

1938

Picozzi-Rey

1951

1970

Benchmarks

2010

Figure 6.1 Information used in the reconstruction project.

1951. A couple of years later Fenoaltea achieved a first reconstruction of GDP at constant 1911 prices for the period 1861–1913. The state of the art for Italian historical data at the time we began working on our project is summarized by the solid items in Figure 6.1. The initial year in the new project was shifted back to the very origin of the Italian state, 1861.

A Sketch of Our Methodology Supply-Side Estimates To obtain reliable historical data reconstructions, two information sets are required: information about variables’ levels is needed; then, information about dynamics is required to capture the cyclical movement of the economic variable to be estimated. The first information set allows the construction of national accounts for particular years, considered as benchmarks, years for which sources are deemed more reliable than for other years, or simply years for which sources have been more deeply studied in the literature. The benchmark years are connected exploiting the second information set, which provides information about dynamics (i.e., about movements that occurred over the period between two benchmark years). Benchmarks are like pillars of a bridge that bear a suspended path, the path travelled by the variable to be estimated. Indeed, our reconstruction strategy starts from benchmark years, which are treated as cornerstones. We had four of them (1891, 1911, 1938, and 1951), conceived as the pillars of a bridge spanning from the industrialization years, at the end of the nineteenth century, to the much better known scene of post–World

national accounts, 1861–2011

161

War II. After 1951, the path was a much smoother and easier one to the most recent national accounts. Because we wanted to extend our reconstruction to the earliest economic history of Italy, we built a new benchmark for 1871, even though users of our data should keep in mind that quality of statistical information is not high for the very first years after unification. We also wanted to get closer to the most recent years in our history so we included the new 1970 benchmark, built by Luisa Picozzi in Rey et al. (2012).3 Since 1970, official national accounts were provided by the Italian statistical office (Istat)4. The data conform with the ESA 1995 standard, which is based on concepts, classifications, and definitions that are quite different from the pre-ESA ones, on which the benchmarks are based. As a result, to obtain the 150-year-long series we had to adopt a two-step approach. First, we looked for appropriate sources to interpolate the benchmarks to produce time series consistent with the pre-ESA standard, over the 1861–1970 period. This is the bearing structure of our project. The statistical material, the first layer of our project, is shown in Figure 6.1. For each sector we consider in our reconstruction, it provides us with information about value-added at current prices and/or value-added at constant prices and/or implicit deflators. These three aggregates are linked and, given two of them, the third is uniquely determined: in each case we let the least reliable be determined by the other two. For example, for the period 1861–1911 the level of Fenoaltea’s industry data at 1911 prices was taken as a given, whereas Ercolani’s (1978) implicit deflators were used to get information only about price changes and not about levels: in fact, we use them as proxies to get deflators that are consistent with the benchmarks and with Fenoaltea’s data. Thus, we can find ourselves in one of the four typical cases described in Table 6.1, where we explain our interpolating strategy as depending, for each sector and each between-benchmarks period, on the answer to two questions about time series’ levels reliability and, when at current prices, about their being or not being equal to benchmark years’ estimates. Some specific information about our sources and methodology can be found in Appendix 6.2, whereas the simple method used for interpolating by means of proxy time series is described in Appendix 6.3. Second, after the pre-ESA 1861–1970 reconstruction was implemented, we had to connect those series with the official Istat 1970–2010 national accounts series. We decided to take the levels of 1951 benchmark and of 1970 official data (not equal to the 1970 pre-ESA benchmark) as given: pre-ESA estimate dynamics were used to interpolate the two. Before starting our work, we had to solve a classification consistency problem: our eleven pre-ESA sectors were not consistent with the ESA 1995 classification on which the official Istat time series (1970–2010) are based. The definitions are very different and it is not a trivial task to homogenize such classifications. We therefore had to refer to more highly aggregated series that provide a common ground between the two national accounts standards. We had to cope with a trade-off between the degree of sectoral detail and the length of the series: to obtain a 150-year-long series we

162

sources of growth and welfare

Table 6.1 Proxies and benchmarks: our interpolating strategy Is the current prices time series reliable “in level” and consistent with the benchmarks?

Yes

Yes

No

Case 1

Case 2

(1) Use current prices and constant time series

(1) Calculate implicit deflators for benchmark years as the (2) Calculate the deflator time ratio between current prices values and constant prices series as the ratio between current prices and constant (2) Interpolate the prices value-added time series “benchmark” deflators using available deflators as a proxy (3) Use the deflator in (2) to obtain a current prices time series

Is the constant prices time series reliable “in level”? No

Case 3

Case 4

(1) Use the current prices time (1) Interpolate the benchmark series values using the current prices time series as a proxy (2) Use a reliable deflator to obtain constant prices

(2) Use a reliable deflator to obtain constant prices time series

Note: The table refers to a hypothetical sector and to a given period, limited by two benchmark current prices estimates.

had to lose some information concerning the 1861–1970 period.5 Our solution is summarized in Table 6.2, where column 1 lists the twenty-five pre-ESA sectors used in the benchmarks, the second column reports the aggregated sectors used to connect the benchmarks, and the third column displays the four aggregated sectors used for the “long” series 1861–20106: (1) agriculture; (2) industry including energy (i.e., total industry minus constructions); (3) construction; and (4) services. The main elements of our supply-side reconstruction are reported in Appendix 6.2.

Demand-Side Estimates We considered the results obtained for the supply side as constraints for the demand-side estimates. In working on the demand side we could not rely on solid recent work as we had for the supply side. We decided to be guided by the strongest. Given the GDP level at current market prices, we obtained an estimate of total internal uses as total sources (GDP plus imports of goods and services)

163

national accounts, 1861–2011 Table 6.2 Sector classifications Benchmarks

Between benchmarks (pre-ESA classifications): 1861–1970

1861–2010 Series

Agriculture

Agriculture

Agriculture

Mining

Mining

Foodstuffs Tobacco Textiles Clothing Leather Wood Metal-making Engineering

Manufacturing

Industry including energy

Minerals Chemical Coal and petroleum products Rubber Printing Other manufacturing Utilities Construction Commerce, hotels, and restaurants

Utilities Construction

Construction

Commerce, hotels, and restaurants

Transportation and communications

Transportation and communications

Banks and insurance

Banks and insurance

Miscellaneous services

Miscellaneous services

Public administration

Public administration

Location of buildings

Location of buildings

Services

minus external uses (exports of goods and services)7. The various elements of internal uses, estimated independently, are then imposed to meet this constraint by consistently rescaling their sum. In the end, our demand side is made up of consumption (public and private) and some categories of investment goods (see

164

sources of growth and welfare

Table 6.3 Investment classifications 1861–1970 aggregated estimates Construction

1861–1970 disaggregated estimates

1861–2010

Housing

Housing

Public works Nonresidential construction

Nonhousing

Plant machinery and transport

Plant machinery and

Plant machinery and

equipment

transport equipment

transport equipment

Other investment goods

Other investment goods

Inventory variations

Inventory variations

Other investment goods

below). For the demand-side reconstruction we also moved from the estimates provided by the benchmarks, interpolated by using the indicators presented in Appendix 6.4. For investments, the benchmarks provide information about the amount of investment goods produced by each sector. Our aim, however, was to estimate investment series classified by categories of goods. Thus, we took the construction and mechanical engineering sectors because they evidently produce two identifiable and meaningful categories of investment goods: “construction” and “plant, machinery, and transport equipments,” respectively. We aggregated all the other investment goods in a residual category (see Appendix 6.4, section 3). We ended up with three categories of investment goods: (1) construction; (2) plant, machinery, and transport equipments; and (3) other investment goods. As with the supply side, we started by estimating time series for the period 1861–1970, framed in the pre-ESA national accounts standard. We then interpolated the 1951 benchmark with the first (1970) official year using the growth rates of pre-ESA time series as proxies. In our main estimates, changes in inventories are not separated from fixed investments. An estimate of inventories has been made for practical reasons, but we regard it as a less robust result.8 Similarly, practical reasons have brought us to provide a disaggregation of investment in construction, which is not included in our main estimates. As in the case of the supply side, we had to cope with different classifications of the series in the pre-ESA and ESA standards. In particular, although in the pre-ESA frame (1861–1970) investment in construction is divided into housing, public works, and nonresidential construction, in the ESA frame only housing and nonhousing are reported (nonhousing being the sum of public works plus nonresidential construction). However, in the official (1970–2010) tables, investments are presented net of inventory variations; the total value of the latter is separately reported. Thus, to link our 1861–1970 estimates with the official 1970– 2010 investment data, we had to use our fixed investments estimates, providing a

national accounts, 1861–2011

165

separate series for inventories variation. Table 6.3, column 1, displays the investment classification used in our main estimates; columns 2 and 3 contain the more disaggregated items for 1861–1970 and 1861–2010 estimates, respectively. Our main supply and uses estimates (1861–2010) are reported in Table A1 (current prices) and Table A2 (constant 2010 prices) in the data appendix in this volume. In line with our “open source” philosophy, other tables are included in the NA150 Excel file published on the World Wide Web, including information about calculations and intermediate procedures (see references).

Deflator Estimates Deflators were obtained for GDP and for supply- and demand-side items. In our estimates, reference years vary across subperiods and, in particular, 1861– 1911 is based on 1911 prices, 1911–1951 on 1938 prices, 1951–1970 on 1963 prices, and 1970–2010 on 2010 prices. This means that, for each item (i.e., sectoral value-added on the supply side; and consumption and investment items on the demand side) deflator time series are segmented by subperiod. This feature must be kept in mind when using our series at constant prices. In particular, when a base year for all 150 years is calculated, the resulting series at constant prices suffers from potentially serious nonadditivity problems (on this issue see Appendix 6.5).

The New Series: An Overview (1861–2010) A broad picture of the main features of the new data is provided by the GDP time series. Figure 6.2 shows the log of GDP at constant prices, normalized to zero in 1861. It can be interpreted as an approximation of rates of change with respect to 1861. Given the deep trough of World War II and subsequent acceleration, the earlier events described by the series are of a smaller magnitude. In particular, it took almost eighty years to double the initial per capita GDP: the highest GDP level since unification was reached in 1939, one year before Italy entered the war. It doubled again in the early 1960s. The evolution of Italy’s economic structure is described by the sectoral shares, a calculation that is now possible thanks to one of the new features introduced by our reconstruction (i.e., the estimates of sectoral value-added at current prices).9 In Figures 6.3–6.5, the structure of the economy is shown for three periods: (1) 1861–1911, (2) 1911–1951, and (3) 1951–1970.

166

sources of growth and welfare 350 300 250 200 150 100 50

18 6 18 1 66 18 7 18 1 76 18 8 18 1 86 18 9 18 1 9 19 6 01 19 0 19 6 11 19 1 19 6 21 19 2 19 6 3 19 1 3 19 6 4 19 1 4 19 6 5 19 1 5 19 6 6 19 1 6 19 6 71 19 7 19 6 81 19 8 19 6 91 19 9 20 6 0 20 1 06

0

Figure 6.2 Constant prices GDP per capita (semilog scale; 1861 = 0).

In the first period, the structural change in the Italian economy, as broadly described by the shares of value-added of the three large sectors, shows a stable share for industry, a slowly declining share for agriculture, and a growth in the weight of the tertiary sector. This evolution hides an interestingly mixed contribution of changing quantities and relative prices. Again, our reconstruction of sectoral deflators (one of the new, characteristic elements of the new national accounts database) allows us to identify quantity and price contributions to changes in shares. This issue is dealt with in the following section, referring to the period 1861–1911. Turning to the demand side, Figure 6.6 shows the downward trend of the share of total resources channelled to private consumption, with a fall during 0.5 0.45 Agriculture 0.4 0.35 Services 0.3 0.25

Industry

18 61 18 63 18 65 18 67 18 69 18 71 18 73 18 75 18 77 18 79 18 81 18 83 18 85 18 87 18 89 18 91 18 93 18 95 18 97 18 99 19 01 19 03 19 05 19 07 19 09 19 11

0.2

Figure 6.3 Sectoral values added (percentage shares, 1861–1911).

167

national accounts, 1861–2011 0.55 0.5 0.45 Services

0.4 0.35 Agriculture

0.3 0.25 0.2

Industry

0.15

19 11 19 13 19 15 19 17 19 19 19 21 19 23 19 25 19 27 19 29 19 31 19 33 19 35 19 37 19 39 19 41 19 43 19 45 19 47 19 49 19 51

0.1

Figure 6.4 Sectoral values added (percentage shares, 1911–1951).

World War I and after 1929. Figure 6.7 shows the accumulation rate (investment/ GDP ratio), which fluctuates around 5–6 percent during the first decade after unification, a value typical of a backward economy. It increases from the early 1870s until 1873 (the so-called triennio febbrile, the feverish triennium); after ups and downs around 10 percent, it drops back to 7–8 percent after 1887; and investment activity begins to accelerate sharply at the end of the century until the crisis of 1907 when Italy’s accumulation rate reached a peak that was only attained again during the second postwar “economic miracle.” An all-time peak in the accumulation rate was reached in 1963.

0.5 Services 0.4 Industry

0.3

0.2

Agriculture

0.1

53 19 54 19 55 19 56 19 57 19 58 19 59 19 60 19 61 19 62 19 63 19 64 19 65 19 66 19 67 19 68 19 69 19 70

52

19

19

19

51

0

Figure 6.5 Sectoral values added (percentage shares, 1951–1970).

168

sources of growth and welfare 92.0% 87.0% 82.0% 77.0% 72.0% 67.0% 62.0%

18 6 18 1 6 18 6 71 18 7 18 6 8 18 1 86 18 9 18 1 96 19 0 19 1 06 19 1 19 1 16 19 2 19 1 2 19 6 3 19 1 3 19 6 4 19 1 4 19 6 5 19 1 56 19 6 19 1 66 19 7 19 1 7 19 6 8 19 1 8 19 6 91 19 9 20 6 01 20 06

57.0%

Figure 6.6 Private consumption/GDP ratio.

Industrializing Italy (1861–1913) Before Fenoaltea’s reconstruction of industrial data for the period 1861–1913, the prevailing interpretations depicted Italian industrialization as a sudden and revolutionary structural break occurring at the end of the nineteenth century, the effects of which spanned the Giolitti period (1898–1913) with an interruption only during World War I. Fenoaltea’s time series are the backbone of our reconstruction, which thus largely shares the time profile of his time series.

34.0%

29.0%

24.0%

19.0%

14.0%

9.0%

18 6 18 1 6 18 6 7 18 1 76 18 1 18 8 86 18 9 18 1 9 19 6 0 19 1 0 19 6 1 19 1 1 19 6 21 19 2 19 6 31 19 3 19 6 4 19 1 4 19 6 5 19 1 5 19 6 6 19 1 6 19 6 7 19 1 7 19 6 8 19 1 8 19 6 91 19 9 20 6 01 20 06

4.0%

Figure 6.7 Accumulation rate (investments/GDP, 1861–1970).

169

national accounts, 1861–2011 280.000 260.000 240.000 220.000 200.000 180.000 160.000 140.000 120.000

18

61 18 63 18 65 18 67 18 69 18 71 18 73 18 75 18 77 18 79 18 81 18 83 18 85 18 87 18 89 18 91 18 93 18 95 18 97 18 99 19 01 19 03 19 05 19 07 19 09 19 11

100.000

base 1871

base 1881

base 1891

base 1901

base 1911

Figure 6.8 GDP levels (1861 = 100). The underlying series are calculated at the following base years: 1871, 1881, 1891, 1901, and 1911.

This new pattern of growth introduces a completely new view in the commonly accepted historiographic perspective of Italy’s industrialization. There seem to be no Rostowian or Gerschenkronian take-offs or big spurts. Natura non facit saltus, so Fenoaltea argued. Is this a conclusive result? I will not try to give an answer here, but rather point out that deflators for sectoral values added, which is one of the main novelties of our reconstruction, may widen the debate, which has so far been confined to 1911-prices time series.10 If we assume that innovations pushed the prices of some industrial sectors down and that, for a given demand curve, growth in the output of those products is observed, then we can expect that aggregating sectoral values added with 1911 relative prices may underestimate the total growth in value-added. This is nothing new; it is known that Alexander Gerschenkron taught economic historians about this index number problem as early as the 1950s.11 In fact, the 1911 reference year, which is more or less located at the end of an important wave of industrialization, may bias downward the Giolittian acceleration in the growth rate. This intuition seems to be corroborated by using our new deflators. Deflators are weights we use to aggregate quantities, so I used different base years deflators to aggregate our eleven sectoral values added. The results of the exercise are summarized in Figure 6.8, where GDP estimates from 1861 to 1911 are provided at 1871, 1881, 1891, 1901, and 1911 prices. The lines in the graph show that changes in base years produce observable effects only from the 1880s onward; before that time the change in base year does not seem to matter. If we take this index number bias as a measure of economic development (Ames and Carlson

170

sources of growth and welfare 0.36

1.05

0.34

1 0.95

0.32 prices

0.9

0.3 0.85 0.28 0.8 0.26

0.75

0.24

0.7

quantity

0.65

0.2

0.6

18 6 18 1 6 18 3 6 18 5 67 18 6 18 9 7 18 1 73 18 7 18 5 7 18 7 79 18 8 18 1 83 18 8 18 5 8 18 7 89 18 9 18 1 93 18 9 18 5 97 18 9 19 9 0 19 1 03 19 0 19 5 0 19 7 09 19 11

0.22

Figure 6.9 Prices and quantity contributions to industry’s shares of total value-added. d I, t QI, t Industry’s share of value-added is equal to dTOTT, t QTOTT, t , where d = deflator, Q = value-added in volume, I = industry, and TOT = whole economy. The lines in the d Q graph represent, respectively, the factors d I, t and Q I, t . TOT, t

TOT, t

1968), the preliminary analysis seems to indicate that some acceleration in the path toward industrialization already took place during the 1880s, whereas a second more robust wave occurred during the Giolitti period. On this point, a future line of research should aim to reconstruct deflators at a higher level of disaggregation, with special focus on the manufacturing sectors, of which we could estimate only an aggregate deflator. Deflators also can be profitably used to gain insight from the analysis of structural changes. In Figure 6.3 we noted the stable share of industry over the fifty postunification years. Using sectoral deflators we can decompose the shares at current prices, considering relative prices and quantity effects. Figure 6.9 shows that quantity tended to push up industry’s share throughout almost the entire period, whereas relative price movements offset the expansion in real production. Interestingly, the quantity effect is greatest in the 1880s and in the Giolitti period, which provides hints for a periodization of Italy’s industrialization. Some other interesting new features of our data emerge if we look at the demand side. In the old Ercolani series, private consumption per capita fell sharply during the third war of independence (1866) and then stagnated at those low levels until the end of the century when a recovery led it back close to its initial level. By contrast, our new series, after a similar fall in 1866, shows a strong recovery from 1873 to 1887, when consumption per capita reached a level similar to the 1865 peak. It fluctuated around a slightly lower level until

171

national accounts, 1861–2011 0.52 0.5 0.48 0.46 0.44 NEW 0.42 0.4 Rossi-Sorgato-Toniolo (1993)

0.38

Ercolani-Fuà (1978)

0.36

67 18 69 18 71 18 73 18 75 18 77 18 79 18 81 18 83 18 85 18 87 18 89 18 91 18 93 18 95 18 97 18 99 19 01 19 03 19 05 19 07 19 09 19 11

65

18

63

18

18

18

61

0.34

Figure 6.10 Private consumption per capita (millions of 1911 lire).

1900, when a new strong recovery started, lasting until 1909, which was a new historical high (Figure 6.10). New features also characterize our investment series. In Figure 6.11 the new series of investments in plant and equipment at constant 1911 prices is compared with earlier estimates: it is interesting to note the faster pace of our series in important historical periods, such the beginning of the 1870s, the 1880s, and above all the Giolitti period until the 1907 crisis.

4000 3500 3000

New

2500 2000 1500 1000 500 Ercolani-Fuà (1978)

Rossi-Sorgato-Toniolo (1993)

18 71 18 73 18 75 18 77 18 79 18 81 18 83 18 85 18 87 18 89 18 91 18 93 18 95 18 97 18 99 19 01 19 03 19 05 19 07 19 09 19 11

69

18

67

18

65

18

63

18

18

61

0

Figure 6.11 Investments in plants, machinery and transport equipment (millions of 1911 lire).

172

sources of growth and welfare

Two Wars and a Great Depression (1911–1951) Numerous studies populate the difficult interwar period in Italy, many of which tend to stress the inadequacy of available quantitative macroeconomic evidence. New national accounts reconstructions were eagerly awaited by many scholars and we have tried to fulfil their expectations. Unlike our estimates for the period 1861–1911, for which we could use the data of Fenoaltea and Federico, the new time series for the four decades 1911–1951 are based on the use of hitherto little-exploited statistics. The two ovals in Figure 6.12 highlight the main changes introduced by our work. The one on the left refers to Italy’s economic performance during World War I. Broadberry (2005) noted that available national accounts series “all show an extremely large increase in Italian GDP during World War I, which is hard to square with both the experience of other countries during World War I and the generally pessimistic tone of the literature on the Italian war economy” (305). The new GDP series in the period examined by Broadberry follows a much more moderate path. The main innovation is based on the work of Patrizia Battilani, Emanuele Felice, and Vera Zamagni as part of this project.12 In particular, their estimate for general government value-added at current prices is much lower than, for instance, that of Ercolani (1978): the new data imply a 200 percent growth between 1913 and 1918, compared with 800 percent in the earlier estimates. An even greater difference emerges if we consider constant price estimates: according to the calculations of Baffigi and Brunetti,13 the growth in general government value-added at 1938 prices over the period 1913–1918 is a mere 13 percent, whereas the growth rate in Ercolani’s data is 580 percent. The oval on the right in Figure 6.12 refers to Italy’s performance during the Great Depression. The revision with respect to earlier estimates is entirely attributable to another member of our team, Ferdinando Giugliano, who revised recent estimates of industrial value-added by Carreras and Felice (2010). He makes an interesting critical use of such sources as employment data and an industrial production index produced by the Ministero delle Corporazioni.14 Overall, if we look at GDP movements, we notice a sharp fall in activity in the years immediately after 1929, not unlike earlier estimates. However, our data show a very different timing for the ensuing recovery: now it takes some years for the recovery to begin. Only in 1937, after an illusory blip in 1935, did GDP reach the same level as in 1929. Private consumption per capita seems to reflect this movement in economic activity: after 1929 it declines steadily until 1936 (the only partial exceptions being 1932 and 1935) when a recovery begins, which is suddenly broken off by the outbreak of World War II (Figure 6.13). Investment in plant and equipment

173

national accounts, 1861–2011 236000.000 216000.000

Rossi, Sorgato, and Toniolo (1993)

196000.000 176000.000 156000.000 136000.000

Ercolani and Fuà (1978)

116000.000 NEW 19 11 19 13 19 15 19 17 19 19 19 21 19 23 19 25 19 27 19 29 19 31 19 33 19 35 19 37 19 39 19 41 19 43 19 45 19 47 19 49 19 51

96000.000

Figure 6.12 Market prices GDP levels (millions of 1938 lire).

followed a similar path, reaching a peak in 1929 and then falling sharply; it was only in 1938 that a higher level was reached (Figure 6.14).

Final Considerations A GDP series covering Italy’s entire history has been estimated, together with sectoral value-added and, for the demand side, disaggregated consumption and 3.3

2.8

2.3

1.8

1.3

17 19 19 19 21 19 23 19 25 19 27 19 29 19 31 19 33 19 35 19 37 19 39 19 41 19 43 19 45 19 47 19 49 19 51

19

15

13

19

19

19

11

0.8

Figure 6.13 Household consumption per capita (1911–1951; millions of 1938 lire).

174

sources of growth and welfare 40000 35000 30000 25000 20000

Rossi, Sorgato, and Toniolo (1993) Ercolani and Fuà (1978)

15000 10000

New

5000

19 11 19 13 19 15 19 17 19 19 19 21 19 23 19 25 19 27 19 29 19 31 19 33 19 35 19 37 19 39 19 41 19 43 19 45 19 47 19 49 19 51

0

Figure 6.14 Fixed investment in plant, machinery, and transport equipment (millions of 1938 lire).

investment. All the time series are available at current prices and in volume. This is the output of the project presented in this chapter. The importance of this update of sources for Italy’s economic history cannot be overstated. The period 1861–1951 has been dissected and heavily revised by our work; it was then connected to the ensuing years (1951–1970, less heavily revised) and, finally, to the official 1970–2010 Istat data to obtain a quantitative documentation of united Italy’s entire macroeconomic history. As emphasized in the introduction, we did not work in a vacuum. The obsolescence of the old historical national accounts was caused by the results obtained by scholars working over the last thirty years. New results have been added by ad hoc studies forming part of our project, namely the new estimates of value-added time series, at current and constant prices, for services (1861–1951), for industry for the period 1928–1938, and a new benchmark for 1871 of supply and uses. Old and new results were the input of our reconstruction work, the first layer of the project. Elaboration, homogenization, and reconciliation of this corpus of quantitative research was the task of the second layer of the project. We brought all available data into a common conceptual space provided by the national accounts framework. We have constructed deflators, either revising old ones or estimating new ones ex novo, and we have estimated supply- and demand-side series covering 150 years of Italy’s history. Important elements have been added to the economic historian’s toolbox. After this task has been accomplished, it is natural to think about possible refinements, or about new lines of research to build on the results just obtained. A disaggregation of the manufacturing sector could bring some important new information, in particular about the first industrialization period between unification and World War I. Detailed sector value-added data do exist

national accounts, 1861–2011

175

(see Fenoaltea 2006 and bibliography therein), but they are all at 1911 constant prices. No deflators are available for this period. An effort to construct more disaggregated deflators for the manufacturing sector would be extremely useful for a better understanding of the patterns of industrialization in Italy. Another interesting research topic is related to demand-side estimates, which are not based on statistical material as robust as that available for value-added estimates. Many archival sources are awaiting discovery and could be incorporated advantageously in our overall framework, similarly to the ones we have already used to estimate investment in means of transport for 1911–1951. Notwithstanding potential refinements and improvements, which are encouraged by the transparency of our methodologies and assumptions, Italy too can now boast a complete and consistent historical national accounts framework that beckons new and more accurate interpretations of the country’s 150-year-long history.

Appendix 6.1. Original Work by the “150 Anni” Team Besides already existing time series our project includes some new reconstruction work: 1. Patrizia Battilani (University of Bologna), Emanuele Felice (Autonomous University of Barcelona), and Vera Zamagni (University of Bologna) produced a new current prices series for service sector value-added from 1861 to 1951. New series have been produced sector by sector (i.e., for commerce, hotels and restaurants, inland transportation, maritime transportation, air transportation, communications, insurance, miscellaneous services, public administration, and buildings). 2. Notwithstanding the small weight of banks’ value-added in the entire tertiary sector, the importance of bank data for today’s economies led us to pay particular attention to this sector. Riccardo De Bonis, Fabio Farabullini, Miria Rocchelli, and Alessandra Salvio (all colleagues at the Bank of Italy) undertook this task and, besides estimating the financial sector value-added, provided long time series (1861–2010) of the main items in the banks’ balance sheets. Finally, based on the same set of quantity data used by Battilani et al., Baffigi and Brunetti constructed new series of services sector value-added at constant prices. 3. Battilani et al. paid particular attention to 1871 value-added, which is their contribution to the construction of the new benchmark for that year developed by Alberto Baffigi and Ivan Triglia: it is made up of a 25 x 25 input-output matrix and 25 “uses and sources” accounts at current prices.

176

sources of growth and welfare

4. Ferdinando Giugliano (University of Oxford), improving on a recent paper by Carreras and Felice (2010), conducted research on value-added in Italian industry during the Great Depression. His main result shows that, contrary to the indications of previous data, the recession was severe in Italy, as in other industrial countries. Earlier and new time series are presented in Figure 6.1, where the red items are new estimates, part of our project. 5. The forthcoming Picozzi-Rey 1970 benchmark (Figure 6.1) is not part of our project but the authors kindly shared with us their data.

Appendix 6.2. The Supply Side (1861–1970): Methods and Sources Given the benchmark value-added at current prices, the interpolating method, explained in Appendix 6.3, is applied for each sector to one of the three mutually interdependent time series (deflator, value-added at constant prices, and value-added at current prices), depending on the reliability and the availability of the existing series. In the remaining part of this appendix the various cases are listed by period and by sector with reference to the four cases presented in Table 6.1.

1861–1911 Agriculture (Case 4) (a) Value-added at current prices is estimated by imposing Federico’s (2003c) data through the benchmark levels, when actually minor differences occur.15 (b) The price deflator is Federico’s (1911 = 100). (c) Value-added at 1911 prices is obtained by applying (b) to (a).

Industry (Case 2) (a) Value-added at 1911 prices is given by Stefano Fenoaltea’s (see Fenoaltea 2006, 2011b, and the articles cited therein) various sectoral estimates, which have been made consistent with present and historical boundaries, from those of 1911. (b) Price deflators of value-added of the different industrial sectors have been calculated for the benchmark years (1871, 1891, and 1911) as the ratio of the (current price) benchmark estimates and the corresponding constant prices estimates provided by Stefano Fenoaltea.16 The time series of the price deflators was then completed by interpolating the benchmark values of the deflators on the basis of the Ercolani (1978) (henceforth EF) implicit price deflators. For the years to

national accounts, 1861–2011

177

1871, the set of new deflators was calculated by retropolating the benchmark price deflators on the basis of the dynamics of the EF deflators. (c) Current price value-added time series are obtained by applying the price deflators to Fenoaltea’s estimates of value-added at constant prices.

Services (Case 1) (a) Current prices value-added series for seven tertiary sectors are given by the new work by Battilani et al. included in this project. (b) 1911 prices value-added series have been estimated by Baffigi and Brunetti by retropolating 1911 valueadded, based on quantity data consistent with the work of Battilani et al. (c) Price deflators have been obtained as a ratio of (a) to (b).

1911–1938 Agriculture (Case 4): Estimates of Value-Added at Current and Constant Prices For the years 1913–1938, the value-added of agriculture at current prices has been estimated by interpolating the benchmark levels (the interpolation has been carried out on the basis of the EF series). As for the constant price series, it has been calculated on the basis of the EF deflator.

Industry: Estimates of Value-Added at Current and Constant Prices First Step (Case 2): Estimates of the Current Price Value-Added for 1928–1938 The value-added of industry at current prices for the years 1928–1938 has been estimated on the basis of Giugliano’s 1938 price estimates and Carreras-Felice’s 1938 price deflator.17

Second Step (Case 4): Estimate of the Current Price Value-Added for 1911–1928 The value-added at current prices for the period 1911–1928 has been estimated by interpolating the benchmark year 1911 with the 1928 values as calculated in the previous step, using the Carreras-Felice’s industry current price value-added time series as proxy series.

Third Step (Case 4): Estimates of Constant Prices (1938) Value-Added for 1911–1928 The value-added of industry at constant prices for the years 1911–1928 has been estimated on the basis of the current price value-added of industry (see second step) and the Carreras-Felice 1938 price deflator.

178

sources of growth and welfare

Services (Case 1) As in 1861–1911 estimates.

1938–1951 Agriculture (Case 4): Estimates of Value-Added at Current and Constant Prices The procedure used is the same as for the years from 1911 to 1938.

Industry (Case 4): Estimates of Value-Added at Current and Constant Prices The procedure used is the same as for agriculture.

Services As in the 1861–1911 estimates (Case 1); because of lack of data, the constant price value-added of transport and communications has been estimated based on Ercolani’s data (Case 4).

1951–1970 All estimates for this period are interpolations between the 1951 benchmark (Rey 2000, 2002) and the new 1970 benchmark by Luisa Picozzi (2012). Interpolations have been performed based on Istat (1973) sectoral dynamics (Case 4). Deflators have also been taken from Istat (1973).

Appendix 6.3. The Interpolating Algorithm The set of benchmark values has been interpolated based on the information on the evolution of the annual rates of change of available proxy time series.18 More precisely, let Z0, Z1, . . . , ZT denote the time series of a variable of interest in years [0,T], and f1, f2, . . . , f T, the corresponding rates of change: ft = (

Zt − ) = πt − 1 Zt −1

179

national accounts, 1861–2011

^

^

If Z 0 and Z T are assumed to be the updated estimates of the variable Z in the first and last year of the time interval concerned (updated benchmark values), the problem arises as to how the in-between levels of Z can be recalculated to be coherent with the new information. The problem concerns the interpola^ ^ tion of the benchmark values Z 0 and Z T , based on the information already available on Z. Notably, the reconstruction of the variable of interest, no matter how it is obtained, implies the revision of the series of its rates of change, f t. From this point of view, a possible solution to the problem at issue is to evaluate a constant correction coefficient to the growth factors πt . Given, ^

ZT ZT α = ^ / Z0 Z0 T

Our correction coefficient is: ^

ZT ZT 1 α = ( ^ / )T Z0 Z0

Appendix 6.4. From Supply to Demand: Reconstructing National Account Demand Side. Italy, 1861–1970 The Procedure at Current Prices Our procedure follows four steps. The first step consists of calculating aggregate internal uses (i.e., consumption plus investment: Ct + It) at current prices and at current borders as GDPt + Mt – Xt (t = 1861, . . . 1970), where GDPt is our estimates at market prices; Mt = import ( from 1861 to 1951: see endnote 7; from 1951 to 1970: Istat 1973); X = export (ibidem). In the second step, we estimated the following seven demand components (details on sources and methods are given in Section 3 of this appendix)19: Cg t = public consumption Cpt = household consumption Imt = investment in machinery and transport equipments Ict = investment in construction Iat = investment in agricultural goods Iot = other investment (investment goods produced by other sectors) Iit = investment in intangible goods

180

sources of growth and welfare

In the third step, the estimates obtained in Step 2 are modified so that their sum complies with the following constraint:

(1) C + I = Cgt + Cpt + Imt + Ict + I at + I ot + I it. Compliance with constraint (1) is imposed by multiplying the demand components by the following coefficients: (a) for t = 1861, . . . ,1911 Ct

I t = Ctg

−1911 m −1911 a −1911 o Ctp + β1861 I t + I tc + β1861 I t + β1861 It t t t

β1861 t

1911

−1911 i + β1861 It t

(2a)

where −1911 β1861 = [(Ct t

I t ) − Ctg

I tc ] /[Ctp

I tm + I ta + I to + I ti ]

(2b)

(b) for t = 1912, . . . . ,1951 Ct

I t = Ctg

−1951 a −19511 o Ctp + I tm + I tc + β1912 I t + β1912 It t t

β1912 t

1951

(3a)

−1951 i + β11912 It t

where −1951 β1912 = [(Ct t

I t ) − Ctg

I tc − I tm ] /[Ctp + I ta

I to ]

(3b)

(c) for t = 1952,. . . . , 1970 Ct

−1970 g I t = β1952 Ct t

β1952 t

1970

−1970 m −1911 1 Ctp + β1952 I t + β1861 I tc β11861 t t t

1911 a t

I

(4a)

1 −1911 o 1861 t t

+β

I

where −1970 β1952 = [Ct t

I t ] /[Ctg

Ctp + I tm + I tc + I ta

I to ]

(4b)

In the fourth step, historical boundaries estimates are converted into present boundaries data by multiplying each series by appropriate coefficients, of which detailed information is provided in the NA150 Excel file cited in the references.

The Procedure at Constant Prices The constant prices time series reflect those calculated in the supply-side estimates20:

national accounts, 1861–2011

181

1911 prices for 1861–1911 1938 prices for 1911–1951 1963 prices for 1951–1970 2010 prices for 1970–2010 The procedure is symmetric with the one at current prices: 1. Calculate Ct + It at constant prices as GDPt + Mt – Xt, deflated with appropriate deflators; 2. Apply a deflator to each single demand side series; 3. In this case, too, estimates must comply with a constraint formally identical to (1) we came across in treating the problem at current prices. In this case, coefficients, calculated as in (2b), (3b), and (4b) were applied to each series throughout the 150-years time span.

The Current Price Estimates of the Demand Components and their Deflators The six benchmark years (1871: Baffigi et al. 2011; 1891; 1911; 1938; 1951: Rey 2002; 1970: Picozzi 2012) data are our starting point. We take time series or proxy time series for each of the six demand components and we constrain them through the benchmark year estimates. The time series we use to implement the procedure are listed in this section along with the associated deflators. For consistency with the supply-side calculations, the base year of Ercolani’s (1978) deflators, always 1938, has been changed to 1911 when used for the 1861–1911 period.

Public Consumption Time Series 1861–1951: To obtain a current price time series, apply Ercolani (1978) public consumption deflators (1938 = 100: Tav.XII.4.1.B, 434–435) to the 1938 constant price public consumption series (Ercolani, 1978, Tav.XII.4.1.A, 432–433). 1951–1970: Istat (1973), Tav.37, 96.

Deflators 1861–1951: Ercolani (1978), Tav.XII.4.1.B, 434–435 (1938 = 100; the base year for the period 1861–1911 has been changed to 1911). 1951–1970: Istat (1973), Tav.37, 96 (1963 = 100).

182

sources of growth and welfare

Private Consumption Time Series 1861–1951: use consumption goods import data from Federico et al. database to interpolate benchmark data. Missing data for 1861, 1947, and 1951 have been recovered by extrapolation and interpolation. 1951–1970: Istat (1973), 60–63.

Deflators 1861–1951: Istat (2009a).

Investment in Plant, Machinery, and Transport Equipment Time Series 1861–1911: net import data (1862–1880) of “Machinery and transport equipment” (Standard International Trade Code = 7), from the database of Federico et al. (2011) concatenated with Warglien (1985) quantity index. 1911–1951: new time series.21 1952–1970: Istat (1973).

Deflators Ercolani (1978), Tav. XII.4.14.B, 452–453 and Istat (1973).

Investment in Construction Time Series 1861–1913: Fenoaltea (1987), Table 5, column 1, 23–24; data are at 1911 prices. To obtain current prices time series Ercolani’s (1978) deflator (Table 12.IV.22, 460– 461) is applied. 1914–1970: Lupi and Mantegazza (1994) growth rates from 1914 to 1970 are used to extend Fenoaltea (1987) time series at current prices.

Deflators Ercolani (1978), Table 12.IV.22, 460–461.

Investment in Agriculture Goods Time Series These are animals (mainly horses) for urban services. Until 1951 the benchmark fixed investment was interpolated using the time series of transport

national accounts, 1861–2011

183

and communications as a proxy. The 1951–1970 time series was obtained by interpolating the 1951 and 1970 benchmarks using a quantity index (inflated with the agriculture value-added deflator) based on livestock data reported in Rey (1991, Table. 1.05, 116).

Deflators Agriculture value-added deflator.

Other Investments (Investment Goods Produced by Other Sectors) Time Series This is a broad and heterogeneous category including all goods produced by branches not explicitly considered: mining, tobacco, textiles, clothing, leather, wood, metal-making, minerals, chemical, coal and petroleum products, rubber, printing, other manufacturing, and utilities. The time series of this aggregate has been estimated as a percentage of the plant, machinery, and transport equipments investments. In particular, such ratios have been first calculated for the benchmark years, then they have been linearly interpolated; the 1861–1870 was set equal to the 1871 benchmark value.

Deflators Istat (2009a).

Investment in Intangible Goods Time Series We retropolated the 1970 official Istat estimate by using the growth rates of total services value-added.

Deflators Deflator for total services value-added.

Appendix 6.5. Nonadditivity of Series in Volume Close attention must be paid in using our constant prices series because of the possible bias arising when deflators with different base years are linked.

184

sources of growth and welfare

Linking deflators with different base years normally implies nonadditivity of the deflated series. In our estimates, reference years vary across subperiods and, in particular, 1861–1911 is based on 1911 prices, 1911–1951 on 1938 prices, 1951–1970 on 1963 prices, and 1970–2010 on 2010 prices (to follow the example below, please keep in mind these subperiods and the related base years). This means that, for each item (i.e., sectoral value-added on the supply side and consumption and investment items on the demand side) deflator time series are segmented by subperiods. Within each subperiod, deflated total value-added is always equal to the sum of deflated sectoral value-added. This is not necessarily true if we calculate constant price series with the base year outside the subperiod considered in the analysis. Suppose that we want to construct a total value-added series at 2010 prices over the entire period 1861–2010. We need to link all subperiod deflators to obtain a deflator series valid over all 150 years, with 2010 as reference year. Now, as a first step, linking the deflator series for 1970–2010 with the deflator series for 1951–1970 requires the latter to be shifted up or down, depending on the relative level of the deflators in the common year (i.e., 1970). Similar operations are required going backward in relation to the earlier subperiods. Suppose that we need to calculate also sectoral value-added at 2010 prices. Obviously, a similar backward linking procedure should be applied for each sector. Logically, the nonadditivity problem comes out here: there is no guarantee that, given the modified levels of the sectoral deflators within the period 1951–1970, the total value-added at 2010 prices will be equal to the sum of its components. A moment’s reflection reveals that additivity would be preserved only if the ratios between the sectoral deflators in 1970 were the same at 1963 and 2010 base years, which could happen only by chance. One important implication of the nonadditivity problem is that sectoral comparisons in levels at constant prices are not meaningful if the deflator’s base year is outside the subperiod considered in the analysis. Strictly speaking, the informational content of our series at constant 2010 prices, over the entire 1861–2010 period, is limited to their variations over time.

Notes 1. Baffigi and Brunetti worked at the second layer. Massimiliano Iommi (Istat) joined the project to contribute to the demand-side estimation. 2. The book by Fuà (1978) contains a statistical section prepared by Paolo Ercolani to which I refer in the following pages (Ercolani 1978). 3. We thank Guido Rey, Sandro Clementi, Paolo Piselli and Luisa Picozzi for providing us with their work before it was published, in particular with the data concerning the new benchmark for 1970. The article is now published as Rey (2012).

national accounts, 1861–2011

185

4. In this chapter I could not take into account Istat’s revision of the 1990–2011 national accounts data. However, data published on the Bank of Italy’s Web site will be regularly updated to the most recent Istat releases and to the most relevant quantitative results in the literature (see link in the references). 5. However, all results obtained and procedures used for the 1861–1970 period estimates are available for researchers in the already mentioned NA150 Excel file (see link in the references). 6. This is a tedious but important technical note. The differences between the two blocks of estimates (i.e., 1861–1970 and 1861–2010) are not limited to the disaggregation levels. Other differences arise for two reasons. The first and most important is related to the 1951–1970 period and is caused by the difference between the pre-ESA and ESA 1995 values for 1970. The second reason is caused by different treatment of the “financial intermediaries services indirectly measured” (FISIM) in the two estimates. Consistent with the benchmarks, in the 1861–1970 estimates FISIM are assumed to be consumed by a conventional unit, to avoid double counting; in the 1861–2010 estimates, consistent with the ESA 1995 standard, they are allocated to each sector in proportion of the importance of financial services for its activity. In the 1861–1970 case, to obtain GDP we have to deduct FISIM from the sum of sectoral values added; in the 1861–2010 case, GDP at factor costs is exactly equal to the sum of sectoral values added. Whereas the GDP level is the same in the two cases, sectoral values added are lower in the 1861–2010 estimates. 7. We estimated imports and exports of goods and services by interpolating the values of the benchmarks. Interpolations were achieved by using the data of imports and exports of goods reconstructed by Federico et al. (2011). 8. We assumed procyclicality of inventories, which were thus estimated with a common Baxter-King detrending procedure. 9. Fenoaltea (2011a) obtained reasonable estimates of the sectoral shares over the first five decades after unification by modifying his sectoral value-added shares at 1911 prices, taking into account the 1891 and 1911 benchmarks of Rey–Bank of Italy. Estimates of the shares for the period 1861–1880 are obtained by assuming similar productivity growth in the three sectors. His ingenious estimates of the sectoral shares do not require estimates of price deflators. 10. All Fenoaltea’s reconstructions are at constant 1911 prices. In 1911 the first census for Italy’s industry and commerce took place (see Fenoaltea 2011a). 11. Gerschenkron (1947). See also Ames and Carlson (1968), Jonas and Sardy (1970), and Scott (1952). Nicholas Crafts dedicates some important pages to this problem in his book about the British industrial revolution (Crafts 1985, 17 ff.). 12. De Bonis et al. (2011) contributed to the estimates of value-added of services with a study on the financial sector. 13. Baffigi’s and Brunetti’s calculations of value-added of services at constant prices are largely based on precious indications of Vera Zamagni. 14. These sources were not new in the literature (Mattesini and Quintieri 1997), but they had not been exploited for a new estimate of industrial value-added. 15. The work by Federico (2003c) explicitly focuses only on long-run variations in value-added. The implication is that movements at cyclical frequencies are not captured by the series. Federico’s series, however, is the best available reconstruction in line with the historiographic approach we have adopted in our project, not to mention the critical work on the sources behind that work. Also see Cerrito (2003) for a completely opposite view.

186

sources of growth and welfare

16. Fenoaltea’s estimates of value-added at constant prices (1911) are basically quantity indices, weighted with 1911 value-added per physical production unit. They are not obtained by deflating current price values. 17. Here point (2) of CASE 2 has been skipped because we could use Carreras and Felice’s deflators with no further elaboration. 18. This Appendix is part of a more general methodologic note on interpolation written by Alessandro Brunetti. 19. An estimate of investments in intangible goods was needed to link 1861–1970 with the 1970–2010 time series. 20. For 1970–2010, strictly speaking, we do not have constant prices times series. The 2010 base year is obtained by elaborating on the official Istat series at previous year prices. 21. The new series for investment has been built using the following sources. Road vehicles: 1928–1951, administrative register of road vehicles (Pubblico Registro Automobilistico, data available from the Web site of UNRAE, table “Il mercato degli autoveicoli dal 1920 al 2010, elaborazioni UNRAE su dati ACI e Ministero delle Infrastrutture e Trasporti”); 1911–1927, production of road vehicles (Rey 1991). Rolling stock: 1911–1951, production of rolling stock (Rey 1991) integrated with data from Amministrazione delle Ferrovie dello Stato, “relazioni per l’anno finanziario – various years.” Ships and boats: 1911–1951, launch of new boats and ships (“Navi varate”; Rey 1991). Airplanes: 1926–1951, business air traffic of Italian airlines companies (“Traffico aereo commerciale delle società di navigazione aerea italiane,” Istat 1986, tav. 14.13). Machinery and equipment: 1938–1951, investimenti fissi in Impianti attrezzature ecc. (Istat 1986, tav. 8.28). 1911–1938, imports of investment goods (imports of goods classified in the code 7 in the SITC classification, excluding codes 78 and 79).

Chapter 7

PRODUCTIVITY stephen n. broadberry, claire giordano, and francesco zollino

Introduction At the time of its unification in 1861, Italy was one of the poorest countries in Western Europe, after a long period of decline that lasted from the late Middle Ages to the nineteenth century (Malanima 2007). Whereas the center of economic gravity within Europe in 1300 had clearly been in the Mediterranean region and particularly in the city states of northern Italy, during the centuries after 1500, it had shifted northward, first to the Netherlands and by the nineteenth century to Great Britain, where the Industrial Revolution ushered in the transition to modern economic growth (Kindleberger 1996). As the Industrial Revolution spread to other parts of Europe and the New World, there was a danger that Italy would fall further behind. This Chapter examines Italy’s growth and productivity performance over the 150-year period since 1861, first in isolation, and then in an international comparative perspective. Italy is compared with the old and new technologic leaders, the United Kingdom and the United States, with a similar “late-unifier,” such as Germany, and with two Asian countries, India and Japan. The study makes use of the new estimates of Italian value-added, provided in Chapter 6 by Baffigi, broken down into ten sectors, so as to capture the dynamics of structural change. New estimates of labor and capital inputs are constructed by the authors, so as to identify the proximate sources of growth. The growth accounting exercise is carried out within an international comparative framework. After the country’s political unification, Italy achieved modest rates of per capita income and productivity growth. Structural change remained limited

188

sources of growth and welfare

before World War II, and Italy made little headway in catching up to the technologic leaders of the time. After the first twenty postunification years of stagnant growth, caused largely by a weak productivity performance in agriculture, Italy’s first growth spurt (1881–1911) was driven by manufacturing industry and by services (mainly trade, transport and communications, and credit and insurance). During and between the two World Wars, Italian labor productivity growth rates remained subdued. In particular, the Great Depression years were characterized by low industrial labor productivity growth across the board. Italy’s Golden Age began after 1945 and was propelled by manufacturing. Strong productivity growth was also registered in all other sectors, thanks to spill-over effects and to new technology generated in the industrial sector. A crucial factor here was the significant release of labor from agriculture, which moved into industry and services. Only Japan registered higher growth rates than Italy during this Golden Age. After a long period of catching-up, Italy overtook the United Kingdom in aggregate labor productivity terms during the 1970s, although there is some uncertainty about the precise year of il sorpasso.1 Since 1993, however, Italy has registered a striking productivity slowdown, compared with other countries and Italy’s past. Industrial growth has lost its previous impetus, but perhaps of more significance is the decline in service sector productivity growth at a time when services have come to dominate economic activity. In a sense, then, Italy seems to have come full circle: whereas in the first twenty years of its unified history low growth rates in the large agricultural sector held back aggregate growth rates, now the services sector is playing a similar role. It is tempting, looking at the aggregate data, to draw the conclusion that the slowdown was inevitable after Italy had exhausted its potential for catching up. However, the sectoral analysis gives more cause for concern. Structural factors seem to be at work here, with Italy failing to follow other countries in making effective use of information technology in services, which shows up in weak labor productivity growth in this sector and weak total factor productivity (TFP) growth in the economy as a whole. The Chapter proceeds as follows. The second section analyzes the time series evidence on Italy’s sectoral labor shares and labor productivity growth rates. An exercise in shift-share analysis helps disentangle the contributions of sectoral labor productivity growth and structural change to overall productivity growth. The third section illustrates the differences in the sectoral distribution of the labor force and in labor productivity growth rates in Italy, the United Kingdom, the United States, Germany, Japan, and India. It also focuses on comparative levels of labor productivity calculated at purchasing power parity (PPP), with the United Kingdom as the numeraire country. The fourth section performs a growth accounting exercise to gauge TFP’s contribution to aggregate growth. TFP growth rates are then compared with those registered in the other countries of the sample. The fifth section provides a brief discussion concerning

189

productivity

Italy’s most recent performance, in labor and TFP terms, when using the latest official data release of October 2011 (ISTAT 2011b), not included in the previous sections. The sixth section draws some conclusions. Finally, the Appendix explains how the comparative levels of productivity were constructed and considers ways of cross-checking the results.

The Contours of Italian Economic Growth Italy’s Labor Data The starting point for this comparative study is an analysis of Italy’s annual sectoral accounts on employment. We have compiled an entirely new dataset on labor, published in full in the Data Appendix at the end of this book.2 We provide two alternative measures: the headcount of total workers and the number of full-time equivalent workers. The territory covered by our data refers to Italy at current boundaries. Both measures are built at a ten-sector level of disaggregation3 and span the overall period 1861–2010.4 At the time of Italy’s political unification in 1861, half of the country’s total population was engaged in working activities.5 By 2010 this participation rate had dropped to around 41 percent (43 percent in 2007, before the recent crisis). However, as pointed out also by Daniele and Malanima (2011b), the decline was not monotonic. A minimum of 36.7 percent was attained in 1972 after which a weak recovery ensued, until current rates. The sectoral labor force shares changed even more dramatically. Figure 7.1 shows the percentage distribution of employment in agriculture, industry, and services over the 150 years considered. For ease of exposition, in this section we only discuss results based on our full-time equivalent measure, the theoretically preferable measure.6 At the onset of its unified history, nearly two-thirds of the total labor force worked in agriculture, and the remaining workers were similarly distributed between industry and services. Whereas until World War I the exodus from agriculture was limited, the 1930s, but mostly World War II, witnessed a significant shift of the labor force toward the nonfarm sectors. By the early 1950s the sectoral labor force shares had converged to the most balanced structure over the entire period, its “three thirds” point, as defined by the literature. Agriculture was still in the lead, however, with about 40 percent of the labor force devoted to it. By the end of the 1960s, the services sector was instead clearly dominant and continued to increase in importance until current times, in which it engages over two-thirds of workers against less than a third working in industry and with a tiny fraction engaged in agriculture—a complete reversal

190

sources of growth and welfare 100% 90%

Sectoral FTE shares

80% 70% Services Industry Agriculture

60% 50% 40% 30% 20% 10% 2001

1991

1981

1971

1961

1951

1941

1931

1921

1911

1901

1881

1891

1871

1861

0%

Years

Figure 7.1 Full-time equivalent labor shares in Italy, 1861–2010 (percentage shares). Source: see our FTE labor estimates in the Data Appendix at the end of this book.

of the picture relative to 1861. The most significant trend is the contraction in employment devoted to agriculture, coupled by a steady increase in importance of the services sector. Industry instead showed an inverted U-shape pattern, first steadily rising in importance and then declining after the 1970s oil shocks. By breaking down our data further, we find that industrial labor was nearly all employed in manufacturing. The construction industry was the second largest industrial sector, moving progressively, although not continuously, from 10 to 30 percent circa over the 150 years considered. The extractive and utilities industry were and remain tiny, only accounting for three percentage points at their peak. Employment within the services sector was instead more diffused. Trade and personal services were the largest sectors from the onset, roughly accounting for a total 60–80 percent of the aggregate services sector over the whole period. Transport and communications were also quite stable within a range of 10–20 percent. Labor engaged in the credit and insurance sector grew from approximately zero to the current 4 percent. Finally, government services peaked at 33 percent in 1971 and are currently around 22 percent.

Sectoral Labor Productivity Growth in Italy Our labor input figures, together with value-added data, may be used to calculate indices of labor productivity by major sector and for the aggregate economy. All labor productivity figures are based on an output per employee measure, because reliable data on hours worked are not available on a consistent basis at a sectoral level for much of the period under consideration.7 The sectoral value-added and gross domestic product (GDP) data, covering the period 1861–2010, on which we built, are the new estimates described in

191

productivity

Table 7.1 Italy’s output per full-time equivalent worker and GDP per head growth rates, 1861–2010 (percentage changes; yearly average in periods) Years

Agriculture

Industry

Services

Total Economy

GDP per Head

−0.1

0.2

0.6

1861–1881

0.3

0.4

1881–1911

1

2

1.9

1.5

1

1911–1938

0.8

0.5

0.1

0.6

0.9

1938–1951

1.9

2

−0.1

1.7

2.2

1951–1973

4.7

5.3

4.9

5.9

5.1

1973–1993

5.2

3.4

0.4

2.3

2.3

1993–2010

2.9

0.6

0.9

0.9

0.6

Source: Elaborations on VA estimates are from Chapter 6, FTE labor estimates are ours (see the Data Appendix.) Notes: The benchmark years chosen until 1951 coincide with selected census years, for which estimates are more robust. VA is computed net of housing and net of government services.

Chapter 6. However, in this section, value-added of the services sector and of the total economy has been calculated by us net of housing (for consistency reasons with our input data) and of government services (because of the methodology underlying the construction of value-added in this sector, which hinges on the cost of labor and hence makes the output per worker measure rather meaningless). We again use our full-time equivalent labor series. Average annual growth rates of sectoral labor productivity for selected subperiods of Italy’s history are thus presented in Table 7.1, alongside GDP per capita growth rates. Malanima (2007) claims that, from the Middle Ages (approximately 1300) until Italy’s unification, labor productivity steadily declined, until its lowest level was reached between 1810 and 1820. Our data show that after 1861 aggregate labor productivity had indeed begun growing again, confirming the trend reversal. However, in the first two decades after Italy’s unification, labor productivity growth was very modest, in fact the lowest ever achieved over the 150 years considered, because of weak growth rates in agriculture and industry, and of a negative productivity performance in the services sector, where employment growth outstripped that of value-added. Furthermore, in this period, overall working population grew faster than total population, hence explaining the higher GDP per capita growth rate. In 1881, nearly 52 percent of total population was active; this peak was never again attained in Italy’s unified history. Table 7.2 adopts a similar periodization to Table 7.1 to present annual average growth rates of labor productivity within the industrial and services sectors. The latter sectors immediately stand out as being highly diverse in terms of their labor productivity performance, thus underlining the importance of a more finely disaggregated analysis. In the immediate postunification years, manufacturing, utilities, and transport and communications,

Table 7.2 Italy’s industrial and services’ output per full-time equivalent worker growth rates, 1861–2010 (percentage changes; yearly average in periods) Years

Mining

Manufacturing Construction

Utilities

Trade

Transport and Communications

Credit and Insurance

Personal Services

1.2

−0.5

−0.6

1861–1881

−1.3

0.7

−4.6

1.9

−0.1

1881–1911

0.8

1.9

1.8

−0.1

2.5

1

1

0.3

1911–1938

−0.2

0.6

−4.2

7.7

−0.8

1.3

1.3

0

1938–1951

4

1.5

3.2

−0.8

2.6

2

−1.7

1951–1970

9.7

6.2

0.5

4.6

4

5

3.3

4

1970–1993

0

4.4

1

−0.3

1.2

2.2

−1.6

−0.7

1993–2010

−2.1

1

−1

2

0.2

1.8

1.5

−0.1

Source: VA estimates are from Baffigi (2011) for 1861–1970 and ISTAT (2011a) for 1970–2010; FTE labor estimates are ours (see the Data Appendix.) Notes: The periodization differs slightly to the one presented in Table 1, given the two different datasets here used for disaggregated VA.

0.2

productivity

193

recorded positive labor productivity growth rates. Yet all other sectors had no, or negative, productivity growth, thus explaining the overall performance of Italy’s economy. In the following period 1881–1911, Italy achieved its first productivity spurt with an overall labor productivity growth rate of 1.5 percent. Whereas the growth rates in agriculture increased to 1 percent per year, industry and services were even better achievers relative to the previous period, registering impressive rates of 2 and 1.9 percent per year, respectively. If one breaks the period down further (1881–1901 and 1901–1911), to account for different political regimes and to isolate the so-called Giolittian Era (i.e., the decade in which the Italian statesman Giovanni Giolitti was Prime Minister; see Chapter 2), one finds that the overall productivity growth rate was higher in the second subperiod, reflecting a faster growth rate especially in the industrial sector (2.5 percent), but also in services (2.2 percent). The acceleration, relative to 1861–1881, can anyhow already be seen across the board in the first subperiod. The three decades under study included a period of trade tariffs (1887–1894) and trade wars with France (1887–1898), yet Federico and O’Rourke (2000b), and James and O’Rourke in Chapter 2, found that Italian protectionism only affected total agricultural output marginally, by less than 5 percent. With a growing agricultural output and a declining engagement of labor in this sector, labor productivity growth rates in agriculture rose as a result. Transports in the services sectors, which were greatly enhanced by a boom in railway construction until 1895, by their backward linkages stimulated the extraction and construction industries and, after 1895, the manufacturing industry for the maintenance, repairs, and improvements of the railroads (Fenoaltea 2006, 196–199). When focusing only on the Giolitti period, public utilities played an important role, displaying annual average growth rates of 6.9 percent. These were the years in which the electrification of the country received a significant impulse.8 The shift in energy use from steam to electricity was particularly relevant in a country, such as Italy, in which coal was painfully lacking but in which water was abundant. The use of “white coal” loosened up Italy’s binding energy constraint and favored the country’s industrialization. However, the transformation of industrial processes as a result of the introduction of the new technology was neither immediate nor costless, and required the construction and extension of power transmission networks over the whole country.9 It also called for an adequate legislative framework that regulated the new technology and its transmission nationwide (Mori 1992). Productivity gains showed up with a lag only in the early twentieth century, and set the stage for Italy’s second productivity spurt after World War II. Finally, in the services sectors, from 1881 through 1911, annual average labor productivity growth rates were positive across the board, a result that has never since been repeated in Italy’s 150 year history. The three leading sectors were trade, transport and communications, and credit and insurance, sectors that traditionally accompany the production and consumption of industrial goods, the process of industrialization and urbanization, undertaken

194

sources of growth and welfare

by Italy in those years. Although the literature has mainly focused on industrialization as the main way out of economic backwardness (e.g., Williamson 2011a), the figures here presented show how the growth of the services sector also played a part in this period, especially if compared with its relative performance in previous years. Between 1911 and 1938, Italy’s economy slowed down once again and registered little labor productivity growth. Agriculture suffered the least, whereas industrial and services labor productivity growth fell, respectively, to 0.5 and 0.1 percent per year. If one breaks the period down further (1911–1929 and 1929– 1938) so as to set apart the Great Depression years from World War I and the 1920s, industry turns out to have performed well in the first subperiod, but very poorly in the 1930s. Conversely, labor productivity growth rates in agriculture and services were low (negative) during the war and the 1920s, but picked up (became positive) in the 1930s. Manufacturing and the already mentioned public utilities contributed to the positive, albeit low, overall industrial productivity growth rate in the three decades 1911–1938. However, zooming in on the Great Depression years, all industrial sectors suffered in productivity terms, relative to the previous subperiod (1911–1929), with only public utilities holding the fort. Within the services sectors, transport and communications and credit and insurance confirmed their leadership in productivity terms, even during the troubled 1930s. These were the years in which horse-drawn carts were gradually being replaced by trucks and lorries (Battilani, Bertagnoni, and Vignini 2008). Because of swift and “secret” bailouts of Italy’s main mixed banks during the 1930s, the country’s financial system was saved from collapse (Toniolo 1980); this is confirmed by no significant changes in credit and insurance productivity outcomes in those years. Productivity growth in trade was instead negative, especially during the Great Depression years, thus contributing to the aggregate negative growth rate. Conversely, the period 1938–1951 saw an increase in the overall labor productivity growth rate, which reached a yearly rate of 1.7 percent to which agriculture and industry positively contributed. The war years were actually years of negative growth in all three sectors, which makes the post-1945 growth even more remarkable. The substantial post–World War II increase in agricultural labor productivity growth rates confirms Federico and Malanima’s (2004) view of productivity soaring because of mechanization, but also because of substantial migrations from the countryside, reviewed in the next section. In industry, value-added increased vis-à-vis an unchanged aggregate workforce. In the services sector, productivity in the credit and insurance sector instead fell, mainly caused by a downturn in banks’ value-added during the war as a result of a high number of bank failures (De Bonis et al. 2012). The 1951–1973 period was Italy’s Golden Age, as the data clearly show. Labor productivity growth rates reached a hitherto unprecedented overall average yearly rate of 5.9 percent. This period markedly summarizes a success

productivity

195

story, relative to Italy’s economic record over the entire 150-year period, but also, as we shall see, in an international context; a success story that was propelled by industry. Agriculture and services too registered strong yearly growth rates in these two decades. The productivity boom provided a strong foundation for rapid improvement in living standards: GDP per head reached its highest ever yearly growth rate (5.1 percent). Breaking up the period further, so as to gauge the changed macroeconomic setting—full employment, rise in trade union strength, wage increases, inflation, balance of payments deficit—and to evaluate the policy shift—restrictive monetary and fiscal policies—in 1963, one finds that the acceleration in all sectors was even greater in the first subperiod (1951–1963) relative to the second subperiod (1963–1973). When examining the disaggregated data, all subsectors’ productivity grew at exceptional rates (e.g., manufacturing at 6.2 percent per year and transports and communications at 5 percent), with the only exception of the construction industry (0.5 percent). Mining, manufacturing, and transport and communications were the leaders. This result confirms the traditional view of manufacturing activities being the key to post–World War II growth, and stressing the rising importance of personal services with a collective nature, tied to health and education among others. Finally, not only transport and communications, but also distribution (wholesale and retail trade), strongly benefited from organizational and technologic transformation, made possible by significant road construction. During the twenty years after the energy crisis and the breakdown of the Bretton Woods system of fixed exchange rates (1973–1993), growth in agricultural labor productivity continued to increase, reaching 5.2 percent growth per year, mainly because of the shift of the labor-force away from this sector. Productivity growth, however, slowed down in industry and collapsed in the services sector, by then the largest sector even in headcount terms. The aggregate labor productivity growth rate (2.3 percent) was thus negatively affected by this composition effect. Manufacturing remained by far the most important driver of industrial productivity growth until 1993. Within the services sector, credit and insurance and personal services registered negative productivity growth rates. Transport and communications and, to a lesser extent, trade offset the negative performance of the former sectors.10 Finally, in the most recent period (1993–2010), whereas labor productivity in agriculture continued to grow at a strong rate (2.9 percent), although slower than in the previous periods, probably because of the exhaustion of the gains from the rationalization of this sector, growth in industry and services was far from impressive (0.6 and 0.9 percent, respectively). Industrial productivity growth was underpinned by the fastest-growing but small public utilities sector, and by the slower-growing but large manufacturing sector. In the services sector, we again find considerable heterogeneity in labor productivity growth rates, which further confirms the statement by Timmer et al. (2010, 13) that: “the treatment of the services sector as a homogenous and stagnant sector, in

196

sources of growth and welfare

Average annual growth rates

contrast to dynamic manufacturing, is no longer warranted.” Productivity in transport and communications and credit and insurance grew annually at 1.8 and 1.5 percent, respectively, compared with a 1 percent rate in manufacturing. Lackluster growth in trade and negative growth in personal services were two striking features of the period. The overall yearly output per worker growth rate dropped to 0.9 percent, whereas GDP per head growth fell to 0.6 percent per year, the same rate as that registered in the postunification years, the lowest ever. If one truncates the period considered at 2007, to net out the effects of the recent negative downturn, agriculture continues to be the leading sector (3.3 percent average yearly growth). Industry and services perk up a bit (1 and 1.2 percent, respectively). In particular, in the services sector, trade triples its labor productivity growth rate (0.6 percent), but the growth rate of personal services’ productivity actually becomes more negative (−0.2 percent). The overall economy 1993–2007 average yearly labor productivity growth rate only increases to 1.3 percent, a full percentage point lower per year compared with the previous period. A finer analysis points, however, to the turn of the millennium as leading to a true break in productivity trends. Annual average labor productivity growth rates in agriculture between 1993 and 2000 were in line with those registered in the previous two decades, but then dropped to 1.3 percent in the period 2000–2007. Industrial labor productivity growth had already slowed down in the 1990s (1.8 percent), but then fell to −0.1 percent in the 2000s. The overall economy labor productivity and GDP per head growth rates largely mirrored those recorded in the nonhousing private services (approximately 2 percent in 1993–2000 and 0.2 percent in 2000–2007). Finally, in the three years 2007–2010, labor productivity growth was positive only in agriculture. Industry suffered the most (−1.4 percent per year); the services sector was only marginally more virtuous (−0.6 percent). Overall labor productivity growth annually declined by 0.9 percentage points and GDP per head by 0.8. 7.0% 6.0% 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% 1950–1973

1973–1995

1995–2007

Years Value added per hour worked

Value added per FTE worker

Figure 7.2 Post–World War II labor productivity growth rates: Value added per full-time equivalent worker and per hour worked (percentage changes; yearly average in periods). Sources: See Table 7.1 for the VA and FTE labor estimates’ sources; The Conference Board (2011) for hours worked. Note: Benchmark years are those reported in Chapter 3.

productivity

197

Our labor productivity estimates have been thus far computed on an output per worker basis. However, this definition may be misleading in a context in which yearly work-time varied significantly over the decades.11 Yet, because of data availability we can only account for these changes for the most recent years, and at an aggregate level. Therefore, for the post–World War II period, we can compare our total economy labor productivity growth rates with output per hour worked growth rates, as presented in Chapter 3 and in turn based on Conference Board (2011), to verify the presence of any major discrepancies. As Figure 7.2 shows, the two growth rates are reassuringly quite similar.

Structural Change in the Italian Economy What has emerged from the previous section is that until the end of World War II, Italy grew very little in terms of per capita GDP and labor productivity; an outstanding acceleration in growth rates was instead attained thereafter, when Italy evolved from an agricultural to an industrial and, after the 1970s, into a services economy. In explaining the slow growth rates before the 1950s, a sectoral analysis once again comes to our aid. Positive correlations between increasing levels of productivity in agriculture and the rise of an urban industrial economy are well understood. Before 1951, in Italy, the low agricultural labor productivity we have seen documented in Table 7.1 was coupled with high land productivity and low land endowment per worker. Figures in O’ Brien and Prados de la Escosura (1992, 529) point to a more or less stable land/labor ratio of 3.3 in Italy until 1950, significantly lower than that registered in other countries, such as the United Kingdom. The same source points to output per hectare in Italy being instead relatively high. Italy’s long-term growth was constrained by a combination of high population density and an unfavorable land/labor ratio: low labor productivity in agriculture can thus be attributed to overcrowding and congestion, and not to the scarcity of fertile land. This conclusion implies that Italy could only grow if its primary sector were to shed labor. We have already seen the evolution of sectoral employment shares in Italy in Figure 7.1. In contrast to more developed countries, the higher share of agricultural labor force in Italy in the late nineteenth century meant that the latter country had larger scope for net gains linked to the reallocation of resources. Given the lower level of value-added per employee in agriculture and the higher level in industry and services, structural change was the key to kick-start Italy’s development process. In the present section, we time and quantify the structural change of Italy’s economy over 150 years by performing an exercise in shift-share analysis. The aggregate growth rate of labor productivity is not merely the average of sectoral productivity growth rates with constant weights; it is the result of changes

198

sources of growth and welfare

of the labor-productivity ratio within individual sectors, but also of systematic shifts of employment shares between sectors. The basic approach is derived from Nordhaus (1972), in which the aggregate productivity growth is broken down to disentangle its two structural components, in turn modified by Broadberry (1998).12 The resulting decomposition is the following: ^ 0

/

0

∑

i {A , I ,T }

∑

^

i

(VA VAi / VA V o)

i {A,II ,T }

^

VA V Ai / VA V o (Li / Li

^

Lo / Lo )

(1)

Where: ^

^

^

α i = X i / Xi − (L o Li ) ^

α i = X i / Xi And: X0 = level of aggregate labor productivity; L0 = aggregate employment; i = sector of origin (A = agriculture; I = industry; T = tertiary sector); Si = share of employment in sector I; Xi = productivity level in sector I; VAi = value- added in sector I; Li = employment in sector I; and hats above variables denote time derivatives. The first term on the right-hand side of equation (1) is the “pure productivity” (Nordhaus 2001), “direct productivity” (Stiroh 2002), or “within effect” (Antonelli and Barbiellini Amidei 2007). It is a weighted average of the productivity growth rates in component sectors, where the weights are period-average nominal value-added shares of each sector. As the productivity in one sector grows, aggregate productivity rises in proportion to the sector’s size. The second term is the “reallocation effect” (Stiroh 2002) or the “Denison effect” (Nordhaus 2001), after Edward Denison, who was the first to point out how the shift from a low-productivity-level sector to a high-productivity-level sector raises productivity even if the growth rates in the two sectors are the same (Denison 1967). Broadberry (1998)’s correction, summarized by the variable α, implies that, in a declining sector, the actual productivity growth rate is reduced by the difference between the growth rate of the aggregate labor force and the growth rate of the labor force in the particular sector, whereas in expanding sectors the actual productivity growth rate is used.13 The results of the modified shift-share calculations for Italy over key subperiods are given in Figure 7.3. The corrections related to the modified shift-share analysis are large mainly in the 1973–1993 and 1993–2010 periods, when shares in not only agriculture but also in industry were declining. The shift from industry to services characterized the process of structural change since the 1970s.

199

Annual average growth rates

productivity 4.50% 4.00% 3.50% 3.00% 2.50% 2.00% 1.50% 1.00% 0.50% 0.00%

19 3 99 –1

3 97 –1

0 01 –2

73

51

93

19

19 1 95 –1

8 93 –1

1 91 –1

1 88 –1

38

11

81

61

19

19

18

18

Years Internal productivity growth

Structural change

Figure 7.3 Shift-share analysis of Italy’s labor productivity growth rates, 1861–2010 (percentage change; yearly average in periods). Sources: See Table 7.1. Note: This analysis is the result of a standard shift-share exercise, modified as in equation (1). VA here includes government services.

What we can tease out from Figure 7.3 is that in periods of low productivity growth (i.e., the first twenty years after Italy’s political unification, the World War I and Great Depression period) structural change accounted for the bulk of aggregate labor productivity growth. Therefore shifts of employment away from agriculture to higher labor productivity level sectors managed to boost the overall productivity rate in these critical subperiods. The years of Italy’s first spurt (1881–1911) and of its “economic miracle” (1951–1973), which actually began in 1945, were instead characterized by large direct productivity growth in industry and services sectors. The 1973–1993 period growth was instead characterized by an equally balanced internal growth and structural change. In the most recent period, structural change accounted for only 17 percent of overall productivity growth, given the exhaustion of gains obtained from a shift of labor force out of agriculture. In the long run (1861–2010), structural change explained approximately one-fourth of Italy’s labor productivity growth.

Italy’s Productivity Performance in International Perspective So far, we have focused on Italy’s growth performance during different periods since 1861. In this section, we place that performance in an international perspective by making comparisons with a sample of other countries. Because

200

sources of growth and welfare

it is widely recognized that economic backwardness provides scope for relatively fast catch-up growth, it is important to consider levels of productivity and growth rates (Gerschenkron 1962a; Abramovitz 1986; Baumol 1986). Because levels and growth rates of productivity may vary among agriculture, industry, and services, we need to consider performance in all three major sectors, and the total economy-wide performance (Broadberry 1998, 2006).14 The sample of countries chosen includes the United Kingdom, the United States, Germany, Japan, and India. The United Kingdom was the first industrialized country, the per capita income leader during the nineteenth century, and the richest European country for much of the twentieth century, and has thus been widely used as the numeraire country in international comparisons of productivity, particularly those involving European countries. As the productivity leader during the twentieth century, the United States is included to represent the technologic frontier. Germany attained its political unification in the second half of the nineteenth century (in 1871), similarly to Italy. However, in contrast to Italy, Germany quickly emerged as a major industrial power to challenge Britain’s industrial dominance and remains today Europe’s largest industrial producer and exporter. In addition, we have included two Asian countries to provide a more global perspective. Japan was the first nonwestern country to industrialize, after the Meiji Restoration in 1868, which can be seen as a major institutional shock similar to Italian unification in 1861. India provides an example of a much less developed country, notwithstanding its recent impressive growth performance and its large size in the world economy as a consequence of its massive population (Bosworth and Collins 2008). For data, we rely mainly on the historical national accounts of the previously mentioned countries, which we explain in detail in the Data Appendix of Broadberry, Giordano and Zollino (2011), to which we refer. In our international comparisons we use the gross services and total economy value-added figures, as reported in Chapter 6, together with our headcount labor data, for reasons of international comparability. As we shall see, the resulting labor productivity growth trends are not significantly different from the ones presented in the previous section.

The Structure of Employment Before we analyze productivity performance in the different parts of the Italian economy in international comparative perspective, it is instructive to note the changing structural balance of our six sample economies since the late nineteenth century. The data showing the breakdown of the labor force among agriculture, industry, and services are given in Table 7.3. The Italian data show the classic pattern of development noted by Clark (1951) and Kuznets (1974), with the economy dominated by agriculture at low levels of development, followed by a phase of industrial-led development and leading ultimately to a dominance of services. Note that agriculture still accounted for more than half of

Table 7.3 Sectoral shares of employment (headcount), 1870–2007 (percentage shares) A. Italy

Agriculture

Industry

Services

B. United Kingdom

Agriculture

Industry

Services

1871

68.1

15.8

16.2

1871

22.2

42.4

35.4

1911

59.1

23.5

17.4

1911

11.8

44.1

44.1

1921

59.1

22.5

18.4

1924

8.6

46.5

44.9

1931

53.8

25.4

20.8

1930

7.6

43.7

48.7

1936

52

25.6

22.5

1937

6.2

44.5

49.3

1951

44.3

31

24.8

1950

5.1

46.5

48.4

1973

17.7

38.4

43.9

1973

2.9

41.8

55.3

1993

6.6

31.3

62.2

1990

2

28.5

69.5

2007

3.9

27.9

68.2

2005

1.4

18.4

80.2

C. United States

Agriculture

Industry

Services

D. Germany

Agriculture

Industry

Services

1870

50

24.8

25.2

1871

49.5

29.1

21.4

1910

32

31.8

36.2

1913

34.5

37.9

27.6

1920

26.2

33.2

40.6

1925

31.5

40.1

28.4

1930

20.9

30.2

48.9

1930

30.5

37.4

32.1

1940

17.9

31.6

50.5

1935

29.9

38.2

31.9

1950

11

32.9

56.1

1950

24.3

42.1

33.6

1973

3.7

28.9

67.4

1973

7.2

47.3

45.5

1990

2.5

21.8

75.7

1990

3.4

39.7

56.9

2005

1.5

16.7

81.8

2005

2.1

25.5

72.4

Table 7.3 (cont.) E. India

Agriculture

Industry

Services

F. Japan

Agriculture

Mining/ Construction Manufacturing

1875

73.4

14.5

12.1

1891

75.8

9

1.4

1910/1911

75.5

10.3

14.2

1920

55.4

16.2

2.8

1929/1930

76.1

9.1

14.8

1950

48.3

17.6

4.3

1950/1951

73.6

10.2

16.2

1973

16

27.3

9.3

1970/1971

73.8

11.1

15.1

1990

9.2

23.5

9.2

1999/2000

64.2

13.9

21.9

2007

5.1

17.4

8.4

F. Japan

Facilitating Industry

Services

1891

1

12.8

1920

3.6

22

1950

5.1

24.7

1973

6.3

41.1

1990

6.2

51.9

2007

6.4

62.7

Source: For Italy, headcount labor estimates are ours (see the Data Appendix at the end of this book); for other countries see the Data Appendix in Broadberry, Giordano and Zollino (2011).

productivity

203

all employment in 1936, whereas industry continued to increase its share until 1973. This general pattern of development can also be seen in the data for the United Kingdom, the United States, Germany, and Japan, but only to a much lesser extent in the later-developing India. Despite the general pattern, there have been some substantial differences between countries in the timing of the release of labor from agriculture. Italy’s structural transformation away from agriculture occurred much later than in the United Kingdom, where the share of employment in agriculture in 1871 was just 22.2 percent. Italy took almost another century to reach this level of development. Italy’s pattern of structural change was more similar to that of the United States and Germany, where agriculture continued to account for around half of all employment in 1870–1871. The similarity becomes even closer when Italy’s development trajectory is compared with that of Japan, where agriculture continued to account for around half of employment until after World War II. Finally, Italy’s pattern of structural change clearly looks much more developed than that of India, where agriculture continued to account for nearly two-thirds of employment at the end of the second millennium. It is also worth noting in Table 7.3 some differences in the relative importance of industry and services as labor shifted out of agriculture. As the first industrial nation, the United Kingdom accounted for a large share of world industrial exports and production in the nineteenth century and the first half of the twentieth century, and hence redeployed a large share of its labor force into industry. As Germany industrialized from the late nineteenth century, it also built up a large export business and hence transferred a large share of its labor force from agriculture to industry. Although the United States also enjoyed industrial export success, exports accounted for a smaller share of economic activity than in the more open European economies. Combined with the high levels of labor productivity achieved in US industry already by the late nineteenth century, this meant that industry did not account for as large a share of employment as in Germany or the United Kingdom. The sectoral breakdown is slightly different in the case of Japan, where it is not possible to provide a clean break between industry and services before World War II because of the inclusion of gas, electricity, and water together with transport and communications in facilitating industry. Nevertheless, the growing success of Japanese industry in export markets is reflected in the rising share of mining and manufacturing in employment. All these countries seem to have followed an industry-led development process until at least World War II. After 1950, the share of industry began to decline in the United States and the United Kingdom, with services becoming the most dynamic sector. However, in Germany and Japan, industry continued to expand its share of employment until 1973, and Italy also followed this pattern. Increasingly, services have come to dominate the employment structure. Services were already the largest sector in the United Kingdom by 1930 and in the United States, where industrial labor productivity was exceptionally high, as early as in 1870. In Germany, services came to employ more people than

204

sources of growth and welfare

industry only after 1973, and Italy was closer to the German than the UK case, with services becoming larger than industry just before 1973. The cases of India and Japan provide an interesting contrast with the European economies considered here, with both economies showing relatively large service sectors at early stages of development.15 A number of conclusions follow immediately from this evidence on the sectoral distribution of labor. First, because agriculture dominated economic activity in most economies during the late nineteenth century, low productivity growth in agriculture at this time must mean low productivity growth in the economy as a whole. Second, although industry never came to play as dominant a role in total employment as agriculture, its importance did clearly increase in the first half of the twentieth century, so that achieving high productivity growth in industry became an important determinant of overall productivity growth performance. Third, during the second half of the twentieth century and into the twenty-first century, high productivity growth in services has become essential for high productivity growth overall, as services have come to exercise the kind of dominance over economic activity exerted by agriculture during the nineteenth century.

Labor Productivity Growth by Sector The overall pattern of Italian labor productivity growth has already been noted. During the pre–World War II period, growth of total value-added per employee was generally quite modest, but with some periods slightly faster than others. The fastest annual labor productivity growth in Italy before World War II was recorded between 1881 and 1911. There then followed a period of very rapid productivity growth between 1951 and 1973. Since then, the Italian labor productivity growth rate has declined, particularly since 1993. In assessing which sectors account for these variations in Italy’s overall productivity growth in panel A of Table 7.4, it is important to bear in mind our previous observations on the relative size of the major sectors. Before 1881, agriculture experienced the fastest labor productivity growth in Italy and was also the largest sector, so there can be little doubt that industry and services played a subsidiary role during this period. Between 1881 and 1973, Italian industry recorded the highest rate of labor productivity growth in three of the four subperiods. Because industry was increasing its share of employment substantially in Italy during these years, this is suggestive of a period of industry-led development. Since 1973, the slowdown in Italian productivity growth can be accounted for by a sluggish labor productivity growth rate in services, which has seen a dramatic increase in its share of employment. However, before we rush to conclude that Italian agriculture performed well before 1881, or praise Italian industry between 1881 and 1973, or indeed condemn the performance of Italian services since 1973, we need to place this

205

productivity Table 7.4 Output per worker (headcount) growth rates, 1870–2007 (percentage changes; yearly average in periods) A. Italy

Agriculture

Industry

Services

Total Economy

1861–1881

0.5

−0.1

0

0.4

1881–1911

0.8

1.9

1.6

1.3

1911–1938

1.3

1.1

−0.1

1.2

1938–1951

1.9

2.1

1.6

2.1

1951–1973

5.7

6.9

3.8

6.2

1973–1993

6.4

3

0.5

2.1

1993–2007

3.2

1

0.2

0.6

B. United Kingdom

Agriculture

Industry

Services

Total Economy

1861–1881

1

1.8

0.5

1.3

1881–1911

0.2

0.5

0.3

0.4

1911–1938

1.7

1.9

0.1

1

1938–1951

2.7

0.9

0.5

0.8

1951–1973

5

2

1.3

2.5

1973–1993

3

3

0.8

1.8

1993–2007

2.6

2

1.8

1.8

C. United States

Agriculture

Industry

1869–1879

1.7

1

0.9

1.9

1879–1909

0.8

1.6

1.1

1.4

1909–1937

1.4

1.8

0.2

1.2

1937–1950

4

2.4

1.8

2.4

1950–1973

5.6

3.1

1.4

1.9

1973–1990

4.4

0.8

0.5

0.4

1990–2007

2.2

2.5

2.1

1.9

D. Germany

Agriculture

Industry

Services

Total Economy

1871–1881

0.3

1.5

0.5

0.8

1881–1911

1.3

1.7

1

1.6

1911–1937

1.0

0.9

0.6

1

1937–1950

−0.4

0.1

0

0.1

Services

Total Economy

206

sources of growth and welfare

Table 7.4 (cont.) D. Germany

Agriculture

Industry

Services

1950–1973

6.3

4.9

3.1

4.2

1973–1990

6

2

1.5

1.8

1990–2007

4.1

3.9

1.2

2

E. India

Agriculture

Industry

Services

Total Economy

1872/1873–1900/1901

0.4

1.1

0

0.4

1900/1901–1946/1947

0

1.4

1

0.5

1950/1951–1970/1971

0.9

3.4

2.8

1.9

1970/1971–1999/2000

0.9

2.7

2.3

2.5

F. Japan

Agriculture

1891–1920

2.3

3.2

0.3

4.6

1920–1950

0.4

1.4

1.3

−0.2

1950–1973

4.9

8.9

4.3

1973–1990

2.3

4

1990–2007

2.5

3.4

Total Economy

Mining/ Construction Facilitating Services Total Manufacturing Industry Economy 0.3

2.6

1

1

7.7

3.1

6.6

1.5

2.6

1.9

2.8

−1.9

1.4

0.9

1.5

Sources: For Italy, see the Data Appendix at the end of this book for VA estimates (described in Chapter6) and for our headcount labor estimates; for the other countries, see the Data Appendix in Broadberry, Giordano and Zollino (2011).

Italian experience in international perspective. It is easier to achieve rapid growth while the rest of the world is booming rather than while it is stagnating or contracting. The first point to note is that labor productivity growth in Italian agriculture between 1861 and 1881 was not exceptional by international standards. Indeed, it was substantially exceeded by the United Kingdom and the United States. Second, labor productivity growth in Italian industry stands out as being noticeably higher than in other countries only during two subperiods. Italy recorded the highest rate of labor productivity growth in industry between 1881 and 1911, although German industrial labor productivity grew almost as quickly at this time. Furthermore, although industrial labor productivity growth during the period 1951–1973 was substantially higher in Italy than in the United Kingdom, the United States, India, and Germany, it was nevertheless surpassed by the even more impressive performance of Japanese mining and manufacturing. A third point worthy of comment is that Italy’s labor productivity growth in services since 1973 has been strikingly slower than in all other countries, particularly since 1993. Given the growing dominance of services in economic activity, it is difficult to avoid the conclusion that this must

productivity

207

have been a major factor in Italy’s overall slower labor productivity growth at this time. However, before seeing this as a failure of Italian services, we need to consider comparative labor productivity levels, because it is also widely accepted that economic backwardness opens up opportunities for rapid catch-up growth. The flip side of the coin is that starting from high levels of productivity makes it harder to achieve rapid rates of productivity growth.

Comparative Labor Productivity Levels Table 7.5 provides an overview of Italy’s comparative labor productivity level between 1871 and 2007, broken down by sector. The United Kingdom is the numeraire country, with the UK labor productivity level taking a value of 100 in all years and in all sectors. The Appendix to this Chapter explains how we constructed these comparative productivity levels. An important part of the methodology involves the use of more than one benchmark to ensure the consistency between information on comparative levels and growth rates of labor productivity. The GDP column in panel A of Table 7.5 informs us that at the level of the economy as a whole, Italy failed to catch-up to the United Kingdom at all between 1871 and 1936, had embarked on a catching-up trajectory by 1951, and had overtaken it by 1973. Since 1993, however, the United Kingdom has been catching-up to Italy. Note that this levels analysis, even at the aggregate level, already takes some of the sting out of the more critical commentaries on recent Italian productivity performance. Slower productivity growth since 1993 can be seen as a result of the end of catching-up, with Italy reaching the technologic frontier. Furthermore, the breakdown by sector reinforces this conclusion, because services, where the labor productivity growth rate has most obviously lagged behind rates achieved in other countries, had achieved the largest productivity lead during the 1970s. This optimistic reading of recent Italian productivity performance needs to be qualified in light of the US/UK comparison, which is presented in panel B of Table 7.5. Here we see that at the aggregate level, the US labor productivity lead over the United Kingdom remains substantial. Furthermore, the US productivity lead in services has remained substantial, whereas the Italian productivity lead over the UK has evaporated since 1993. One factor often used to explain the high level of labor productivity in US services is the widespread use of information and communications technology, which has sometimes been seen as relatively slow to diffuse in Italy (Timmer et al. 2010).16 The US/UK comparison also helps to put the achievements of Italian industry during the twentieth century into perspective. During most of the first half of the twentieth century, the average Italian industrial worker produced less than half the output of his or her British counterpart, who in turn produced around half the output of the average American industrial worker. This large transatlantic labor productivity gap has widely been attributed to differences in technology, with high-throughput

208

sources of growth and welfare

Table 7.5 Comparative labor productivity levels by sector, 1870–2007 (UK = 100) A. Italy

Agriculture

Industry

Services

Total Economy

1871

39.5

44.6

49.3

37.6

1881

40.8

33.8

46

35

1901

38.6

38.5

56.8

36

1911

42.4

47.7

64

41.3

1921

37.6

41.2

55.8

36.7

1931

36.5

39.2

61.1

39.5

1936

29.4

35.4

53.9

35.4

1951

30.1

42.2

68.5

46.5

1963

36.6

82.7

92.2

77.5

1973

34.5

117.4

119.4

101.6

1993

66.7

117.2

112.6

106.4

2007

73.1

102.2

90

89.6

B. United States

Agriculture

Industry

Services

Total Economy

1869/1871

86.9

153.6

85.9

89.8

1879/1881

98.1

149.8

87.9

95.9

1889/1891

102.1

164.1

84.2

94.1

1899/1901

106.3

174.7

104

108

1909/1911

103.2

193.2

107.4

117.7

1919/1920

128

198

118.9

133.3

1929

109.7

222.7

121.2

139.4

1937

103.3

190.6

120

132.6

1950

126

243.5

140.8

166.9

1960

153.1

250.4

137.7

167.9

1968

156.7

248.1

139.6

164.2

1973

131.2

214.8

137.4

152.3

1979

156.1

186

137.2

145.5

1985

146.9

161.1

134.1

134.8

1990

151.1

163

129.6

133

2007

196.4

166.2

125.1

127.7

209

productivity

C. Germany

Agriculture

Industry

Services

Total Economy

1871

55.7

91.7

62.8

59.5

1881

54.7

93.7

61.3

57.3

1891

53.7

99.3

64.4

60.5

1901

67.2

105

71.9

68.4

1911

67.3

127.7

73.4

75.5

1925

53.8

92.3

76.5

69

1929

56.9

97.1

82.3

74.1

1935

57.2

99.1

85.7

75.7

1937

59

96.9

89.4

79.2

1950

41.2

91.8

83.2

74.4

1960

47.8

117.9

102.6

94.5

1968

48.6

121.9

115.9

107.1

1973

50.8

121.1

120.1

114

1979

65.5

132.8

131.8

126.5

1985

62.1

114.8

131.6

120.9

1990

75.4

111

134.9

125.4

2007

92.4

105.2

117.5

114.6

D. India

Agriculture

Industry

Services

Total Economy

1871/1873

11.2

18.2

18.1

15

1881/1883

11.3

16.8

15.9

14.1

1890/1891

10.4

17.3

15.6

13.8

1900/1901

10.5

18.6

15.6

13.2

1910/1911

11.1

24.2

17.7

14.4

1920/1921

9.8

21.1

21.1

13.4

1929/1930

8.3

25.3

25.2

14.2

1935/1936

7.1

21.8

23.2

12.8

1946/1948

7

18.1

23.5

11.7

1950/1951

5.4

14.6

17.5

9.3

1960/1961

4.3

16.4

20

9.7

1970/1971

2.3

17.3

22.6

8.9

1980/1981

1.6

16.1

29.3

10.2

1990/1991

0.9

18.3

33

11

1999/2000

1

15.8

32.8

11.4

210

sources of growth and welfare

Table 7.5 (cont.) E. Japan Agriculture Mining/ Construction Facilitating CommerceManuf Industry Services

Total Economy

1891

20.5

14

75

32.4

33.5

16.5

1901

24.7

16.6

64.7

50.8

36.6

19.7

1911

27.8

20.5

94

65.5

33.5

22.5

1920

29.7

30.9

74.4

95.9

31.9

29.1

1929

28

35.5

59.3

121

33.1

31.6

1935

25.2

35.9

89.8

124.2

30.5

32

1950

15.5

25.4

75.1

49.2

46.1

28.8

1960

18.9

46.6

109.4

80.7

45.2

43.6

1973

15.1

86.8

147.3

109.2

76.4

75.5

1979

13.2

98.1

156.4

107.7

83.7

83.9

1990

12.7

94.6

162.8

108.5

97.9

92.4

1997

14.2

89

92.9

79.6

95

86.2

2007

13.4

96.2

94.5

65.9

88.1

84.9

Sources: See Table 7.4.

or mass production techniques developed in the United States, but difficult to apply in European conditions because of differences in factor endowments and demand conditions (Hounsell 1984; Broadberry 1997a; Chandler 1990). The comparison of Germany with the United Kingdom suggests that the scale of the Anglo-Italian productivity gap in agriculture may be explained more by a strong UK performance than by a weak Italian performance, because the AngloGerman productivity gap in agriculture also remained quite large until very recently. Furthermore, Italian productivity in agriculture has been substantially higher than in Asia throughout the period under consideration. It is perhaps not surprising that agricultural labor productivity has been very low in India, but the scale of the agricultural productivity gap in Japan has also been very large.

Accounting for Italian Economic Growth In this section we perform a standard growth accounting exercise, with a two-fold aim. First, further insight into labor productivity dynamics can be

211

productivity

provided by examining the evolution of Italy’s capital intensity and TFP over the period 1861–2010. Second, the computed residual, in broad terms, reflects the development of Italy’s ability to innovate, and its organizational and institutional changes. Together with evidence of changing contributions coming from primary inputs to GDP growth, the analysis provides additional insights into the restructuring process of the Italian economy in a historical perspective. This could provide a benchmark against which the dismal performance in more recent times can be compared and better appraised in its intensity and duration.

The Theoretical Framework We have followed the standard theoretical reference, namely the basic neoclassical growth model described in Solow (1957), to disentangle the contributions of labor, capital, and technologic change to the development of the Italian economy. In particular, we have assumed that output is described ^by a standard neoclassical Hicks-neutral production function. TFP growth, A/ At (where hats denote time derivatives), is thus computed as a residual, as the difference ^ between output growth ( Y / Yt ) and the weighted average of the growth rates of factor inputs, labor (Lt), and capital (Kt), as presented in equation (2): ^

A / At

^

Y / Yt − [

^

t

Lt / Lt + (

^

t

) K / Kt ]

(2)

where αt = w tLt/Yt is the labor share of output, w t is the unit wage, and (1-αt) is capital’s share of output under the assumptions of perfect competition and constant returns to scale. Although it is agreed that, for example, distortions from imperfect competition, externalities, omitted inputs, and nonconstant returns to scale confound the interpretation of this residual as a pure technology measure, “it remains a useful indicator of the underlying technological factors” (Stiroh 2001). Basu and Fernald (2002), for instance, find a high correlation between a traditional Solow residual and a more sophisticated index of technology that controls for market imperfections. However, a further caveat must be stated in our exercise, because of the very large time-span we consider: since 1861, the market structure and the institutional setting have undergone radical changes in Italy, as everywhere in the world, hence potentially adding noise in the residual approach to TFP measurement. To implement the growth accounting exercise, first we have constructed a historical series of physical capital stock in Italy over the period since 1861. We refer to the Data Appendix in Broadberry, Giordano, and Zollino (2011) for a discussion of the sources and methodology used. We computed net and gross capital stock, at current prices and at chained values, for four different assets (machinery, infrastructure, and equipment; means of transport; nonresidential

212

sources of growth and welfare

construction; and housing), published in full in the Data Appendix at the end of this book. Because sectoral investment data do not exist for the overall period, we could only compute the stock of capital for the overall economy, government services included. Based on our estimates, we find that Italy experienced important changes in capital composition as economic development deepened. In the early stage, asset substitution took place mostly from construction, in particular nonresidential structures, to machinery and equipment, and to a lesser extent, to means of transport. In the first decade of the twentieth century a housing upsurge began, against a continuing drop of the share of nonresidential construction in total capital and a roughly stable profile of the other assets. This pattern changed in the late 1960s, because machinery, equipment, and means of transport resumed a positive trend, which was offset by a declining share of housing as the downward correction of nonresidential structures came to a halt. However, for the purpose of the growth accounting exercise, we focus solely on productive assets, thus ruling out the housing sector from the output (once again) and the input sides. Secondly, following Jorgenson (2001) we estimated the rental price of single productive assets to control for the possible trend in the quality of productive services they provide over time. For this purpose, for each asset i we have first computed the rental price uit 17 and then calculated the changes in capital input as a Divisia index: n

kt / kt

∑


/ si ,t

(3)

i ,t i ,t

i =1

Where: vi ,tt

1⎛ u S 2 ⎝ i ,t −1 t

∑u

i ,t

St −

ui ,t St

∑u

⎞ S⎟ ⎠

i ,t t

and si is the log of the chained values of the net stock of asset type i (Si), vit is the respective share on total returns to capital, and hats once again denote time derivatives. Rather than simply summing up net stocks of different assets, the Divisia index of capital input controls for the possible upgrading in the quality of capital as it implicitly assigns relatively larger weights to changes in the more productive (or short-lasting) assets than to the less productive (or long-lasting) ones.18 Thirdly, we computed the factor shares, filling the gap between 1861 and 1951, the latter being the first year for which data are available in the received literature. For this purpose we have constructed a series of unit wages by aggregating all information available for single activities over different periods (again refer to the Data Appendix of Broadberry, Giordano, and Zollino [2011] for the primary sources used). Based on unit wages and total employment, we retrieved the wage and profit shares in total value-added as the two sum to one under the standard assumptions of constant returns to scale and perfect competition.

213

productivity

Our data show that the profit share in Italy started off high in the immediate postunification years, at a time when the capital-output and capital-labor ratios were low. Although with some fluctuations, particularly pronounced in war times, the profit share tended to stabilize after the late 1890s at around half its original value. A further downward correction started in the early 1970s, which was temporarily, and only partially, reversed during the 1990s. In the meantime, the capital intensity of production and the ratio of capital to income have shown a clear, positive trend.

Aggregate TFP Development in Italy Focusing on the total economy, and netting out housing rental services on the output side and residential stock on the input side because of their negligible contribution to growth potential, our results in Table 7.6 show that in the first twenty years of unified Italy capital accumulation provided the greatest contribution to GDP growth, partly thanks to a sizeable upgrading in the quality of productive assets. Employment creation also contributed positively to output growth, against the drag caused by a decline in TFP. Because the latter usually proxies disembodied technical progress and efficiency gains in the production process, its negative development could signal the transitory distress caused by the revision of organization and management practices in the face of the increasing migration of workers out of agriculture and the expanding labor force. Indeed, TFP growth gained momentum thereafter and explained Table 7.6 The sources of growth of the Italian economy, 1861–2010 (percentage changes; yearly average in periods) Contribution Contribution of which asset Changes in substitution TFP of labor of capital services

Years

Changes in GDP

1861–1881

1.3

0.7

1.1

0.4

−0.5

1881–1911

1.8

0.3

0.9

0.1

0.6

1911–1938

1.7

0.6

−0.3

−0.1

1.3

1938–1951

3.1

0

1

0.2

2.1

1951–1973

6

1

1.7

0

3.3

1973–1993

2.5

0.5

0.9

0.1

1.2

1993–2010

1.1

0.2

0.6

0.1

0.3

1861–2010

2.5

0.5

0.8

0.1

1.2

Sources: See Table 7.1 for GDP and labor estimates; see the Data Appendix for our capital stock estimates. Note: GDP and capital are both net of the housing sector, but include government services.

214

sources of growth and welfare Capital deepening

Changes in TFP

4.00 3.00 2.00 1.00 0.00

0 01 –2 61 18

0 01 –2 93 19 3 99 –1 73 19 3 97 –1 51 19 1 95 –1 38 19 8 93 –1 11 19

–1.00

1 91 –1 81 18 1 88 –1 61 18

GDP per FTE worker average annual growth rates (% changes)

5.00

Years

Figure 7.4 The contribution of TFP and capital deepening to labor productivity dynamics (percentage changes; yearly average in periods). Sources: See Table 7.6. Note: GDP and capital are net of residential structures, but include government services.

one-third of total GDP growth in the years from 1881 to 1911; at the same time, the contributions of labor and capital levelled off, although capital accumulation continued to provide the largest contribution to growth (0.9 percentage points per year). Seen from another perspective (Figure 7.4), the greater labor productivity growth of these three decades (1.4 percent per year) was spurred on by capital deepening (0.8 percent) and by TFP growth (0.6 percent). Italy’s Golden Age, heralded in the years just after the end of World War II, mirrored a brisk recovery in capital accumulation and, even more so, in TFP growth. This was evident already in the subperiod 1938–1951 because of the strong recovery after 1945, but was even more pronounced in the years 1951–1973, when capital accumulation and TFP growth provided positive contributions to growth as large as 1.7 and 3.3 percentage points, respectively. As employment creation also became robust (with a contribution to growth of 1 percent), the Italian economy entered a rapid growth phase that was extraordinary in terms of its intensity and duration: even abstracting from the flourishing housing activities, GDP grew by 6 percent per year. Labor productivity grew by almost 5 percent per year, benefiting from higher capital deepening but even more so from an acceleration in TFP growth. As the catching-up of the Italian economy was rapidly progressing and structural bottlenecks began to show up, between the first oil shock and the crisis of the early 1990s GDP growth decelerated to 2.5 percent. More worryingly, the deceleration in TFP was particularly pronounced, and its growth rate dropped to 1.2 percent. Capital accumulation also lost momentum (falling roughly by a half), but still explained almost one third of the total growth of the Italian economy. Moreover, capital deepening, despite its moderation, continued

productivity

215

to contribute significantly to labor productivity growth, which decelerated to around 2 percent per year mostly caused by the TFP dynamics. Finally signs of structural weaknesses in the Italian economy have become more severe since the middle of the 1990s, with particular reference to the dismal performance of TFP growth, which hit a low across the several periods we considered (0.3 percent; 0.6 if the years since the start of the contraction in 2008 are excluded). Compared with long-term developments, proxied by the average changes over the full period since 1861, in recent years the gap in GDP growth (1.1 percent against 2.5) is almost completely explained by the virtual stagnation of TFP, in addition to the slightly lower contribution coming from labor and capital inputs.19 The gap in labor productivity is less severe (0.6 percent against 1.7), because capital deepening proceeded in the latest years broadly at the same intensity on average as in the full history of unified Italy (0.5 percent compared with 0.6 percent). Once again, however, if we break down the period 1993–2010 into three subperiods, the worsening of Italy’s productivity performance stands out in the 2000s, exacerbated by the end-of-the-decade recession. Whereas TFP growth rates were 1.3 percent in the 1990s, they later turned negative in the period 2000–2007 (−0.1 percent) and declined even further in the last three years of the decade (−0.5 percent). Evidence in Table 7.6 and in Figure 7.4 lead us to conclude that after Italy’s political unification in 1861 factor accumulation predominated in the country’s early stages of development, whereas a switch to TFP growth became the driving development factor after World War II, potentially pointing to a more sustainable path. This result is similar to the one found by Collins and Bosworth (1996) for East Asia and by Prados de la Escosura and Rosés (2009) for Spain. In the most recent period, however, TFP growth has gone back to being the weakest proximate cause of Italy’s growth, which is causing great concern. Once again, however, one must turn to the international context to see if these features of Italy’s growth process, especially since 1993, are common to other countries and if our concern is fully justified.

TFP Growth in an International Perspective Because of the lack of wage data throughout the period under study for all the countries in our selected sample, in this section we assume factor shares to be fixed at 0.65 for labor and 0.35 for capital. Similar assumptions have already been made in cross-country studies (Broadberry and Gupta 2010, Carreras and Josephson 2010, and Crafts and Toniolo 2010, just to mention some of the most recent). Although criticism may be expected for this assumption, we are reassured by the fact that studies such as Aiyar and Dalgaard (2005) find that a Cobb-Douglas production function with a constant capital share of one-third is a very good approximation of more general production functions that relax the restrictive neoclassical assumptions. Furthermore, comparing our two estimates for Italy (variable versus fixed

216

sources of growth and welfare

Table 7.7 GDP and TFP growth in a sample of countries (percentage changes; yearly average in periods) A. Italy

GDP

1861–1881

1.3

1881–1911

TFP

B. United Kingdom

GDP

TFP

0

1871–1891

1.8

0.6

1.7

0.4

1891–1911

1.7

0.3

1911–1938

1.7

0.7

1911–1950

1.3

0.6

1929–1938

1.1

−0.8

1929–1937

2.3

1.1

1938–1951

3.2

2.6

1950–1973

2.7

1.2

1951–1973

5.8

3.3

1973–1990

1.1

0.3

1973–1993

2.6

1

1990–2007

2.6

0.7

1993–2007

1.7

0.3

C. United States

GDP

TFP

D. Germany

GDP

TFP

1869–1889

4.3

0

1871–1891

2.4

0.7

1889–1909

4.2

0.8

1891–1911

2.1

0.8

1909–1950

3

1.3

1911–1950

−0.3

0.6

1929–1937

0.6

0.3

1929–1935

0.1

0.7

1950–1973

3.6

1.4

1950–1973

5.4

7

1973–1990

1.5

0

1973–1990

4.6

2.3

1990–2007

3.1

0.9

1990–2007

0.6

1.5

E. India

GDP

TFP

F. Japan

GDP

TFP

1890/1891 to 1900/1901

0.4

−0.7

1891–1911

2.9

1.1

1900/1901 to 1946/1947

0.9

0

1911–1950

2.4

0.4

1950/1951 to 1970/1971

3.8

1.2

1929–1935

2.3

−0.3

1970/1971 to 1999/2000

4.8

1.5

1950–1973

8.7

4.2

1973–1990

3.8

0.8

1990–2007

1.4

1.1

Sources: For Italy see Table 7.4 for VA and labor and Table 7.6 for capital stock estimates; see the Data Appendix of Broadberry, Giordano and Zollino (2011) for the other countries. Note: For international comparability reasons, in this Table we assume factor shares to be fixed at 0.65 for labor and 0.35 for capital. GDP here includes residential structures.

shares) we do not see striking differences in the main trends. We therefore proceed in this section with constant shares and gauge the TFP growth rates for the total economy for Italy, the United Kingdom, the United States, Germany, Japan, and India, presented in Table 7.7, alongside GDP growth rates.20

productivity

217

The stagnation of TFP in the first twenty years after Italy’s unification looks less worrying when compared with the similar performance of the United States in the same years. Later developments in TFP growth were also not too dissimilar to those registered in all the other countries of our sample, India excluded. Interestingly, Italy and Japan were the only two countries to register negative TFP growth in the Great Depression years.21 Although Italy, Germany, and Japan were defeated countries in World War II, their GDP recovery was remarkedly fast, as the reported GDP growth rates already in 1938–1951, but mainly in 1951–1973, show and as explained by Boltho in Chapter 4. The latter period again stands out as being one of exceptional TFP growth in Italy’s history, yet is overshadowed by Japan and Germany’s TFP performance in the same years. By contrast, during the most recent years Italy’s disembodied technologic and organizational advance has been very low by international standards (with TFP growth at 0.3 percent per year), the lowest in our sample, India once again excluded. Whereas the United Kingdom and the United States never saw any particularly dramatic acceleration in their TFP growth rates throughout the 150 years considered, Germany continued to display strong growth rates (of 2.3 percent) even in the two decades after its “economic miracle,” growth rates that stabilized at 1.5 percent per year in 1990–2007. The less-developed India again shows a different development pattern with negative (or approximately zero) TFP growth rates, which only became positive after 1950, and thereafter contributed to one-third of overall growth until the year 2000. Italy’s TFP performance can thus be roughly summarized as being in line with the average of our sample until 1951, above the average during its golden years, and strikingly below the average in recent years. Whereas the low TFP growth in Italy’s first twenty years of unified history does not stand out in the international comparison as particularly alarming, the most recent downturn is an outlier relative to the other countries and turns out to be the main cause of Italy’s modest GDP growth.

A Brief Update on Italy’s Most Recent Period (1992–2010) In October 2011 the Italian National Institute of Statistics (Istat) released new national accounts data for Italy that revised the existing figures for the period 1992–2010. This new vintage of data (Istat 2011b) was not taken into account by Baffigi in Chapter 6. Yet, given the relevance of the correction, in this section we provide some preliminary evidence of the impact of the new data on our productivity estimates.

218

sources of growth and welfare 0.13 0.12

Changes in labor productivity

0.11

Contribution by capital deepening

0.10

TFP

0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0.00 1992–2010

1992–1999

2000–2010

Figure 7.5 Italy’s total economy performance in the most recent years (differences in percentage points between estimates based on the new and old vintage of national accounts; yearly average in periods). Source: See Figure 7.4 and our elaborations on ISTAT (2011b). Note: Total economy VA and capital stock are net of residential structures. FTE labor is used.

Figure 7.5 plots the difference in percentage points between estimates built on Istat (2011b) and the estimates described until now in this Chapter. Upward corrections in the overall period 1992–2010 have increased labor productivity growth by 0.08 percent per year, mainly because of a greater contribution of TFP growth (+0.075 percent). The contribution of capital deepening has altered only marginally. Yet, when breaking down the period into two separate decades, as done in the right-hand side of Figure 7.5, it is clearly the years 2000–2010 that have undergone the major revision. Labor productivity growth increases by nearly 0.13 percent per year relative to the estimates presented in our previous sections, once again driven by faster TFP growth (+0.12 percent). This implies that in the most recent decade (2000–2010) Italy’s total economy TFP growth has stagnated rather than declined (which is what the previous vintage of national accounts pointed to). A sectoral breakdown of labor productivity growth, at this stage, is too demanding. Yet, preliminary calculations suggest that the main improvement has been achieved in industry, in particular in manufacturing. The annual average growth rates of value-added (including housing and government services) per full-time-equivalent worker are 2.8 percent for agriculture, 0.7 for industry, 0.6 for services, and 0.8 for the total economy.22 In particular, the manufacturing productivity growth rate rises to 1.4 percent per year, and transport and communications and credit and insurance rates increase to 2.1 percent. All other subsectors’ labor productivity growth rates improve slightly (except for agriculture, utilities, and personal services), because of this recent revision, but to a very minor extent. Once again, when dividing the overall period 1993–2010 into three relevant subperiods, sectoral labor productivity growth rates in 1993–2010 do not change

productivity

219

significantly relative to the ones presented previously. Instead, in the later seven years, industrial labor productivity growth was found to be positive, although sluggish (0.3 percent per year). Finally, in the recession years (2007–2010), labor productivity declined in industry and services, but at a lower rate (−1.0 and −0.1 percent, respectively) than what the previous data revealed, leading to a fall in total economy labor productivity growth of “only” −0.3 percentage points per year. Italy’s new national accounts therefore lead to a less gloomy portrait of Italy’s last twenty years, and last ten in particular. However, a slowdown relative to Italy’s past, and to other countries in the same years, continues to stand out, deserving great attention.

Conclusions: A Sectoral Analysis of Italy’s Development (1861–2010) This Chapter contains a growth accounting exercise, which helps to shed light on Italy’s aggregate growth performance in terms of variations over time and comparisons with other countries. Italian labor productivity growth was modest before World War II, despite Italy’s economic backwardness in 1861, so that although a long period of decline was decisively reversed, there was little catching-up on the leading economies of the time. The 1881–1911 period fared relatively better than the average (1901–1911 even more so), but the country’s growth was soon halted, and reversed, by thirty years of war and recession. Modest growth was largely the result of slow TFP growth before World War I and inadequate capital formation during the interwar period. The war and interwar years are difficult to interpret in a comparative perspective because of problems with the data, but it is noteworthy that Italy’s TFP growth rates turned negative during the 1930s, similarly only to Japan. Because the United States was forging ahead of the United Kingdom at this time, Italy was falling further behind the technologic frontier. In common with other European countries, Italy was hampered in adopting US high-throughput technology in industry, because of the abundance of cheap labor and the fragmentation of markets. Rapid growth of labor productivity was achieved during the Golden Age of the 1950s and 1960s, driven primarily by TFP growth, although there was also a strong contribution from capital. Together with Germany and Japan, Italy represented a success story in those years. By 1973 Italy had caught up with the United Kingdom in the economy as a whole, and was beginning to forge ahead, particularly in industry. At this point, growth rates slowed down, as would be expected within a catching-up framework. However, although Germany in particular continued to grow at a sizeable rate, Italy’s success was rapidly reversed

220

sources of growth and welfare

after the 1970s oil shocks. After 1973, Italy’s labor productivity growth slowed down, primarily as a result of slowing TFP growth, although the contribution of capital also declined. The deterioration in TFP growth rates in the most recent period, especially in the last decade, is even more striking, although slightly less worrying when considering Istat’s most recent data release. Factor accumulation has once again become the main engine of Italy’s growth, as was the case in the early stages of Italy’s development. A full understanding of Italy’s productivity performance, however, requires a consideration of sectoral developments. At this stage, without sectoral data on capital inputs, the sectoral analysis has to be conducted in terms of a partial productivity indicator (labor productivity) rather than TFP. Before World War II, although labor productivity growth was positive, Italy made little headway in catching up to the United Kingdom because of low productivity growth rates in agriculture, the largest sector of the time. Industry was no doubt the main engine in Italy’s first bout of acceleration (1881–1911), but the services sector too saw an interesting increase in its productivity growth rates in those years. Italian labor productivity growth increased dramatically during the Golden Age of 1951–1973, with industry once again in the lead, as Italy caught up with the European productivity leaders. A major factor in the convergence process during this period was a structural shift of labor away from a low-level labor productivity sector, such as agriculture, to high-level labor productivity sectors, such as industry and services. After 1973, Italian labor productivity growth slowed down. To some extent this was to be expected as Italy approached the technologic frontier. However, a sectoral analysis raises some concerns, particularly for the post-2000 period. A substantial productivity gap remains between Italy and the United States, particularly in services, where Italy seems to have shared in the productivity bonus from the application of information and communications technologies to a much lesser extent than other advanced countries. Drilling deeper into Italy’s labor productivity dynamics, manufacturing was a relevant driver of Italy’s industrialization, and growth process in general, throughout the country’s 150-year history. In contrast, the nonmanufacturing sectors presented fluctuating performances, but contributed less to overall economic performance given their size. Transport and communications was the only services sector that registered positive labor productivity growth rates over the whole period, with intense accelerations in particular after 1938. Interestingly, this is a services sector that, after a certain degree of development, bears greater resemblance to industry than the others in respect of the scale of production and of the capital intensity of technology. Trade and tourism also performed well after this date (and in 1881–1911), with the exception of the most recent period. Labor productivity in credit and insurance alternated between bursts of positive growth and bouts of negative growth. In particular, the latter coincided with the first twenty years after the country’s unification, when the banking and financial system

221

productivity

was highly fragmented and underdeveloped; the World War II years; and the 1970–1993 subperiod. Credit and insurance is, however, the smallest of the sectors considered, hence contributing little to the aggregate services productivity dynamics. Finally, personal services’ productivity registered high growth only in one subperiod (1951–1970), but has been declining since 1970, contributing heavily to the productivity slowdown of the most recent period (1993–2010), given its large size. Trade too is also to blame for current labor productivity dynamics in the services sector. Finally, it is appropriate to sound a note of caution about the data. The analysis presented here can only be as good as the available statistics, and imperfections remain despite the best efforts of official statisticians and historical researchers. Nevertheless, we think it unlikely that future revisions will overturn the basic findings presented here, particularly the slowness of Italian catching-up before World War II, the dramatically improved productivity performance of Italy during the Golden Age of the 1950s and 1960s, and the slowdown of Italy’s growth since the early 1990s. Even Istat’s latest revision of the Italian national accounts (Istat 2011b) has only marginally brightened up the country’s productivity performance over the past ten years.

Appendix As well as growth rates in different countries, in the third section we also presented comparative levels of labor productivity. In this Appendix we explain the methodology used to construct these levels in index number form. To pin down the comparative labor productivity level, we used a cross-sectional benchmark for 1997, derived from the EU KLEMS database (Timmer, Ypma, van Ark 2007). The benchmark is estimated from data on nominal value-added (i.e., in national currency), deflated by relative sector-specific price ratios adjusted for PPP, per person engaged in each country.23 This deflation procedure is necessary because the exchange rate cannot be assumed to accurately reflect differences in prices between different countries, especially at the level of individual goods and services, or particular sectors. In principle, price discrepancies converge to zero in sectors open to international trade, yet different degrees of monopoly power, lags in response to exchange rate movements, barriers to trade, and so forth may fuel persistent differences. Furthermore, exchange rates have been known to be subject to substantial short-term fluctuations and international capital movements, thus becoming misleading converters to a common currency, even for tradeable goods and services.24 In the case of cross-country comparisons, value measures must be corrected for differences in relative prices between countries. Furthermore, sector-specific PPPs are to be used, because large cross-sector differences in PPPs can be shown to exist (Inklaar and Timmer 2008, 16–17).25 The

222

sources of growth and welfare

PPP for services in Italy is, however, believed to be underestimated because it leads to a very high labor productivity level for that sector in the international context, especially in the nonmarket services’ sector, which is difficult to justify given the existing literature on the subject.26 The EU-KLEMS PPP for services has thus been replaced by the geometric mean of the PPP for industry and of the PPP for agriculture. In turn, the PPP for the total economy has been computed as a weighted average of the three sector-specific PPPs. The time series of labor productivity are then projected, backward and forward, from the 1997 benchmark thus built. A number of studies (Prados de la Escosura 2000; Ward and Devereux 2003) have questioned the use of time series projections from a single benchmark over long periods of time, the methodology used here, which potentially raises index number problems. The issue was the subject of debate between Broadberry (2003) and Ward and Devereux (2004). The problem may be mitigated by using additional benchmark estimates—for earlier and later years—to check for consistency with the comparative labor productivity levels suggested by the time series projections. Broadberry (1993) had already suggested the use of additional benchmarks to provide cross-checks in a study of comparative productivity in manufacturing, whereas Broadberry (1997a; 1997b; 1998; 2006) applied the method to full sectoral productivity comparisons over the period 1870–1990 for the United Kingdom with the United States and Germany, and found broad agreement between the benchmarks and time series evidence for those countries. Broadberry and Irwin (2006, 2007) find similar agreement between time series projections and benchmarks for the United Kingdom compared with the United States in the nineteenth century and the United Kingdom compared with Australia over the period 1861–1948. More recently, Broadberry and Gupta (2010) have applied the same methodology to an Anglo-Indian comparison between 1870 and 2000 and Broadberry et al. (2012) to an Anglo-Japanese comparison between 1870 and 2007. We thus proceed in the same way here to provide some additional corroboration and conclude that, although index number problems do exist, with careful treatment of the data time series projections from a fixed benchmark can tell a consistent story. In particular, we compare our time-series projections with existing direct benchmarks built in previous studies, presented in Table 7.8. The first benchmark in order of time refers to 1905 and is taken from Broadberry and Klein (2012). The PPPs are in turn taken from a study by Williamson (1995) and are based on the prices of a basket of consumption goods. This is not ideal because national income includes other items of expenditure besides consumption; yet, consumption is the most relevant item. Furthermore, the benchmark refers to per capita income and not precisely to aggregate labor productivity. However, the two measures are known to be strictly related. The difference between the benchmark and our projection is relatively small, which is reassuring given the claims of inevitable large disagreements between time series projections and direct estimates made by Ward and Devereux (2003). O’Brien and Toniolo (1991) produce a direct estimate for agriculture in 1910. It is based on comparative full-

223

productivity Table 7.8 Benchmark Cross-Checks on Time-Series Projections Years

Direct benchmarks

Time series projections

1905

43.7

38.5

1910

43

42.4

2007

83.8

89.6

Sources: For the direct benchmarks, Broadberry and Klein (2012) for 1905; O’Brien and Toniolo (1991) for 1910; OECD (2011c) for 2007; our estimates for the time-series projections. Note: The first and third are direct estimates of GDP per head; the second is a direct estimate of male FTE labor productivity in agriculture.

time equivalent labor input in United Kingdom/Italy. It is, however, assumed to be similar to the headcount labor input comparative ratio in the same sector. As Table 7.8 shows, in 1910, too, the time projection differs only 1.5 percent from the direct estimate. In conclusion, our time series projections from a 1997 benchmark substantially agree with existing direct estimates in 1905 and 1910, which further corroborates our projection from a fixed benchmark technique. Finally, for recent years, OECD (2011c) provides data on internationally comparable real GDP per capita. We can thus compare our comparative Italy-UK labor productivity estimate for 2007, the only overlapping year, with these data. The comparison is again presented in Table 7.8, from which it is clear that our forward time-series projection is also relatively close to the OECD results. In conclusion, although time-series projections from a fixed benchmark are not devoid of problems, especially given the long time-span of our study, we are reassured by the fact that our resulting estimates are similar to existing direct benchmarks or to other internationally sourced estimates.

Acknowledgment The authors thank their discussant Leandro Prados de la Escosura, and Federico Barbiellini Amidei, Nicholas Crafts, Herman de Jong, Stefano Fenoaltea, Matteo Gomellini, Paolo Malanima, Guido Maria Rey, and Gianni Toniolo for their valuable comments and suggestions. They are also grateful to Giovanni Federico, Ferdinando Giugliano, and Roberto Golinelli for sharing their data on request. Finally, they thank all participants of the Conference “Italy and the World Economy, 1861–2011” held at Banca d’Italia, Rome, in October 2011, and the Workshop “Italy and its International Economic Position, 1861–2011” held at Banca d’Italia, Perugia, in December 2010, for their useful questions and comments. The views expressed herein are those of the authors and do not necessarily reflect the views of the institutions represented.

224

sources of growth and welfare

Notes 1. Our figures, which make use of substantially revised Italian output data, point to 1973, but with the GDP figures available at the time, the more conventional dating of il sorpasso is 1979. 2. Similarly to Baffigi’s estimates of Italy’s value-added presented in Chapter 6, for the period 1970–2010 our figures are based on Istat (2011a) and hence do not take into account the revisions published in Istat (2011b). 3. The ten sectors considered are (1) agriculture, forestry, and fishing; (2) mineral extraction; (3) manufacturing; (4) construction; (5) utilities; (6) trade and tourism; (7) transport and communications; (8) finance, insurance, and real estate; (9) social and personal services; and (10) government services. 4. Complete details of data sources and the methodology behind the construction of the series are reported in the Data Appendix of Broadberry, Giordano, and Zollino (2011). 5. Participation rates are here computed as the headcount measure of labor divided by resident population in Italy at present boundaries. 6. Note that the headcount labor shares are very similar to the ones here described; general patterns and evolutions are the same. 7. For the more recent years for which data on hours worked are available, a rapid comparison of output per worker and output per hour worked growth rates will, however, be drawn. 8. The first electric power station was built in Milan in 1883. 9. David (1990) effectively claimed the existence of a “diffusion lag.” 10. In the literature, it is quite common to distinguish between the services subsectors prone to domestic and international competition and those that instead are more protected and hence presumably have fewer incentives to increase their efficiency or employ marginal workers, therefore registering lower labor productivity growth rates. The communications sector, being characterized by the presence of a State monopoly, seems to go against this theory prediction, yet our data point to this sector as being one of the most productive over the 150 years considered. In particular, Pellegrini (1991, 16) explained the high labor productivity growth in this sector in the 1980s by the technologic progress it underwent in these same years and by the buoyant demand for these services. 11. See Daniele and Malanima (2011b) for a discussion of this issue. 12. Broadberry (1998) argues that a major problem with Nordhaus’s (1972) orthodox shift-share approach is that it assumes that productivity growth rates in each sector would be unaffected by the absence of structural change. If Kindleberger’s (1967b) assumption that surplus labor was being drawn from agriculture and reallocated to nonfarm activities with little or no loss of agricultural output is accepted, as is reasonable, then restoring labor to agriculture would not have positively affected output, but simply lowered labor productivity growth rates. However, the shift of labor away from nonfarm activities would not only have lowered labor, but also output, leaving labor productivity growth rates unaltered. Therefore, had agriculture continued to employ an unchanged share of workers, because of an absence of structural change, labor productivity growth rates in agriculture would have been lower. This effect is taken into account in Broadberry’s modified equation.

productivity

225

13. These calculations are to be regarded as upper-bound estimates of the effects of structural change (Broadberry 1998). This modified shift-share calculation has also been adopted by Crafts and Toniolo (2010). 14. To achieve international comparability we are obliged to revert back to a three-sector disaggregation. 15. Especially in those subsectors where entry barriers are low, such as petty trade and personal services, because of the strong pressure of the active population that cannot find employment in agriculture or industry, the services sector can act as a sponge, absorbing the excess work-force. 16. The exploitation of opportunities offered by information and communications technology in the United Kingdom has been put into relation with the de-regulation of services, which are intensive in the use of information and communications technology (e.g., finance and retail trade). See, for example, Crafts (2011a). 17. Under the usual assumption that rental price does not vary across vintages of a (qi ,t qi ,t ), where qit is the capital asset i it is measured as, ui ,t = qi ,t − rt + qi ,t δ i ,t − (q market price of the productive asset i, rt is a measure of the opportunity cost that we proxy by the nominal long-term interest rate on public bonds, δit is the same depreciation rate adopted in estimating the capital stock, and the terms in brackets stand for the expected revaluation of the asset, that we compute as a three-term moving average of the market price. 18. Because of data limitations on wage and employment composition by skill or education level, we could not perform the same calculations for labor input, which we continue to measure using our estimates of full-time equivalent workers. 19. However, Bassanetti, Torrini, and Zollino (2010) show that, by taking into account the prolonged wage moderation and intense process of privatization since the start of the 1990s, which resulted in a fall in the share of wages over value-added, TFP in Italy would have marked a more moderate increase between the mid-1980s and the end of the 1990s, followed by a less negative dynamics from 2000 to 2005, the last year covered in the analysis, compared with what emerges from the official Istat national accounts. Interestingly, the finding was not matched in the other leading countries in the Euro area, confirming that part of the recent delay in productivity growth in Italy was caused by the competition structure of the markets and their development over time. 20. Again, for comparability reasons, we use the headcount measure for labor and GDP inclusive of residential structures. 21. By looking at our Table, the reader can gauge that TFP growth rates in the United States in the period 1929–1937 were low (0.3 percent per year). However, a recent interpretation of the US Great Depression, offered by Alexander Field, who focused on the longer period 1929–1941, led the scholar to conclude that those years were actually “the most technologically progressive of any comparable period in the US economic history” (Field 2011, 19), with TFP growth rates of 2.3 percent (43). Simply “Kendrick [i.e. Kendrick (1961) whose data we too use for the US] led us astray by choosing 1937 as a reference peak in calculating productivity growth in the 1930s” (Field 2011, 41). The cause of this exceptional surge in TFP growth was a still rapid progress in manufacturing coupled with the spillover effects in transportation and distribution resulting from the construction of the surface road network. This novel light shed on the US performance in the 1930s does not alter our conclusions on Italy’s relative performance in the same years.

226

22.

23.

24. 25.

26.

sources of growth and welfare

Similar innovations were adopted in Italy mainly after World War II, when Italy’s TFP growth rates soared. Another noteworthy point is that in the period 1911–1938 the differences between TFP growth rates in Italy as computed in Tables 7.6 and 7.7 are the largest. This is caused in particular by the Great Depression years which, weighting labor and capital by constant shares, creates some distortion. The estimate presented in Table 7.6 is thus the more reliable of the two. Unfortunately the figures for services and total economy are only partially comparable with those presented in the last row of Table 7.1 because valued-added here has not been netted of the housing sector and of government services. PPP can be defined as “the number of currency units required to buy the goods equivalent to what can be bought with one unit of currency of a base country” (Kravis, Heston, and Summers 1982). See for example Taylor and Taylor (2004) for a review of the debate on PPP. In particular, the PPPs provided by EU KLEMS include production PPPs for agriculture, mining, manufacturing (except high-tech), transport, communication and trade industries and expenditure PPPs for all remaining sectors. See Timmer, Ypma, and van Ark (2007) for a discussion on the advantages and drawbacks of the different types of PPP. See for example Bripi, Carmignani and Giordano (2011) for a survey of studies on the (poor) quality and efficiency of public services in Italy in recent years in an international context.

Chapter 8

STANDARDS OF LIVING andrea brandolini and giovanni vecchi

Introduction In 1861, a newborn Italian could expect to live for another twenty-nine years. One and a half century later, life expectancy at birth has increased to eighty-two years, 84.5 for females and 79.4 for males. With a gain of fifty and more years over 150 years, Italians have climbed to the top of the country ranking by life expectancy, and Italy stands among the best performers worldwide. There is hardly any other indicator that is as effective as life expectancy to gauge in a single number the progress of a population. On this account, Italy since unification in 1861 is undeniably a success story. Yet, this is hardly the end of the story. Human well-being is a multifaceted concept. Living longer is an achievement by itself, but it also matters how people live. Access to consumer goods and leisure time, for instance, is important. The capability of choosing one’s own life is even more important. However significant, life expectancy cannot account for these different aspects of well-being, nor can per capita income, another popular indicator of progress. Moreover, any assessment based on average well-being is bound to ignore its distribution across the population, whereas our evaluation of the advancements in, say, the prevention of avoidable morbidity is likely to depend on whether it is spread across the entire population or is instead concentrated among a few wealthy individuals.

228

sources of growth and welfare

Multidimensionality and distribution are the two keywords in tracing the path of the well-being of Italians since the country’s unification. We begin with an overview of the historical changes in income per capita and integrate this description with information on its distribution. Income is a good measure of the command that people have over resources, but it cannot capture all the diverse dimensions of human well-being. Here, we go “beyond GDP” (European Commission 2009; Stiglitz, Sen, and Fitoussi 2009) by considering also the secular evolution of working time, health outcomes, and educational achievements. This broad set of variables allows us not only to depict a nuanced view of the human progress in unified Italy, but also to delve into its dynamics. Was it a smooth unfolding of long-run, mechanical, and inescapable tendencies, or was it the outcome of a sequence of episodes (Atkinson 1997), alternating advancements to stasis and, possibly, regression? Did Italy sail with the international tide or did it succeed to surf the long wave of economic well-being because of the ability of expert skippers (Toniolo 2007)?

Income and People It is far-fetched to identify well-being with income, but there is little doubt that command over resources is a fundamental determinant of living standards. As observed by Anand and Sen (2000), “in an indirect way—both as a proxy and as a causal antecedent—the income of a person can tell us a good deal about her ability to do things that she has reason to value. As a crucial means to a number of important ends, income has, thus, much significance even in the accounting of human development” (100). Likewise, at the aggregate level, the total amount of incomes produced or enjoyed in a country, expressed in per capita terms, is an important measure of the prosperity of its population. National accounts offer several aggregate income concepts but, because of data availability, we focus here on gross domestic product (GDP). Our reference series is drawn from the reconstruction of the Italian historical national accounts documented in Chapter 6 (see also Vecchi 2011), but for international comparisons we rely on the series elaborated by Maddison (2001, 2010) and updated by Conference Board (2012).1 The growth of average income was fairly slow after unification, and it took about eighty years for GDP per capita to double. After the destructions of World War II, the speed rose considerably, and income doubled once during the 1950s and a second time in the following fifteen years. Qualitatively this historical pattern is not different from that of other advanced countries, but Italy lags behind the top performers. Considering Maddison’s estimates (Figure 8.1), in 2011 the Italian GDP per capita equaled 59 percent of the US level, just above the ratio for Spain (55 percent), but well below those of other Organization for Economic Co-operation and Development (OECD) countries, such as Germany

229

standards of living 140 120 100 80 60 40 20 0 1861 1871 1881 1891 1901 1911 1921 1931 1941 1951 1961 1971 1981 1991 2001 2011 Italy

France

Germany

Spain

Sweden

UK

Figure 8.1 GDP per capita in selected countries (index: US = 100). Source: authors’ elaboration on data from Maddison (2010) and Conference Board (2012); underlying values are in 1990 international Geary-Khamis dollars. Data for Italy in 1915–1920 are adjusted on the basis of the new reconstruction of national accounts exposed in Chapter 6.

(70 percent), France (71 percent), the United Kingdom (72 percent), or Sweden (84 percent). Although some changes have occurred in the rank order of countries, most notably the relative decline of the United Kingdom, Italy has overall shared the ups and downs of other countries (Crafts and Toniolo 2010). There is little indication of Italian-specific growth miracles in comparative terms. Dividing aggregate income by the number of persons provides only a rough indication of average living standards. Two people living together do not need an income twice as large as the income they would need were they living alone. Living in a household generates economies of scale in consumption, because certain goods, such as housing space and heating, can be shared. Moreover, needs differ by age, with children and the elderly typically requiring less than adults, at least in terms of calorie intake. Thus, for a society as for a household, the demographic structure affects the standard of living achievable with a given income. Over the last 150 years changes in age structure and average household size have been dramatic. The share of persons younger than fifteen years fell from 34 to 14 percent of the total population between the end of 1861 and the end of 2009, whereas the share of people older than sixty-four quintupled from 4 to 20 percent. The “ageing index,” calculated as the number of persons sixtyfive years old or over per hundred persons aged fourteen or less, has increased by a factor of twelve between 1861 and 2011 (from 12 to 145 percent), “making [Italy] the world’s ‘oldest’ major country” at the onset of the third millennium (Kinsella and Phillips 2005, 7). Household size decreased from 4.5 persons per unit in 1881 to 2.4 in 2010 (Figure 8.2).

sources of growth and welfare 50

6.0

45

5.5

40

5.0

35

4.5

30

4.0

25

3.5

20

3.0

15

2.5

10

2.0

5

1.5

No. of persons per household

Thousands of 2010 euros

230

1.0 0 1861 1871 1881 1891 1901 1911 1921 1931 1941 1951 1961 1971 1981 1991 2001 2011 GDP per capita

GDP per equivalent adult

GDP per equivalent person living alone

Household size (right scale)

Figure 8.2 GDP per capita and per equivalent person in Italy. Source: authors’ elaboration on Istat data. The number of equivalent adults is computed by weighting individuals by their age-dependent relative nutritional needs (drawn from Istat 2009b). The number of equivalent persons living alone is computed by weighting individuals by a factor capturing the economies of scale from cohabitation (set equal to the square root of the household size).

To account for these demographic trends, we compute the number of “equivalent persons,” which is the number of individuals made comparable in terms of needs (Coulter, Cowell, and Jenkins 1992), and calculate GDP per equivalent person rather than per capita (Figure 8.2). For the lack of suitable household-level information, we carry out separate adjustments for age and household size, but in either version GDP per equivalent person exceed GDP per capita. When we adjust for the age structure, this happens because children and the elderly are weighted less than adults. Despite the remarkable changes in the age composition, the difference between GDP per capita and GDP per equivalent adult is very small; it only widens slightly in the last decades, but not in a way sufficient to alter the overall income profile. The secular changes in household arrangements seem to be far more important. The gains from pooling and sharing goods and services within the household, which explain the difference between equivalized and nonequivalized GDP, diminish over time, since the 1920s. Thus, the eleven-fold increase of GDP per capita between 1881 and 2010 would turn into an eight-fold rise of GDP per equivalent person. This result, however, should not be overplayed. First, at earlier stages of development, expenditure is higher on goods characterized by low economies of scale, such as food, than on goods characterized by greater economies of scale, such as housing (Deaton 1997). Should we assume that economies of scale in consumption rise over time, rather than being constant as in the calculation underlying

standards of living

231

Figure 8.2, the discrepancy between equivalized and per capita GDP would be proportionally lower at the beginning than at the end of the period, and the conclusion would be different. Second, and more importantly, the secular movement toward a smaller household size reflects a people’s choice. Thus, the lower economies of scale in consumption must have been more than offset, in welfare terms, by the greater independence allowed by living in smaller family units. Exit from the family of origin and household formation reflect economic feasibility and cultural factors (Reher 1998; Giuliano 2007). In sum, although the fall in household size is an important feature of the evolution of quality of life in Italy, as in other advanced countries, its welfare implications are not easily captured by any mechanical application of equivalence scales.

Worked Time and Child Labor As early as 1952, Kuznets denounced that the long-run reduction in hours worked, that is the secular increase in the consumption of leisure, was one of those aspects that economists had “ . . . overlooked, or, if recognized, dismissed lightly because they lie outside the boundaries of economic discipline, as narrowly defined. The neglect or dismissal of these problems is likely to be more detrimental to the understanding of the process of economic growth than even crude attempts to deal with them” (1952, 63–64). In the same vein, Nordhaus and Tobin later remarked that “the omission of leisure and of nonmarket productive activity from measures of production conveys the impression that economists are blindly materialistic. Economic theory teaches that welfare could rise, even while NNP [net national product] falls, as the result of voluntary choices to work for pay fewer hours per week, weeks per year, years per lifetime” (1972, 9). The reduction in worked hours may take many forms: later entry in the labor market, as years spent in full-time education rise; shorter working days or weeks; longer holiday leaves; and early retirement, relative to residual life expectancy. Together with the rise in the length of life, all these changes have gone in the direction of reducing the share of lifetime devoted to work, diminishing effort and increasing leisure. Data are currently insufficient to carry out a systematic adjustment of GDP for the reduction in worked hours along the lines suggested by Kuznets and Nordhaus and Tobin, but two aspects can be examined: child labor and working hours. Available evidence suggests that child work was widespread in most European countries during the nineteenth century (Cunningham and Viazzo 1996). Italy was no exception, with about 80 percent of children aged ten to fourteen years being classified as economically active at the time of the country’s unification (Vecchi 2011). Industrialization was, however, accompanied by a decline

232

sources of growth and welfare

in child work, unlike in other countries (Goldin and Sokoloff 1982; Horrell and Humphries 1995; Tuttle 1998). This decline proceeded rapidly throughout the entire period, except for a stasis between the two world wars, and even an increase in the aftermath of the Great Depression from 1931 to 1936. These estimates, which are based on census results, contradict the pessimist conclusion of the Parliamentary Commission of Inquiry into Unemployment that “ . . . in the course of Italy’s economic development, the employment of children in workplaces has steadily increased during its early phases” (Spesso 1953, 171). They are also considerably lower than the figures reported in some international comparisons of child work (International Labor Organization 1996; Basu 1999). As to working time, it is not easy to reconstruct its historical movements, not least for the large variability across sectors and occupations. The selection of data reported in Figure 8.3 mostly refers to manual workers in industry, including construction. Time spent at work did not vary much during liberal Italy, so that the gains in well-being as measured by the GDP per capita were not offset by a concomitant increase in the effort required to produce it. The calculations by Zamagni (1975, 1994) show a temporary rise of work-time during World War I, followed by a sharp fall around 1919, when the eight-hour workday was introduced. Working time slightly rose in the aftermath of World War II, stabilized during the 1950s, and then declined until the early 1980s. Thereafter, the Italian National Institute of Statistics’ (Istat) national accounts 3,500 3,250 3,000 2,750 2,500 2,250 2,000 1,750 1,500

Huberman Zamagni Ministero del lavoro Istat, SCI Istat, NA Conference Board

1,250 1861 1871 1881 1891 1901 1911 1921 1931 1941 1951 1961 1971 1981 1991 2001 2011

Figure 8.3 Annual hours of work in Italy. Source: authors’ elaboration on data from Huberman (2004, Tab. 6, 982), industrial manual workers; 1911–1918, Zamagni (1975, Tab. 1, 532), industrial manual workers, hours of work per day multiplied by 6×30/7×12; 1919–1939, Zamagni (1994), industrial manual workers, hours of work per month multiplied by 12; Ministero del Lavoro’s survey of industrial establishments with at least ten employees (fifty since 1978), industrial manual workers; Istat’s survey of firms with at least fifty employees, industrial manual workers (SCI); Istat’s national accounts (from EU-KLEMS for 1970–1979), all industrial employees (NA); Conference Board (2012), total economy, all employed.

standards of living

233

figures for all employees, rather than blue-collar workers only, indicate a flat trend, except for the abrupt drop during the Great Recession of 2008–2009. The Conference Board’s (2012) series for the total economy exhibits a substantially similar pattern, but at a higher level because it also covers the self-employed, who typically work longer hours. Italy’s historical record follows, by and large, the pattern of most European countries (Ausubel and Grübler 1995), although nowadays the average employed person works more in Italy than in the United States, the United Kingdom, France, or Germany. When we consider the work effort relative to the entire population rather than the employed only, we may expect to observe similar long-run tendencies, but reinforced by a decline in labor market participation (Zamagni 1987; Daniele and Malanima 2011a). The reduction in child labor and the ageing of population, when not accompanied by a corresponding elevation of retirement age, imply an additional fall in hours of work per person, besides that caused by shorter work-time. Kuznets (1966, 75) calculates a significant long-run decrease, until the early 1950s, in the number of man-hours per capita (as opposed to per worker) in thirteen advanced economies, ranging from 1.1 percent per decade in Great Britain (1870–1952) to 4.5 percent in the Netherlands (1900–1952), and to an “exceptional” 7.5 percent in Italy (1901–1953). According to the Conference Board’s (2012) data, the fall continued until the 1970s; it was then reversed in the mid-1990s, as labor market participation began rising again. The postwar trend in hours of work is in Italy similar to those of the other main European countries, although somewhat less pronounced.

Beyond Average Incomes GDP per person ignores the distribution of income across people. However, the welfare of a community depends not only on total income, but also on the way in which it is distributed among its members. As put by Dalton (1920, 349), “the objection to great inequality of incomes is the resulting loss of potential economic welfare.” In Dalton’s utilitarian approach, social welfare equals the sum (or mean) across individuals of identical concave utility functions, where income is the argument. Social welfare is maximized when a given income is equally distributed, and inequality can be identified with the shortfall from this maximum value. Atkinson (1970) suggests that this welfare loss can be measured in the income space by means of the “equally-distributed equivalent income,” that is the level of income that would give the same level of social welfare as the given distribution when equally assigned to all individuals. By construction, the equally-distributed equivalent income is lower than mean income and represents a measure of the economic welfare allowed by a certain income level that scales down the mean for the inequality of its distribution among people. Sen

234

sources of growth and welfare

(1976) derives from different nonutilitarian assumptions a measure of economic welfare that is analytically identical, except for using the Gini index rather than the Atkinson index to measure inequality. In this section we document the secular changes in the distribution of income, and assess their implications for economic welfare by constructing distributionally adjusted GDP per capita series as suggested by Atkinson and Sen.

Data Problems and the Italian Household Budget Dataset Modern household surveys began after World War II, once the statistical theory and method to construct representative samples had been developed and legislators had created national statistical offices (Stapleford 2009). Before World War II, the individual- or household-level data required to estimate the evolution of the personal distribution of income were scarce and inadequate, because of their limited coverage of population and income. This is the case of the income data drawn from tax files, originally used by Pareto (1895) to study the revenue curve. The adjustment to income and population totals derived from external sources, developed by Kuznets (1953) and recently revived by Atkinson and Piketty (2007, 2010), improves the representativeness of income tax data, but does not eliminate all problems, including that inequality can only be measured by the income shares of top income earners. Other sources and proxies have been imaginatively used (Williamson and Lindert 1980; Williamson 1985; Milanovic, Lindert, and Williamson 2010), but despite the eclecticism and ingenuity, the paucity of suitable data has curbed the analysis of long-run changes in income inequality. For Italy tax-based information has limited time coverage. However, we can rely on a novel and innovative database, the Italian Household Budget Database, constructed from the thousands of household budgets collected, in various forms, since the country’s unification (Niceforo 1933; Somogyi 1959; Vecchi 1994). The first version of this database comprised about 4,700 household budgets over the years 1881–1961 and focused more on expenditure (Rossi, Toniolo, and Vecchi 2001); the new version, used here, covers also income and includes as many as 20,000 household budgets (Vecchi 2011). The individual budget data are used to construct “samples” centered on census years. Census data provide the information needed to compute the weights (“expansion factors”) used to poststratify the income information and correct for the potential biases arising from the lack of a probabilistic survey design (Holt and Smith 1979; Little 1993).2 We use the information from the Italian Household Budget Database for the period 1861–1931, together with the data collected by Istituto Doxa for 1948 and from the Bank of Italy’s Survey of Household Income and Wealth from 1967 to 2008 (tabulated data until 1976 and microdata thereafter; on these sources, see Brandolini 1999). The Gini index and the Atkinson index for the distribution of per capita incomes are estimated by fitting, for each available year, a four-

235

standards of living

parameter functional form, the Generalized Beta of the Second Kind, discussed by Jenkins (2009).

Long-Run Dynamics of Inequality and Distributionally Adjusted GDP Per Capita Figure 8.4 shows the secular trend of income inequality in Italy, as measured by the Gini and the Atkinson indices. The calculation of the latter requires specifying the parameter ε that measures the aversion to inequality: the higher ε, the more weight we attach to transfers at the bottom of the distribution relative to transfers at the top; in the extreme case where ε tends to infinity, only the income of the poorest individual matters. We consider three values of ε (0.125, 1, and 2), which cover a wide range of social preferences. There are local divergences in the movements of the four indices, in particular between the Atkinson index most sensitive to changes at the bottom of the distribution (ε = 2) and the others. However, except for a notable divergence in the period 1921–1931, the overall story seems to be fairly consistent across the four indices. During the first seventy years since unification the Gini index fluctuated within a relatively narrow band, between 45 and 50 percent; these variations were statistically not significant. Eighty years later, it has decreased by almost a third to around 33 percent, although the historical minimum was reached, at 30 percent, in the 1980s. Even allowing for statistical errors, the long-run trend is unambiguously downward: Italy’s modern economic growth has been 0.60 0.50 0.40 0.30 0.20 0.10 0.00 1861 1871 1881 1891 1901 1911 1921 1931 1941 1951 1961 1971 1981 1991 2001 2011 Atkinson index ε=0.125

Atkinson index ε=1

Atkinson index ε=2

Gini index

Figure 8.4 Gini and Atkinson indices of per capita income in Italy (percent). Source: authors’ elaboration. Dashed lines help to identify the three main sources: (1) the Italian Household Budget Database until 1931, (2) the survey of Istituto Doxa for 1948, and (3) the Bank of Italy’s Survey of Household Income and Wealth from 1967 onward.

236

sources of growth and welfare

accompanied by a narrowing of the distribution of income. This process has not been linear, however, and it is better described as a sequence of periods alternating declines to stasis, or sharp drops to sudden rises, consistently with Atkinson’s observation that “ . . . it is misleading to talk of ‘trends’ when describing the postwar evolution of the income distribution,” whilst “ . . . it may be better . . . to think in terms of ‘episodes’ when inequality fell or increased” (1997, 303). During the early stages of industrialization the Gini index stalled. The increase between 1871 and 1901 is not negligible, but not statistically significant and by and large offset by the reduction between 1901 and 1921. There is no evidence of the inverted U-shaped curve that Kuznets (1955) suggested should emerge along the process of industrialization. Rather, the evidence seems consistent with the hypothesis of a “benign industrialization” conjectured by Toniolo (2003) and Vecchi (2003). In this respect, Italy’s experience may have differed from those of the United States and the United Kingdom as summarized by Lindert (2000, 169), “income and wealth inequality definitely rose over the first 150 years of US history. Britain may also have had an early period of rising inequality, but the most likely period of rising inequality (1740–1810) was earlier than most writers have imagined” (but see Feinstein 1988 for a different view about the UK experience). Although income distribution tended to narrow during the two episodes of more intense economic growth (the Giolitti’s years before World War I and the post–World War II years of the so-called economic miracle), it was between the end of the 1960s and the early 1980s that the fall in inequality was more pronounced, with a sharp drop of the Gini index by more than 10 percentage points. This fall was driven by a strong compression in the distribution of labor incomes, which was reinforced by the expansion of the welfare state (see also Chapter 3). This “egalitarian phase” was not specific to Italy: changes in the same direction, if not of the same size, took place in France, Germany, Sweden, and Finland (Brandolini and Smeeding 2008). The downward trend reversed during the 1980s. Inequality rose especially during the severe economic and currency crises of the early 1990s, when the Gini index abruptly went back to the levels of the late 1970s; it did not change much in the subsequent period, including the period of the Great Recession of 2008–2009. The tendency of inequality to rise since the 1980s is, however, less pronounced than in other countries, particularly the United States and the United Kingdom, but also Sweden and Finland. The fall in inequality amplifies the extent of the economic progress of Italians since unification. As shown in Figure 8.5, the overall increase of economic welfare is always larger for the distributionally adjusted measures of real income. The improvement is modest when one takes Atkinson’s equally distributed equivalent income with ε equal to 0.125, which is not surprising in light of the modest weight assigned to inequality in the social welfare function. For higher values of ε or for the Gini index, the increase of the distributionally

237

standards of living 20 18 16 14 12 10 8 6 4 2 0 1861 1871 1881 1891 1901 1911 1921 1931 1941 1951 1961 1971 1981 1991 2001 2011 μ

μ(1–A(0.125))

μ(1–A(1))

μ(1–A(2))

μ(1–G)

Figure 8.5 Distributionally adjusted measures of real income in Italy (index: 1861 = 1). Source: authors’ elaborations as described in the text. μ is GDP per capita, A is the Atkinson index calculated for the value of ε indicated in parentheses, and G is the Gini index. The lines connect the years for which inequality statistics are available.

adjusted GDP per capita is far more conspicuous. Sen’s measure, embodying the Gini index, suggests an eighteen-fold increase of economic welfare between 1861 and 2008 instead of the thirteen-fold rise indicated by GDP per capita.

Life Expectancy and GDP-Augmented Measures We have already mentioned how the advancement in the Italians’ life expectancy since unification is enormous, and remarkable even in the international comparison (Figure 8.6). In the liberal period, from 1861 to 1913, the progress was rapid and sizeable, and Italy almost caught up with France. Improvements are in line with those in other countries during the Fascist period, but after World War II they were more substantial than elsewhere, and Italy overtook most countries. According to the last estimates by the World Health Organization (2012), Italians can expect to live longer than any other world citizen except for the Japanese. Until the 1960s the improvement in life expectancy was driven by that in infant mortality. In Figure 8.7 we compare for selected countries the long-run series of the infant mortality rate, defined as the number of deaths within the first year of life per one thousand live births. In 1863, this rate equaled 290 in Italy, slightly less than in Germany, but considerably more than in France and

238

sources of growth and welfare 90 80 70 60 50 40 30 20 1861 1871 1881 1891 1901 1911 1921 1931 1941 1951 1961 1971 1981 1991 2001 2011 Italy

France

Germany

Spain

Sweden

UK

US

Figure 8.6 Life expectancy at birth in selected countries. Source: authors’ elaborations on data from Vecchi (2011) for Italy and from University of California Berkeley and Max Planck Institute for Demographic Research (2012) for other countries. The UK figures refer to England and Wales only from 1861 to 1921.

350 300 250 200 150 100 50 0 1861 1871 1881 1891 1901 1911 1921 1931 1941 1951 1961 1971 1981 1991 2001 2011 Italy

France

Germany

Spain

Sweden

UK

US

Figure 8.7 Infant mortality rate in selected countries (deaths within first year of life per 1,000 live births, three-year centered moving average). Source: authors’ elaboration on data from Istat (2012), University of California Berkeley and Max Planck Institute for Demographic Research (2012), and Mitchell (2007). The series are from the Human Mortality Database for France, Sweden, and the United Kingdom (England and Wales for 1861–1921); the series from the Human Mortality Database are retropolated backward with the series from Istat (2012) for Italy (1963–1971) and from Mitchell (2007) for Germany (1863–1955), Spain (1861–1907) and the United States (1915–1932).

standards of living

239

England and Wales. Infant mortality fell quickly after the unification, mainly thanks to the reduction in mortality caused by infectious diseases, but in 1910 it was still above the levels observed today in the poorest countries of the world, such as Sierra Leone, Liberia, or Angola. The infrastructure of public health was still underdeveloped at the end of the nineteenth century, and this likely prevented Italy from absorbing more rapidly the excess mortality, especially among the youngest cohorts: deaths caused by diarrhea and enteritis (negatively correlated with the presence of public health facilities) were considerably higher in Italy than in other countries (Caselli 1991). The convergence toward the French and British values slowed down during world wars, and in the mid-1970s the infant mortality rate still exceeded those of other richer countries. In 2008, this rate was equal to 3.3, a value higher than in Japan, Sweden, or Finland (around 2.5), but lower than in Germany (3.5), France (3.6), the United Kingdom (4.7), and the United States (6.7) (University of California Berkeley and Max Planck Institute for Demographic Research 2012). Until recently, infant mortality has been persistently higher for out-of-wedlock births, which may be seen as evidence of less adequate health conditions in poorest classes, where these births are more common (Manfredini and Pozzi 2004; Tizzano 1965). Data on people’s height can offer further insight into the well-being of a population, particularly in periods for which there is a lack of more informative sources (Fogel, Engerman, and Trussel 1982). Although the variation of individual heights is dominated by randomly distributed genetic potential, variations over time or across socioeconomic groups are driven by systematic differences in diet, disease environment, workload, and health care. The evolution of heights provides an indirect confirmation of the improvement in living conditions in Italy (Vecchi 2011; Federico 2003a; A’Hearn 2006): the average height of male conscripts steadily grew by almost a centimeter per decade between the cohort born in 1861 and the cohort born in 1980, a gain in line with the experience of other European countries (Hatton and Bray 2010). Interestingly, Italy does not seem to have experienced any period of declining heights, such as those that have been associated with industrialization in the United Kingdom and the United States (Williamson 1981; Komlos 1998; Steckel 2008). How can we combine the improvements in life expectancy with those on material living standards to arrive at a more comprehensive evaluation of the progress in well-being? Usher (1973) suggested that the value for individuals to live longer derives from the greater lifetime consumption that they can expect to enjoy. Thus, the value of longevity can be seen as the difference between the (higher) lifetime consumption that would have been necessary to generate the same level of utility if mortality rates had remained constant, and the actual lifetime consumption. By taking an isoelastic utility function, the increase in economic welfare can then be proxied by the growth in GDP per capita augmented by a term capturing the rise in longevity relative to a base year.3 Using the Istat life tables as assembled in the Human Mortality Database (University of California Berkeley and Max Planck Institute for Demographic Research

240

sources of growth and welfare

2012), it is possible to compute Usher’s GDP measure adjusted for longevity for Italy. For standard values of parameters, the measured progress in economic welfare would rise considerably relative to considering solely GDP per capita. The impact is strongest during liberal Italy, because of the fast improvement in life expectancy: between 1873 and 1913, adjusted GDP per capita grew twice as much as GDP per capita. The increase in life expectancy contributed less to welfare growth during Fascism and after World War II. These conclusions hold even looking at estimates net of the endogenous improvements in mortality attributable to rising incomes (Crafts 1997a).

Educational Achievements and the Human Development Index Education is a further important dimension of human well-being. School qualification is a major determinant of earnings capacity and social status, but it also captures, in a broader sense, the capacity of a person to function, to comprehend, and to act in the outside environment to do the things that he or she values. Italy’s current record in educational achievement is comparatively mediocre, be it measured by school attainment levels or by proficiency scores in surveys, such as the Programme for International Student Assessment or the Trends in International Mathematics and Science Study (see Chapter 9). According to Morrisson and Murtin (2009), the situation was even worse in 1870, when Italy stood out as a laggard among most other comparable countries: with an average of less than one year of schooling in the population aged fifteen to sixty-four, Italy was far away from the leaders (United States, 5.6 years; Germany, 5.5 years; and France and the United Kingdom, 4 years) but also from backward countries (Spain, 2.4 years; Japan, 1.7 years; and Greece, 1.5 years). This picture is consistent not only with many studies in the literature (Checchi 1997; Zamagni 2002a), but also with the testimony of contemporary commentators. The British historians Bolton King and Thomas Okey remarked in 1901: “Education is the gloomiest chapter in Italian social history, a chapter of painful advance, of national indifference to a primary need, of a present backwardness, that gives Italy (next to Portugal) the sad primacy of illiteracy in Western Europe” (1901, 233). Italy’s disappointing performance can be appreciated by comparing the new estimates of the gross enrolment rate for primary education made available by Vecchi (2011) with those collected for a number of countries by Lindert (2004). Over the years 1860–1930, Italy’s position is pretty stable at the bottom of the ranking, far below France, Germany, the United Kingdom, and the United States, but also behind Spain, despite the higher level of GDP per capita. During this period, Italy was outperformed even by Greece, whose GDP per capita was

standards of living

241

about half that of Italy in 1860, that succeeded in catching up and eventually overcoming Italy. King and Okey had little hesitation in identifying the causes of this poor record: “There have been thirty-three Education Ministers since 1860, each eager to distinguish himself by upsetting his predecessor’s work. Money has been stinted, and State and communes, lavish in all else, have economized in the most fruitful of national investments” (1901, 233). The Casati law, providing for compulsory and free primary education since the first day of unification, remained a dead-letter. “Prosecutions for non-attendance are probably unknown, and a head-inspector reports that he has never heard of one” (King and Okey 1901, 235). We lack long-run time series on attendance rates, but the evidence gathered by Vecchi (2011) suggests that approximately 20 percent of enrolled pupils did not attend school at the end of the nineteenth century, with small differences between boys and girls. The insufficient resources of local administrations played a crucial role, but review of the reports by ministerial inspectors reveals that inertia and lack of incentives also mattered, particularly in the South; in the Center-North poor infrastructure and climatic conditions were a further important impediment to school attendance. The evidence on literacy confirms this picture. At the time of unification illiteracy rates approached 80 percent, with a gender gap of about 14 percentage points in favor of males, and it took a surprisingly long time for Italy to catch up with other countries. According to the 2001 population census, illiteracy has been eradicated from the country, although many observers draw attention to other forms of “functional” illiteracy, which are more appropriate to assess the present situation (De Mauro 2010). A measure of human progress that accounts for educational achievement together with income and life expectancy is the Human Development Index (HDI). Since its introduction in 1990 (United Nations Development Programme 1990), the HDI has become a popular way of capturing essential dimensions of well-being into a single index easily comprehensible to the general public, despite controversies about its conceptual foundations (Klugman, Rodríguez, and Choi 2011; Ravallion 2012). According to the United Nations Development Programme (2011, Table 1), in 2011 Italy ranked twenty-fourth out of 187 countries, behind the United States, Canada, Japan, South Korea, and most Western European countries, including Ireland and Spain, but ahead of Luxembourg, the United Kingdom, Greece, and Portugal. Italy’s position in the international ranking has barely changed during the previous thirty years. The HDI takes a simple unweighted mean of three indicators: (1) income per capita (in logarithms); (2) life expectancy at birth; and (3) education. Its exact formula has changed a few times since its first release, and alternative formulae have been discussed in the literature. In this Chapter we have reconstructed the annual series of the HDI for Italy since unification using the formula proposed by Prados de la Escosura (2010), to be able to compare our series with his estimates for the OECD countries (Western Europe plus Australia, Canada, Japan,

242

sources of growth and welfare

New Zealand, and the United States), Latin America, Asia (excluding Japan), and Africa in selected years.4 All series are shown in Figure 8.8. (The sharp drops in the Italian series broadly correspond to world wars, whose periods are not covered by the other series.) In 1870, Italy’s human development gap relative to the OECD countries as a whole was remarkably large, about 50 percent, far more than the comparison of GDP per capita would suggest. Going beyond GDP emphasizes Italy’s backwardness at the onset of modern economic growth. Italy’s secular convergence to top performers is evident in Figure 8.8, but its pace is slow. It takes more than forty years to halve the gap, and there are no further improvements for the next twenty-five years, from 1913 to 1938. The post–World War II catchingup narrows the gap further, but it only falls below 10 percent after 1980. Italy reached in 1921 the level currently (2005) observed in Africa; in 1955 that of Asia (excluding Japan); and in 1971 that of Latin America. Achievements in the three constituents of human development, however, differ. In 2005, the nonOECD countries taken as a whole had approximately the same value of the HDI estimated for Italy in 1954, but that Italian value was achieved by virtue of a far higher income level (38 percent) than the one currently observed in the non-OECD world and by a correspondingly lower educational achievement (life expectancy was virtually the same). More than half of the HDI overall increase since Italy’s unification is caused by the improvement in life expectancy, almost four-fifths by the rise in educational levels, and only a tenth by the growth of income levels.

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 1861 1871 1881 1891 1901 1911 1921 1931 1941 1951 1961 1971 1981 1991 2001 2011 Italy

Africa

Asia (exc. Japan)

Latin America

OECD countries

Figure 8.8 Improved Human Development Index in Italy and world regions. Source: authors’ elaboration on data from Vecchi (2011) and Istat (2012); OECD countries (Western Europe, Australia, Canada, Japan, New Zealand, and the United States), Africa, Asia (exc. Japan), and Latin America Prados de la Escosura (2010, Table 3).

standards of living

243

When did It Occur that Italians became Well-Off? The discussion of the HDI illustrates the extent to which nonincome dimensions can modify an assessment of Italy’s secular progress based exclusively on GDP per capita. The problem with HDI, and with the GDP-augmented measures discussed previously, is that they assume a specific degree of substitution among the different dimensions: the same level of well-being can be achieved by exchanging x units of GDP per capita with y units of life expectancy, and so forth. These exchange rates reflect value judgments that can vary over time and be legitimately different across people. Rather than pursuing further these lines of investigation, in this section we use our dashboard of indicators to identify the different stages in the evolution of the Italians’ well-being since unification. We measure progress by the proportionate increase over the range zero to one, where zero is the value of the well-being indicator in 1861 and one is its value in the most recent available year. Thus, the value of the normalized indicator in a given year represents the proportion of the overall absolute increase accomplished between 1861 and that year. In Figure 8.9 we plot these normalized wellbeing indicators against normalized GDP per capita. The shape of the resulting curves reveals the extent to which progress is “balanced.” A curve (or section of curve) close to the forty-five-degree line, the line from the bottom-left to the top-right corner, indicates that the movements of the welfare indicator on the vertical axis parallel those of GDP per capita on the horizontal axis; an approximately concave curve suggests that the welfare indicator improves more rapidly than GDP per capita, whereas the opposite happens when the shape is approximately convex. The top-left panel compares the progress in life expectancy with that in GDP per capita. Using jubilees as benchmark years (the vertical lines in the graph), it stands out that the conquest of longevity was much faster than the conquest of economic prosperity: by 1911 GDP per capita had only achieved 5 percent of the overall increase between 1861 and 2010, whereas life expectancy had gained about 30 percent; by 1961 the progress had risen to 25 and 75 percent, respectively. The graph highlights how the improvements in GDP per capita underestimate those in well-being in the first century after Italy’s unification, because they fail to capture the swift gains in longevity of the population. The graph also leads to revaluate the gains in well-being after the Golden Age: exactly half of the total distance covered by GDP per capita during the past 150 years is covered between 1971 and today. (The curve for infant survival, which is not shown here, would closely match that for life expectancy, confirming the correlation between these two variables.) As to education, we consider adult literacy and school enrolment. Eradication of illiteracy among adults required an extraordinary long time to be completed.

244

sources of growth and welfare 1911

1961

2011

1911

1961

2011

1911

1961

2011

1911

1961

2011

Figure 8.9 The time patterns of selected welfare indicators versus GDP per capita in Italy. Source: authors’ elaboration on data drawn from various sources, as described in the text. Each panel compares the dynamics of a welfare indicator (vertical axis) with that of GDP per capita (horizontal axis); all variables are normalized according to the formula (x–x1861)/(x2010–x1861), where x1861 is the value in 1861 and x2010 is the value in the latest available year.

In 1951, still 14 percent of Italian adults could not read and write, and only four-fifths of the total progress had been accomplished; it took another fifty years to bring illiteracy close to zero. The pronounced concavity of the curve shows that this process was largely independent of income growth. Gross enrollment rates tell a different story, depending on the school level. Primary school enrollment rates improved impressively faster than income from 1861 to 1931. After World War I the enrollment of the population in primary school age (six to ten years) has become universal; the rates exceeding 100 percent for many decades, from the 1920s until the early 1980s, indicate the presence of pupils repeating grades and the school attendance by persons older than school-age. The enrollment rates for secondary education exhibit a strict positive correlation with GDP per capita. Conversely, advances in tertiary education enrollment rates seem to have lagged behind those in GDP. At 5 percent, the

standards of living

245

enrollment rate for tertiary school had in 1961 only accomplished one-tenth of the overall progress, compared with 25 percent for GDP per capita. A sharp acceleration in university participation has, however, occurred in the last two decades (see Chapter 9). The last chart, in the bottom-right corner of Figure 8.9, compares the dynamics of equality (the complement to one of the Gini index) with the dynamics of GDP per capita. The jagged curve reflects the discrete nature of the equality indicator. It shows that more than half of the secular advancement in equality was achieved during the 1970s, and that about a fifth of that advancement was eroded in the subsequent period. The timing of the progress varies if information on the level of income is integrated with that on its distribution.

Concluding Remarks “Backwardness” is probably the term that best summarizes Italy’s socioeconomic conditions at the time of the country’s political and administrative unification (Gerschenkron 1962a; Toniolo 1988). The well-being of Italians was not only far-off from the level reached today, but also from that enjoyed by the citizens of the most developed countries. Progress has been considerable in all dimensions ever since, but has not developed smoothly. The findings reviewed here are rich in implications for liberal Italy (1861– 1913). First, the well-being of Italians is severely underestimated if measured by GDP per capita. Many indicators, including health and education outcomes, improved dramatically vis-à-vis a gentle increase in average national income. Second, improvements were both in absolute terms (e.g., Italy in 1911 versus Italy in 1861) and in relative terms (e.g., Italy versus other comparable countries). We found no evidence of “dark satanic mills” during Italy’s initial stage of industrialization (Williamson 1981), at least in the aggregate indicators. Third, taking GDP as the yardstick of the living standard, Italy seems to have poorly exploited the unprecedented movements of capital and labor across national frontiers of the late nineteenth century globalization (Williamson 1996) to converge with the leading countries. However, looking at some nonmonetary welfare measures the conclusion may be reversed: Italy did succeed in catching up with the leaders in life expectancy and infant mortality. The question “convergence of what?” is therefore a nontrivial one when investigating gainers and losers in the process of globalization. Fourth, the average living conditions of Italians improved without a widening of income distribution; the improvements in living conditions spread across the country’s regions, including Mezzogiorno (Vecchi 2011). All this addresses some key issues raised by Federico (1996) and sheds new light on Italy’s economic development: we can no longer say, with Federico, that Italy’s success story is a little-known one.

246

sources of growth and welfare

World War I stopped Italy’s catching up. During the interwar years, Fascist Italy opted for autarchy, thereby partaking in the world-wise process of deglobalization. The well-being of Italians did not benefit from this choice; although GDP per capita continued to grow roughly in parallel with other countries, the march toward better living standards slowed down. In fact, now and then it even came to a halt: child work, for example, increased during the recovery from the Great Depression (1931–1936). Inequality seems to have risen too, but only slightly according to the estimates based on households’ incomes, more sizeably in terms of consumption (Rossi, Toniolo, and Vecchi 2001). The modest growth of household incomes recorded between 1921 and 1931 benefitted only people from the fiftieth to the ninetieth percentiles, with all others persons (in the bottom half of the distribution and at the very top) suffering an income fall; the gap between the North and the South widened in life expectancy at birth, infant mortality, and conscripts’ heights (Vecchi 2011). The combination of two world wars, the most severe economic crisis ever experienced in modern times, and the autarchic strategy had negative consequences, but all in all the well-being of Italians kept rising, if more slowly. Of course, this was not the case for civil liberties and political rights, dimensions that should be part of a fuller notion of well-being (Dasgupta and Weale 1992; Crafts 1997b). In the immediate aftermath of World War II, standards of living in Italy deteriorated dramatically. Hunger became widespread: according to Vecchi (2011), half of the population was undernourished during the years 1946–1947. The results of a Parliamentary Commission set up in 1951 depicted a picture of the diffusion and nature of poverty in the country as gloomy as in late nineteenth century investigations, irrespective of the indicator used to measure the incidence and depth of destitution. Italy had undoubtedly tumbled in the past. The recovery was impressively fast, however, and was paralleled by a profound transformation of society and the lifestyles of Italians. In the 1950s and 1960s, economic performance was extraordinary by historical standards, less so by international standards. The North-South income gap narrowed significantly, this being the only episode of convergence during the 140 years for which regional GDP estimates are available (Vecchi 2011). The findings of this chapter also lead to an interesting reappraisal of the 1970s. Despite harsh social conflicts, terrorism, political turbulence, two major oil crises, and high international instability, economic growth remained well above 3 percent per year and was characterized by a distinctly “pro-poor” bias; income inequality and absolute poverty fell, although inflation and unemployment rose. However, the 1970s left many problems unresolved, most importantly the structural weaknesses that would have impaired economic growth in subsequent decades (Rossi and Toniolo 1996; see Chapters 3 and 4). The rise of welfare expenditures (Ferrera 1984) sustained further improvements of living standards in the 1980s, but at the price of a public debt soaring from 51 percent of GDP in 1982 to 102 percent in 1990. Current well-being was being traded against the well-being of future generations.

247

standards of living

The currency crisis of 1992 marked a turning point. After 130 years of economic growth accompanied by an overall narrowing income distribution, Italy entered a phase of low growth, persistently high inequality, and increasing sense of vulnerability (Boeri and Brandolini 2004; Rossi and Vecchi 2011). The level of well-being achieved after 150 years since unification may not be necessarily enjoyed by future generations. The sense of accomplishment for the success in heading the country “from the periphery to the centre” (Zamagni 1998b) mixes with the concern that today’s achievements need not be forever (Ciocca 2007).

Acknowledgments The authors thank for useful comments Nicola Amendola, Tony Atkinson, Luisa Minghetti, Gianni Toniolo, and participants in the conference “Italy and the World Economy, 1861–1911,” Banca d’Italia, Rome, October 12–15, 2011. The views expressed here are solely the authors’; in particular, they do not necessarily reflect those of the Bank of Italy.

Notes 1. The newly reconstructed series for GDP per capita and the Maddison series show some local differences, but a similar dynamics in the long run. The main difference is around World War I, when the former does not show the sharp up-and-down exhibited by the latter. Hence, in the period 1914–1922 we have adjusted the dynamics of the Maddison series to mimic that of the new series. 2. The weights are the population shares resulting from the breakdown by area of residence (Northwest, Northeast, Center, South, and Islands) and sector of occupation of the breadwinner (agriculture, industry, and services). 3. The latter term is discounted by the inverse of the constant elasticity of utility to consumption. The reason for the discounting is that the more elastic is utility, the smaller is the reduction in the consumption flow necessary to keep the utility level unchanged, and hence the measured value of longevity. 4. Like the HDI since 2010 (United Nations Development Programme 2010), this formula takes the geometric mean of the three elementary indicators instead of the arithmetic mean originally used, to remove the assumption of perfect substitutability and penalize unbalanced developments in the three dimensions. The formula used here follows the old HDI practice to construct the education indicator as the average of adult literacy, with a two-third weight, and gross enrolment in primary, secondary, and tertiary schools, with a one-third weight. In the new HDI introduced in 2010 the education indicator combines, with equal weights, the

248

sources of growth and welfare

mean years of schooling and the “expected” years of schooling (number of years of schooling that a child of school entrance age could expect to receive if prevailing patterns of age-specific enrollment rates remained constant). All elementary indicators are normalized by taking the proportional country’s achievement over a prefixed scale; for nonmonetary indicators this distance is also subjected to a logarithmic transformation. After methodologic differences are taken into account, our calculations suggest a progress slower than that estimated by Crafts (1997a), but roughly in line with that found by Felice (2007b).

Chapter 9

HUMAN CAPITAL giuseppe bertola and paolo sestito

Introduction Like many other aspects of Italy’s historical experience, its human capital accumulation and economic development were shaped by the new country’s location at the boundaries of Western Europe’s industrialization process, and by the interplay of national rules and reforms with the heterogeneous characteristics of its constituent regional states. Aiming to characterize the educational system’s role in different historical phases of Italy’s development, this chapter compares its structure and performance over time with those of a group of countries with common European historical roots, and influenced by broadly similar external shocks. We also examine regional differences within Italy, which are informative because the country’s unification was itself an important economic integration experiment. Variation across regions and over time of education’s structure and effectiveness casts useful light on the mechanisms linking human capital accumulation and economic performance. Education can foster economic and cultural growth and convergence (or divergence) through several well-understood theoretical mechanisms (Checchi 2006). Schooling contributes to economic productivity not only through provision of specific skills, but also of general skills that ease on-the-job training and, at the primary level, of communication and relational skills that make it possible for individuals to function in a large society with extensive division of labor. The latter type of schooling plays an essential role in ensuring the political sustainability of national entities, because mandatory, free, and largely

250

sources of growth and welfare

uniform elementary education can replace family- and village-level solidarity with patriotic values. Later in life, routing youths to different tracks may help develop specialized skills and allow the best performers to access elite education. Such differentiated schooling may or may not improve the average effectiveness of education, but is very likely to increase the persistence of inequality, because the risk of failure in more demanding tracks is a stronger concern for children, particularly for those from more disadvantaged households. When examining the consequences and, especially, the motivation of Italy’s educational policies and institutions, one needs to keep in mind such distributional concerns. These were shaped by variable social and distributional objectives at the national level, and their implications for economic growth in Italy and across its regions were significantly influenced by different initial conditions and different local circumstances. The second section summarizes the evolution of Italy’s education system, in terms of relevant legislation and as regards practical details of organizational aspects (funding, regulation and implementation, access criteria and selectivity, tracking) that bear on a schooling system’s ability efficiently to produce suitable human capital. The third section discusses broad trends over time, relative to comparable countries and across Italy’s regions, of the educational system’s inputs (enrollment and expenditure) and effectiveness (as measured by literacy statistics and, over the most recent period, by standardized test scores), and interprets this evidence in light of the frequent and often only partially implemented reforms that shaped the Italian education systems, driven in turn by politico-economic mechanisms. The fourth section discusses the role of the resulting human capital stock’s dynamics and regional heterogeneity as a determinant of Italy’s economic performance, outlining in particular how some aspects of the Italian economy’s recent stagnation may reflect its education system’s structure and performance.

Structure and Evolution of Italy’s Education System Over its 150 years, Italy’s educational system has featured an evolving mix of features from preexisting regional states, on the one hand, and influences from developments and reforms in neighbor nation states, on the other hand. This section offers a necessarily brief and incomplete history of its structure and evolution, sketching also some relevant features of other European nations’ experience, and emphasizing key discrepancies between the letter of the national law and its actual implementation, which was often highly heterogeneous across the country’s regions.

human capital

251

The Situation in 1860 Two features of educational institutions in Italy’s constituent states shaped the new country’s evolution in the following periods. The first is that, overall, they were much less developed than in most other Western countries (Lindert 2004). The second is the presence of wide regional gaps (D’Amico 2010). Only in Austrian-ruled Lombardy and Veneto, in Tuscany, and in the Piedmont and Liguria parts of the Savoy Kingdom had primary education started to near the standards of Prussia, where eight years of primary education had been made compulsory in 1763, the Netherlands (where primary education was already under state control as early as 1806), and Scandinavian countries that soon followed suit. Austria had brought elementary schools under state control in 1774, even though it would wait until 1869 to effectively provide eight years of elementary education. The other regional states were much behind in terms of elementary school provision. Regional differences were less pronounced at the secondary school level. Licei had been introduced throughout the peninsula at the beginning of the nineteenth century, during the Napoleonic domination; in the North, Austrian-type ginnasi were also present, not only in Lombardy and Veneto but also in the northwest regions ruled by the Savoy king. Some heterogeneity could, however, be observed as regards the more or less academic orientation of these and other secondary schools’ curricula (D’Amico 2010; Genovesi 2010). In the Kingdom of Naples, secondary schools were mostly run for the elites by the Jesuit order, whose prerogatives had been restored in 1814; in Tuscany and Piedmont, the government had a stronger grip on the system and relied for outsourcing on less ideologic and more practically oriented religious orders, such as the Scolopians (see D’Amico 2010 on the different pedagogic orientation of the several religious orders). At the tertiary level, Italy was home to some of the Western World’s most ancient and prestigious universities. Only one was located in the continental South, in Naples (but three were in Sicily).

The Origins of Italy’s Education System (1860–1900) The Casati law, issued in November 1859 (when Lombardy, Parma, and Modena had already voted in favor of annexation to the Savoy Kingdom) was meant as a blueprint for the new Nation’s educational institutions, rather than as a regional state’s reform, and indeed became applicable in March 1861 to all territories of the newly established Kingdom of Italy. Its wide ranging scope shaped Italy’s education system for a very long time, and it was explicitly inspired by the German system of nationally directed education (Canestri and Recuperati 1976; Zamagni 2002a). As a matter of principle, and in contrast to the British system of no government involvement in education or the Belgian system of

252

sources of growth and welfare

public-private competition, privately organized establishments would be allowed to coexist with public ones, but all would be subject to a common regulatory framework. At the primary level, the law envisioned two years of entirely free and compulsory education. This was fully consistent with Italy’s new national ambitions and with the need to provide at least some knowledge of the new nation’s official language: in 1861, less than three percent of the new country’s citizens were Italian speakers (Scotto di Luzio 2007). In 1877, the Coppino law extended compulsory schooling to three years (with provisions for enforcement and fines for noncompliant parents) and introduced a five-year elementary school curriculum. In practice, however, Italy did not develop a truly common and mandatory primary education system during this period. Funding of primary education was almost entirely left to municipalities; even formally, rural municipalities (those with less than four thousand inhabitants) were only mandated to organize the first two years. This resulted in huge differences in the quality and even the existence of primary schools, which depended on local resources, and on the local sociopolitical propensity to devote resources to general education (A’Hearn, Auria, and Vecchi 2011). For the first half century of life of the new state, this crystallized the huge regional differences inherited from the preunitary period. At the secondary level, education was envisioned to be fully regulated and to a larger extent funded by the new nation’s central power. State control was meant to foster cohesion, and to diminish the role in education of the Catholic Church, hostile to a new kingdom that had removed the Papal States from its control. The new state focused attention mostly on the academic licei track, whose antecedents (and remaining competitors) were the Jesuit schools’ Latin-based curricula. Licei were established and run by the State in all provinces. With five years of gymnasium and three of liceo classico, they offered an academic curriculum meant to shape the culture of the new nation’s elite. In keeping with its German inspiration, the Casati law also provided for a less academic middle and high school parallel track. Despite its scuole tecniche denomination, it aimed at supplying students with general human capital rather than practical techniques; in 1872, four-year istituti tecnici were introduced at the high school level, and one of their four curricula was centered on mathematics and physics. Unlike licei, scuole tecniche offered some financial help to students, were largely funded by private or local sources, and were administered by the Ministry of Agriculture, Industry, and Commerce rather than by the Ministry of Education. A more clearly vocational track was organized in the form of scuole professionali, which remained under Ministry of Agriculture, Industry, and Commerce control even after the Ministry of Education was put in charge of technical schools and institutes in 1878. Supply and demand of practical secondary-level education was much more influenced by local cultural and socioeconomic features and funding sources, and scuole professionali were

human capital

253

very differently relevant across regions and sectors (Zamagni 1996). In urban areas, working-class youths could also enroll in new religious schools, such as those run by the Salesiani, that offered some practical and cultural education. As to tertiary education, the Casati law tasked the new nation’s ancient and reputable academic institutions to transmission of specific, practical, professional knowledge, and to the maintenance and advancement of culture; no universities were closed, and Naples remained the only university in the South of the peninsula. The law provided again for an applied educational track, parallel to traditional university-level education, instituting scuole tecniche superiori engineering schools; scuole commerciali for business studies were also instituted in 1868, a tertiary-level agriculture school opened in Portici in 1872, and others followed. The educational system built in the new nation’s institutional structure, and its uneven implementation, arguably reflects the “elite democracy” social and political texture of the new nation. Because lower classes had little or no political voice and power, the State’s public funds were mostly devoted to secondary and tertiary academic education, whereas education was mostly funded and organized locally at the elementary level and in applied fields.

Broadening and Tracking (1900–1970) After some political turbulence over the 1890s, the political climate shifted in more progressive directions. The 1904 Orlando law widened the scope of primary education (in terms of years of mandatory schooling and tasks assigned to all municipalities) and further enlarged the financial support from the State to the municipalities. After the results of the influential 1908 Corradini enquiry, the 1911 Daneo-Credaro law finally fully shifted the cost of all personnel and materials for primary education to the central State budget, leaving local governments in charge only of providing adequate buildings (which remains their obligation to this day). The cultural shift that produced this important reform of the Italian educational system’s funding also spurred debates regarding its organization and syllabus. After the war, modernization tendencies eventually resulted in the Gentile reform enacted in 1923 by the newly established Fascist regime. Many aspects of the new design were similar to those originating from the Casati law; others instead aimed (rather controversially, also because of Minister Gentile’s strong personality) at strengthening the role of the central State, the early tracking features of the Italian system, and the pivotal role in it of liceo Classico as an elite track characterized by a difficult syllabus and tough assessment standards (Scotto di Luzio 2007). The elementary school curriculum was shortened to five years, and the State role was extended beyond funding with the institution of administrative teacher control and adoption of a common syllabus. These new features aimed at fostering political loyalty and national homogeneity.

254

sources of growth and welfare

Mandatory schooling was extended to age fourteen but, after elementary school, the tracked structure of the system was strengthened by instituting a dead-end scuola complementare track, which did not allow its students to proceed beyond lower secondary school, alongside an academic scuola media that made it possible to enroll in upper-secondary institutions (including istituti tecnici, which were again brought under control of the Ministry of the Economy). Possibly also with the aim of strengthening the elite connotation of liceo classico, the reform introduced new types of secondary schools, notably a four-year liceo scientifico (a diluted version of liceo classico that would in time become the most popular upper secondary track) and a liceo for female students (that would disappear). The standards (not only of the elite academic track, but of all schools) were enforced by central controls, because performance was monitored and certified by State-run examinations at the end of elementary, middle, and high school cycles. The resulting system was thus strongly segmented at very early ages. Strictly ranked secondary school alternatives were meant to maintain the high quality and purity of liceo classico and university students who, in Gentile’s words, were supposed to be “few but good.” The Gentile reforms also aimed at segmenting tertiary education, distinguishing the elite academic university degrees from professionally oriented education. Licei degree holders could enroll in any university degree program, whereas istituto tecnico degrees allowed enrolment only to programs in specific subject areas, where qualifications were assessed by national Esame di Stato examinations. The timing and sharpness of tracking proved very controversial, even under Fascism. Already in 1928 a new minister, Belluzzo, replaced scuola complementare with an avviamento al lavoro lower secondary school that made it possible for students to access technical schools. Further reforms were envisioned by the Bottai plan, which included tracked education in the Fascist corporatismo blueprint for Italian society, but had not been implemented as the war brought an end to the totalitarian regime. After the war, accordingly, the Italian education system was essentially still shaped by the Casati law’s blueprint of a centrally managed, early tracked structure. That structure had been stiffened in the interwar period, however, by stronger administrative tools, such as the definition and central enforcement of primary school teacher qualifications (Ruolo unico nazionale dei maestr i law 1942/675), and especially by the institution of State examinations at the end of each course of study. The structure envisioned by legal provisions, however, was still far from being fully and effectively implemented. Schooling was supposedly mandatory until age fourteen, but was effectively provided up to that level only when, in 1962, the lower secondary level was unified in a scuola media unica – after the five-year primary education cycle – providing for some postponement of the early tracking system. In 1968 the State also introduced scuola materna at the preprimary level, where education had until then been provided privately or,

human capital

255

in some cases, by municipalities (which still represent most of the pre-primary services’ providers). The relationship between the new scuola media and the unreformed elementary and upper secondary segments of the resulting educational system was not made explicit by the 1962 reform: although different unions lobbied for or against allowing teachers qualified for the elementary level to teach in the new lower secondary program, legislators (and Italian society) remained ambivalent as regards the preferred intensity and features of tracking. At the upper secondary level, academic licei and mildly less demanding istituti tecnici continued to coexist with a variety of professional institutes, intertwined with local (and national) industry, and with schools organized internally by such large firms as Olivetti from the 1950s, patterned on American corporate training institutions.

Maturity and Deterioration (1970s to Present) The transition to a fourth period of institutional evolution is marked by the period of student revolts and worker strikes that erupted in 1969. By 1973, sweeping reforms had allowed free access by any secondary school graduate to any tertiary degree program. University enrollment boomed; degrees earned also began to grow a few years later, but the labor market inflow of university graduates was reduced by high drop-out rates and long degree completion times because universities were poorly equipped to educate large masses of unscreened students. Free access to universities was supposed to be followed by suitable university reforms, but only in 1980 a reform began to be implemented, replacing the previous German-style institutes, professors, and assistants with three faculty ranks, organized in departments. Almost all of those who in the 1970s had been employed on temporary teaching and research positions were tenured, without proper screening, and this made it difficult for subsequent younger cohorts to start university careers. The administration of university personnel and the structure of degree programs remained fully centralized until the 1990s, when many new universities were established all over Italy. Around 2000, universities were finally granted broad and uncontrolled freedom to hire and (more frequently, because of obvious evaluation biases and smaller budgetary costs) promote their faculty, and to design new courses in the three-plus-two harmonized European format. At the upper secondary level, the system based on licei, istituti tecnici, and istituti professionali until recently maintained its prewar configuration. Comprehensive reform of the upper secondary tracks remained pending (and hotly debated) for about fifty years. Most of the local connections between technical- and vocational-oriented tracks and the entrepreneurial world have

256

sources of growth and welfare

been severed since the mid-1970s. Since the early 1990s school-level sperimentazioni were allowed to introduce fairly substantial changes in the admittedly obsolete upper secondary school curriculum. Despite their name, these innovations were neither introduced in a properly experimental setting, nor evaluated. They did not substantially change the academic orientation of the licei’s stream, which was increasingly unsuitable to their wider and less selected population of students. Although until the 1980s the rise in high school enrollment had been driven by technical and (initially) professional tracks, since the 1990s enrollment has favored licei, perceived as more prestigious and a better gateway to university. A new simplified, and less expensive, structure has only recently been introduced, and has begun to be implemented in 2010. The fact that all upper secondary schools would have to prepare students for potential university access further complicated the inconclusive debates that followed the 1962 scuola media unica reform of lower secondary schools. Lower secondary schools’ mission was not fully clarified. They moved from being preparatory schools arranged along already differentiated tracks to provision of a universal and mandatory education, but it was unclear whether they should serve as completion of the primary segment, or as an introduction to an upper secondary level that was also gradually becoming almost universal. Today, lower secondary teachers seem to fare worse than those in other schools’ segments in job satisfaction surveys (Argentin and Cavalli 2010). In the primary segment, where teachers up to very recently were not required to have a tertiary degree, teacher numbers increased faster than enrollment with the introduction (in 1971) of full-time sections and the assignment (in 1977) of additional teachers to help problematic students who were previously confined in remedial sections; between 1985 and 1990, the elementary school curriculum was reformed assigning multiple teachers to each class. As a baby bust followed the baby-boom’s peak reached in 1964, the pupil-to-teacher ratio declined along with enrollment rates: in public elementary schools, it was 27.3 in 1950, 22.2 in 1960, 21.4 in 1970, and 15.7 in 1980 (Brunello and Checchi 2005); in 1992, it was at 10.8 the lowest of the Organization for Economic Cooperation and Development International Indicators Project data. Not only the mismatch between unreformed curricula of different school levels and mass school attendance, but also administrative and legislative innovations may have reduced the effectiveness of Italian schools’ organization and operation. Hiring and retention of teachers gradually lost any remaining selective feature, and the lack of selectivity at the access gate may have been further compounded by the self-selection implications of the erosion in the social prestige and the economic appeal of the teacher profession. Increasingly rare and over-crowded official competitions for permanent teaching positions (kept below the numbers required by current enrollment to avoid longer run budgetary costs) implied that the many would-be teachers who had earned degrees in the newly accessible universities could only obtain temporary employment. Because access to tenured position could result from accumulation of seniority

human capital

257

in such positions, rather than official competitions, teaching became an occupation for individuals who either are very motivated and willing to wait for a long period of time before getting a permanent position, or lack alternative employment opportunities. When teachers finally obtain a tenured position, they earn the right to move to other schools on the basis of seniority, without any performance evaluation. As a result, teacher/student ratios and teacher seniority are higher in the South and, within each local district, in schools with better socioeconomic catchment areas. In the areas and schools with more pressing needs, teachers are not only fewer and less experienced but also, because they hope and expect to move soon, poorly motivated to supply effort (Barbieri, Rossetti and Sestito 2011). The implications of inefficient teacher allocation mechanisms are compounded by changes in student assessment procedures in the direction of weaker standards and reduced comparability of examination results. The process began with the reform of the Esami di Maturità in the early 1970s and possibly peaked between 1995 and 2006, when students who failed only a few subjects were no longer required to make up their shortcomings during the summer break. Already in June, teachers had to decide either to have them repeat the grade, or to pass them to the next grade; the latter was the more agreeable option for teachers in an environment where anxious parents, rather than Ministry inspectors, would be questioning their choices. From 1997 until 2005, the Esame di Stato at the end of high school was administered by the students’ own teachers, with only one Ministry-appointed external examiner.

Education Inputs and Outputs Italy’s institutional evolution offers a very rich setting for evaluating the determinants and effects of education quantity, quality, and funding modes. In this section, we first examine the behavior over time and across regions of indicators of the Italian education system’s input and output, aiming to detect the implications of the institutional features reviewed previously. Then, we compare them to those observed in other countries, preparing the ground for the next section’s analysis of education’s role in explaining Italy’s economic development.

Long-Run and Convergence Dynamics The consequences of the institutional developments briefly reviewed previously are readily apparent in the expenditure and enrollment data shown in Figure 9.1. Public expenditure remained low until primary schools shifted from local to State funding; enrollment in primary schools took off only after the

258

sources of growth and welfare .10 1 .9 .8 .7 .6 .5 .4 .3 .2 .1 0

.08 .06 .04 .02 0 1880

1900

1920

1940

1960

1980

Education expenditure/GDP, left axis: Municipalities + other local government + State Enrollment rates, right axis: Primary students/6−10 population Secondary students/11−18 population

Figure 9.1 Source: Checchi (1997).

provision of State funding, and approached 100 percent only in the 1930s. Even though enrollment increased continuously, attendance was lower and actually declined between 1881 and 1893 and between 1926 and 1936 (A’Hearn, Auria, and Vecchi 2011). Over time, public spending in education increases almost always faster than GDP. For the initial portion of Italy’s history, this reflects the country’s growing population and standard of living, and the resulting increasing propensity to invest in a rich-country’s educational system. More recently, school expenditure was buoyed by a tendency to overstaff schools even as school-age youth began to decline, and economic growth to slow down. Table 9.1 displays summary indicators of long-run dynamics and regional differences in primary school effectiveness, which was poor as indicated by low literacy rates that until the early 1900s also remain persistently different across macro regions. Both features are largely explained by local funding of primary schooling. In the late 1800s restricted democracy polity of most European countries, the land-owning classes endowed with political power were not much interested in raising the literacy of their prospective salaried workers (Lindert 2004). Their propensity to fund education was even lower at the local level, where it would have spurred political awareness and wage-increase pressure, also through out-migrations. Although these factors were likely relevant across the whole country, in the South they operated more strongly: local resources (as measured by gross domestic product [GDP] per capita) were slightly lower than elsewhere and this naturally reduced municipalities’ propensity to fund

259

human capital Table 9.1 Illiteracy and schooling over time and across regions Regions

1871

1901

1921

1951

1951

1971 2001 2010

(I)

(P)

(I)

(P)

(I)

(P)

(I)

(P) Average years of schooling of 25–64 population

Northwest

42.4

63.3

21

67.5

8.2

71.1

2.9

68.7

5.2

5.7

8.9

11

Northeast

68.2

41.1

40.7

58.7

18

72.7

6.3

72.2

4.6

5.3

8.6

11

Center

70.5

31.8

49.3

48.2

28.8

61

11.3

70

4.4

5.3

8.8

11.3

South

83.9

22.2

69.6

33.4

47.2

45

24.2

68.3

3.1

4.1

7.4

10.2

ITALY

68.8

38.5

48.5

49

27.4

57.6

12.9

68.3

4.1

5

8.3

10.8

Note: (I) Illiteracy rate: percent ratio of illiterate individuals to population older than six years of age. (P) Primary schooling enrollment: percent ratio of primary school enrollment to six to fourteen years old population. Average years of schooling: computed on the basis of minimum number of years necessary to attain the highest educational title held. Source: Regional indicators up to 1951 are taken from Felice (2007), where they refer to the total adult population; here, they are rescaled to the age structure and overall average from the 1971 census. Later data are taken from the census (between 1971 and 2001) and from the Labor Force Survey for 2010.

education as it was seen as a relative luxury; A’Hearn, Auria, and Vecchi (2011) show that in regions with better-funded primary education a higher share of municipal tax receipts were spent on education, and note that regional diversity was likely rooted in cultural, political, and economic differences. The demand for literacy and numeracy skills was related to the presence of modern industrial firms, hence much weaker in the South, where the dominance of landlords in local politics was possibly stronger; the stronger presence of Catholics and their fears vis-à-vis public schools may also have played a role. In the case of males, military service taught basic literacy and numeracy at scuole reggimentali (founded in the 1830s in Piedmont, expanded throughout the independence wars period, and only abolished in 1892; Della Torre 2010). Because the draft was dodged by many in southern regions, the northern regions literacy advantage was especially high for males, who also had a prominent role in formal economic activities, thus amplifying North’s human capital advantages. Regional differentials persisted almost unchanged until the early 1900s. Although primary enrollment caught up quickly by 1921 and was more or less uniform by the 1950s, differences in illiteracy rates in the overall population and, more broadly, in the average years of schooling of adult population were still wide in 1951. Today, most of them have been eliminated, at least among younger cohorts. The dynamics of postprimary enrollment lag behind those of primary enrollment in Figure 9.1. For a long time, secondary schools and their students were not many but high-quality. They mostly focused on classical studies, with a strong presence of technical schools. A quantitatively significant vocational

260

sources of growth and welfare

Table 9.2 Regressions for ten-year average schooling years of twenty-five to thirty-four population at the municipal level T

M

F

Initial average years of schooling, twenty-six to thirty-four population

−0.349

−0.383

−0.428

−0.364

−0.419

−0.488

0.004

0.005

0.005

0.006

0.006

0.006

1991 and 2001dummy

−0.162

−0.364

0.345

−0.147

0.014

0.015

0.016

0.014

0.015

0.017

2001dummy

0.366

0.418

0.484

0.359

0.374

0.469

0.011

0.011

0.014

0.011

0.012

0.014

0.02

0.054

0.082

0.005

0.005

0.006

Initial average years of schooling, thirty-five to sixty-four population Constant

Observations R2

T

M

−0,347

F

0,402

4.369

4.588

4.754

4.384

4.601

4.758

0.025

0.029

0.025

0.025

0.029

0.025

24,168 0.51

24,168 0.52

24,162 0.43

24,168 0.51

24,168 0.53

24,162 0.43

Note: T = all individuals, M = males, F = females. Standard errors below point estimates. Source: Computations by Federico Giorgi on 1971–2001 census data.

secondary track began to grow in the 1930s, and continued to develop after the World War II especially in the Northwest, but also in the Eastern and Central Adriatic regions that caught up with the traditional industrial triangle (Zamagni 1996). Since the 1970s, enrollment has been continuously rising in the unreformed secondary school institutions that were awkwardly positioned between scuola media unica and universities. In the 1970s and 1980s, higher enrollment was concentrated in the more vocationally oriented tracks; since the 1990s, however, the enrollment share of academic tracks, especially liceo scientifico, has increased more strongly. A quantitative assessment of trends and convergence phenomena is provided in Table 9.2, where we regress ten-year changes in the educational attainment of young adults (the twenty-five to thirty-four years old) on their starting level and a set of time dummies (the first three columns, respectively, for the total, males and females). The following three columns also add as a control the educational attainment of the previous generation. The average years of schooling of thirty-five to sixty-four years old individuals at initial the census date has a positive coefficient, indicating that intergenerational spillovers increase the persistence of past historical heterogeneity. The lagged dependent variable

human capital

261

coefficient remains large and significant, however; convergence is strong, and stronger among females, who also experience faster and accelerating trend growth. The evolution of Italy’s human capital was also influenced by the in- and out-migrations discussed in more detail by Chapter 10. Until the last quarter of 1900s Italy was an out-migration country. Because migrant individuals were on average less educated than those who remained in Italy (and than of residents of destination countries), migrations in all likelihood tended to decrease Italy’s education gap vis-à-vis the other Western countries. This phenomenon was particularly intense in the decades before World War I and after World War II. Although migrants were relatively uneducated persons, they were among the most entrepreneurial in their cohorts and the acquisition of a minimum threshold level of education was very often a trigger factor allowing them to migrate. In the more recent decades Italy has become an immigration country, and migratory flows no longer reduce the Italian education gap vis-à-vis other developed countries: not only because immigrants to Italy are less educated than Italian natives and the relatively few but well-educated Italian emigrants, but also because they are less educated than immigrants to other developed destination countries. The Italian socioeconomic system seems less able than other developed countries to attract qualified (and legal) immigrants; most immigrants in Italy are initially irregular and working in the black economy (until they are regularized by amnesties, taking place every three to five years in recent experience). Also within Italy the last two decades have seen a reversal of trends that previously contributed to regional convergence in human capital. Before World War I, relatively uneducated migrants were leaving the less developed and least endowed (in terms of human capital) regions in the Mezzogiorno and the Northeast (individuals from the northwestern regions were likely to migrate only temporarily in France and Switzerland). Although external and internal migrations were blocked during the Fascist period, migrations boomed again in the 1950s and 1960s, with less educated individuals leaving the Mezzogiorno and settling either in Northern Europe or in the northwest (to work as bluecollars in the booming manufacturing sector). These flows decreased in later decades, and their composition changed: since the 1980s individuals leaving the Mezzogiorno are among the most educated (see Cannari, Nucci, and Sestito 2000; Mocetti and Porello 2010), and internal migration contributes to maintain the regional gap in human capital endowment.

International Comparisons Figure 9.2 offers a comparative perspective on Italy’s educational attainment, plotting the average years of education of the working-age population in a group of comparable Western European countries. Italy is clearly a laggard in this group, which is otherwise rather homogeneous apart from Finland, initially at

sources of growth and welfare Average school years of 15−64 population

262 15

10

5

0 1860

1885

1910

1935

1960

1985

Austria

Belgium

Finland

France

Germany

Italy

Netherlands

Norway

Sweden

Switzerland

United Kingdom

2010

Denmark

ALL

Figure 9.2 Time path of average school years of working-age population in twelve Western European countries (source: Morrisson and Murtin 2009), and in the aggregate of these countries (weighted by population data from Angus Maddison Web site, http:// www.ggdc.nl/maddison/).

the bottom of the pack and rapidly getting to the top. In quantity terms Italy’s human capital intensity is always lower than the average, consistently with the relatively low level of Italy’s income. It does catch up rather steadily (albeit not as quickly as Finland’s, that overtakes Italy in 1950); even between 1990 and 2000 the relative education intensity of Italy’s potential labor force is still growing fast, if not as fast as in the nineteenth century. Not only the quantitative size of the Italian educational system (in terms of financial resources, students, and teachers), but also its quality needs to be compared with that of other countries, and across different historical periods. Before 1900, the quality of the enrollment and expenditure inputs was not high, because attendance was not enforced, and elementary teachers’ qualifications were not defined and monitored effectively. The changes since 1970 of the organization of teacher careers and student assessments may also have had negative implications for the extent to which formal educational attainment translated in actual competencies. Theory and empirical evidence suggest that efficient use of schooling resources is fostered by autonomous choices made by individual schools but only insofar as standardized students assessments provide for the right amount of accountability of those choices (Woessmann 2007). Thus, the evolving institutional and organizational features discussed in the previous section are potentially much more important than the number of years spent in school as a determinant of the skills and competencies that are measured by standardized tests and that may in turn determine economic productivity.

263

Cognitive skills, all math and science (test score)

human capital Switzerland 5.1

Finland

Austria France

5 Denmark

Germany

Netherlands

Belgium Sweden United Kingdom

4.9 Norway 4.8 Italy 4.7 .88

.9

.92

.94

.96

Share of students reaching basic literacy

Figure 9.3 Horizontal axis: fraction scoring above basic level in literacy tests. Vertical axis: average mathematics and science test scores. Both standardized across all results available (in various years) for countries considered in Figure 9.2. Source: Hanushek and Woessmann 2009.

International surveys of student competencies (PISA for the fifteen-year-old students, PIRLS and TIMMS for the fourth and eighth grade students) show wide differences across countries (and within countries) in the actual skills of individuals with similar formal education. Figure 9.3 displays the standardized average of all available tests, reported in Hanushek and Woessmann (2009), for a set of comparable Western European countries (in the complete data set, only Luxembourg, Greece, and Portugal do uniformly worse than Italy, whereas Japan is the best performer). Most of the available evidence on the quality of students’ competencies refers to the most recent years. To examine whether such a gap has been a constant of the Italian history we looked at the admittedly more sparse evidence coming from the earliest international comparison exercises. Table 9.3 reports data from 1971, when three different grades were tested, and 1991, when only two of those three were covered. Because relatively few countries were participating in both exercises, Table 9.3 reports Italy’s scores in the original metric and Italy’s scores rescaled with respect to either the European Union countries participating in each exercises, or the subgroup of five European Union countries (including Italy) participating in both. In the 2000s data (discussed in detail by Cipollone, Montanaro and Sestito, 2012), Italian students fare poorly in the lower and upper secondary schools covered by the TIMSS and PISA surveys, but have better results in the TIMSS and PIRLS tests taken at the primary school level. The table shows that this was already the case in 1971, when Italian upper secondary students were a small and select group. Over time, the survey administered in 1991 shows a particularly pronounced

264

sources of growth and welfare

Table 9.3 Italian students’ test performance at difference ages and times 1971 Survey

1991 Survey

Score

EU = 100 (a)

5EU = 100 Score (b)

EU = 100 (a)

5EU = 100 (b)

Ten years old

19.9

104.1

103.4

529

103.5

100.6

Fourteen years old

27.9

105.3

105.9

515

99

98.4

Seventeen to eighteen years old (finishing secondary school)

23.9

83

86.2

—

—

—

Note: Overall reading comprehension average score of Italian youth, normalized (a) by the average of all EU students taking the same test and (b) by the average of students from the five EU countries participating in both 1971 and 1991 surveys: Finland, Sweden, the Netherlands, Belgium (Flemish regions in 1991, all regions in 1971), and Italy. Source: Elaborations on data from Thorndike (1973) and Elley (1994).

deterioration in the performance of fourteen-year-old students at the end of the lower secondary school. Another piece of evidence concerning the possible deterioration in the quality of the output of Italy’s education system may be drawn from the International Adult Literacy Survey. The information, collected in 1994, 1996, and 1998, traces the education (in terms of number of years of formal schooling) and skills (in terms of performance on a literacy test) of adult individuals whose schooling experience dates back to many years in the past. This makes it possible for Meschi and Scervini (2010) to construct summary statistics along these dimensions for the cohorts that, in each country, are adequately represented in the sample, and were born at various times over the last century. Comparing Italy with the other Western European countries represented in the International Adult Literacy Survey sample one sees not only that Italian cohorts lag behind their contemporaries in terms of years of education, but also that literacy is higher elsewhere than in Italy at almost all education levels in terms of years spent in school. Strikingly, the Italian cohorts born in 1955–1959 achieve more years of education than those born in 1950–1955, but are not more competent. Although this type of evidence can only be suggestive, it does offer intriguing indications that some of the institutional developments reviewed previously had measurable negative implications for schooling quality implications. The data also indicate that education is more heterogeneous within Italian cohorts in terms of schooling years and test performance. Sacrificing the pursuit of excellence to more equitable results in education might have been a reasonable and deliberate choice of the institutional evolution described in the second section, which was indeed largely driven by equity concerns. However, Italy’s education system seems to fare rather poorly also in terms of equity. Actual competencies vary widely within Italy, along regional lines and across

human capital

265

schools within each region. Southern regions have almost completely caught up in terms of years of education, but still lag significantly below northern regions in terms of student competencies. A portion of the geographical test score variation is related to parental background and schooling inputs variation (Bratti, Checchi and Filippin, 2007). The teacher allocation mechanisms discussed previously tend to amplify differences in such observable respects, however, because it tends to assign relatively ineffective and poorly motivated teachers to difficult teaching environments (see Barbieri, Rossetti, and Sestito 2011, for further discussion and some evidence). Differences in the expectations and goals of students may be relevant: in the South (where public employment is relatively more important) obtaining a formal degree through locally administered examinations may be more important than achieving the competences assessed by standardized tests.

Human Capital and Economic Development Taking stock of the previous sections’ review of institutional developments and of evidence regarding the resources, size, and quality of education, this section discusses whether and how the country’s checkered education history may have played a role in the 1861–1896 and 1995 to present periods of slow growth, and in the 1901–1915 and 1950–1980 periods of strong growth. To foster economic growth, formal education should be available in suitable quantity and quality, and its production should be organized efficiently so as to minimize its cost. In all phases of the evolution of Italy’s educational system, the quantity, quality, and cost of its human capital production was shaped by changing and controversial views as regards the desirability of equal outcomes, in a setting where opportunities were far from equalized and the linkages between efforts and results rather weakly defined. Because the same sociopolitical factors that shaped production of human capital in schools and universities also influenced the organization of factories and offices, however it is not easy to conclusively disentangle the specific role of human capital accumulation processes from these broader cultural and socio-political factors. What is clear is that Italy’s economy has generally been much more productive that one would expect on the basis of its human capital intensity. Figure 9.4 plots the observations of average years of schooling of the working age population and of domestic production across all the countries and periods covered by the Morrisson and Murtin (2009) data. The relationship between the two indicators for Italy is located on the bottom-right boundary of the data cloud. Compared with Western European countries with similar historical roots, and similarly variable exposure to international trade and technologic trends, Italy

266

sources of growth and welfare

Average school years of 15−64 population

15

Italy2010 Italy2000

10

Italy1990 Italy1980 Italy1970 Italy1960 Italy1950 Italy1940 Italy1930 Italy1920 Italy1910 Venezuela1960 Iraq1980 Italy1900 Venezuela1950 Italy1890 Venezuela1940 Italy1880 Iraq1970 Italy1870 Iraq1960 Iraq1950

5

0 −1

0

1

2

3

4

Log of real GDP per capita

Figure 9.4 Horizontal axis: GDP per capita in purchasing power units (source: Angus Maddison Web site, http://www.ggdc.nl/maddison/). Vertical axis: average school years of working-age population (source: Morrison and Murtin 2009). All available countries and periods are plotted. Darker symbols mark observations for the twelve Western European countries considered in Figure 9.2. Observations for Italy, and some with even lower education/income ratios, are labeled with country name and year.

was typically producing more than should have been implied by its workers’ human capital. To assess the extent to which the country’s educational system was suitably configured to exploit the country’s economic structure and changing comparative advantage, we focus next on the interaction between regional heterogeneity and the common forces shaping the various stages of economic development.

Education and Localized Growth in Early Italy The better performance of northwestern Italy starting in the 1890s may be partly traced to human capital. The regions that led Italy’s industrial revolution were also those with the highest preunification literacy rates, and this is not a coincidence. Some evidence along these lines is in Felice (2011), who reports the results of conditional convergence growth equations estimated across sixteen Italian regions over different time periods: human capital differentials are positively and statistically significantly related to growth over the 1891–1911 and 1911–1938 subperiods, and when all periods are pooled together. The human capital differentials inherited from the past, and affected by the uneven implementation of the Casati law, are likely to have interacted with broader developments: political unification in 1861 slashed tariffs across regions and, temporarily, vis-à-vis

human capital

267

international partners. This arguably made industrialization easier in the slightly more developed and better endowed northwestern regions, that would later also benefit from external trade protection, enacted since the late 1870s, and from availability of cheap hydroelectric power. All these possibly small advantages were amplified by the presence of sufficient human capital in the northwestern regions (Gagliardi and Percoco 2011); other regions specialized in agricultural production for the internal market, which in some southern regions, such as Campania and Sicily, replaced export-oriented and labor-intensive modern cultivations (Cohen and Federico 2001). As a result, Italy’s industrial revolution not only generated the 1900–1915 growth acceleration, but also widened previously shallow regional differences in the level and sector composition of income and production (Fenoaltea 2003).

Education and the Postwar Boom Finding that Italy stands out as peculiarly poorly educated among its industrialized peers, we need to wonder what may explain the country’s relatively high standard of living and its especially fast growth in the period after World War II. The 1950s–1960s economic boom was largely driven by growth of relatively low-skill manufacturing industries, which could exploit the interwar expansion and improvement of primary and secondary education. To produce high and increasing per capita incomes, however, low-skill labor needs to be combined with other factors of production, and Italy’s physical capital stock is not higher than that of comparable countries in available data sets, such as those compiled by Nehru and Dhareshwar (1993) for the 1950–1990 period, or those included in the 5.6 release of the Penn World Tables. The country’s income/education ratio, in data such as those displayed in Figure 9.4, is higher than that of all but some oil-producing countries. This suggests that Italy’s output may be produced by factors that, like natural resources, are unrelated to formal schooling and may be more or less strongly demanded by the global economy. In Italy’s case, the diffused knowledge and habits accumulated in centuries of civilization may well have made it possible to organize production efficiently through informal organizational arrangements and knowledge transmission channels. The ability to do so is a type of “social capital” that is related to generalized trust and respect of the spirit of the law (see Chapter 11), but need not coincide with it; in Italy, generalized trust is generally low, and informal arrangements more prevalent than in other countries, but informal organization and generalized trust seem to be positively correlated across regions of Italy. An indication of the small role of formal education in Italy’s postwar success story is the fact that, in census data, the lowest incidence of university education at the province level is in Biella and Prato, where high incomes were and still are the result of successful textile industries organized in tight distretto

268

sources of growth and welfare

industriale clusters of small firms. More generally, there is a negative relationship between educational achievements and the prevalence of such production structures also for furniture, shoes, and other typical “Made in Italy” premium quality light manufactures. This is an unsurprising implication of the fact that, during the boom decades, youths who could easily obtain employment and onthe-job training had little reason to enroll in secondary and tertiary schooling. On the labor demand side, industrial organization might in principle itself be endogenous to the lack of formal education, but more plausibly make it redundant by providing alternative and more effective channels of intergenerational knowledge transmission. These include apprenticeship and on-the-job training, and the less-than-high-school vocational track which, as discussed in Zamagni (1996), followed much the same time trajectory and regional pattern as the manufacturing sectors that fuelled Italy’s post-war boom.

From Growth to Stagnation Italy for a long time managed to sustain a high and increasing, if geographically unequal, standard of living despite relatively low levels of formal education, which did not prevent the early twentieth century infant-industry growth (but did concentrate it in the region with the highest literacy), and was not an obstacle (even across regions) to the postwar economic miracle. In international comparisons, this apparently gravity-defying model may well be explained by the fact that Italy’s comparative advantage privileged sectors in which academic and scientific knowledge play unimportant roles. Informal mechanisms of information transmission and training, within highly specialized industrial districts and in the scuole professionali vocational track, were in fact especially well suited to foster local productivity in such sectors also across regions within Italy. Whatever kept Italy growing without education ceased to be effective over the period when detailed growth-accounting estimates are available from the Conference Board Total Economy Database (http://www.conference-board.org/ data/economydatabase/) for the twelve-country comparison group of Figure 9.2. In those data, Italy’s 1990–2009 growth performance, at less than 1 percent yearly average, is the second slowest (only ahead of Switzerland); the contribution to it of labor composition, measured in terms of years of education, is the third smallest (only ahead of Switzerland and Germany); and the largest negative contribution to Italy’s poor performance is made by capital deepening in information and communication technologies, and, especially, by total factor productivity, which declines by almost 5 percent over the two decades. Of course, many factors that lie beyond this chapter’s scope may have determined the dismal performance of the Italian economy. The evidence and mechanisms we have reviewed, however, do suggest that the role of education may have been important, in two related respects.

human capital

269

First, Italy’s formal education gap vis-à-vis other Western countries may have become more relevant than in the past. Over the last twenty years, expansion of trade opportunities between Western Europe and emerging countries has reduced Italy’s comparative advantage in low-skill sectors, and information and communication advances have highlighted Italy’s difficulties in even adopting (let alone introducing) new technologies. In the postwar boom, the informal, apprenticeship-based processes of skill accumulation and transmission exemplified by distretti, and the related vocational schooling provided by scuole professionali, were key to success. More recently, and especially since the 1990s, formal education may have become increasingly necessary to exploit Italy’s comparative advantage. It is not easy, because of data comparability problems and of the presence of many confounding factors, to seek empirical support for this plausible hypothesis in patterns of economic and educational development across geographic units and over time. The association between low educational achievements and strong economic performance, discussed previously as regards distretti industriali and more generally apparent in data that single out the prevalence of traditional “Made in Italy” sectors, does seem to become weaker over time. For the most recent period, some supportive evidence is provided by Schivardi and Torrini (2011), who find that the broader access to tertiary education triggered by the new organization implemented in the 1990s and 2000s is positively related across Italy with firm-level productivity and restructuring indicators. Second, Italy’s education has increasingly suffered from a qualitative gap. We discussed how various institutional and organizational developments may have worsened the effectiveness of the country’s schooling system since the 1970s, and we have seen that even as Italy’s education quantity has continued to catch up, its quality seems to be poorer than in other countries, in terms of standardized student tests and of adult skills per year of completed schooling. Because the generations that attended school after the mid-1970s came of age just as growth ceased in the 1990s, the two phenomena may be related: even as education kept on growing in terms of the years-in-school measure used in the growth accounting exercises, the low residual total factor productivity estimate may reflect its declining quality. Although quality is harder to measure than quantity, it is very interesting to find that Italy displays not only the lowest standardized test scores, but also one of the most negative among the 1975–2000 test-score trends that Hanushek and Woessmann (2009) show to be significantly associated with changes in per capita GDP growth. Thus, some of the recent marked slowdown of total factor productivity growth may stem from its education sector and in particular from mechanisms linking organizational features to cognitive skills (Woessmann, 2007). The lack of comprehensive reforms of the secondary sector, where many new differentiated curricula have proved expensive and confusing, may have deprived that segment of the selective orientation that would have maintained quality as student numbers increased. The progressive weakening of assessment standards,

270

sources of growth and welfare

rooted in turn in the lobbying power of underemployed teachers, was facilitated by the progressive dismantling of ministry-based inspective services, and by the suppression or dilution of external examinations at the end of each schooling cycle. It is hard to deny that more and better human capital would make a useful contribution in Italy’s current conditions. In the early twenty-first century, low educational attainment and poor quality of education may again be an obstacle to socioeconomic progress in a country that seems to be poorly equipped to increase the quantity and quality of its human capital, and that may see resilient regional heterogeneity grow stronger as a federalist reform of the constitution introduces shared competence in education, backtracking over the single most important educational milestone of the country’s 150 years of history: national funding and administration of primary education.

Acknowledgments The authors gratefully acknowledge helpful comments on earlier versions of this Chapter by Gianni Toniolo, Daniele Checchi, and by the discussants (Cormac Ó Gráda and Daniele Terlizzese) and other conference participants. A longer working paper version is available at http://www.bancaditalia.it/pubblicazioni/ pubsto/quastoeco/quadsto_06/Qse_06.pdf.

Chapter 10

MIGRATIONS matteo gomellini and cormac ó GRÁDA

Introduction Throughout history people have always moved to better their lot, but before the nineteenth century the extent of such movements was severely constrained by transport costs and by fear of the unknown. Mass migration from Europe to the New World began in the 1840s. In the early decades it was mainly confined to migrants from northwestern Europe, and included few Italians. International migration within Europe was also limited before the 1880s. Italy’s emigration rate rose from five per thousand (of population) in 1876 to nearly twenty-five per thousand in 1913.1 Between 1876 (when data on Italian emigration first become available2) and 1975, twenty-six million emigrated. The outflow was disproportionately a pre–World War I phenomenon: between 1876 and 1914 nearly fourteen million left. On the whole, more than half headed for destinations elsewhere in Europe; over 6 million reached the United States and Canada; and over 4 million chose Argentina and Brazil (Table 10.1). Before the beginning of the last century, Italian migrants headed mainly for Europe and Latin America. Thereafter, caused in part by the dynamism of the US economy and in part by an ongoing transport revolution that made overseas trips safer and cheaper, there was a big surge of emigration to the United States that lasted until the war (Hatton and Williamson 1998a). After a temporary halt because of World War I, emigration resumed, showing a progressive shift from overseas to continental destinations, mainly because of the restrictive laws on immigration passed in the United States (Timmer and Williamson 1998). In 1927 the Fascist regime, in turn, enacted legislation to restrict emigration from Italy.3

272

sources of growth and welfare

Table 10.1 Italian emigration by main destinations Average rate of emigration (per thousand)

Europe (million)

United States and Canada (million)

Brazil and Argentina (million)

Total (million)

1876–1900

6.7

2.5

0.8

1.6

5.3

1901–1913

17.6

3.5

3.5

1.4

8.1

1920–1938

5.1

1.9

0.9

0.7

3.8

1948–1973

5.2

4.9

0.9

0.6

6.8

Year

Source: Our elaboration from Istat (various years).

Because of a combination of these restrictions and of the Great Depression, only 2.5 percent of the population emigrated in the following decade and the ratio of return migration to gross emigration fluctuated between 60 and 80 percent.4 The post–World War II emigration was definitely a European one: the overseas share of emigration dropped to an average of 10 percent of the total. Gross flows were nonetheless sustained: 8.5 million people emigrated in this period, 7.3 million of them before 1975. The numbers returning were also high. Although most migrants remained abroad for good, a significant but varying proportion always returned. A sense of the relative importance of return migration in the cases of the United States and Argentina may be obtained by comparing gross migration flows and the numbers of Italian-born residents as recorded in the census (Table 10.2). Thus, a gross migration of more than 0.6 million Italians during the 1890s led to an increase in the number of Italianborn of only 0.3 million in the United States between 1890 and 1900, whereas a gross outflow of 1.2 million in the 1910s increased the number of Italian-born by less than 0.3 million between 1910 and 1920. In Argentina, by comparison, gross outflows of 0.6 million in 1876–1895 and 1.2 million in 1896–1914 yielded increases in the numbers of Italian-born of 0.4 million in 1869–1895 and more than 0.4 million in 1896–1914.5 Italiani brasiliani were more likely to remain because of their faster assimilation (Trento 2002, 3–23). After 1923 the proportion of returnees from Europe surpassed that from the United States. This was mainly caused by the huge numbers of returnees from France, whose economy virtually stagnated between 1924 and 1927. The ratio of returnees to leavers rose until World War II, driven by the reduction in outflows. In the second part of the twentieth century, the ratio of return to total emigration reached 60 percent in 1963. Then, in the first part of the following decade the ratio was stable, whereas in the second part gross outflows decreased and the return to total emigration reached unity in 1973. As far as regional aspects are concerned, over the entire period about half of the emigrants came from northern regions, two-fifths from southern regions, and one-tenth from the regions in the center (Figure 10.1). Although the share of emigrants from regions in central Italy remained steady during the entire period, the shares of the South and North fluctuated

273

migrations Table 10.2 Italian emigration and Italian-born residents in the United States and Argentina, 1880–1920 Year

Italian-born in United States

Decade

Gross Migration to United States

1880

44,230

1876–1879

16,345

1890

182,230

1880–1889

267,660

1900

484,027

1890–1899

603,761

1910

1,343,125

1900–1909

1,930,475

1920

1,610,113

1910–1919

1,229,916

Year

Italian-born in Argentina

Period

Gross Migration to Argentina

1869

71,403

1895

492,636

1876–1895

590,125

1914

912,209

1896–1914

1,197,029

Sources: Historical Stats of the U.S., pp. 66, 56–7, and Commissariato Generale della Emigrazione Italiana (1927).

considerably. In particular, in 1876 emigration was an almost entirely Northern phenomenon. By 1900, the shares of North and South had converged and remained similar until the early 1920s. After the US Quota Act of 1924, the Southern share fell, losing 10 percentage points: its average during the Fascist period would be around 30 percent. After World War II the situation was reversed, with the Southern share rising from 45 percent during the 1950s to 75 percent in the crisis year of 1963. 100%

80%

60%

40%

20%

South

North

Centre

0% 1876 1881 1886 1891 1896 1901 1906 1911 1916 1921 1926 1931 1936 1947 1952 1957 1962 1967 1972

Figure 10.1 Emigration: Shares from the North, Center, and South of Italy (percentages). Sources: Our elaborations from Istat (various years).

274

sources of growth and welfare

During the 1970s emigration flows decreased substantially, notwithstanding the high level of unemployment in Italy during that period. It seems that the gains made during the years of the “economic miracle” were mainly responsible for this apparently counter-economic choice. Emigration today is very low and restricted mainly to highly skilled and specialist workers. Almost from the outset Italian migrants spread themselves widely over a range of destinations. The choice was rarely accidental; the immigrant to Australia who declared that it had never occurred to her “that Australia was not in America” (Choate 2008, 23) was quite atypical. It will be argued later that, by and large, before 1914 swings between destinations—mostly in Europe, North America, and Latin America—reflected shifting relative prospects in the different receiving countries, although the sharpness of such swings was attenuated by the size of preexisting migrant stocks. Migration to Brazil totaled about one million between the early 1880s and the early 1900s, but declined rapidly thereafter, whereas migration to Argentina reached 0.7 million in the 1900s. The increasing preference for Argentina (where in 1914 one inhabitant in nine was Italian-born and where over half the population today can claim some Italian ancestry) over Brazil is accounted for by the relative decline of the latter’s economy. After 1914, war and immigration policy mattered: with access to the United States severely limited, Europe became the most important destination of Italian migrants. This chapter focuses on some economic aspects of the phenomenon (overlooking other important features) and is organized as follows. The second section describes migrant characteristics, with particular attention given to the issue of self-selection: were those who left the “best and the brightest”? The third section focuses on the determinants of emigration. A particular concern is whether emigration as a whole raised the living standards of those who stayed and whether it promoted interregional convergence within Italy. In the fourth section we do simulation exercises and calculate the short- and long-run effects of emigration on per capita gross domestic product (GDP); the contribution of emigration turns out to be significant with respect to the growth of living standards and to regional convergence. Finally, in the recent past Italy has become a country of large-scale net immigration. The fifth section offers a review of that immigration. It focuses on the links between migration, economic activity, the labor market, balance of payments, crime, and public opinion.

Emigrant Characteristics Generalizations are always necessary for analysis to proceed. Although it makes sense to model Italian migration as consisting of unskilled workers because

275

migrations (a) Age-distribution of Roma passengers, 1902–05

(b) Age at arrival of Italian-born US residents 1900

25

25

20

20 15

15 %

% 10

10

5

5

0 0

10

20

30

40 Age

50

60

70

80

0 0

10

20

30

40 Age

50

60

70

80

Figure 10.2 Age and gender distribution of migrants. Source: Ellis Island Ship Database [at: http://www.ellisisland.org/search/ship_year.asp?lett er=r&half=2&sname=Roma]; IPUMS 1900 U.S. census sample.

most in periods of mass migration, such as 1880–1914 were unskilled (see below), Italians were also associated at one time or another with niche occupations, such as picture-framers, plaster figure makers, barbers, fish and chip merchants, ice-cream makers, and tunnel workers. Migrant characteristics may be inferred from the many publications of Italian scholars quoted previously (see note 4), but also from quantitative sources, such as census data, shipping records, and official inquiries, and also from qualitative sources. Three of the richest sources include the massive report of the Dillingham Immigration Commission (United States Congress 1911), a by-product of nativist concerns about the social and economic impact of immigration into the United States; the Annuario statistico della emigrazione dal 1876 al 1925, drawn up by the Commissariato Generale per l’Emigrazione, an institution founded in 1901 by the Luzzati Law (n. 23, January 31), a by-product of Italian concern for emigrant welfare6; and the Ellis Island archives detailing immigrants into New York from 1892 onward. Ship passenger lists offer valuable insights into the age and gender aspects of migration. As noted, when the fixed cost of migration is high, there is a presumption of a bias in favor of young adult males and infrequent return migration. Figure 10.2a describes the age and gender distributions early in the last century of thirty thousand or so Italian emigrants on the steamship S.S. Roma, which made the crossing from Naples to New York several times a year between the early 1900s and the 1920s. Figure 10.2b describes the age at arrival of Italian-born males and females as reflected in the US census of 1900. Some features of the migration are clear. First, males were much more likely to leave than females: in the period in question more than 70 percent of the emigrants were male. Over half the males were aged fifteen to twenty-nine years, although the significant proportion of older males on board (more than 30 percent were aged thirty or above) is also striking. The Annuario Statistico reveals that migrants from the South were more likely to be temporary or seasonal migrants. Thus, between 1905 and 1920 more than one-third

276

sources of growth and welfare (a)

SEI by age group (15–19 to 55–59): Italian (black line) and US born, 1920

40 35 30 25 20 15

15

20

(b)

25

30

35

40

45

50

55

EDSCOR50 by age-group (15–19 to 55–59): Italian and US born, 1920

18 16 14 12 10 8 6

15

20

25

30

35

40

45

50

55

Figure 10.3 Earnings and occupational status of Italian immigrants and U.S. born first-generation Italian-Americans. Note: SEI and Edscor50 are two-widely used IPUMS proxies for earnings/job status. SEI measures occupational status based on the income level and educational attainment associated with each occupation in 1950, while EDSCOR indicates the percentage of people in the respondent's occupational category who had completed one or more years of college. Source: http://usa.ipums.org/usa/.

of migrants from the regions of Lazio, Abruzzo, and further south returned, whereas the proportion from regions to the north was about 10 percent. The exceptions were Sardinians (who were very reluctant to migrate in this period but, when they did, mostly left for good) and Ligurians (of whom over half returned). In the 1900s an unskilled male worker from the Italian south might have hoped to earn the equivalent of 500 lire annually at home or 2,000–2,500 lire ($400–$500) in, say, New York. He would have weighed such numbers against the duration of the voyage (seven to ten days); the cost of getting from his home village to the port of embarkation; the 170–190 lire fare for a steerage or third-class passage to the United States in an iron steamship carrying hundreds of passengers; and the uncertainty of gaining employment on landing (Commissariato Generale dell’Emigrazione 1927; Fenoaltea 2002; also Keeling 2007). The cost could have constrained his initial move outward, but the

migrations

277

frequency with which some males crossed the Atlantic (notably the “golondrine” or “birds of passage,” who travelled as seasonal migrants) implies that it was not a constraint for them. Figures 10.3A and 10.3B capture one aspect of immigrant adaptation by comparing the socioeconomic and educational status of Italian immigrants and first-generation Italian-Americans (i.e., those with at least one Italian-born parent) in 1920. Based on the IPUMS census sample, they track the mean values of SEI and EDSCOR50, two widely used IPUMS proxies for earnings/job status, by age-group (from fifteen to nineteen years, to fifty-five to fifty-nine years). Both measures indicate that the immigrants acquired few skills after arrival, but the children of immigrants fared better. Moreover, those who arrived when young were at an advantage, even after controlling for language and literacy (compare Hatton and Williamson 1998a, 137–138). Italian immigrants fared relatively better in the northeastern states than the US-born, perhaps because their skills were more readily recognized where they were most numerous. The returns to age were much greater for those born in the United States than for immigrants. This could mean that it was easier for the US-born to acquire experience because they changed jobs less frequently, or else (as Hatton has suggested) that the market did not fully recognize the value of increases in immigrant human capital.7

Migrant Selection in the Age of Mass Migration and after World War II Migrants are not randomly selected from the population of their countries of origin. Each generation in the host country believes that the latest wave of immigrants is of poorer “quality” (slower to assimilate, more criminal, less industrious) than those that preceded it. In sending economies, however, the worry has long been that the departures of their best and brightest created a “brain drain.” Public opinion in the United States a century ago, reflected in the report of the 1907–1911 Immigration Commission, notoriously distinguished between “old” and “new” immigrants to the United States and argued that the latter (which was heavily Italian) were less educated and slower to adapt to American culture than their Northwestern European predecessors. That conviction was partly responsible for the literacy tests stipulated in the 1917 US Immigration Act, harbinger of a series of restrictive measures seeking to screen newcomers. For Robert Foerster, however, author of a classic work on the Italian diaspora, the self-confidence of Italian emigrants who braved the Atlantic indicated that they were “of the race of Columbus still” and, furthermore, their “energy and prowess” placed them well ahead of their stay-at-home neighbors (Foerster 1919, 419). Foerster’s account was largely deductive and impressionistic, however, with little hard evidence in support of his assertion of positive selection. The nature and direction of selection in migrant populations remains controversial. In one respect, the presence of self-selection is clear: emigrants

278

sources of growth and welfare Table 10.3 Literacy by age and gender Arrivals in Census Year 1880 Age-Group

1900 M

1910

M

F

F

M

F

10–19

67.9

73.9

58.4 56.6 64.7

74.7

20–29

73.8

60.1

47.1

43.3

60.5

52.5

30–39

72.8

53

44.7

36.2

47.2

44.6

40–49

70.2

56.1

46

34.3

46

45

Note: the “arrivals” in 1880 include all those resident in the U.S. in that year. Source: our elaborations on IPUMS.

tended and tend to be disproportionately young and healthy. In the past, the gender bias toward males entailed a reduction in labor force productivity in the sending country. Common sense suggests that those with most to gain left, unless prevented from doing so by legislation. Were those who left also better schooled, more self-confident, and less risk-averse than their peers? These aspects of human capital are less easily identified. One proxy often used to signal positive self-selection is, for example, upward mobility within and across generations. Hatton (2010) invokes the superior labor market performance of the children of immigrants (and this certainly holds for Italian-Americans in the United States before 1914) as evidence that they inherited valuable characteristics that their disadvantaged parents had failed to capitalize on; Ferrie and Mokyr (1994) infer positive selection bias from the overrepresentation of immigrants among US entrepreneurs. Again a priori, the high cost of migration from Eastern and Southern Europe during the pre-1914 age of mass migration and the large gap between incomes in sending and host economies tip the scales toward positive selection bias (Hatton and Williamson 2005, 14). In general, the greater is the gap between incomes in the sending and receiving economies, the greater the presumption that the more skilled will leave. There is little doubt that emigrants were relatively uneducated (Sori 1979, 205). Moreover, the literacy rate among Italian emigrants was low compared with that of other immigrant groups in the age of mass migration (in line with Italian disadvantage in the population at large; see Chapter 9), and the illiteracy rate was highest for those from Southern Italy (Cipolla 1971, 93). Nonetheless, there is some evidence based on different indicators that points to a positive selection of Italian emigrants. Table 10.3 reports the percentages of Italians arriving in the United States in 1900 and 1910 and of all Italians resident in the United States in 1880 described as literate in the census. The relatively small number of arrivals in the United States before the mass migration of the 1880s and later were clearly in a different

279

migrations

league than the immigrants of later decades. Fewer than half of males and females arriving in their twenties in 1900 were literate, although the proportions that were literate rose thereafter. Still, the post-1880 emigrants were more literate than the average in the south whence most of them came. Those data refer to the United States; the literacy rates of Italian-born brides and grooms in Uruguay a century or so ago—89.9 percent for grooms and 65.8 percent for brides in 1907–1908 (Goebel 2010, 221)—imply stronger positive selection among emigrants there. The mean height of a community or population is a widely accepted measure of its health and nutritional status (A’hearn, Peracchi, and Vecchi 2009). A recent anthropometric study of Italian-Americans based on Massachusetts naturalization records offers a second indication of positive selection bias in the pre-1914 period. It finds that early twentieth-century Italian immigrants in the United States were taller than the mean in the regions they left: 165 centimeter on average for adult, mainly southern, males born after 1880, three to four centimeter taller than military recruits from the south born around the same time. The outcome is striking, even if the sample is rather small and perhaps biased by the tendency for self-reporting to exaggerate heights somewhat (Danubio, Enrica, and Vargiu 2005). A third piece of evidence, less supportive of positive selection bias, invokes wage data on skill differentials. The Roy hypothesis (see Borjas 1987) states that emigrant selectivity is a function of relative skill premia in receiving and sending countries. The trends in skill premia in Italy and the United States (Betran and Pons 2004) described in Figure 10.4 implies that, ceteris paribus, between 1880 and 1930 the departure of unskilled workers lowered the skill premium in Italy and increased it in the United States (although the line is slightly upward sloping between 1900 and 1915). 2 1.9 USA 1.8 1.7 1.6 Italy 1.5 1.4 1.3 1880

Source: Bertran and Pons (2004)

1885

1890

1895

1900

1905

1910

1915

1920

1925

Figure 10.4 Skill premia (Italy versus United States), 1880–1930.

1930

280

sources of growth and welfare

The ensuing loss of human capital to the Italian economy was mitigated by three factors. The first is emigrant remittances, discussed later in this chapter. The second stems from the impact of emigration on the literacy of the stay-athomes. An increasing probability of leaving must have led to increased investment in schooling by all those who had some prospect of emigrating, and not simply those who left (Coletti 1911, 257–259; Sori 1979, 207; Stark, Helmenstein, and Prskawetz 1997; Faini 2003; Williamson 2011b). The third relates to return migration: migrants may well have returned endowed with human capital accumulated abroad (Dustman, Fadlon, and Weiss 2010). The continuing ambiguities and controversies about migrant selection underline its Janus-faced character. Anti-immigrant commentary in Italy today closely mirrors that of anti-Italian commentary in the United States a century ago. In practice, selection from the middle of the income distribution is a stylized fact about unfettered migration that has almost never been violated since 1492, and Italian migration is unlikely to be exceptional in that regard. After World War I a series of developments (restrictive US quotas from 1924, the rise of Fascism, global economic depression, and World War II) combined to reduce Italian emigration greatly. When emigration resumed after World War II its main focus was no longer the New World, but rather Western Europe. The proportion of emigrants opting for the latter destination rose from 55 percent in the 1950s and 81 percent in the 1960s. The numbers involved were large: 2.8 million in the 1950s, 2.9 million in the 1960s. As always, the size and direction of the outflow was sensitive to macroeconomic conditions and host-country regulations. Emigration begin to decline from the late 1960s, as Italian wages converged on those further north and the economic recession of the 1970s led to the closure of labor markets. Although Switzerland remained the favored destination of Italian emigrants throughout this period (by the mid-1960s there were nearly 0.5 million Italian nationals, mostly male, working there) France ceded second place to Germany (Venturini 2004, 10–17) when the bilateral Italian-German Gastarbeiterprogramm negotiated in 1955 prompted the migration of more than one million Italians between 1955 and 1973. Even to a greater extent than earlier the outflows were accompanied by significant return migration: thus in the 1960s a gross outflow of 2.9 million led to a net emigration of only 0.8 million (Venturini 2004, 10, 16). The Gastarbeiterprogramm targeted low-skilled workers. The findings of a recent analysis of the children of guest-workers who stayed in Germany are consistent with negative selection bias (Dronkers and de Heus 2010). However, the particularly poor PISA8 science scores recorded by the children of Italians may have been partly caused by their parents’ origin in Southern Italy, where scores were strikingly lower than in the North (ranging from 436 to 450 in the poorest southern regions to 520 to 540 in the richest northern regions). In the case of children of Swiss-Italian immigrants, the analysis revealed positive selectivity, perhaps because Switzerland practiced a strict policy of repatriating unemployed immigrants.

migrations

281

The Determinants of Emigration What are the determinants of migration flows? Are push or pull forces (e.g., employment opportunities at home and abroad) important, or does the gradual formation of migrant networks dominate? The pre-1914 mass migrations from Italy, Tsarist Russia, and “Other Central Europe” to the United States were uncannily similar in size and short-term fluctuations. The correlations between first differences in the three flows between 1880 and 1914 ranged from 0.7 to 0.8. This implies that all three were prompted by the relative dynamism of the US economy in that era, but common features affecting the sending countries (late-comer industrialization and rapid population growth) undoubtedly played their part too (Williamson 2004). At its simplest the economic analysis of migration proceeds with a representative individual who moves from sending country (S) to host country (H ). The likelihood of migration, usually represented by the ratio of annual gross migration to population, is then modeled as a function of the difference between expected earnings in S and H. If H represents a choice between countries, then the model predicts that with the highest expected wage. Such a model captures the overwhelmingly economic motivation of most migrant flows. For added realism, allowance may be made for the cost of migration, the occupational flows between S and H (Hatton and Williamson 1998b). The economic motivation of the migration is likely to influence its age and gender composition, and its regional distribution. Finally, if S is a very economically backward region, the response may be constrained by low levels of human and physical capital, and “income growth [ . . . ] will lead to more migrations” (Faini and Venturini 1994a, 86). Hatton and Williamson (1998b; henceforth, Hatton and Williamson) is one of the few quantitative attempts (together with Faini and Venturini 1994a, 1994b; Moretti 1999) at investigating the determinants of pre-1914 Italian emigration. Following Todaro (1969) and Hatton (1995), they use a framework in which potential migrants decide whether to leave by comparing expected incomes at home and abroad. In addition, they account for dynamics by using adaptive expectations, and introduce migrant stock to catch a network effect.9 Hatton and Williamson (1998b, 118–120) have also attempted to model the choice of destination across Italian provinces between the United States, Argentina, and Brazil early in the last century. We replicated the estimates of Hatton and Williamson (1998b) using a different dataset. We collected data on gross emigration for sixteen Compartimenti of origin and nine destination countries, from 1876 to 1925. Our specifications follow Hatton and Williamson (1998a) and Faini and Venturini (1994a): MigRatet,i,j = α + β1Hactivityt-1,i,j + β2Sactivityt,i,j + β3 (Hy/Sy) t,i,j + β4 MigStockt,i,j + +β5DPassport + β6WPopt-1, i + Dregion+ Dhost + Dyear

282

sources of growth and welfare

where i = 1, . . . 16 is the region of origin; j = 1, . . . 9, is the country of destination, y is the per capita income; Ds are dummy variables for region of origin, country of destination, Hactivity and Sactivity are the deviations of the logs of host and sending country per capita GDP from trend; Wpop is the population in working age; DPassport is a dummy for post-1900 emigration boom. We experimented with per capita income and wage ratios.10 Measures of host and sending country economic activity are obtained as a deviation of GDP per capita from trend (calculated using a Hodrick-Prescott filter). Migrant stock (MigStock) is the cumulated flow; WPop is the one year lagged share of population at working age (fifteen to sixty-four years) obtained using census data of the population share at working age within each region. The structure of our data allows us to introduce fixed effects for each region of origin, country of destination, and year. When we account for these fixed effects, our results confirm in some respects those obtained by other authors but differ in some others. The level of host country activity is significant and has the expected impact (in particular when relative wages are considered) domestic activity is often not significant. Relative wages, relative per capita incomes, and network effects (proxied by previous migrants) are the variables that explain most; population pressures count mostly in the pre1900 period. As far as macro areas are concerned, southern regions exhibit a lower elasticity with respect to the ratio of per capita incomes and to the relative wage relative to North-Center regions, but show a stronger network effect (Table 10.4). On average, income and wage effects are a little lower than in previous studies. In particular, Hatton and Williamson (1998b) estimated that a 10 percent increase in the wage ratio would have raised the emigration rate by an average of 1.3 per thousand a year. Our analysis implies an increase of 0.5 per thousand in the migration rate in response to a 10 percent increase of the per capita GDP ratio, whereas a fall of 10 percent in the relative real wage reduces the migration rate by 0.9 percent. Our estimates capture an important chain-network effect that is nearly 30 percent higher than previously found: Hatton and Williamson estimated that a net increase in the migrant stock of a thousand people would lead to an increase of eighty-three migrants a year, whereas we find the same net migrants (i.e., taking into account return migration) would draw another 109 abroad.11

On the Consequences of Emigration Who really gained from migration? The most obvious winners were those who left; they were usually the envy of their brothers and sisters who stayed at home. However, the latter benefited too, because their bargaining power as laborers was stronger and they stood a better chance of renting land. Taylor and

283

migrations Table 10.4 FE estimates* Yratio

Stock

S_activity −0.11

H_activity

WPop

0.05

−0.32

DPassp R-squared

South

0.40

0.97

0.10

(N. obs: 1240)

0.000

0.000

0.027

0.411

0.003

0.627

North-Center

0.52

0.95

0.004

0.08

0.20

0.37

(N. obs: 1652)

0.000

0.000

0.861

0.059

0.000

0.007

Italy

0.46

0.96

0.02

0.15

0.21

(N. obs: 2892)

0.000

0.000

0.205

0.528

0.000

0.073

Transocean

0.94

0.93

−0.06

0.02

0.16

0.008

(N. obs: 1552)

0.000

0.000

0.002

0.000

0.000

European

0.40

1.02

0.02

0.16

0.05

(N. obs: 1340)

0.000

0.000

0.002

0.000

0.000

0.87

Pre-1900

−0.30

0.98

−0.03

0.046 −0.17 0.000 −0.05 0.569

0.03

(N. obs: 935)

0.000

0.000

Post-1900

0.83

0.96

(N. obs: 1827)

0.000

0.000

0.054

0.461

Wratio

Stock

S_activity

H_activity

−0.07

−0.06

0.458 0.03

−0.05

WPop

1.00

(N. obs:375)

0.000

0.000

0.615

0.000

North-Center

0.85

0.91

0.13

0.03

(N. obs: 573)

0.000

0.000

0.216

0.000

Italy

0.83

0.96

0.06

0.03

(N. obs: 948)

0.000

0.000

0.451

0.000

0.000

Transocean

1.1

0.87

0.10

0.08

0.17

(N. obs: 543)

0.000

0.000

0.233

0.000

0.204

European

0.8

0.97

0.04

—

(N. obs: 405)

0.000

0.000

0.073

—

0.300

0.66

0.950

—

0.085

South

−0.14

0.04

−0.38

−0.14

0.88

0.90

0.88

0.89

— 0.87 — 0.87

0.306 −0.12

0.87

DPassp R-squared

0.552 −0.14

0.87

— 0.84 — 0.87 —

0.92

* Dependent variable, log of emigration rate. Emigrants per 1,000 population. P-values in italics. Dummies for region-of-origin and years. Variables: Yratio is the log of foreign over domestic per capita GDP; Wratio is the log of foreign over domestic wage; Stock is the stock of previous migrants; H and S_Activity are the deviations of the logs of Host and Sending country per capita GDP from trend; Wpop is the population in working age; DPassp, dummy for post-1900 new legislation. Our elaboration (data sources are in the text).

Williamson (1997; henceforth TW) find that international migration accounted for 70 percent of the convergence in wages between the Old World and the New World between 1870 and 1914, and that it increased real Italian wages by 22 percent.12 To be more precise, unskilled workers and their families who remained

284

sources of growth and welfare

gained, whereas landowners and capitalists were likely to lose. Opponents of Italian emigration in the past articulated the interests of such people. Conversely, the Stolper-Samuelson theorem predicts that unskilled workers in the United States would have been the most strenuous opponents of Italian immigration. The perception (exaggerated though it certainly was) that immigrants were crime-prone, charity-dependent, and unwilling to integrate into the cultural mainstream broadened the coalition that inspired the anti-immigrant report of the Dillingham Commission in 1911. The report found that Italian and other new immigrants had brought few skills with them, competed for employment with the most disadvantaged natives, and were slow to integrate. Articulating the concerns of the labor movement, it argued that “the development of business may be brought about by means which lower the standard of living of the wage earners,” and accordingly called for slower economic growth (U.S. Congress, 1911, I:45). Goldin (1994) and Hatton and Williamson (1998a) support claims that immigration reduced unskilled wages in the United States The former found that a 1 percent increase in the foreign-born share of the population reduced wages by 1–1.5 percent, whereas the latter reckoned that real wages in the United States would have been 11–14 percent higher in the absence of immigration after 1870 (1998a, 173). Immigrants concentrated in the urban northeast and in jobs in slow-growing sectors, with the result that they had an important negative impact, “just as the Immigration Commission argued” (1998, 174). Allowing for migrant selection complicates the outcome; the departure in disproportionate numbers of the healthier and more resourceful among the unskilled could be damaging to the sending country, whereas the converse could hold for the receiving country. The human capital characteristics of the migrant flow are therefore very important but often difficult to pin down. Did out-migration in earlier eras, such as before 1914 and after 1945, penalize those Italian regions where it was heaviest? Some critics of emigration argued that its age-structure imposed an additional significant capital loss on the Italian economy. The cost of the emigrants’ upbringing and education was borne by Italy, whereas host economies benefited from the arrival of “ready-made” adults equipped for immediate employment. The loss was compounded by the return in their twilight years of migrants likely to impose a further economic burden on the mother country. The claim builds on a life-cycle view of production and consumption, whereby the young and the old are net consumers and those aged in between net producers. Vilfredo Pareto reckoned the annual cost to the Italian economy of this life-cycle character of migration at between 400 and 450 million lire (or about 2 percent of GDP) in the early 1900s (Pareto 1905; Choate 2008, 93).

“A Fantastic Rain of Gold”: Remittances Their close links to home and their high probability of returning permanently increased the likelihood that Italian migrants would remit some of their

285

migrations 7

6

5

4

3

2

1

0 1861 1868 1875 1882 1889 1896 1903 1910 1917 1924 1931 1938 1945 1952 1959 1966 1973 1980 1987 1994 2001 2008

Figure 10.5 Remittances as a share of GDP (%). Sources: Borghese (2010) and Baffigi (2012).

earnings to the home country. The most recent estimates put remittances’ share of GDP in 1876–1913 at 2.7 percent (ranging from 0.3 percent to 5.8 percent: see Borghese 2010 and Figure 10.5). Details on the number of remitters are lacking, but data from other sending countries imply that they may numbered as many as three million annually in the 1900s (Esteves and Khoudour-Castéras 2009, Appendix B). The impact was clearly much greater in low-income regions of high emigration, and so helped reduce the regional disparity in incomes, if not in productivity. There remains some controversy as to how the remittances were spent. Were they spent or invested (Balletta 1978; Sori 1979)? Either way, the benefits to the sending regions must have been significant. Even not allowing for the savings of “birds of passage” who travelled seasonally across the Atlantic, between 1880s and 1900s the role of remittances in financing the deficits in the balance of payments on current account was of the utmost importance (Massulo 2001; Esteves and Khoudour-Castéras 2009). Foerster (1919, 448) linked Italy’s ability to convert its foreign debt in 1906 and the lira’s strength on the foreign exchanges before World War I to remittances and its subsequent weakness to the decline in “the export of labor services” (on this point see too Cesarano, Cifarelli, and Toniolo 2009). Remittances found their way back through a variety of channels, formal and informal, but as the size of the outflow increased, the scope for institutionalizing the remittance business also rose. The Dillingham Commission reported that Italians remitted money through 2,625 banks in 1907. The sense that remitters in the United States were subject to systematic exploitation by predatory banchieri was highlighted in the US press, and this prompted Italian Treasury Minister Luigi Luzzatti in 1897 to propose a plan to replace the banchieri

286

sources of growth and welfare

by a reliable and low-cost means of transferring remittances. After a highly acrimonious debate, the still privately owned Banco di Napoli was given the privilege in 1901 (De Rosa 1980). The role of the Banco di Napoli in reducing the risk and cost of making transfers was notable, although hardly dominant. In its first year it remitted 9.3 million lire, and it was sending back nine times that amount, mostly from the United States, by 1914. Boosted by wartime inflation, remittances through the Banco di Napoli peaked at nearly one billion lire in 1920 (Choate 2008, 79–80). Remittances had also an important positive impact on financial development. Savings banks became important hosts for repatriated savings, helping facilitate the spread of financial know-how. In Italy as a whole, the volume of deposits in post office savings accounts rose from 323 million lire in 1890 to more than two billion lire in 1913, whereas the share of emigrants’ savings rose from 0.03 to 4.4 percent (Esteves and Khoudour-Castéras 2010). Southern savings banks, in particular, benefited from remittances. Because the income gap between sending and receiving economies is greater, today emigrant remittances are probably a more important factor in the economic well-being of the world’s poor economies than in Italy before 1914. They average about 6 percent of GDP in the poorest two-thirds, compared with 1 percent in the top third. In 2009, without emigrant remittances, the current account deficits of Romania and Poland would have been over half higher than they were. In the 2000s non-EU immigrants in Italy remitted an annual average of €2,500–€3,000 per head, most of it to developing countries.13 In 2007, Senegal’s 63,000 Italian immigrants remitted an estimated total of €252 million, the equivalent of about 2 percent of Senegalese GDP. The proportionate loss in host economies was not commensurate: in 2007–2008 in Italy, for example, workers’ remittances by immigrants were worth about 0.4 percent of GDP (€6 billion of a GDP of €1,600 billion).

Evaluating the Economic Impact of Migration: A Counterfactual Analysis Massive immigrant flows invariably generate fears about their impact on the host economy. Such fears, amply represented in print and in economic studies, often focus on the impact of immigrants on unemployment, wages, and on the long-run budgetary position of the host country (because of difference between costs and contributions to the welfare system; Coppel, Dumont, and Visco 2000). The literature on the economic effects of emigration on the source country is sparse by comparison, although emigration can have a major economic impact. The sign of this impact is not ex ante determined and depends on the interplay between different factors: the amount and use of remittances; the

287

migrations

2200

2000

1800

1600

1400 1870

1880

1890

1900

1910

Year treated unit

synthetic control unit

Figure 10.6 Actual (treated) and counterfactual (synthetic) GDPs. Our elaborations. Sources in the text.

characteristics of the population outflow in terms of gender, age and skills, brain drain/brain gain effects; and the impacts on the labor market and on real wages. Are the potentially positive aspects of emigration outweighed by the loss of human capital entailed by the emigration of skilled workers? We approach these issues by performing two simulation exercises that consider hypothetical no-emigration scenarios. In the first, we perform a counterfactual analysis (using the methodology developed by Abadie and Gardeazabal 2003)14 and try to evaluate the effects of Italian mass emigration in the early twentieth century on Italy’s per capita GDP. We construct a sort of “synthetic” Italian per capita GDP, estimated using per capita GDPs of other European countries15 and we obtain a counterfactual Italian per capita GDP, as if the early twentieth-century boom in emigration had not happened. Figure 10.6 shows actual and counterfactual GDP per capita (Y1 and Y*1) for the period 1880–1910. After 1900 the lower line represents the synthetic Italy obtained as a weighted combination of the per capita GDP of some European countries (that did not experience the post-1900 boom in migration). The upper line is the actual Italian per capita GDP. This counterfactual scenario implies that had Italy not experienced its post-1902 surge in emigration its GDP would have been around 5 percent lower (1902–1913 average). This would have deprived it of nearly 10 percent of its entire GDP per capita growth over the first decade of the twentieth century. In the second exercise, we draw on TW’s simulation analysis, one of the few estimates of the effect of emigration on per capita GDP, and important from an Italian standpoint (see also Sori, 2009). TW study the effects of migration on convergence in Europe, using a partial equilibrium model to calculate counterfactual per capita GDPs (and wages, and GDP per worker) for several European

288

sources of growth and welfare

countries. Their results in the case of no emigration show, for Italy, a level of per capita GDP 12 percent lower than the actual in 1910.16 We apply their methodology using new data and allowing for the impact of remittances.17 We show how the outcome varies with alternative values of the model’s parameters and we extend the analysis in space and time by considering regional aspects and the post-World War II migration flows. Using a Hicks-neutral production function where migration affects long-run equilibrium per capita output through population increase and its influence on labor supply, and allowing for the contribution of remittances to capital accumulation, we derive the following reduced form equation for per capita GDP growth: Y* – POP* = θL μ γ M – θK [(1/λ) ∙ R ∙ r∙(θK)-1]- M = (μ γ θL – 1)M –(1/λ)∙R∙r (2)

The additional rate of growth of per capita GDP from nonemigration depends on the cumulative emigration rate (M), on the labor share (θL), on the ratio of total remittances to GDP (R) and investment (λ), on the relative labor participation rate of migrants (γ), on the rental price of capital (r), and on a relative productivity parameter (μ) that accounts for self-selection. In performing simulations based on this accounting framework, we set γ = 1.318; μ at 1.2 for the early twentieth century and at 0.8 from 1945 onward; on the basis of employment data we assume that 10 percent of the counterfactual population in working age would have stayed unemployed in the absence of emigration after World War II.19 The θL parameter was calculated using new GDP data.20 Our simulations imply that per capita GDP in 1910 would have been 2.4 percent lower without emigration. This would account for 6.6 percent of total growth during 1880–1910 (Table 10.5). Adding half of the remittances (through investments) would increase the contribution by 0.5 percent. If we cumulate the effects of migration until the early 1970s the simulation shows that per capita GDP would have been 4.2 percent lower and would have accounted, on the whole, for 4.2 percent of all growth. Limiting the calculation to the post–World War II period, and imposing negative self selection (μ = 0.8) and a 10 percent of unemployment on nonmigrated workers, yields a contribution of net migration to GDP per capita growth of 3.8 percent, with a per capita GDP level 2.5 percent higher (Figure 10.7). We performed the same exercise to evaluate the long-run contribution of net migration on per capita GDP by region, adopting, where possible, region-specific parameters. The contribution of net migration to per capita GDP varied significantly by region. The results range from very low for Liguria, Lazio, Sardegna, and Puglia to very high for Basilicata (where per capita GDP without emigration would have been 20 percent lower in 1970), Abruzzo and Molise, Sicilia, and Calabria.21 If we subdivide our sample into macro areas, international emigration seems to have positively contributed to long-run convergence22: in the North

Table 10.5 Differences in per capita GDP (levels and growth) in a counterfactual scenario with zero emigration (percentage) TW* calibration

New calibration

Differences in per capita GDP levels 1910

3.4

2.4

Contribution to per capita GDP growth 1880–1910

9

Differences in per capita GDP levels 1970

With remittances

With half remittances

Abadie and Gardeazabal

3

2.7

5.6

6.6

7.2

7

4.7

3.4

4.9

4.2

—

Contribution to per capita GDP growth 1880–1970

5.2

3.8

5.1

4.4

—

Differences in per capita GDP levels in 1970 (starting from 1950)

1.3

1

2.4

1

—

Contribution to per capita GDP growth 1945–1970

2

1.5

3.8

1.5

—

Differences in per capita GDP levels 1970**

5.4

6.2

7.8

7

—

Contribution to per capita GDP growth 1880–1970**

5.9

6.7

8

7.4

—

Differences in per capita GDP levels in 1970 (starting from 1950: negative self-selection)

1.5

1.7

3.2

2.5

—

Contribution to per capita GDP growth 1950–1970 (negative self-selection)

2.3

2.7

4.9

3.8

—

Our elaboration (data sources are in the text). *Taylor and Williamson (1997). ** Hypothesis on self selection: positive in 1880–1913; negative in 1950–1970.

10.7

290

sources of growth and welfare (a)

Actual

3.4

Counterfactual, no remittances 3.2

Counterfactual, with remittances

3

Counterfactual, with half remittances

2.8 2.6 2.4 2.2 2 1880 1882 1884 1886 1888 18901892 1894 1896 1898 19001902 1904 1906 1908 1910 (b) 16 Actual 14

Counterfactual, no remittances Counterfactual, with remittances

12 Counterfactual, with half remittances 10 8 6 4 1952 1954 1956

1958

1960

1962

1964

1966

1968

1970

1972 1974

Figure 10.7 Actual and counterfactual per capita GDP, (a) 1880–1913 and (b) 1952–1974 (thousands of euro on vertical axis). Our elaborations. Sources in the text.

net emigration accounted for 4.1 percent of GDP per capita growth (Veneto had the highest contribution); 5.1 percent in the Center (fuelled by the strong impact in Marche); and 7.9 percent in the South (Table 10.6).

From Emigration to Immigration In Italy as elsewhere, immigration is currently a contentious issue. In a 2009 poll it ranked second after the economy as “the most important issue facing Italy today,” well ahead of crime and education. Opponents worry about the impact of immigration on wages, unemployment, public finances, and social conditions, whereas supporters see in it a possible solution to specific labor shortages and to an ageing population (see Coppel, Dumont, and Visco 2000;

291

migrations Table 10.6 Differences in per capita GDP (levels and growth) in a counterfactual scenario with zero emigration: regional analysis (percentage) Differences in GDP per capita levels: actual versus counterfactual in 1970

Long run contribution of migration to per capita GDP growth. 1905–1973

Piemonte

4.2

4.3

Lombardia

4

3.5

Veneto

7.4

7.7

Liguria

0.9

0.9

Emilia

4.2

4.2

Toscana

3.8

4.4

Marche

9

8.3

Abruzzo e Molise

13.8

13.7

Umbria

5

5.6

Lazio

0.2

1.9

Campania

2.8

3.8

Basilicata

15.8

15.3

Calabria

10

9.9

Puglia

2.3

2.2

Sicilia

9.1

9.8

Sardegna

0.5

0.6

Nord

4.1

4.1

Centro

5.3

5.1

Sud

6.8

7.9

ITALIA*

5.7

6

Our elaboration (data sources are in the text). *Unweighted average.

Einaudi 2007). Today’s anti-immigrant attitudes in Italy and elsewhere have their close parallels in fears in the 1900s of southern Italians “seeking vengeance . . . with the stiletto” in the United States or in the 1920s of “Dagoes and Aliens” displacing “white” workers in Australia (Mayo Smith 1890, 166; Richards 2008, 105–106). In 1960, Italy’s foreign-born population numbered 63,000; two decades later it was still only 0.3 million. Since then Italy has become a country of significant net immigration, with the recorded proportion of foreign residents rising from just 0.5 percent of inhabitants in 1980 to 2.5 percent in 2000 and 7.6 percent (or 4.7 million) in 2011. Clandestini may have accounted for a further

292

sources of growth and welfare

0.5–0.75 million in 2011. The increase in the number of foreign-born residents accounted for all of Italy’s population growth between 1980 and 2010. Like Italian emigrants in the New World a century ago, migrants to Italy in recent decades have been disproportionately concentrated in high-wage, high-employment regions.23 In the late 2000s the regional distribution of Italy’s foreign-born closely mirrored the regional distribution of living standards. Every extra one thousand euros in provincial GDP per head was associated with about 0.4 percentage points in the proportion of foreign-born in 2008, ranging from less than 2 percent in the south and islands (where GDP per head in 2008 was less than twenty thousand euros) to nearly 10 percent in Veneto-Lombardy-Emilia-Romagna (where GDP was more than thirty-two thousand euros). Settlement patterns varied by country of origin; Albanians, Moldovans, and Macedonians were most likely to be located in high-income regions; high numbers of Ukrainians and Poles were attracted to the regions of Lazio and coastal Campania. Immigrants from afar (Indians, Chinese, Filipinos, and Ecuadorans) are clustered in a small number of regions. Whereas host countries a century ago obtained most of their migrants from a relatively small number of countries of origin, today’s Italian immigrants arrive from virtually every corner of the globe. Although the top three sending countries (Romania, Albania, and Morocco) accounted for more than two-fifths of the all foreign-born residents in early 2009, the top sixteen sending countries accounted for only three-quarters of the total. With the remaining quarter representing 175 nationalities, the immigration of the recent past has been truly global (Caritas Migrantes 2010). Is the immigration a mirror image of the emigration of an earlier era in other respects? This section offers a brief overview of immigration in the light of our findings on earlier Italian emigration.

Size and Impact of Immigration: Actual Versus Perceived First, size matters. Italian migration in the 1876–1914 period dwarfed present-day flows as a proportion of Italian population and of the population of receiving countries. Between 1900 and 1914 when it was at its most intense, the annual net outflow reached 2 percent of the population. The impact on the demography of receiving countries varied. In Argentina, where it was greatest, Italians by 1895 already accounted for almost one resident in eight, a share they maintained until 1914. In the United States in the 1900s Italians added 0.86 million to the population, more than any other sending country but still less than 1 percent of the total population. Immigration into Italy just a century later was unprecedented by Italian standards, rising from insignificant levels in the 1990s to a peak net migration

migrations

293

rate of over ten per thousand inhabitants in 2003, and falling gradually thereafter. It was modest, however, compared with the rates experienced by the United States and Argentina before World War I. Moreover, even in the 2000s the net migration rate into Italy was considerably less than that into Spain and Ireland. The gap between the actual share of immigrants in the total population and popular perceptions of that share is striking. In 2009, immigrants constituted about 7 percent of Italy’s population. Yet according to the Transatlantic Trends poll (2010), the estimated migrant share varied from one-fifth according to those with a college education to one-quarter according to those with only an elementary education. Females, people with only elementary education, and those affiliated to the political center and right were more likely to exaggerate the migrant share. So were the young: whereas those aged eighteen to twenty-nine years estimated the migrant share at 27 percent, those aged sixty-five and older estimated it was 20 percent. Moreover, most Italians believed that most migration was illegal, whereas in reality only a small fraction was illegal. The poll results summarized in Table 10.7 reflect the prevailing negativity about immigration in Italy in the late 2000s, but also highlight how attitudes varied across age, education, gender, socioeconomic group, and political outlook. In 2008–2009 those more favorably disposed toward migration, legal or illegal, were more likely to be well-educated and in white-collar occupations. They were also more likely to be male and left-leaning in politics. The link between age and attitudes to migration was not straightforward: although the young were most likely to deem most migration illegal, they were also most likely to agree that migrants filled job vacancies that Italians did not want. Although most respondents in all categories in Table 10.7 professed worry about illegal immigration, less than half of college graduates professed similar worry. Although almost three in five of those with little education believed illegal migrants should be sent home, only one-third of graduates did. There was a similar cleavage between left and center-right voters on these same questions. The impact of immigration on wages and unemployment in the host economy remains a hotly contested issue. At first glance it is difficult to imagine why unskilled immigrants might not displace or reduce the wages of native unskilled labor. Surely, “the labor demand curve is downward sloping” (Borjas 2003). This is certainly what TW (1997, 36, 40) found for the United States and Argentina before 1914 (although wages grew between 1870 and 1910, TW reckon that they were 27 percent less than they might have been in the absence of immigration). Such impacts on wages explain why unskilled Argentine and US workers would have been hostile to immigration, whereas capitalists and skilled workers would have welcomed immigration. However, these “headline” results assume, crucially, that migration did not induce complementary flows of capital that could have reduced the impact of immigration on wages considerably (see Taylor 1997).

294

sources of growth and welfare

Table 10.7 Italian attitudes to immigration 2008–2009 “Most “I worry “Immigrants take jobs “Immigrants fill jobs migrants are about illegal away from natives” Italians don’t want” illegal” immigration” Strongly Strongly Strongly Strongly agree disagree agree disagree Education Elementary school

71

86

10

34

28

8

Some high school

67

83

13

37

34

4

High school graduate

63

81

5

43

33

6

College graduate

49

72

0

58

47

4

Left

57

70

6

52

42

7

Center/Right

71

89

10

33

27

5

18–24

75

79

7

39

20

4

25–34

66

87

8

35

32

7

35–44

63

87

8

36

29

5

45–54

66

84

9

46

34

7

55–64

66

80

9

37

33

6

65+

69

82

9

40

36

7

Male

64

82

8

41

33

6

Female

69

89

8

36

30

6

Self-employed, etc.

55

85

8

38

40

7

White-collar

65

82

5

44

38

5

Manual

70

88

15

34

25

7

Politics

Age Group

Gender

Occupation

295

migrations

“Immigration Legal will influence migrants my vote a lot” should be temporary

Legal migrants should be given a chance to stay

Immigrants in Illegal migrants Italy are Too A lot Should be sent home many but not too many

Should be allowed stay

Education Elementary school

20

23

66

60

32

59

26

Some high school

23

18

75

55

35

54

37

High school graduate

23

17

74

49

40

51

36

College graduate

17

10

79

28

53

33

49

Left

16

13

80

35

48

35

48

Center/Right

22

23

66

62

31

62

31

18–24

23

17

17

42

48

47

44

25–34

22

22

22

51

40

56

33

35–44

22

19

19

59

31

58

31

45–54

19

19

19

50

37

52

32

55–64

22

23

23

56

36

53

31

65+

21

18

68

63

28

56

28

Male

20

18

73

50

37

54

33

Female

23

21

68

58

36

54

31

Self-employed, etc.

26

17

73

49

34

55

29

White-collar

20

14

76

45

43

49

37

Manual

19

25

65

57

34

57

30

Politics

Age-group

Gender

Occupation

Source: Transatlantic Trend database. 2008 and 2009 data combined; the entries are percentages of the relevant sub-group replies.

296

sources of growth and welfare

Although recent immigration into Italy has not prompted comparable inward movements of capital, research into the labor market impact of that immigration finds little evidence of a negative effect on the wages and employment prospects of native workers. Gavosto, Venturini, and Villosio (1999) found that immigration impacted positively on the wages of native unskilled labor, whereas Venturini and Villosio (2002, 2006) found that immigrant share had no effect on the transition from employment to unemployment for native workers and that immigration had a positive effect on wages. Staffolani and Valentini (2010) differentiate between “clean” and “dirty” unskilled work and find that native workers’ wages always rise with immigration. Less surprisingly, such results are in line with some research on the labor market impact of recent and contemporary immigration in Italy, Europe, and the United States (Dustman, Frattini, and Halls 2009; Brandolini, Cipollone, and Rosolia 2005; Ottaviano and Peri 2008; Cingano and Rosolia 2010; D’Amuri, Ottaviano, and Peri 2010; contra, see Borjas 2003). The specialist literature on migration offers several possible explanations for such an outcome. Ottaviano and Peri (2008) claim that immigrant and native unskilled workers have been complements rather than substitutes. A second possibility, raised by Shapiro and Stiglitz (1984), is that the threat from immigrants caused domestic labor to shirk less. A possibility explored for Italy is that immigration did not threaten the wages of native workers because of the weak assimilation of nonnative workers into the Italian labor market (Venturini and Villosio 2008). Hatton and Williamson (2008) surmise that the reduced importance of immobile factors may explain why wage effects today are lower than they were in 1880–1914. Another major difference between Italian emigration a century ago and Italian immigration today is the income gap. A century ago, Italian GDP per head was over half that of the United States and over three-quarters that of Argentina, the two main destinations for Italians at the time. Today GDP per capita in the main source countries of Italian immigration is less than one quarter the Italian level. Perhaps this helps explain why today’s immigrants are more reluctant to leave in times of high unemployment than Italians were a century ago. The gap may also help explain why the rates of immigration to Italy in recent years are more responsive to fluctuations in the unemployment rate than those to Ireland and the United Kingdom. Italian migration has been less responsive to the economic downturn of the late 2000s than migration elsewhere in Europe (Gomellini and Ó Gráda 2011). The much bigger gap between incomes in host and sending economies—4:1 for Italy today as against less than 2:1 for Ireland, for example—may help account for the different response in those two destinations. However, even in Italy there are signs that the rate of growth of the foreign-born population is decelerating (Koehler et al. 2010).

migrations

297

Some Remarks on “Welfare Shopping” and Crime Milton Friedman’s claim that “you cannot simultaneously have free immigration and a welfare state” has been interpreted as an argument against immigration by some commentators and an argument for curbing or harmonizing welfare systems by others. In the era of mass migration before World War I the dilemma did not arise, but the growth of the welfare state during the twentieth century has prompted fears in host countries that some immigrants seek welfare rather than work. In Italy, immigrants cannot claim social security benefits unless they have been paying for them, and healthcare for immigrant nonclaimants is restricted to emergency hospital treatment. The belief that “welfare tourism” accentuates the fiscal burden on taxpayers has fuelled demands for restrictions on immigration.24 Experts disagree as to the macro fiscal impact of immigration, although most find that it is small (Borjas 1990, 152–153; Borjas 1999; Lee and Miller 2000; Rowthorn 2008, 2009; Dustman, Frattini, and Halls 2009). Nonetheless, its perceived importance has been the most important cause of hostility to immigration, followed by fear of job market competition and crime (Boeri 2009). The richer and more unequal the host country and the more extensive its welfare system, the greater is the hostility toward immigrants (Hatton and Williamson 2005). If today’s Italy were to be deprived suddenly of all its immigrants the population share of those aged 65 years and older would increase from 20.2 to 21.6 percent, and the dependency ratio from 52.2 to 54.4 percent. Such increases may seem modest, but Italy’s combination of an aging population and a low fertility rate has rather alarming implications for the future viability of its social security regime and, indeed, its economic well-being. A recent prediction reckons that, in the absence of further migration, Italy’s population will fall from today’s 60 million to 55.55 million by 2030, and its elderly dependence ratio25 rise from its current 32.2 percent to 50.6 percent. Maintaining migration at current levels would keep numbers from falling (Billari, Graziani, and Melilli 2010), as migration has done since the early 1990s; even so, the elderly dependency ratio would still rise to 44.8 percent by 2030. Immigration is thus not a panacea against population aging, although it can moderate its impact. The hostility toward immigration, in Italy and elsewhere both today and a century ago, stems also from the perceived link between immigration and crime. According to the Transatlantic Trends 2010 poll, Italian perceptions distinguished sharply between legal and illegal immigration; whereas two respondents in three denied that legal migrants increased crime, nearly four in five believed that illegal immigrants did so. Moreover, those with little education were more inclined to link migration, legal and illegal, with crime.26 Yet again, the available evidence suggests that the link between the two is less straightforward than as reflected in the media and public opinion. A century ago in the United States, the widespread belief that Italian immigrants

298

sources of growth and welfare

were particularly crime-prone, which was heightened by a number of horrific high profile cases (including the murder of the chief of police in New Orleans in 1890 by the Mafia, which led to the lynching of eleven Italians accused of his murder) fuelled anti-immigrant prejudice. Yet, the Dillingham Commission, no friend of immigration, was forced to concede in 1911 that no “satisfactory evidence” proving a link between immigration and an increase in crime was forthcoming. On the contrary, the statistics of crime and population indicated that immigrants were less prone to commit crime than native Americans. Moehling and Piehl (2009) corroborate, finding that a century ago the relative youthfulness of immigrants and their geographic distribution accounted for their edge in conviction rates. Controlling like with like dissipates much of the immigrants’ greater criminality.27 The significant overrepresentation of nonnationals in Italian prisons sustains the widespread sense that immigration and crime are correlated. Again, the gender and age structure of the immigrant population, its low socioeconomic status, and limited opportunities account for a considerable proportion of the raw gap in criminality (Bianchi, Buonanno, and Pinotti 2008; compare Caritas Italiana 2009). Mastrobuoni and Pinotti (2011) highlight the link between legal status and immigrant crime in Italy. Their finding that the achievement of legal status reduces the probability of recidivism by more than half may argue for a more relaxed immigration policy, but at the likely cost of higher immigrant inflows. So is the fear of crime-prone immigrants “irrational”? Is this another instance where the gap between popular perception and a less dramatic statistical reality matters, and has informed voter choice?28 Regression analysis using the Transatlantic Trends 2010 poll (see Gomellini and Ó Gráda 2011) indicates that the perceived link between illegal immigration and crime was significantly boosted by personal characteristic variables representing poor education, living in the South, low skill levels, and conservative political preferences. Research and controversy about the impact of immigration will continue. In the meantime, majority opinion among researchers into the issue implies that the impact on native wages and employment, fiscal burden, and crime are modest. What remains to be explained is the apparent disjunction between public opinion, political rhetoric, and reality.

Conclusions Migration has loomed large in the history of Italy since the foundation of the state. In the era of mass emigration, beginning in the 1880s and ending in the 1970s, its reach was truly global; since the 1980s, the Italian magnet has attracted immigrants from far and near.

299

migrations

Here we have focused on the economic aspects of migration, ignoring important social, cultural, and political ramifications. We began by investigating emigrant characteristics in the age of European mass migration and we found that different indicators points to a positive self-selection of Italian migrants. Then we investigated the determinants of the early twentieth-century Italian emigration. Our study confirms the finding of earlier scholars that migration has a strong economic rationale. In our analysis, the role that networks (formed by those previously migrated) had in shaping Italian emigration emerges more strongly than in previous studies. We also found that before World War I and in the postwar era, emigration contributed significantly to rising living standards and in all likelihood pushed toward a reduction of the North-South economic gap. According to our findings, over the long run the contribution of the huge outflow of Italians to per capita GDP growth ranged from 4 to 5 percent of all growth; the impact in the South was double that in the North. Italy’s recent history of migration is almost a mirror image of its role during the era of mass migration before 1914; the fears that stoked anti-immigrant resentment in the United States a century ago find an echo in Italy today. Yet, now as then, the gains from immigration outweigh the losses. Public opinion tends to exaggerate the link between immigration, on the one hand, and crime, fiscal pressure on the welfare state, and living standards, on the other, and may turn out to have underestimated the responsiveness of the return migration rate to economic downturns.

Acknowledgment The authors thank Alberto Baffigi, Federico Barbiellini Amidei, Claudia Borghese, Alfredo Gigliobianco, Claire Giordano, Ferdinando Giugliano, Kevin O’Rourke, Lorenzo Prencipe, Roberto Tedeschi, Gianni Toniolo, and Jeff Williamson for comments and guidance. The views expressed in this chapter are those of the authors and do not reflect necessarily those of the Bank of Italy.

Notes 1. See Toniolo (1990) for an economic history of Italy from 1850 to 1918. 2. For a reconstruction of Italian emigration flows from 1869 to 1876 see Carpi (1887). Studies of Italian emigration, in particular by Italian scholars, are numerous (e.g., Rosoli [1978], Sori [1979], Bevilacqua, de Clementi, and Franzina [2002], Del Boca

300

3.

4.

5.

6.

7.

8. 9.

10.

11.

sources of growth and welfare

and Venturini [2003], and Corti and Sanfilippo [2009]). Rosoli and Ostuni (1978) is an extremely rich bibliographic essay on Italian emigration. The present chapter does not aim to review these works. Although the law also sought to dampen internal mobility and urbanization as potential threats to the regime, de facto there was high intraregional internal mobility. The available official data on return migration (lacking until 1905) imply that the ratio of return to gross emigration cannot have exceeded half in the pre-1914 period. Giusti (1965) offers an alternative estimate of net migration flows based on censal and natural increase data. Gomellini and Ó Gráda (2011) suggest caution in using the Giusti data; hence our choice of postponing elaborations on return migration for a future study. See Bandiera, Rasul and Viarengo (2010). This tallies with the high correlation (0.67) across regions between the proportion of all emigrants returning in 1905–1920 and the proportion choosing the United States in 1876–1910. One of the many aspects not addressed here is the evolution of legislation referring to migration, including the important Legge Luzzatti of 1901 (Coletti 1911). For the legislation on immigration, see Einaudi (2007). A recent study of migrants in Lombardy, toward which more skilled and educated immigrants have tended to gravitate, finds that the return on immigrant investment in schooling was very low and virtually nonexistent in the case of illegal migrants and immigrants from Latin America (Accetturo and Infante 2010). This outcome points to the nontransferability of human capital accumulated in the home country. Discrimination is one likely reason for this, although the low quality of that capital may also play a role. Programme for International Student Assessment. The data refer to 2006. Such models are geared to periods when migration was unrestricted. In practice, public policy also influenced the size and destinations of migrant flows. Before 1914 and after World War II Italian policy toward emigration was broadly supportive; the focus was on migrant welfare rather than on discouraging people from leaving (Choate 2008). During the interwar Fascist era, policy was ambivalent, disapproving of but not prohibiting emigration outright (Cannistraro and Rosoli 1979; Choate 2008, 230–231). Restrictions in the United States, the main host country, exemplified by literacy restrictions and quotas introduced by the Immigration Acts of 1917 and 1921, hit would-be Italian migrants hard in the 1920s. Reduced demand for immigrant labor in the New World reduced migration further in the 1930s. Gross migration to the United States fell from 350,000 in 1920 to annual averages of 42,000 in 1921–1930 and 12,000 in 1931–1939. Internationally comparable regional GDPs per capita are obtained by using Maddison (2009b) and the regional indices proposed by Daniele and Malanima (2007). Internationally comparable regional wages (from 1905) are obtained from Williamson (1995) and the indices of regional wages proposed in Arcari (1936). Of special interest is the relation between the emigration rate and the level of per capita income. At low income levels a “poverty trap” could emerge because the cost of emigration to overseas destinations could be beyond reach and for low levels of income one can find emigration growing when income grows. After a certain income threshold (see Faini and Venturini 1994a) the relationship between income and emigration reverses. For a recent analysis that points to the existence of a poverty trap in the age of mass migration, see Gomellini and Ó Gráda (2011).

migrations

301

12. We estimated the effect of migration on real wages using different data than Taylor and Williamson (1997). Our computations suggest that the emigration-induced increase in real wages was about 12 percent. 13. http://ec.europa.eu/economy_finance/publications/publication6422_en.pdf; http:// www.mandasoldiacasa.it. 14. The authors examine the effects of terrorism in the Spanish Basque region in terms of induced per capita GDP loss. They found that terrorism reduced per capita GDP in the Basque region by about 10 percent. 15. Ferenczi and Wilcox (1929). See Gomellini and Ó Gráda (2011) for a detailed discussion of the methodology and for the interpretative caveats we have to bear in mind while adapting Abadie and Gardeazabal (2003) to our analysis. 16. This seems to be a very high value. It implies that emigration alone would have accounted for more than one-third of per capita GDP growth between 1870 and 1910. In fact, the 1870 level of GDP per capita was 1,244 (Geary Khamis $ [GK$]); the actual level in 1910 was 1,933 GK$. The counterfactual level was 1,692 GK$, 12 percent lower than the actual. In terms of growth, the actual growth was 55 percent. The counterfactual is 36 percent, so 19 percentage points can be attributed to emigration (i.e., a share of 35 percent of the entire growth rate [19/55 = 0.345]). In the same context they estimate the effects of migration on real wages finding that in 1910, in a counterfactual no-migration scenario real wages in Italy would have been 22 percent lower. Our estimates, not presented here, point to a virtual zero-migration real wage 12 percent lower than the actual. 17. The simulations make some neutrality assumptions on the fertility rate of nonmigrant people in the regional analysis (see Livi Bacci 2010, 60); the possible migration-induced human capital accumulation (on human capital, see Chapter 9); and the possible effects on foreign trade and productivity (for analyses on trade, productivity, and immigration, see Peri 2010 and Peri and Requena 2010). We use new GDP estimates from Baffigi (see Chapter 6), wage data from Fenoaltea (2002), and remittances data from Borghese (2011). For details of the simulation model, see Gomellini and Ó Gráda (2011). 18. Rather than TW’s 1.65, on the basis of census data and data drawn from Commissariato Generale per l’Emigrazione (1927), which reports age and gender structure of migrants (also by region of departure). 19. For the mass emigration period assuming an unemployment rate is not very meaningful. If people wanted to work they could, but in a structural condition of underemployment (Alberti 2010; Luzzato 1953, 4:5). 20. TW set θL = 0.485; our computations produce a θL ranging from 0.567 to 0.650. 21. These results are correlated with the size of the emigration rates, although not perfectly. A finer calibration using regional specific parameters might improve our results. 22. For short-run analysis, see also Chapter 20. We have also estimated the contribution of internal migration to per capita GDP growth using the same model. Our very preliminary findings suggest that internal migration was a win-win game. Migrants would have stayed unemployed in the regions of departure while they contributed to productivity growth in the regions of destination (see Peri 2010). Adding internal migration to the scenario, the 9.8 percent long-run contribution of Sicily (a region of outflows) shown in Table 10.6 would become around 17 percent and for Piedmont (a region that had inflows) the long-run contribution would double from 4.3 to almost 9 percent.

302

sources of growth and welfare

23. The correlation between the foreign-born proportion of the population and regional GDP per head in 2008 was +0.87. 24. Over two-thirds of a recent sample of Italians who thought immigrants were “too many” believed they constituted a fiscal burden, only one-fifth of those who declared that immigrants were “not many” believed they were a burden (GMF database). 25. The elderly dependency ratio is defined as [population aged 65+]/[population aged 15–64]. 26. In July 2003 a poll of Italian teenagers found that nearly half of them believed “immigrants make cities less safe” (http://www.angus-reid.com/polls/27215/ racism_present_in_some_italian_teens/). 27. Compare Bell, Machin, and Fasani (2010) on the United Kingdom in the 1990s. 28. Kaplan (2001) reports that according to the Survey of Americans and Economists on the Economy nearly half of noneconomists think “too many immigrants” is a major reason why the economy is underachieving, whereas four-fifths of economists think it is “not a reason at all.” In Italy similarly wide gaps in perceptions may be found between those with more and less schooling. Thus although most respondents to the Transatlantic Trends 2009 poll professed worry about illegal immigration, less than half of college graduates did. Although almost three in five of those with little education believed illegal migrants should be sent home, only one-third of graduates did.

Chapter 11

DEMOCRATIZATION AND CIVIC CAPITAL Luigi Guiso and Paolo Pinotti

If the democratic model of the participatory state is to develop in these new nations it will require more than the formal institutions of democracy—universal suffrage, the political party, the elective legislature. These in fact are also part of the totalitarian participation pattern, in a formal if not functional sense. A democratic form of participatory political system requires as well a political culture consistent with it. Gabriel Abraham Almond and Sidney Verba, The Civic Culture: Political Attitudes and Democracy in Five Nations

Introduction To explain why differences in economic success across countries tend to persist over the course of centuries, during the past decade a new literature has drawn attention to the role of institutions in affecting economic performance and its persistence. One strand of literature led by the work of Acemoglu and others (Acemoglu, Johnson, and Robinson 2001; Acemoglu and Johnson 2005), has focused on the role of formal institutions. Protection of property rights and

304

sources of growth and welfare

limitations on the power of the executive, which they claim are essential to the development process, are built into the formal institution of a county and tend to persist over the centuries. A parallel strand, linked to La Porta et al. (1998), has emphasized the importance of legal traditions and the beneficial effects of British legal origin for the development of financial markets and ultimately for economic development. Dismantling dysfunctional formal institutions and replacing them with more economy-friendly ones, such as institutions and political processes that constrain the power of the executive, may be seen as a recipe for economic development. One important issue is then what makes an institutional reform successful. A second strand of literature has called attention to the role of informal institutions, in particular the prevalence of cultural values of cooperative behavior and beliefs of mutual trust (Tabellini 2010; Guiso, Sapienza, and Zingales 2004, 2008; Knack and Keefer 1997) and values of fairness and individualism that instill a sense of self confidence in economic actors (Roland and Gorodnichenko 2011a, 2011b) promoting achievement and innovation, in the tradition of Max Weber. Although it has often been recognized that there could be complementarities and mutual influences between informal cultural norms and formal institutions (e.g. Tabellini 2008), these two strands have mostly moved in parallel. Yet, because cultural norms in general evolve more slowly than formal political or legal institutions (Williamson 2000), they may act as a constraint (Roland, 2004) and may affect the success of institutional reform and ultimately the type of institutions that actually prevail in a country. This was indeed the path-breaking contribution of Robert Putnam (1993)—showing that the performance of (formal) regional governments instituted in Italy in 1970 depended on the (informal) set of values and beliefs (i.e., the civic traditions) prevailing in each region. More recently, Murrel and Schmidt (2011) argue that the important institutional change brought about by the Glorious Revolution of 1688 was made possible by the cultural changes that preceded formal institutional reforms. This chapter examines how preexisting cultural norms of civic behavior interact and affect the success of a major institutional reform—a democratization reform through the extension of voting rights from a restricted elite to a vast segment of the population. We draw data from the Italian 1912 enfranchisement, which granted voting rights, once restricted to a minority based on income, to almost all adult males. We document that when voting was restricted to the elite participation in political elections was significantly and consistently higher in the Southern regions of the country, those that according to Putnam (1993) and more recently Guiso, Sapienza, and Zingales (2004, 2008, 2010) are less endowed with civic capital. This feature may sound puzzling precisely because electoral turnout itself has been used to document the prevalence of norms of cooperation and civic engagement in the North of Italy. After the enfranchisement participation drops in the North and in the South but much more in the latter than in the former so that there is a reversal.

democratization and civic capital

305

We argue that the pattern in the data is consistent with a simple theoretical framework where voting in political elections is driven by a private rent-seeking motive, because political power can give rise to private benefits, and by the diffusion of cultural norms of civic engagement, which regard voting as a duty. In principle, the importance of such motives differs across individuals within the population. In particular, the benefits of political power are likely greater for the members of the elites, who comprise both sides of political connections (politicians and entrepreneurs, respectively); moreover, civic engagement differs across communities, strong in the North and weak in the South in obedience to the different civic traditions of the country initially documented by Putnam (1993); Banfield (1958); and, more recently, Guiso, Sapienza, and Zingales (2008). These assumptions imply that the enfranchisement of nonelites decreases voting turnout, the more so in less civic regions. We then test this empirical prediction on electoral data across Italian regions during the period between unification and the advent of Fascism (1861–1921). We find indeed that the 1912 enfranchisement lowered electoral turnout across the board but much more so in the low civicness regions: the drop is estimated to be between 14 and 20 percentage points larger in the Southern regions than in those in the North, and this effect is strong enough to reverse the higher political participation among the elites in the South. This result is not caused by differences in civic capital reflecting differences in literacy or in economic development between Northern and Southern regions at time of enfranchisement, because it persists when we control for these factors. We also control for the relative size of the newly enfranchised nonelites (i.e. the “treatment intensity” of enfranchisement), as approximated by the change in the electorate between the last preenfranchisement and the first postenfranchisement election. This evidence is consistent with the view, first explored in the seminal and influential work of Almond and Verba (1963), that democratization to be effective requires that nonelites be ready and willing to exercise their voting rights when reforms grant them voting power. This willingness and readiness depends in turn on whether a culture of civicness is already diffuse. To examine this issue more directly we study how nonelites voted after they were granted democratic rights. Arguably, a democratization reform is successful if nonelites make extensive use of these rights and if their votes are not captured by preexisting elites but are used to strengthen their representativeness. That is, the reform is successful if it generates a substantial reallocation of political power away from elites. We find that after the enfranchisement, radical socialist and communist parties whose manifestos better reflected nonelite interests gained a significant consensus only in the civicness-rich Northern regions, where these parties reached almost 50 percent of total votes in the last elections before Fascism. Thus, lack of civic culture threatens the success of democratization reforms because fewer members of nonelites choose to use their voting power and because those who do so are strongly influenced by previous elites in their choices.1 These findings add to parallel work by Larcinese (2011), who

306

sources of growth and welfare

also investigates the effects of the 1912 enfranchisement on the electoral success of radical parties. Overall, these results show that de jure concession of political rights taking place with the enfranchisement resulted at the time in a de facto reallocation of political power only in the regions where the nonelites were willing to actually exert those rights. Did this difference disappear over time as people voted repeatedly and possibly learned how to benefit from extended democratic rights? Did the formal democratic rights contribute to the development of what in the Almond and Verba’s original study are the values and attitudes that sustain democratic institutions and that relate to the manner in which people within a polity view their relationships with others vis-à-vis their own interests? Put differently, did the 1912 enfranchisement and the subsequent extension of voting rights to the whole population independently of gender after World War II bridge the gap in people’s civicness across regions of the country? A simple comparison of participation in political elections after World War II reveals a persistent gap between North and South, suggesting that democratization did not bridge the gap. The rest of the paper is organized as follows. First we describe the historical background and present descriptive data about participation before and at the time of enfranchisement, particularly the dramatic reversal in voting turnout between Northern and Southern electoral districts. Then we discuss a theoretical framework of voting decision, from which we obtain empirical predictions that are then confronted with the data in the fourth section.

Historical Background Preunification Institutions and Civic Traditions in Italy The unification of Italy in 1861 is the end of a long process of political disintegration into a large number of local powers and subsequent consolidation that started around the turn of the first millennium. This process was triggered by the collapse of the Carolingian Holy Roman Empire under which today’s Italian borders where unified. As the central power of the Empire fell a number of local powers emerged to fill the vacuum. The way this lack of power was filled in differs between the North and the South of Italy, and this difference is presumably at the origin of different civic traditions. Because of a number of historical contingences (the greater distance from the headquarters of the Emperor, the relative weak army of the Papal state) the South of Italy was conquered by the Normans, which replaced the authority of the Emperor and established a highly hierarchical system of government that persisted, under different rulers, until Italian unification for about seven hundred years.

democratization and civic capital

307

The North instead saw the affirmation of self-government in cities, a feature that spread over North-Central and Western Europe (Reynolds 1997; McKitterick 2004) but whose development was particularly strong in North-Central Italy between 1100 and 1300.Initially, in response to this vacuum of power and to manage the problems of common interest to which the absenteeism of the Emperor was giving rise (e.g., the maintenance of the city walls, the security of trade routes, and so forth), small groups of individuals in some cities joined forces and formed a pact to provide mutual help and collaboration to solve problems of common interest. In some cities (e.g., Genoa) this alliance involved prominent families of the marine aristocracy who swore to work together to manage the city. From this seed, slowly more stable institutions started to develop. An initial creation was the consulate—a group of citizens that hold power for a year and were then replaced by another group, allowing a vast share of the populations to take charge of local government and acquire familiarity with political engagement. According to Milani, the counsels mark the most visible sign of the rise of the free city as an independent political body (Milani 2005, 23). Over the course of the centuries these initial institutions evolved into a formal parliament of elected members that allowed for wider participation of the population in the decision-making process. With these changes, the independent city-state was fully developed. It provided law and order, collected duties and tolls, took care of the walls that provided the basic protection to the citizens, and was responsible for the maintenance and security of the roads, essential to guarantee trade, which was developing fast. In the mid-twelfth century a new word came into use to describe the independent town communities and their government: commune. As Reynolds (1997) notes “ . . . its use in urban context may derive from communia, meaning common property: not only walls and streets but also cathedrals and churches were increasingly regarded as the responsibility and therefore the property of the local community. Their rights and liberties were also their common property which it was their responsibility to maintain” (170). Indeed, the word commune is a synonym for republic (res publica, common property) and is used with this meaning. This sense of responsibility for the common good that citizens of independent towns developed and consolidated over two centuries of self-government is the civicness Putnam refers to, which seems to characterize the North of the country and which was missing in the South because the hierarchical system of government suppressed any involvement of its populations with the management of power and the responsibility of government. Guiso, Sapienza, and Zingales (2008), who provide a fuller account of the evolution of the free cities, highlight three features of the free system of cities that have contributed to shaping its citizens political culture. First, the political entity was not identified with a single person ruling but rather with the whole community. Second, the ruler ruled in the people name—that is the citizens were the source of power and thus its exercise required people consensus, although obviously prominent families mattered a lot. Third, the commune has

308

sources of growth and welfare

a statute, similar to a constitution, enacted by an assembly and norms apply to the whole population independently of status, as it was in feudal times. In sum, the commune grants legal protection to personal freedoms and thus limits the power of the executive, a feature that as argued by Acemoglu and Johnson (2005) is necessary for development. Finally, all these democratic characteristics of the free city and its emergence were not the result of unilateral concessions from the Emperors but rather a need for their citizens to provide local government services when the central authority of the Empire stopped to function. We will come back to this point later. The system of free cities evolved over time as cities fought against each other and the losers were conquered by the winners; with time, elected bodies were slowly replaced by more centralized forms of government, which evolved into the Signoria—the power of the lord. However, even these centralized powers retained many traits of the free cities, in particular the principle that power originated from people and was exercised in people’s name. When the process of unification started around the Congress of Vienna in 1815, Italy was a politically divided country with many kingdoms but with an even bigger culture divide: that between the South, at the time dominated by the Bourbons, and the North, where the civic culture created with free cities prevailed. In the South citizens were probably well characterized by what Almond and Verba (1963) defined as “parochial” (politically sleepwalker, not involved, with little or no knowledge or interest in the domestic political system);in the North they were probably much closer to the “participant” type(people possessing a strong sense of influence, competence, and confidence in understanding domestic political matters). Indeed, such differences were to be found in the civic capital endowments accumulated many years later, after the unification of the country. Although in general information on civic traditions during the past centuries is not readily available, the Ministry of Agriculture, Industry and Commerce (1888) published regional statistics on the diffusion of voluntarily mutual aid societies and cooperatives, which is one of the main indicators of civic capital considered by Putnam (1993, 139).2 The picture in Figure 11.1, plotting the distribution of such organizations over total population across Italian regions, is consistent with the persistence of a huge cultural divide between North and South even decades after unification of the country.

From Unification to the Enfranchisement of 1912 Political fragmentation of the Italian peninsula finally came to an end during the second half of the nineteenth century with the Risorgimento (the

democratization and civic capital

309

Regions under the Bourbons before 1861

Figure 11.1 Civic capital across regions in 1883, as measured by the diffusion of mutual aid societies. Notes: This figure shows the number of mutual aid societies over population across Italian regions in 1883.Darker colors correspond to greater diffusion.

resurgence). The nationalistic sentiments once confined to the narrow circle of intellectual elites spread out in the aftermath of the Congress of Vienna, which imposed the dominion of the Habsburgs over most of Northern Italy, whereas the Center and South were still controlled by the Church State and the Bourbons, respectively. After a number of crushed insurrections during the period between the 1820s and the European revolutions of 1848, the strategic alliance with Napoleon III allowed the Kingdom of Sardinia (actually centered around the Savoy Monarchy in Piedmont) to overcome Austria in 1859, whereas the Bourbons were defeated with the expedition of Garibaldi in 1860. After the proclamation of the Kingdom of Italy in 1861, another war against Austria for the northeastern regions (1866) and the annexation of the remaining territories of the Church State (1870) completed the reunification of the country. Throughout these events (maybe with the exception of the Garibaldi expedition), the Risorgimento remained essentially an affair between the Savoy monarchy and the urban high class, whereas large segments of the peasant population were at best skeptical about, or even fiercely against, the process of unification. This influenced deeply the institutional architecture of the newborn Kingdom of Italy, which among other things restricted suffrage in elections for the National Parliament to wealthy adult males paying more than forty Italian lire of taxes, about 420 thousand individuals in 1861, or 1.9 percent of total

310

sources of growth and welfare

population. In 1876 the income threshold was lowered from forty to 19.80 lire (and the minimum age from twenty-five to twenty-one), but it was not until 1912 that the census criterion was finally abolished, at least for all individuals older than age thirty; even for those aged between twenty-one and thirty, compliance with compulsory military service became a sufficient condition for voting, extending the suffrage to almost 8.7 million individuals, about the entire adult male population. Interestingly, the enfranchisement of 1912 did not represent a political achievement of the left; rather, it was a concession of the liberal coalition led by Giolitti. Although its political motivations are subject to considerable speculation among historians, the reform was probably aimed at undermining the strength of the revolutionary movement amid increasing social unrest during the first decades of the twentieth century, consistent with the idea of Acemoglu and Robinson (2000) that the risk of a violent revolution may trigger political reforms. Whatever the reason for such a concession, Figure 11.2 shows that the enfranchisement had immediate effects on electoral participation in Italy: participation declined everywhere after the enfranchisement; and the drop was significantly larger in Southern than Northern regions, leading to a reversal in voting turnout between the two areas before and after 1912.Importantly, the differential accumulated in the first decades of the twentieth century is not dissipated during the following one hundred years, characterized instead by a continuous divergence between Northern and Southern regions.

1

.8

.6

.4 1850

1900 Southern regions Fascism and WW2 years

1950

2000 Northern regions

Figure 11.2 Voting turnout in northern and southern regions, 1861–2008. Notes: This figure plots the voting turnout in northern and southern regions during all national elections held since the formation of the Italian state (1861). The vertical line corresponds to the change from restricted suffrage (on an income basis) to the enfranchisement of the entire male population.

democratization and civic capital

311

In the next section we discuss to what extent differences in civic capital endowments between northern and southern regions provide an explanation for these patterns. By focusing on the differential dynamics between north and south we instead neglect the long-run increase in voting turnout that characterizes both areas, which likely reflects a continuous process of accumulation of human capital and other aspects of economic development.

Theoretical Framework When deciding whether or not to vote, individuals weight the benefits of electoral participation against its costs. The latter may include the need for acquiring information about parties and candidates, the “shoe-leather” cost of actually going to the polling station, the time spent in the line, and so forth. The benefits of participation depend instead on two elements: the extraction of private rents and the social value associated with political participation. Private rents might include the returns to connection with politicians; favors obtained from supporting a candidate; the possibility of exercising political control and thus diverting resources in a desired direction; or the labor market premium enjoyed by former politicians, for instance because of their subsequent involvement in lobbying (Fisman 2001; Diermeier, Keane, and Merlo 2005; Ferguson and Voth 2008; and Acemoglu et al. 2010). We assume that access to such rents increases with membership in elite groups (e.g., belonging to the same restricted club may facilitate connections between appointed politicians and entrepreneurs). This is indeed the picture that emerges from the historical portrait of American elites depicted by Mills (1945). Therefore, we should expect that voting turnout decreases, on average, after the enfranchisement of nonelites in period T : (V Vit Vis

ffor s T < t .

(1)

where Vit denotes the voting turnout in region i at time t. Apart from private gains, political participation may also be motivated by a genuine desire to contribute to the functioning of the political system, what we call civicness. For instance, voting in elections allows for an ordered succession of powers in democratic systems: whereas in large elections the probability that each single voter is pivotal tends to zero, nevertheless the individual decision to vote contributes to the existence of an electoral result, in the same way as the participation of infinitesimal agents contributes to the existence of many other public goods. Civicness would increase electoral participation of elites and nonelites, thus mitigating the drop in voting turnout after the enfranchisement:

312

sources of growth and welfare E(Vit Vis | Ci ) < E(V (Vjt Vjjss C j ) for s < T

t and Ci < C j .

(2)

where Ci denotes the level of civic capital in the i-th region. Equations (1) and (2) taken together imply that voter turnout should fall after the enfranchisement of 1912, the more so in less civic regions. We can test this empirical prediction by estimating the following difference-in-difference specification, Vit

αP Post ostt t βP Post t Ci + γ ’ Xiitt +

i

+ εit .

(3)

where Post t is a dummy equal to one in all periods after the enfranchisement, γ ’Xit is the effect of other (observable) determinants of voting turnout, νi summarizes the effect of other (unobserved) region-specific characteristics, including civicness, and εit is a residual term.

Historical Evidence The preliminary evidence presented previously documents stark changes in voting patterns across Italian regions before and after enfranchisement of the nonelites in 1912. The latter was followed by a general decline in voter turnout, which was more marked in Southern regions. We next provide additional econometric evidence in this respect by estimating the equation (3) across Italian regions over nineteen elections held in the period between unification (1861) and Fascism (1921). The data come from the Atlante Storico Elettorale d’Italia (2009), produced by the Istituto Carlo Cattaneo, which provides information on voter turnout and electoral results at the regional level for the entire unitarian period (thirty-seven free elections held between 1861 and 2008, plus three plebiscites during the Fascist period).3

Baseline Results In the first column of Table 11.1 we regress voter turnout in region i and year t on an indicator variable for elections held from 1912 onward. The estimated coefficient quantifies the average drop within each region at 4.6 percentage points, doubling to 9.3 when we include a linear trend (column 2) to control for the long-run increase in voter turnout that is evident from Figure 11.2. When we split the sample to distinguish between different areas of Italy (columns 3 and 4), the average drop is more than four times larger in Southern than in Northern regions. The difference-in-differences specification in column (5)

Table 11.1 Voting turnout before and after the enfranchisement (1)

(2) All regions

Post

(3)

(4)

North

South

(5)

(6)

(8)

(9)

All regions, difference-in-differences specification

−0.046***

−0.093***

−0.042**

−0.176***

−0.209***

(0.014)

(0.016)

(0.016)

(0.016)

(0.016)

North x Post

(7)

0.188***

0.188***

0.134***

(0.018)

(0.015)

(0.020)

North x Trend

0.002*** (0.000)

Associations x Post

0.171***

0.120***

(0.023)

(0.029)

Associations x Trend

0.002** (0.001)

Observations

340

340

207

133

340

340

340

322

322

Region FE

YES

YES

YES

YES

YES

YES

YES

YES

YES

Trend

NO

YES

YES

YES

YES

NO

NO

NO

NO

Year FE

NO

NO

NO

NO

NO

YES

YES

YES

YES

R2

0.383

0.429

0.259

0.624

0.580

0.693

0.708

0.634

0.644

Notes: This table presents the results of OLS estimates of the differential effect of the enfranchisement of 1912 on voting turnout across Italian regions. The unit of analysis is regionelection observations for eighteen Italian regions over nineteen national elections during the period 1861–1921 (two regions did not vote in the first election in 1861). The dependent variable is the fraction of people that voted in each region-election. The main explanatory variables are the indicator variable Post, which is equal to one for all elections held after the enfranchisement and zero otherwise, and two alternative measures of civicness: an indicator variable for regions in the north, which were not under the control of Bourbons before 1861; and the number of voluntarily associations per 10,000 inhabitants in 1885 (this last measure is not available for one region in the sample). The bottom part of the table indicates whether the regression includes in addition a linear trend or region and election fixed effects. Robust standard errors are reported in parenthesis; *, **, and *** denote coefficients significantly different from zero at the 90 percent, 95 percent, and 99 percent confidence level, respectively.

314

sources of growth and welfare

confirms that the estimated coefficients are significantly different from each other and quantifies in almost 20 percentage points the additional drop in Southern regions; such findings are unchanged when we include a full set of year-specific fixed effects (column 6). One threat to the validity of the difference-in-differences approach is the presence of preexisting differential trends between the two groups of regions, which are indeed apparent from Figure 11.2. For this reason, in column (7) we allow for area-specific trends, so that the coefficient of the interaction Northr x Postt captures the postenfranchisement differential drop net of the long-run differential trends in the voter turnout of the two groups of regions. Even under this very requiring specification, the effect of interest remains strongly statistically significant and of the same order of magnitude. In principle one cannot rule out the alternative explanation that cultural values themselves changed around the enfranchisement, determining a reversal in voting behavior even in the absence of any civicness-mediated effects of (changes in) electoral institutions. Although this alternative explanation contrasts with a burgeoning literature on the slow moving and persistent nature of cultural traits (Guiso, Sapienza, and Zingales 2010 for a review), we nevertheless explore this possibility by examining the persistence over time of an observable measure of civic traditions across Italian regions, namely the incidence of voluntary associations already described in Figure 11.1. Therefore, in Figure 11.3 we plot the same measure for years 2001 and 1883, as reported by the last available Census and by the Italian Ministry of Agriculture, Industry and Commerce (1888), respectively; the graph does not suggest any reversal in civicness between Northern and Southern regions. In the last two columns of Table 11.1 we reestimate the same specification plugging the incidence of voluntary associations, as an observable measure of civicness, in place of the crude distinction between Northern and Southern regions. The results are qualitatively similar but lower in magnitude. In particular, the estimated coefficient in column (8) implies that an increase in the incidence of voluntary associations equal to the difference between North and South (0.3 and 0.9 associations per 10,000 inhabitants, respectively) is associated with a differential increase in voter turnout of 10 percentage points after the enfranchisement, which is lower than the corresponding difference-in-differences between the two groups of regions (column 6). This suggests that the difference-in-differences estimates may partly capture variation along other dimensions in addition to civic traditions. To reduce the scope for spurious effects, we thus include a number of additional control variables on the right-hand side of the equation. Besides region-specific and year-specific factors, there are other possible determinants of voter turnout that differ between Northern and Southern regions. In particular, correlates of economic development, such as human capital and industrialization, are likely related to civicness and can also shape the response of nonelites to the enfranchisement.

315

Voluntary associations × 10,000 inhabitants, 2001

democratization and civic capital Vda

6 Tos

5

Ero Mol

4

Lig

Umb Mar

Ven Lom

Bas

Pie

3

2

Cal

Abr Cam Pug Sic Laz

1 0

.5 1 Voluntary associations × 10,000 inhabitants, 1885 Northern regions

1.5

Southern regions

Figure 11.3 Incidence of voluntary associations across Italian regions, years 1885 and 2001. Notes: This figure plots the number of voluntarily associations over population across Italian regions in the year 2001 (vertical axis) against the same variable in the year 1885 (horizontal axis).

Therefore, we have collected literacy rates and measures of industrialization from the 1911 Italian census (i.e., around the period of the enfranchisement) and interact them with Postt on the left-hand side of equation (3); the results are presented in Table 11.2. Controlling for literacy has little or no effect on the difference-in-differences between Northern and Southern regions, whereas it halves the estimated coefficient of interest when civicness is measured by the number of voluntary associations (columns 1 and 5, respectively); the same is true when we control for the level of industrialization, as measured by the number of firms and people employed in industry over total population (columns 2 and 6) or, alternatively, by per-capita energy consumption in each region (columns 3 and 7). However, the drop in the coefficient could be caused by the fact that part of the effect of better civic traditions on voter turnout may go through higher human capital and economic development.4 In any case, the estimated coefficient of civicness remains statistically significant, even when we control for all variables at the same time (columns 4 and 8).

Heterogeneity in Enfranchisement Rates The regional analysis conducted so far confirms that the reversal in voter turnout displayed in Figure 11.2 is statistically significant and robust to control for several other determinants of electoral participation. In principle, however, one

316

sources of growth and welfare

Table 11.2 Voting turnout in northern and southern regions, before and after the enfranchisement (robustness) (1) North x Post

0.173*** (0.022)

(2)

(3)

(4)

(5)

(7)

(8)

0.146*** 0.159*** 0.160*** (0.021)

(0.021)

(0.026)

Associations x Post

0.084*** 0.085*** 0.103*** 0.053** (0.030)

Literacy x Post

(6)

(0.022)

(0.025)

(0.025)

−0.001

0.001

0.004***

0.002**

(0.001)

(0.001)

(0.001)

(0.001)

Firms x Post

Employed in Industry x Post

14.537

11.036

26.560*

25.760

(10.391)

(11.913)

(14.908)

(15.845)

0.576

0.909*

1.016*

0.186

(0.386)

(0.519)

(0.518)

(0.628)

Energy production x Post

0.727**

−0.022

1.496***

0.513

(0.313)

(0.279)

(0.335)

(0.328)

Observations

340

340

340

340

322

322

322

322

Region FE

YES

YES

YES

YES

YES

YES

YES

YES

Election FE

YES

YES

YES

YES

YES

YES

YES

YES

R2

0.694

0.705

0.699

0.706

0.660

0.679

0.663

0.685

Notes: This table presents the results of OLS estimates of the differential effect of the enfranchisement of 1912 on voting turnout across Italian regions. The unit of analysis is region-election observations for eighteen Italian regions over nineteen national elections during the period 1861–1921 (two regions did not vote in the first election in 1861). The dependent variable is the fraction of people that voted in each region-election. The main explanatory variables are the indicator variable Post, which is equal to one for all elections held after the enfranchisement and zero otherwise, and two alternative measures of civicness: an indicator variable for regions in the north, which were not under the control of Bourbons before 1861; and the number of voluntarily associations per 10,000 inhabitants in 1885 (this last measure is not available for one region in the sample). Additional controls include interactions of Post with literacy rates, firms over total population, workers in industry over total population, and percapita energy consumption; the source of all these variables is the Italian census of 1911, the detailed description is reported in the Appendix. The bottom part of the table indicates whether the regression includes in addition a linear trend or region and election fixed effects. Robust standard errors are reported in parenthesis; *, **, and *** denote coefficients significantly different from zero at the 90 percent, 95 percent, and 99 percent confidence level, respectively.

317

democratization and civic capital

Table 11.3 Voting turnout before and after the enfranchisement, controlling for changes in the electorate North x Post

(1)

(2)

(3)

0.188***

0.183***

0.160***

(0.015)

(0.015)

(0.027)

Associations x Post

Ln(Electorate)

(4)

(5)

(6)

0.171***

0.175***

0.062**

(0.023)

(0.030)

(0.028)

−0.029***

−0.022*

−0.033***

−0.022**

(0.010)

(0.011)

(0.010)

(0.011)

Observations

340

338

338

322

320

320

Region FE

YES

YES

YES

YES

YES

YES

Year FE

YES

YES

YES

YES

YES

YES

Additional controls

NO

NO

YES

NO

NO

YES

0.693

0.701

0.711

0.634

0.648

0.692

R2

Notes: This table presents the results of OLS estimates of the difference-in-differences effect of civicness on voting turnout, before and after the enfranchisement of 1912. The unit of analysis is region-election observations for eighteen Italian regions over nineteen national elections during the period 1861–1921 (two regions did not vote in the first election in 1861). The dependent variable is the fraction of people that voted in each region-election. The explanatory variables of main interest are the interaction between the indicator variable Post, which is equal to one for all elections held after the enfranchisement and zero otherwise, and two alternative measures of civicness: an indicator variable for regions in the north, which were not under the control of Bourbons before 1861; and the number of voluntarily associations per 10,000 inhabitants in 1885 (this last measure is not available for one region in the sample). The control variable ln(Electorate) is the log-number of people in the electorate in each election-region. All regressions include region and election fixed effects; the bottom part of the table indicates whether the additional controls in Table 11.2 are also included (interactions of Post with literacy rates, firms over total population, workers in industry over total population, and percapita energy consumption). Robust standard errors are reported in parenthesis; *, **, and *** denote coefficients significantly different from zero at the 90 percent, 95 percent, and 99 percent confidence level, respectively.

cannot exclude that differential changes in voter turnout between Northern and Southern regions reflect omitted variation in the share of newly enfranchised voters. In particular, if the share of nonelites was lower in Northern than in Southern regions, lower treatment intensity (as opposed to higher civicness) could explain the lower drop observed in the former than in the latter. To address this concern, we augment the specification with the log number of people in the electorate in each region-election; after controlling for common time effects and other determinants of voter turnout, differences in enfranchisement rates across regions constitute the main source of within-region changes in the size of the electorate over the sample period. The results in Table 11.3 confirm that a greater expansion of the electorate is associated with an average decrease in voter turnout; still, the coefficients of interest are basically unaffected relative to the estimates reported in Tables 11.1 and 11.2.

318

sources of growth and welfare

Reallocation of Political Power The evidence so far is consistent with the view that, for democratization to be successful, it is required that nonelites be ready and willing to participate in the political process when the reform grants them voting power; however, this willingness and readiness depends in turn on whether a culture of civicness is already diffuse. Lacking the latter, nonelites may fail to benefit from the new potential political influence leaving the distribution of powers unchanged and thus certifying a de facto failure of the democratization reform. To examine this issue more directly we study how nonelites voted once they were granted democratic rights. Arguably, a democratization reform is successful if nonelites make extensive use of these rights and if their votes are not captured by preexisting elites but are used to strengthen their representativeness. That is, the reform is successful if it generates a substantial reallocation of political power away from the elites. Socialist and Communist parties represented the traditional political outlet of nonelite popular masses in Italy. Starting with the beginning of the twentieth century, they went to replace the so-called sinistra storica (historical left), which represented the progressive trends among the elites and dominated the political scene in Italy during the last quarter of the nineteenth century. This process received undoubtedly an impulse after the enfranchisement of the nonelite masses. More precisely, after a slight increase in voter turnout at the first postenfranchisement elections, radical parties doubled their combined electoral share between 1913 and 1919 (Figure 11.4). This was made possible by the change in the electoral system from majoritarian to proportional, which weakened the position of incumbent politicians to the advantage of younger political formations; it was also favored by the spread of radical political ideas among soldiers during World War I (Atlante Storico Elettorale d’Italia, 2009). Yet, the same graph shows that political change was far from uniform. Radical parties gained significant political power (almost half of the votes in the last elections before the Fascist period) in Northern regions only, where enfranchisement resulted in an effective increase in political participation of the nonelites; by contrast, the electoral share of the same parties in Southern regions remained always below 15 percent. These patterns are confirmed in Table 11.4, where we regress the voting share of Socialist and Communist parties on voter turnout. Although higher voter turnout is associated with a greater electoral success of leftist parties throughout all periods (column 1), the relationship is much stronger after the enfranchisement (column 2). After controlling for region and election fixed effects, a differential percentage point increase in participation raises the electoral stake of radical parties by 0.3 percentage points on average, but by twice as much after enfranchisement. One obvious interpretation of this finding is that increases in turnout after enfranchisement were mainly driven by greater participation of the newly enfranchised nonelite masses, which represented the

319

democratization and civic capital .5

.4

.3

.2

.1

0 1860

1880 Northern regions

1900

1920

Southern regions

Figure 11.4 Share of votes to socialist and communist parties in northern and southern regions, 1861–1921. Notes: This figure plots the fraction of voted obtained by socialist and communist parties in northern and southern regions during all national elections held from the formation of the Italian state (1861) to the advent of Fascism (1921). The vertical line corresponds to the change from restricted suffrage (on an income basis) to the enfranchisement of the entire male population.

major constituency of Socialist and Communist parties. These results are unaffected when we control for other correlates of the political support for such parties, namely union density and the incidence of workers’ strikes (columns 3–5).

The Postwar Period The results presented so far show that de jure concession of political rights taking place with the franchise resulted in de facto reallocation of political power only in the regions where nonelites were willing to actually exert those rights. One wonders whether this difference in civicness disappears over time as people voted repeatedly and possibly learned how to benefit from extended democratic rights. Put differently, did the 1912 extended franchise to the male population and the subsequent extension of voting rights to the entire population independently of gender after World War II bridge the gap in people civicness across regions of the country? After all, the main explanation that has been put forward to explain why civicness differs so much between the North and the South of Italy is Putnam’s (1993) story. He was the first to argue that the civic capital divide between the North and the South of Italy arose when history in these two parts of the country bifurcated around the turn of the first

320

sources of growth and welfare

Table 11.4 Electoral support to socialist and communist parties in northern and southern regions, before and after the enfranchisement Voting turnout

(1)

(2)

(3)

(4)

(5)

0.436***

0.299***

0.296***

0.294***

0.294***

(0.071)

(0.080)

(0.079)

(0.079)

(0.079)

0.631***

0.586**

0.568**

0.610**

(0.234)

(0.262)

(0.249)

(0.263)

Voting turnoutx Post

Union density x Post

0.305

−1.276

(0.841)

(1.529)

Strikes x Post

0.813

2.716

(1.318)

(2.304)

Observations

340

340

340

340

340

Region FE

YES

YES

YES

YES

YES

Year FE

YES

YES

YES

YES

YES

0.678

0.687

0.687

0.688

0.689

R

2

Notes: This table presents OLS estimates of the relationship between voting turnout and the electoral results of Socialist and Communist parties. The unit of analysis is region-election observations for eighteen Italian regions over nineteen national elections during the period 1861–1921 (two regions did not vote in the first election in 1861). The dependent variable is the fraction of valid voted in favor of Socialist and Communist parties in each region-election. The main explanatory variable is the voting turnout in each region-election, possibly interacted with the indicator variable Post, which is equal to one for all elections held after the enfranchisement and zero otherwise. Additional controls include interactions of Post with union density and participation into workers’ strikes over total population; the source of these variables is the Italian census of 1911, the detailed description is reported in the Appendix. The bottom part of the table indicates whether the regression includes in addition a linear trend or region and election fixed effects. Robust standard errors are reported in parenthesis; *, **, and *** denote coefficients significantly different from zero at the 90 percent, 95 percent, and 99 percent confidence level, respectively.

millennium. The North experienced for centuries a system of free cities that allowed its citizens to participate in the political process and to learn the benefits of democratic involvement, which crystallized into people’s culture. The South instead ended up being dominated by highly hierarchical governments that never involved nonelites in political decisions and political management until Italy’s unification in 1861. Nonelites, the story goes, have thus internalized norms of dependence from elites and of noninvolvement in the making of collective decisions. Have a century of democratic concessions and formal democratic rights bridged the civicness gap? One simple way to address this question is to compare the gap in participation in parliamentary elections between South and North since the 1912 franchise. The first national elections after the promulgation of the Constitution, held in 1948, were still characterized by higher voter turnout in Northern

democratization and civic capital

321

than in Southern regions; even though the difference between the two areas was initially relatively narrow, the gap widened considerably over the following sixty years (Figure 11.2). On this account, there seem to be no convergence in civicness.

Conclusions Using historical data on democratization and voter turnout in Italy we find that the 1912 franchise led to a lower average participation in political elections but the decrease was much larger in the South than in the North, to the point that a reversal emerged and the gap has since never been bridged. We interpret these results in the context of a simple theoretical framework where participation in politics is driven by a private motive for the potential benefit that involvement with power may give rise to and a social motive induced by preexisting civic culture. These findings have two implications: first, a process of formal democratization through the concession of voting rights to a larger segment of the population is less likely to succeed and may indeed fail to work altogether if the beneficiaries of the new rights lack the civic culture necessary to practice these rights and translate them into effective political participation. Democratic reform supplies the potential for democratization, but its ultimate success rests on the existence of a demand for political participation based on the new voter’s actual willingness to use democratic rights. To create this demand requires a preexisting civic culture, as argued by Almond and Verba (1963).The second point is that democratization seems also to fail to produce civic culture in such a way as to fill the gap between the two areas of the country: after decades of experimentation with democracy and repeated voting in local, national, and European elections, and in referenda, the gaps in political participation and in civic engagement between the North and the South remain unchanged. This raises one important question: can this lack of influence be consistent with the idea, first proposed by Putnam (1993) and supported by the evidence in Guiso, Sapienza, and Zingales (2008), that the difference in civic values between the North and the South reflects difference in exposure to democracy and familiarity with participation in the management of the public good that dates back to the experience of the free cities in the cities of the North? Why did the free city experience of democratic engagement affect people’s civic culture, whereas the concession of democratic rights did not? One possibility is that the free city experience lasted for about three centuries, much longer than the practice of universal suffrage after 1912. The latter may be too short to dismantle a culture of weak involvement that tends to be transmitted from generation to generation and, as such, shows persistence. A second explanation is that although the 1912

322

sources of growth and welfare

suffrage was a concession of elites, perhaps to avoid bigger transfers of power under the threat of a Socialist revolution (Acemoglu and Robinson [2000]), the emergence and invention of democratic participation in the free cities was the result of big historical shocks (the collapse of the Holy Roman Empire and the disorder that followed) that required people’s direct involvement in the management of the public good. This would suggest that a civic culture—“a set of beliefs, attitudes, norms, perceptions and the like, that support participation” (Almond and Verba 1963, 178)—is more likely to emerge when circumstances require that not only elites but the entire population or very large segments take responsibility in the solution of collective action problems.

Appendix: List of Variables Dependent Variables Voter turnout: share of voters over total electorate in each region-election. Source: Atlante Storico Elettorale d’Italia (2009) Electoral support to Socialist and Communist parties: share of votes obtained by such parties over the total valid votes in each region-election. Source: Atlante Storico Elettorale d’Italia (2009)

Explanatory Variables North: indicator variable for Northern and Center regions, which were not under the control of the Bourbons before unification. Post: indicator variable for elections held after the enfranchisement of 1912. Associations: number of voluntarily associations per 10,000 inhabitants in each region in year 1885. Source: Ministry of Agriculture, Industry and Commerce (1888) Literacy: percentage of people older than twenty-one able to write and read in each region in year 1911. Source: 1911 Census Firms: number of firms over total population in each region in year 1911. Source: 1911 Census Employed in industry: share of total population employed in industry in each region in year 1911. Source: 1911 Census Energy production: per-capita energy produced (in kilowatt/hours) in each region in year 1911. Source: 1911 Census

democratization and civic capital

323

Ln(Electorate): log of the total number of people in the electorate in each region-election. Source: Atlante Storico Elettorale d’Italia (2009) Union density: union members over total population in each region in year 1911. Source: 1911 Census Strikes: number of workers involved in strikes over total population in each region in year 1911. Source: 1911 Census

Notes 1. This pattern of behavior is consistent and probably has contributed to shape Mosca’s (1939) sceptical view about the working of democracies where a minority— the elite—ends up exercising the political power and controlling the majority thanks to their greater organizational skills. Lack of civic culture and thus of an understanding of the power of voting among nonelites can be seen as limited possession of “organizational skill.” 2. Mutual aid societies were formed among urban artisans and craftsmen as an insurance device against accidents and economic hardship; similarly, cooperatives provided mutual assistance to their members, be they workers, producers, or consumers. As argued by Putnam (1993), “mutual aid societies provided a locally organized, underfunded, self-help version of what the twentieth century would call the welfare state ( . . . ) represented collective solidarity in the face of the economic insecurities peculiar to the modern age.” 3. The Istituto Carlo Cattaneo is an Italian nongovernmental organization that has been active for several decades in the collection and analysis of electoral data (see www.cattaneo.org). 4. Indeed, faster human capital accumulation is considered one of the main payoffs of higher social capital accumulation (see Knack and Keefer 1997).

This page intentionally left blank

part iii

INTERNATIONAL COMPETIVENESS

This page intentionally left blank

Chapter 12

A LONG-RUN PERSPECTIVE ON COMPARATIVE ADVANTAGE Giovanni Federico and Nikolaus Wolf

Introduction Richard Cobden famously suggested that Italy’s ideal specialization was to grow oranges and produce silk. He was correct at the time, but he did not expect the situation to change in the future. In its 150 years of history as a unified country, Italy has proved him wrong. Since the 1930s, it has exported more manufactures than primary products and currently it is the third largest exporter of manufactures in the European Union and sixth in the world. Exports of industrial products have arguably been the most dynamic sector of its ailing economy in the last decades. This chapter details this long-run change in trade and specialization from 1862 to the present, with a consistent analytical framework and a comparable data set. It complements Chapter 13, which deals with the causes of aggregate performance of Italian exports. We start with a brief survey of the (scientific) literature on Italian trade, stressing how the debate featured alternate waves of optimism and pessimism, which tally well with the cyclical fluctuations of the Italian economy. Then we describe the main macroeconomic trends (degree of openness, trade balances,

328

international competitiveness

share of Italian exports on world market, and so forth), and outline the longterm changes in composition of products and destinations. Whenever possible, we compare Italy with other European countries. Next we deal with the changes in Italy’s comparative advantages. We use the Lafay index (1992), which unlike the revealed comparative advantage or RCA (Balassa 1965), takes into account exports and imports. The next section then explores, with a simple “shift-share” analysis, the relation between the relative growth in exports of manufactures, both total and divided by technological level, and their destination. The last section very briefly sums up the results, puts forward some speculative ideas about the relation between long-term change and overall economic growth, and concludes with an assessment of the prospects for Italian trade in the world markets.

Exports and Economic Growth in Italy The nexus between trade and economic growth has been a prominent topic in Italian economic thinking and policy over the last 150 years. The prevalent mood has been rather pessimistic. For most of the period, Italian elites nurtured a sort of ideologic antipathy for trade. They considered imports a threat to the livelihood of producers, and did not believe that exports could be an engine of growth. Exports were deemed necessary to pay for the essential imports of raw materials, but little more. Exports were reckoned to have spurred growth only in the 1860s and 1870s, and to some extent during the “boom giolittiano,” the twenty years before the outbreak of World War I (Einaudi 1900). This hostility to trade peaked during the 1930s when the Fascist regime regarded independence from foreign trade (autarky) as one of its key economic goals. Optimism gained ground again during the golden years of the European economy after World War II when Italy achieved unprecedented rates of gross domestic product (GDP) growth. Exports were regarded as an essential source of aggregate demand and as a strong stimulus to increase the competitiveness of Italian industry (Ciocca 2007). Furthermore, scholars heralded the change in the composition of Italian exports as evidence of convergence with the more “advanced” European countries, such as Germany and France (Cao Pinna 1965; Miurin and Santini 1971; Conti 1973; Onida 1999). The season of optimism did not last for long. Since the late 1970s, growth slowed and pessimism returned. Few asked for protectionism and many still continued to consider exports potentially as a main source of growth. Regardless, several scholars argued that Italy’s long-term prospects for trade-led growth were poor, because the country still exported mainly low-technology, “traditional” goods, which were subject to increasing competition from emerging markets, rather than the growth-fostering, science-based manufactures (Conti

a long-run perspective on comparative advantage

329

1978; Guerrieri and Milana 1990; Gomellini and Pianta 2007). Indeed, in the 1980s and 1990s, Italy’s specialization changed fairly little and thus the difference with the advanced European countries persisted (Bugamelli 2001; De Benedictis 2005; Di Maio and Tamagni 2008). An alternative, more positive, view of Italy’s export performance in the 1980s and 1990s argues that the aggregate data are misleading. In fact, Italy succeeded in positioning itself on the top of the market for “traditional” goods, exporting high-quality and high value-added products, and thus enjoyed some degree of market power (Fortis 2005; De Nardis and Traù 2005; Lanza and Quintieri 2007). Microeconomic research at firm level shows that the recent success owes a lot to a massive process of restructuring after the euro (Bugamelli, Schivardi, and Zizza 2008). However, the effort has been largely thwarted by the economic crisis. Export pessimism has again emerged as an important component of the widespread narrative about Italy’s long-term stagnation and relative decline. This literature suffers from a two-fold problem. First and foremost the quantitative evidence is scarce for the period before 1939 and confusing for the period after 1950. Italy did publish complete and fairly reliable statistics since 1862 (Movimento Commerciale), but their sheer size (up to three thousand products in the 1930s) and the repeated changes in classification of products have deterred scholars from using them. Thus, the historical literature on the pre-1939 period is based almost exclusively on anecdotal evidence, with very few data and almost no international comparisons. After 1962, scholars can rely on the United Nations data on world trade. However, they use an array of different measures with different levels of product disaggregation and implicitly focus on short- to medium-term trends. Many of their works consider only benchmark years rather than yearly series and none extend the analysis beyond periods of twenty-three years. The only exception to this short-termism is a very recent paper by Vasta (2010), which covers the whole period 1862–2010. However, it deals only with exports and reports original research only for the period before 1939, relying on a survey of the literature for postwar trends. This chapter addresses these shortcomings. It provides a long-term view of Italian trade and specialization from unification to the present, with a consistent analytical framework. For the period before 1939, we rely on the results of a massive research project, funded by the Banca d’Italia (Federico et al. 2011), which has copied the whole set of Italian trade statistics from 1862 to 1949 and reclassified them according to the SITC classification (Rev. 2). The data will soon be available on the Banca website. From 1950 onward, we use the United Nations data as published in their annual Yearbook of Trade Statistics. For the period through 1961 we have to rely on printed summaries (United Nations 1962), whereas afterwards the data are available in a number of databases. We use the NBER-UN world trade data for the period 1962–2000 (see Feenstra et al. 2005 for a description of the data set) and complement this with the UN Comtrade data for the period 2000–2009.

330

international competitiveness

A First Look at Macro-Trends: 1862–2009 Let us start with the degree of trade openness, the ratio of total merchandise trade (imports plus exports) to GDP. Figure 12.1A compares Italy with the two powerhouses of the European continent and Italy’s main trading partners (Germany and France), and with Spain, a smaller and, for most of the period, more backward country than Italy.1 Before the Great Depression, differences between countries were substantial. Italy remained less open than France and Germany, although not less than Spain, throughout all its first fifty years as a unified state. The degree of openness remained roughly constant around 0.20 until the 1890s and then increased to 0.25 on the eve of the war, five points less than France and almost thirteen less than Germany. It remained on that level during the 1920s and, as everywhere, collapsed during the Great Depression. Since the 1950s, the degree of openness has continuously increased in all four countries. In Italy, the growth has been fairly steady, with the exception of a short period of stagnation in the 1980s. Openness exceeded the 1913 level around 1970 and then peaked in the mid-2000s, around 0.55, before declining somewhat during the current crisis. The Italian economy is as open as France’s and Spain’s, but still less than the German economy. This growing openness is a key factor that has shaped the economic history of the Italian (and European) economy. Italy has almost always exported less than it imported (Figure 12.1B). The trade balance was heavily negative after the unification and improved in the 1870s, but stayed negative for more than a century with wide fluctuations between 1 and 8 percent of GDP. Exports only exceeded imports during the “economic miracle” of the late 1950s and early 1960s and for a few years in the early 1990s. The balance turned negative again in recent years. When compared with the three other countries, Italy’s share in world markets at current prices was remarkably stable over this period (Figure 12.2A).2 It declined in the late 1880s, from around 3–3.5 percent to about 2.5 percent, and remained at this level until the early 1950s. As usual, Italy performed worse than Germany, which managed to keep about 10 percent of the market from the 1880s to the 1930s, despite the collapse of exports in the early 1920s. Italy performed marginally worse than Spain, but remarkably better than France, whose share plummeted from 15 percent in the 1860s to less than 5 percent during the Great Depression. Since 1950, Italy’s share of world exports began to rise, as the country as other European ones benefitted from membership in the European Union and from trade liberalization under the GATT and successor agreements, which did not extend to countries still under Socialist

a long-run perspective on comparative advantage (a)

331

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 1860 1865 1870 1875 1880 1885 1890 1895 1900 1905 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

0 France (b)

Germany

Italy

Spain

6 4

0

1862 1866 1870 1874 1878 1882 1886 1890 1894 1898 1902 1906 1910 1914 1918 1922 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006

2

–2 –4 –6 –8 –10

Figure 12.1 (A) Openness ratios (trade/GDP), Italy, France, Germany, and Spain 1862–2009. (B) Italian balance of trade as percentage of GDP, 1862–2009.

regimes. The share peaked around 1990, well above 4.5 percent, before it came down to 3.5 percent in the 2000s as new competitors, most notably China, entered the world market. This fall has raised some concern about Italy’s competitiveness, but it can be viewed more optimistically as a return to its long-run average. One can object that a comparison of absolute shares between economies of widely different size is misleading. Thus, in Figure 12.2B we correct by normalizing with the country’s share of world GDP (Maddison 2009a). The Italian ratio remained below one until the 1950s, with a clear downward trend in the first half of the twentieth century (0.76 in 1870, 0.82 in 1900, 0.74 in 1913, and 0.70 in 1939). Italy exported substantially less than it should have, given its GDP.

332

international competitiveness (a)

25.00

20.00

15.00

10.00

5.00

1862 1867 1872 1877 1882 1887 1892 1897 1902 1907 1912 1917 1922 1927 1932 1937 1942 1947 1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007

0.00 Italy (b)

France

Germany

Spain

2.50 2.25 2.00 1.75 1.50 1.25 1.00 0.75 0.50 0.25 1862 1867 1872 1877 1882 1887 1892 1897 1902 1907 1912 1917 1922 1927 1932 1937 1942 1947 1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007

0.00 Italy

Germany

France

Spain

Figure 12.2 (A) Shares in world trade, Italy, France, Germany, and Spain 1862–2009. (B) Share in world trade relative to share in world GDP, Italy, France, Germany, and Spain, 1862–2009.

The ratio for France and Germany declined before 1939, but their starting point was much higher (1.75 and 1.5, respectively). Spain’s normalized share was lower than the Italian share in 1870, but much higher in 1900 and 1913. With this metric, Italy’s performance is decidedly disappointing. As expected, the postwar trade boom featured a sharp increase in the ratios. By 2008, in all four countries, the ratios were at their postwar maximum: the Italian ratio was at 1.6, as high as France’s and Spain’s, but well below the German one (2.15). In the last half century, Italy has been roughly the European norm, and Germany the outlier. Furthermore, the recent fall in its share in world trade (Figure 12.2A) can be interpreted, in a somewhat less alarming fashion, as an unavoidable outcome of the relative decline of its economy rather than the evidence of a pure loss of competitiveness.

a long-run perspective on comparative advantage

333

Italy’s Foreign Trade Since the pioneering work by Kuznets (1967b), structural change in trade is deemed a key feature of modern economic growth. Backward countries, such as postunification Italy, export primary products and import manufactures. Industrialization increases the share of manufactures on exports and of primary products, most notably raw materials, on imports. More recent work has highlighted the growing role of exchange of manufactures between advanced countries (intraindustry trade), without questioning the outline of changes in the early stages of modern economic growth. The case of Italy tallies only partially with the received wisdom. Figure 12.3 reports the share of manufactures, defined as the sum of SITC (Rev. 2) one-digit categories five to nine, excluding raw silk before 1939.3 Even in the 1860s and 1870s, manufactures accounted for less than half of imports, because Italy lacked most key raw materials including coal and cotton. The need to import them clearly contributed early on to the country’s negative trade balance. The share of manufactures fluctuated between a third and half of total imports until the 1980s and then increased sharply, as a consequence of a significant increase in intraindustry trade. Trends on test export side are closer to the conventional wisdom. After its unification, Italy exported very few industrial products, unlike France and Germany. Manufactures accounted for 50–60 percent of all French exports in the 1860s and for 70 percent of German exports in the 1880s, and for 10 percent and 30 percent of imports, respectively.4 Italy was closer to Spain, where industrial products accounted for about half the imports and a mere 12 percent of exports in 1877 (Tena 2007). However, although the composition of Spanish trade did not change much until 1951, Italian exports of manufactures grew decidedly faster than total trade and on the eve of World War I, their share reached almost 40 percent. The percentage increased further in the 1920s. The war had disrupted the exports of primary products significantly more than of manufactures, and the divergence continued after the end of hostilities. The exports of primary products recovered slowly and they only exceeded their prewar levels in the late 1920s, for few years and by less than a fifth. In contrast, exports of manufactures grew fairly fast throughout the decade. In 1929, they exceeded their prewar level by about 80 percent and Italy’s share on world exports of manufactures reached 3 percent for the first time. The Depression hit all exports to the same extent, but exports of manufactures were boosted by the conquest of Ethiopia in 1936. In 1936–1939, industrial products accounted for 49.6 percent of total exports but for only 43.7 percent exclusive of exports to colonies. Thus, on the eve of World War II, despite its progress, Italy still had not yet caught up with France and Germany. After the war, the change in composition of exports resumed, and continued steadily up to the early years of the twenty-first century, when manufactures exceeded nine-tenths of total exports. This share is roughly as high as in Germany and decidedly above that of Spain and France.

334

international competitiveness (a)

100 90 80 70 60 50 40 30 20 10 1862 1866 1870 1874 1878 1882 1886 1890 1894 1898 1902 1906 1910 1914 1918 1922 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006

0 Imports (b)

Exports

0.25

0.2

0.15

0.1

0.05

1862 1866 1870 1874 1878 1882 1886 1890 1894 1898 1902 1906 1910 1914 1918 1922 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006

0

Figure 12.3 (A) The composition of Italian trade: manufactures as a share of imports and exports, 1862–2009. (B) The composition of Italian trade: Herfindahl Index of export diversification, 1862–2009.

Composition of Italian Exports This structural change came along with a significant increase in the level of diversification of Italian exports. A simple measure is the Herfindahl index, HFI Ft

∑a

2 it

where ait is the share of four-digit SITC (Rev 2) product i in total exports in year t. In the 1860s, this index was above 0.15, because the five most important products (silk, olive oil, sulfur, silk cocoons, and wine) accounted for 65 percent of total exports. Italy’s dependence on a handful of products steadily declined

a long-run perspective on comparative advantage

335

80 70 60 50 40 30 20 10

1862 1866 1870 1874 1878 1882 1886 1890 1894 1898 1902 1906 1910 1914 1918 1922 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006

0

SITC5 (Chemicals) SITC7 (Machinery and Transport Equipment)

SITC6 (Textiles, Metals etc, until 1939 excl Silk) SITC8 (Miscellanus, incl. Clothing, Footwear etc.)

SITC9 (Non Classified)

Figure 12.4 The composition of Italian manufacturing exports (percent shares of SITC Rev 2, one digit), 1862–2009.

to the mid-1900s, when the five most important products, including cotton and silk cloth, accounted for no more than 39 percent of total exports and the index was about 0.08. It then collapsed to about 0.03 on the eve of World War I and declined further to 0.016 in the late 1930s, when the five top products accounted for just 18.3 percent of exports. The index declined further during the golden age of the 1950s and 1960s, down to a minimum around the 1980s. Despite a small rebound in the late 1990s, export concentration remained low at around 0.013. The five most important products, all industrial, accounted for 16 percent of total exports. The discussion so far would not justify the pessimism about the performance of Italian exports. Trends in the share of world markets are broadly consistent with its level of development, and exports show a healthy trend toward diversification. Actually, most misgivings about the long-term prospects of Italian exports, in the past and currently, are related to their composition, and especially to the allegedly excessive share of “traditional” manufactures. Are these concerns justified? As Figure 12.4 shows, the one-digit SITC groups 6 (“Manufactured goods classified chiefly by material”) and 8 (“Miscellaneous manufactured articles”) together account for about 40 percent of total exports of manufactures, whereas the share of chemicals (group 5) hovers around 10 percent. The four most “traditional” industries, that is, textiles and yarns (65 in the two-digit SITC classification), clothing and apparel (84 ), footwear (85), and furniture (82) still supply between a sixth and a fifth of Italian industrial exports. However, the key change is the rise of the share of SITC group 7 (“Machinery and transport equipment”). It became the second largest group, after textiles, in the 1930s, although largely thanks to exports to the colonial captive market, and

336

international competitiveness 45.00 40.00 35.00 30.00 25.00 20.00 15.00 10.00

0.00

1920 1923 1926 1929 1932 1935 1938 1941 1944 1947 1950 1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007

5.00

Low Tech (excl Silk up to 1939)

Medium Tech

High Tech

Figure 12.5 Export shares of low-, medium-, and high-technology goods, 1921–2009.

the largest group in the 1960s. In the last fifty years, this group has contributed about two-fifths to total Italian exports. On a more general issue, one could question whether aggregate data at the one- or two-digit SITC category level are adequately representative, because most of them hide large differences in terms of sophistication and skill-content of exports. Scholars have put forward a number of alternative classifications of goods along these lines, using three- or four-digit categories (Peneder 2003). Several works (e.g., Guerrieri and Milana 1990, Gomellini and Pianta 2007) use the classification by Pavitt (1984) in four groups: (1) science-based, (2) scaleintensive, (3) supplier-oriented, and (4) traditional industries. However, this taxonomy had been conceived to classify firms rather than products and indeed it mixes different criteria (technology level, type of demand, organization of production). In contrast, the more recent taxonomies by Organisation for Economic Cooperation and Development (1996) and Lall (2000) classify products according to only their technologic level. The latter, which we prefer as more precise and detailed, allocates manufactures in three classes: (1) low, (2) medium, and (3) high technology.5 However, we do not deem it suitable to capture the technologies of the nineteenth and early twentieth century and thus we use it only from 1920 onward (Figure 12.5). The share of low-technology manufactures in total Italian exports has remained fairly stable in the long run at around a third, with a hump in the 1980s. In contrast, the share of medium-technology goods has been rising throughout the entirety of the period. The growth has been very fast during the interwar years, but it has continued also after 1950, up to 40 percent in 2009. The share of high-technology products was very low in the 1920s, which is of little surprise given that most of the relevant technologies were only in

337

a long-run perspective on comparative advantage 0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 1862 1866 1870 1874 1878 1882 1886 1890 1894 1898 1902 1906 1910 1914 1918 1922 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006

0

UK

France

Germany

OtherEurope

Other OECD

ROW

Figure 12.6 The geography of Italian exports, 1862–2009.

an early stage. However, the share of these products in all Italian exports has increased fairly little in the last decades. As we detail in the next section, the differences between these groups extend from their shares of exports to the trade balance.

Geographic Reorientation of Trade The change in composition toward manufactures and to ever more diverse and sophisticated manufactures was associated with some quite dramatic geographic reorientation of trade (Figure 12.6). At the beginning of the period, about 90 percent of Italian exports headed towards Europe. One-third of these went to France; one-sixth each to the United Kingdom and Switzerland (probably including a lot of disguised exports to Central Europe), and the rest to other countries. The overall share of Europe declined thereafter, and on the eve of World War I European countries accounted for about 60 percent. However, this apparently steady change conceals the massive shock caused by the outbreak of the trade war with France in 1888. Before 1887, France had absorbed almost a half of Italian exports and afterwards its share dropped to less than a fifth. After 1887, Germany became Italy’s best client and is still the main single destination for Italian exports. The trend toward a greater geographic diversification continued into the 1920s and was boosted in the latter years of the 1930s by the conquest of Ethiopia. Colonies, which previously had been almost irrelevant, accounted for about one-quarter of Italian exports on the eve of World War II. This situation was indeed exceptional, and after the war and the subsequent loss of the colonial empire, the share of Europe rebounded to about two-thirds,

338

international competitiveness

and remained by and large constant until today. Trade with non-European Organization for Economic Cooperation and Development (OECD) countries (i.e., the United States and Japan) only once reached 15 percent and has recently declined to less than 10 percent, whereas trade with developing and emerging markets recently increased to more than 25 percent of total exports.

Italy’s Comparative Advantage There are many ways to measure comparative advantages (Laursen 1998). The workhorse of the empirical work is still the RCA index by Balassa (1965). It compares the share of a given country in world exports for each product to the total market share of that country: a value over one suggests a specialization in that good and vice versa. An alternative approach uses commodity-specific net exports as a measure of comparative advantage. The most common index along this line has been suggested by Lafay (1992), ⎡⎛ x m ⎞ ⎛ X M ⎞ ⎤ ⎛ x + m ⎞ LFI F i = 100 ∗ ⎢⎜ ⎟ −⎜ ⎟⎥⎜ ⎟ ⎣⎝ x + m ⎠ ⎝ X M ⎠ ⎦ ⎝ X M ⎠ , where x refers to exports and m to imports. Lower case letters denote commodity-specific trade, and capital letters denote total trade (i.e., X = Σx).6 In a nutshell, the Lafay index is the difference between the normalized net-balance for the ith product and the total normalized net-balance, weighted with the share of the product on total trade. Thus, a positive value implies a specialization in the ith good. All indexes sum up to zero, with a maximum range from 200 to −200 in the extreme case of complete specialization of both exports and imports (and balanced trade).7 The Lafay index offers three distinct advantages over the RCA (and various modifications thereof): 1. It controls for distortions from an overall net deficit (which was the most common case in Italy) and, above all, it takes the level of imports into account. This is a particularly appealing feature for the analysis of recent trends when outsourcing and intraindustry trade have become increasingly relevant. For instance, Italy would seem to have enjoyed a comparative advantage in iron and steel products in 1998, with a RCA of 1.05 (De Benedictis 2005), whereas actually it was a net importer. As a general rule, the distortion from neglecting the import side is potentially greater the higher the level of aggregation and the share of intraindustry trade. 2. It measures the contribution of different products to changes in total comparative advantage. From this point of view, the RCA can be seriously misleading, especially for minor products. Let’s hypothesize

a long-run perspective on comparative advantage

339

that a country with a 1 percent share of world exports has a monopoly on a product accounting for 0.01 percent of world trade; the RCA for that good would be a whopping one hundred, but the product would still account for only 1 percent of the country’s exports. 3. Last but not least, it needs only national trade statistics. This feature is particularly helpful for a long-run perspective, because data on world exports with a suitable disaggregation are available only after 1962. The figures by Tyszynski (1951) and Yates (1959) are highly aggregated and the implicit classification is quite idiosyncratic and differs markedly from the SITC classification we use for other data. The data in Figure 12.7A show the changes in Lafay indices for all manufactures (defined again as SITC one-digit categories from five to nine). The index turns out to be positive throughout the period (except for a few years in the 1860s) only if raw silk is included. If it is excluded, as we believe it should be, the picture is rather different: Italy apparently had no comparative advantage in industrial products before the turn of the century. The Lafay index was heavily negative after Italian unification but started to grow after the mid-1880s, and reached a positive value for the first time in 1898. It hovered around zero until World War I, but with a short dip around 1907. The index jumped upward during the war and afterward, peaking in 1921 when German firms were hobbled by defeat and hyperinflation. The return of world markets to more normal conditions caused the index to decline, but it remained positive and high throughout the interwar period. The spike of the late 1930s is largely spurious because it reflects manufacturing exports to colonies. Regardless, even when we omit these years, the index remains well above zero and fairly high (8.3 for the period 1923–1936). During the golden years of the postwar period and also in the bleak 1970s, the index continued to grow, increasing to values around or even above twenty until the mid-1980s. Then it started a long decline. In the 2000s, the strong increase in imports of industrial products (Figure 12.3A) drove the aggregate index down to five or below, a level it had not reached since World War I. Figure 12.7B decomposes the Lafay index by type of manufacturing products at the one-digit level. Before World War II, the indices for chemicals (5), engineering (7), and miscellaneous manufactures (8) were low in absolute values and changes were fairly modest despite the on-going industrialization process. The index for engineering remained almost constantly negative. Indeed, Italy was a net importer for most engineering products (if measured at the two-digit level) throughout the 1920s and early 1930s. Only road vehicles (group 78) have remained significantly positive from the early 1900s to World War II. Chemicals are an interesting case. Italy had a sizeable positive balance in the 1870s and 1880s, mostly thanks to exports of traditional tanning materials, soap, and perfumes. The overall balance deteriorated during the “boom giolittiano” and remained slightly negative

340

international competitiveness (a)

25.00 20.00 15.00 10.00

0.00

1862 1866 1870 1874 1878 1882 1886 1890 1894 1898 1902 1906 1910 1914 1918 1922 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2001 2005 2009

5.00

–5.00 –10.00 –15.00 –20.00 Manufacturing

(b)

Manufacturing (excl. Silk)

20 15 10 05

–05

1862 1866 1870 1874 1878 1882 1886 1890 1894 1898 1902 1906 1910 1914 1918 1922 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006

00

–10 –15 –20 SITC5 (Chemicals) SITC8 (Miscellanus, incl. Clothing, Footwear etc.) SITC6 (Textiles, Metals etc, until 1939 excl Silk)

(c)

SITC7 (Machinery and Transport Equipment) SITC9 (Non Classified)

16.00 14.00 12.00 10.00 8.00 6.00 4.00 2.00

1920 1923 1926 1929 1932 1935 1938 1941 1944 1947 1950 1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007

0.00 –2.00 –4.00 LFI_LT (excl. Silk up to 1939)

LFI_MT

LFI_HT

Figure 12.7 (A) Lafay-index for manufacturing exports (sum over SITC one-digit categories five to nine), 1862–2009. (B) Lafay-index for manufacturing exports (SITC one-digit categories five to nine), 1862–2009 (for 1862–1939 category six excluding silk). (C) Lafay-index for different technology classifications, 1921–2009.

a long-run perspective on comparative advantage

341

after the war. Thus, the observed dramatic change of Italy’s comparative advantage toward manufactures reflects changes in group 6—or more precisely in textiles (65), which here exclude silk. Their index shifted from heavily negative values (around or below −7) until the 1890s into positive ones. On the eve of World War I, textiles, most notably cotton cloth, were already the third largest contribution to net exports, after raw silk and fruits. They became the main Italian export after the late 1920s with index values over seven, but they lost this position after World War II. Their index declined steadily throughout the 1950s, down to 1.5–2 in the 1960s. Meanwhile, Italian exports had diversified with substantial net balances in nonmetallic mineral manufactures (66 i.e., marble, stones, and tiles); clothing (84); footwear (85); and to a lesser extent furniture (82). However, the defining trend during the economic miracle was the boom in engineering products. At the end of the 1960s, all two-digit SITC categories of engineering products were positive and five out of ten categories were in the top eleven, if ranked by the Lafay coefficients. “Road vehicles” (78) had the highest index of all sectors at the two-digit level (2.88), ahead of clothing (2.85). This was the moment when Italy came closest to the “European” (i.e., German) pattern of specialization in the whole period covered in this chapter. However, the convergence was short lived. Textiles enjoyed a revival in the late 1970s and 1980s and the competitive position in the so-called “made in Italy” products (clothing, footwear, furniture) remained quite strong until the late 1980s. Then, their index started to decline, as exports stagnated while imports were growing quite fast. The decline was driven by a combination of factors, including a considerable increase in imports of road vehicles from Germany and of such products as articles of apparel and clothing from France, jointly with difficulties in selling these goods in the international market. Thus, by the late 1990s specialized machinery for industries (72 and 74) sat at the top of manufactures if ranked according to the Lafay index. In contrast, the trade balance for road vehicles was negative from 1982 onward, and in the late 1990s their Lafay index was the second lowest of all two-digit categories, after oil products. The reader may infer from these data their own assessment of the technologic level of Italian comparative advantages, but Lall’s (2000) classification scheme adds some precision. As Figure 12.7C shows, Italy has always been very competitive in low-technology products, and it has managed to consolidate a fairly strong position in medium-technology products in the 1960s and 1970s, which it partially lost in the 1990s and 2000s. In contrast, Italy has never been a significant net exporter of high-technology goods, and its competitive position has been markedly deteriorating since the late 1980s when a rise in exports in these products was more than offset by an increase in imports (also compare Figure 12.5). In terms of low-technology products, Italy’s relative strength is driven by textiles and leather products, such as footwear (group 851), which have diminished over the last two decades. The development in medium- and high-technology industries is deeply related to Italy’s position in intraindustry trade. The country’s position as a car manufacturer

342

international competitiveness

(group 781) weakened in the 1970s but recently started to improve again. However, the country is increasingly exporting intermediate goods for the car industry (group 784), but also other highly-specialized machinery and parts thereof (728), and electrical equipment, which keeps the international position in medium-technology industries relatively stable at a Lafay index value around five. Finally, the difficulties in improving Italy’s international position in high-technology products was driven by a significant decline in the export of office machinery accompanied by a massive rise in imports (group 751), in part offset by a rise in exports of telecommunication equipment and parts thereof (groups 764 and 776). Also, exports of medicinal and pharmaceutical products (541) performed quite well, preventing the index in high-technology products from falling more sharply.

Manufacturing Exports and Geography: Where Did Italy Succeed? The previous sections highlight two major trends: the increase in the shares of manufacturing exports, which broadly reflect the industrial transformation of Italy; and the growing geographic diversification of exports. To what extent are these two trends related? Did exports of industrial products increase across the board, or was their success only concentrated in a few markets? In particular, to what extent did Italy manage to penetrate markets for manufacturing goods in advanced economies, such as France, Germany, or other OECD countries, rather than to limit itself to markets in less developed countries? To address this issue, in this section we decompose the change in the share of manufactures in Italian exports between changes in the share of each country on total exports and changes in the share of manufactures on exports to each country. The two can diverge. For example, in 1862 nearly one-third of all Italian exports went to France but only below 1 percent to Germany (Zollverein). In the same year, manufactures (without silk) accounted for 30 percent of exports to Germany, for 18 percent of exports to France, and for 14 percent of total exports. Put differently, in the early 1860s Italy was relatively more successful in exporting manufactures to Germany than to France, even if this latter was a much bigger client. Twenty-five years later, in 1887, total exports to France had increased to 41 percent and exports of manufactures grew more than the total, up to about a quarter. In the same period, Germany jumped from 1 percent to 10 percent of total Italian exports: exports of manufactures did increase in absolute value, but their share on Italian exports to Germany declined to about a fifth. Hence, Italy successfully penetrated the French market with manufacturing products, but less so the German market, which was to become its main client in the next years.

a long-run perspective on comparative advantage

343

In more formal notation, we calculate the contribution of country j to the change in Italy’s manufacturing share in total exports between t and t + 1 as CCCj,Δt:

CCC j∆t

⎡ (CS j t ++1 MS j t +1 ) − (CS j t MS j t ) ⎤ ⎢ ⎥ MSt ⎦, = 100 ∗ ⎣ ⎡ (MSt +1 − MSt ) ⎤ ⎢ ⎥ MSt ⎣ ⎦

where CSj denotes the share of country j in total exports of Italy and MSj the share of manufactures in total exports to this country. This gives us the percentage contribution of each destination to the observed change in the total share of manufactures on Italian exports. The latter increased from unification through 2000; hence, in all of these periods a positive sign in Table 12.1 implies that the exports to the jth country contributed to the overall increase. It thus can be read as a success in market penetration. The interpretation of results would be opposite in cases when we observe a decline in the share of manufactures (e.g., during the last period 2000–2009) and this can be confusing. Thus, in those years we invert signs. Our geographic disaggregation is constrained by the information available in the pre-1939 database. We thus report data on four major countries, the rest of Europe (defined here as sum of Austria-Hungary, Belgium, the Netherlands, and Switzerland), and the rest of the world which we obtain as a residual. Before World War I, Italian industry was successful in all countries but the United Kingdom (1862–1913) and France after 1887, in all likelihood because of the trade war. The interwar years give a mixed picture: the 1920s were still a good period, whereas in 1938 Italian manufactures increased their market share only in Germany. The total share of industrial products grew only thanks to the boom of exports to colonies, which in Table 12.1 belongs to the rest of the world. Unsurprisingly, the export boom during the golden age of the European economy (i.e., the periods 1938–1962 and 1962–1973) was mostly driven by the European market. Italian products fared well in the American market in the 1950s, but not in the 1960s, when the US share on total manufacturing exports and the share of manufactures on this flow remained stable. The period 1973– 1986 is characterized by stagnation, with a poor performance in German and other European markets that could only be partly offset by gains elsewhere. In contrast, from 1986 onward we observe large differences between Italy’s performance in Europe, the United States, and the rest of the world. The composition of exports to advanced countries, its most traditional clients, remained unchanged, but their share on total exports declined. By implication, Italy managed to recoup some of the losses in the rest of the world, notably emerging economies that were growing faster than the old OECD countries. The last line of Table 12.1 considers the contribution of the various geographic areas to the evolution for the whole period. The message is clear: most of the increase of

Table 12.1 Percentage contribution of various countries to the change in the share of manufactures (excluding silk) on total exports Percentage-Contribution to Total Change by:

Share of Manufactures (initial year)

Total Period Change (in percent)

1862–1887

0.17

53

75.5

1887–1913

0.27

52

1913–1929

0.40

1929–1938

United Kingdom

Rest of Europe

14.8

−24.6

12.3

−65.3

12.2

21.8

14.4

2.8

113.8

22

2.9

−15.1

6.2

34.4

27.6

112.7

0.49

4

−97.9

43

−160.3

−45.9

−67.7

429

1938–1962

0.51

41

29

42.9

7.4

34.2

22.6

−36.1

1962–1973

0.74

12

64.3

71.1

1.9

17.1

1973–1986

0.83

7

23.6

−36.1

39.1

−4.1

57.7

19.8

1986–2000

0.88

2

−138.7

−147.1

23.7

−23.9

2.7

430.7

2000–2009

0.92

−3

−54.3

−110.4

181.7

28.7

420

7.8

14.3

0.4

9.7

1862–2009 Source: own calculations.

France

Germany

United States 10

−11

−83 5.1

Rest of World 11.8

−43.4

300.7 63.1

a long-run perspective on comparative advantage

345

the share of manufactures in total exports was driven by the success of Italian industrial products on overseas (non-American) markets. From this point-ofview, the relative increase of manufacturing exports to Europe during the golden age is an exception rather than the rule. Table 12.2 repeats this exercise separately for exports of low-, medium-, and high-technology goods.8 As in Figure 12.5, we limit our attention to the period after World War I, which anyway covers the most sweeping changes in industrial specialization. As Table 12.2 shows, Italy established its trade position in medium-technology products in the interwar period and maintained it until the 2000s. In the more recent period, the main driver of its success was exports to the rest of the world, which includes the emerging markets, whereas trade with OECD countries actually made a negative contribution. The same holds for trade in high-technology goods where the share on total exports stagnated and even slightly declined (from 2001 onward). Italy lost positions on European markets, while it increased its penetration in overseas markets, both in advanced countries (rest of OECD) and in emerging ones (rest of the world). The three lines in bold show the changes in the share of various technology classes in total trade since 1913. The long-run expansion of medium- and hightechnology exports was initially facilitated by trade with Europe, especially during the golden age of growth through 1973. After this we observe that the rest of the world starts to become significantly more important for Italy’s success in exporting medium- to high-technology products.

Conclusion By and large, the results confirm the conventional wisdom about the change in Italy’s specialization, with two important qualifications. First, they highlight the key role of World War I. Before the war, Italy was still mostly a supplier of primary products to advanced countries. After the war, it became an exporter of manufactures, although it sold its wares mostly to the poor markets of the European periphery and to less developed countries, including, after 1936, its new colonies. Second, they show a major discontinuity around 1980. The economic miracle (and the Common Market) had featured a reorientation of exports toward Europe, and a shift in specialization from traditional textiles and low-technology goods toward medium-technology, most notably engineering products (although definitely not toward high-technology goods). For a while, Italy seemed poised to converge to the German model. Since the 1980s, Italy returned to its prewar model, although in the context of a much more globalized world. In relative terms, it lost competitiveness in (many) medium-technology products and in European markets, relying on exports of low-technology “made in Italy” products (plus some specialized engineering

Table 12.2 Contribution of various countries to the change in the share of low-, medium-, and high-technology goods on total exports Technology Classification (following Lall 2000) Low technology (excl. silk)

Period

Total Change Share of Technology-Class (in percent) in Total Exports (initial year)

France

Germany

United Kingdom

Rest of Europe

Rest of OECD

Rest of the World

1913–1929

0.23

42

7.3

−4.1

8.3

−6.2

24.6

70.2

1929–1938

0.33

−18

19.4

2.4

42.5

11

21.4

3.3

1962–1973

0.30

485.8

−140.9

−2.2

−91.5

−480.2

1973–1986

0.31

22

29.9

−13.9

21.7

18.9

28.1

15.1

1986–2000

0.38

−17.4

−41.4

−51.1

−12.9

−9

−13.8

28.2

2000–2009

0.32

−15.4

−5.4

−35.3

−50.7

−19.5

−22.9

33.8

14.7

79.5

63.9

−52.1

77.6

92.2

−161.1

1913–2009 Medium technology

Percentage-Contribution to Total Change by:

2.2

329

1913–1929

0.02

346

2

10.7

−0.9

6.6

5

76.6

1929–1938

0.11

52

−2.6

20.6

−0.9

3.7

5.1

74.2

1962–1973

0.32

11.4

64

37.8

24.4

40.2

27.3

−93.8

1973–1986

0.36

−2

−21.3

−93.5

77.4

−52.9

109.7

−119.2

1986–2000

0.35

10.6

−10.1

8.8

−2.1

36.3

13.2

53.9

2000–2009

0.39

3.4

−62.5

−49.6

−53.7

−107.1

10.4

12.5

4.5

17.5

1913–2009

1,551

−73 12.4

445 42.7

Technology Classification (following Lall 2000) High technology

Period

Total Change Share of Technology-Class (in percent) in Total Exports (initial year)

Percentage-Contribution to Total Change by: France

Germany

United Kingdom

Rest of Europe

Rest of OECD −8.8

Rest of the World

1913–1929

0.01

9

−71.9

−17.1

21.5

−53.8

1929–1938

0.01

239

5.4

2.7

0.1

3.5

0

88.4

1962–1973

0.06

18.3

24.8

21.6

17.9

47.1

−13.1

1.7

1973–1986

0.07

18.9

6.7

20.1

13.9

−7.2

42.5

23.9

1986–2000

0.09

29.8

2.4

−9.3

12.5

32.3

17.4

44.7

2000–2009

0.11

−10.8

−40.5

−14.6

−54.6

−22.2

23.4

8.5

10.8

12.5

8.9

23

20.4

24.4

1913–2009

951

230

Source: own calculations. Rest of Europe pre-1945 is limited to Austria-Hungary, Belgium, the Netherlands, and Switzerland. After 1945 this is the EU15 without France, Germany, and the United Kingdom. Rest of OECD pre-1945 is the United States only; after 1945 it is the OECD as of 1974 excluding Europe.

348

international competitiveness

goods) toward non-European countries. The performance was fairly good until the mid-1990s, but the last decade has been difficult. Italy’s comparative advantage has declined, the share on world markets has stagnated, and the trade balance has been consistently negative. This brief sketch raises three questions: (1) What caused the long-run change in trade patterns? (2) How much can trade explain the long-term growth in the past (or the lack thereof)? (3) What does the past experience tell us about the future? Understanding the causes of long-run changes in specialization is fascinating and rewarding work, but it is well beyond the scope of this chapter. We can only make some very general remarks. The long-run trends in trade are bound to reflect growth at home and abroad and the height of barriers to trade, as determined by transport technology and policies. Empirical studies have found that the income elasticity of exports and the (absolute value of the) elasticity of exports to trade costs are typically just below unity (Anderson and van Wincoop 2004). As a general rule, the specialization of each country reflects its factor endowment and its technical capabilities, but recent work (Deardorff 2004, Harrigan and Deng 2008) has shown that comparative advantage tends to be rather localized. The competition is more intense the closer the trading partners and hence the relevant comparison group for an assessment of comparative advantage is a country’s accessible neighbors rather than the world economy. Our analysis shows that, from this point of view, Italy had fared fairly well to the 1980s, but not afterwards. Our sketch of long-term trends shows a strong coincidence between periods of economic growth, at home and abroad, and good export performance, most notably during the “boom giolittiano” and the economic miracle. Such a coincidence is not really surprising. Most economists believe that an increase in openness is beneficial for economic growth (Krueger 1998; Berg and Krueger 2003; Frankel and Romer 1999; Irwin and Tervio 2002; Yanikkawa 2003; Alcala and Ciccone 2004; Feyrer 2009). However, the mechanisms at work, the welfare effects of trade, and the policy implications are still under discussion (e.g., Rodriguez and Rodrik 2001). Nobody seriously argues that openness has not been beneficial to the Italian economy, but the causation is more difficult to establish, as shown by the failed attempts by Federici and Marconi (2002) and Pistoresi and Rinaldi (2012). Furthermore, there are some doubts about the country’s capacity to extract all possible benefits from its openness because of its comparative advantage. Economists in the 1950s and 1960s regarded a specialization in primary products as a dead end, to be abandoned as soon as possible. This argument has been revived more recently as a “curse of primary products” (Sachs and Warner 2001). Italy succeeded to move away from specialization in primary products at a relatively early stage in its modern economic growth. A recent paper has suggested that this change did not much affect trends or volatility of the terms of trade (i.e., that the “curse” was not that great when Italy specialized in primary products) (Federico and Vasta 2010). However, much

a long-run perspective on comparative advantage

349

of the Italian debate deals with its comparative advantages in manufacturing. One can relate the composition of exports, rather than their aggregate level, which mattered for economic growth, to the theory of biased technical change (Matsuyama 1992; Acemoglu 2002). An opening to trade can clearly affect the incentives to invest in capital-intensive or research-intensive sectors with possibly large dynamic effects. Recently, Hausmann, Hwang, and Rodrik (2007) have provided a simple framework where factor endowments (most notably of capital and human capital) interact with development costs for new products, such that a country’s specialization pattern might be incompletely determined. Empirically, they found a positive relation between an index of export sophistication, which captures the quality of exported goods, and the rate of GDP growth. Along this line, one can decry Italy’s specialization in low-technology goods as a major problem for long-term growth. As mentioned in the second section, this view was widely held in the 1970s and early 1980s and has resurfaced in recent years. However, the policy responses have not been forthcoming. In the 1970s, Italy tried to foster the growth of allegedly advanced sectors, but it ended in a string of expensive failures (Roccas, 1977, Federico, 1999). Calls for such a pro-active industrial policy are fairly isolated in more recent times, but Italy is not investing massively in human capital to change its comparative advantages, as implicitly suggested by Hausmann, Hwang, and Rodrik (2007). Predicting the future is always difficult, but the past performance of Italian exports shows that too much pessimism may not be justified. Although it has been worse than the German performance, it is still in line with the European average. In the long run, Italy has succeeded to prove pessimists wrong time and again. It may well succeed this time, despite the competition by China and other emerging countries.

Notes 1. The data on trade at current prices for France, Germany, and Spain for 1862–1939 are from Annuaire Statistique (1951), Hoffmann (1965, Table 125), and Tena (2007, Table 3); for 1945–1961 from United Nations (1962); for 1972–2000 from Feenstra et al. (2005); and for 2001–2009 from the Comtrade database. Italian GDP at current prices for 1862–2009 has been kindly provided by Alberto Baffigi. For French and German GDP through 1939 we used the series by Toutain (1997), series V41) and Hoffman (1965, Table 248); for Spain through 1969 we used Prados de la Escosura (2004, cuadros A.6.6 and A.12.1). The data for 1950–2002 are from Mitchell (2007) and 2003–2010 from the OECD. 2. For the sources of country data cf. footnote 1. The series of world trade in current dollars for 1862–1913 is obtained by linking the data from Lewis (1981) for 1862–1913, from the United Nations (1962) for 1913–1939, and from the International Monetary Fund for 1948–2009. The country data, in national currencies, have been converted into dollars with exchange rates from www.globalfinance.com.

350

international competitiveness

3. Silk (raw and thrown) is classified in SITC 2 alongside cotton and other industrial yarns. Yet, most of its value (around 80–85 percent for the raw silk and about 70–75 percent for the thrown one) consisted in cocoons—a purely agricultural raw material—and for a number of technical reasons processing of cocoons settled close to the production areas. The case is thus more similar to wine production or smelting of copper than to cotton production. The classification of silk is very important, because it was Italy’s main export from the unification to the 1920s: considering it as an industrial product would give the misleading impression that Italy exported manufactures. Category 9 (“commodities and transactions not elsewhere included”) is added to manufactures because it consists mostly of gold and weapons. 4. See footnote 1 for sources. The figures after 1950 include East Germany. 5. Low-technology products are characterized by stable, well-diffused technologies, mostly embodied in capital equipment. These include leather and textile products, such as cotton fabrics, footwear, glassware, or furniture. Medium-technology products encompass most capital goods and the production of intermediate goods and are typically based on complex technologies, such as cars and engines, but also most products of the chemical industries. The production of high-technology goods require high research and development investments; interaction with research institutions; and highly specialized technical skills, such as optical instruments, and electrical and electronic equipment, such as computers, aircraft, or medical products. 6. The original Lafay index also includes an adjustment for the size of the economy (as proxied by the current GDP), which has been omitted, following Bugamelli (2001) and Zaghini (2003). 7. This appealing property comes with a price, because the index can be falsely positive if (x-m)