New Structural Economics

Transcript

1 New New Structural Economics “This splendid collection of essays, by one of the world’s outstanding experts on economic development, puts to work a newly emerging view, which he has helped to shape, of why in recent decades some countries have prospered while Structural others have languished. Lin’s focus is on countries that were all economically underdeveloped six decades ago, but his analysis offers strong hints about future prospects of the rich world as well. His style is dispassionate and unadorned by drama, which makes the essays all the more moving and illuminating.” Economics — Sir Partha Dasgupta Frank Ramsey Professor Emeritus of Economics, University of Cambridge “ New Structural Economics is a truly important and ambitious book. Justin Lin, A Framework for Rethinking with some help from other distinguished scholars, has succeeded in laying out the complex structural microeconomic dynamics of economic growth, diversification Development and Policy and development, and in capturing the crucial complementary roles of government as investor, regulator, coordinator of activity and expectations, and guide. All of this is set in a global economy that is itself in the midst of massive structural change. This book will become an essential reference for scholars and for policy makers not only in developing countries, but also, increasingly, in developed countries.” — Michael Spence 2001 Nobel Prize in Economics William R. Berkley Professor in Economics and Business, New York University Leonard N. Stern School of Business “The World Bank has long been committed to the goal of achieving a world In this brilliant volume, its Chief Economist, Justin Yifu Lin, lays without poverty. out an economic agenda for how to make this dream a reality. He argues that the successes of China can be achieved elsewhere around the world, and explains clearly and forcefully the structural transformations that will be required and The book the role that government can and must play in that transformation. Justin Yifu Lin will be a landmark in rethinking development. It provides an alternative to the now discredited Washington Consensus policies that guided the Bretton Woods Institutions for years. Justin Lin’s ideas have already stirred discussion and debate. This book will ensure that they will continue to be central in the reexamination of developmental policy.” — Joseph Stiglitz 2001 Nobel Prize in Economics Justin Yifu Lin University Professor, Columbia University ISBN 978-0-8213-8955-3 90000 3 38955 9780821 SKU 18955

2 New Structural Economics

3

4 New Structural Economics A Framework for Rethinking Development and Policy Justin Yifu Lin

5 © 2012 International Bank for Reconstruction and Development / International Development Association or The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org 1 2 3 4 15 14 13 12 This volume is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dis- semination of its knowledge, this work may be reproduced, in whole or in part, for noncom- mercial purposes as long as full attribution to the work is given. For permission to reproduce any part of this work for commercial purposes, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: [email protected] ISBN (paper): 978-0-8213-8955-3 ISBN (electronic): 978-0-8213-8957-7 DOI: 10.1596/978-0-8213-8955-3 Library of Congress Cataloging-in-Publication Data Lin, Justin Yifu, 1952- New structural economics : a framework for rethinking development / by Justin Yifu Lin. p. cm. Includes bibliographical references and index. ISBN 978-0-8213-8955-3 -- ISBN 978-0-8213-8957-7 (electronic) 1. Economic development. 2. Neoclassical school of economics. I. Title. HD75.L56 2012 338.9--dc23 2011050744 Cover design and photo illustration: Critical Stages

6 Contents Contents Acknowledgments ix Abbreviations xi Introduction 1 I New Structural Economics: A Framework for Rethinking Development 11 Comments 48 by Anne Krueger, Dani Rodrik, Joseph E. Stiglitz Rejoinder: Development Thinking 3.0: The Road Ahead 66 81 II The Growth Report and New Structural Economics with Célestin Monga Debate: Should Industrial Policy in Developing Countries Conform to Comparative Advantage or Defy It? 113 with Ha-Joon Chang V

7 VI | Contents Growth Identifi III cation and Facilitation: The Role of the State in the Dynamics of Structural Change 141 with Célestin Monga Comments and Rejoinder 181 with Dirk Willem te Velde, Suresh D. Tendulkar, Alice Amsden, K.Y. Amoako, Howard Pack, Wonhyuk Lim, and Célestin Monga IV Applying the Growth Identifi cation and Facilitation Framework: The Case of Nigeria 215 with Volker Treichel V Financial Structure and Economic Development 259 with Lixin Colin Xu Development Strategy, Institutions, and Economic VI Performance 285 VII Epilogue: The Path to a Golden Age of Industrialization in the Developing World 349 Index 357 Box IV.1: Applying the GIFF: Comparative Value Chain Analysis 228 Figures IV.1: Evolution of Total Factor Productivity 220 IV.2: Total Factor Productivity Relative to the United States 221 IV.3: Prioritization of Value Chains for Further Investigation 235 VI.1: Relative Price of Production Factors and Technique Choice 293

8 Contents | VII 294 VI.2: Product Choice in an Industry VI.3: Industry and Product Choices in an Economy 295 323 VI.4: The TCI and Black-Market Premium VI.5: The TCI and the IEF 324 VI.6: The TCI and Expropriation Risk 325 VI.7: The TCI and Enterprise Autonomy 326 327 VI.8: The TCI and Openness VI.9: Development Strategy and Income Distribution 334 Tables IV.1: Macroeconomic Aggregates, 2003–2009 219 220 IV.2: Real Non-Oil GDP Growth, 2003–2009 222 IV.3: Contribution to Non-Oil GDP IV.4: Labor Force Status 223 IV.5: Types of Employment as a Percentage of the Sample Population 224 IV.6: Types of Wage Employment 224 IV.7: GDP Per Capita PPP in 2009 230 IV.8: Identifying Sectors for Growth: Key Exports of China, India, Vietnam, and Indonesia 232 234 IV.9: Nigeria’s Top Imports, 2010 238 IV.10: Criteria for Screening Potential Subsectors Average Wage, Including Benefi ts, by Industry 239 IV.11: Growth-Inhibiting Cross-Cutting Constraints, IV.A1: Interventions and Expected Outcomes 246 VI.1: Level of Per Capita Income 300 VI.2: Variable Defi nitions and Data Source 329 VI.3: Development Strategy and Economic Growth—Model 1 330 VI.4: Development Strategy and Economic Growth—Model 2 331 VI.5: Development Strategy and Economic Volatility 333 VI.6: The Effect of Development Strategy on Inequality 336 VI.7: Development Strategy and the Performance of Economic Reform/Transition 338 VI.A1: TCI Based on Value Added in the Manufacturing Sector 341

9

10 Acknowledgments This volume presents the key fi ndings of my research program on New Structural Economics (NSE), which I conducted during my tenure as Chief Economist and Senior Vice President of the World Bank from 2008 to 2011. This contribution to development economics and policy would not have been possible without the overall guidance and support of Robert Zoellick, President of the World Bank Group. I am grateful to Bob for his encouragement and valuable advice over my tenure. I am also very thankful to K. Y. Amoako, Alice Amsden, Ha-Joon Chang, Anne Krueger, Wonhyuk Lim, Howard Pack, Dani Rodrik, Joseph Stiglitz, Suresh Tendulkar, and Dirk Willem te Velde for providing insight- ful and detailed comments on the NSE framework. Their contributions fueled a thought-provoking and enriching debate about the framework and are presented in this book. In addition, I would like to thank my many friends and colleagues who provided thoughtful inputs and suggestions on various aspects of the work presented here and throughout the research process. In par- ticular, I am grateful to Shaida Badiee, Gary Becker, Otaviano Canuto, Ha-Joon Chang, Robert Cull, Augusto de la Torre, Christian Delvoie, Asli Demirgüç-Kunt, Shantayanan Devarajan, Hinh Dinh, Doerte Doemeland, IX

11 X | Acknowledgments Shahrokh Fardoust, Ariel Fiszbein, Robert Fogel, Alan Gelb, Indermit Gill, Ann Harrison, James Heckman, Vivian Hon, Jiandong Ju, Auguste Tano Kouame, Aart Kraay, John Litwack, Norman Loayza, Frank Lysy, Shiva Makki, William Maloney, Célestin Monga, Mustapha Nabli, Vikram Nehru, Ngozi Okonjo-Iweala, Howard Pack, Luiz Pereira da Silva, Nadia Piffaretti, Claudia Paz Sepulveda, Brian Pinto, Zia Qureshi, Martin Ravallion, David Rosenblatt, Sergio Schmukler, Luis Servén, Sunil Sinha, Hans Timmer, Volker Treichel, Harald Uhlig, Lixin Colin Xu, Yong Wang, and the many others whom I have had the pleasure and opportunity to collaborate with during the production of the manuscript. I would like to express special thanks to Doerte Doemeland, who worked closely with me in the fi nalization and editing of the manuscript. The research on NSE evolved from my previous work on economic Economic Research development and transition at the China Center for at Peking University. Several papers produced during that period are also included in this volume. I would like to take this opportunity to thank my former colleagues, Qiang Gong, Demin Huo, and Ho-Mou Wu, and for- mer students, Binkai Chen, Shudong Hu, Feiyue Li, Yongjun Li, Zhiyun Li, Mingxing Liu, Peilin Liu, Xifang Sun, Zhaoyang Xu, and Pengfei Zhang for their support and collaboration in the research. Last, the World Bank’s Offi ce of the Publisher provided excellent edi- torial, design, and printing services under the direction of Carlos Rossel. In this context, I would like to thank Santiago Pombo-Bejarano, Patricia Katayama, Aziz Gökdemir, Andrés Meneses, and Martha Gottron.

12 Abbreviations CAD comparative-advantage-defying CAF comparative-advantage-following DC developed countries gross domestic product GDP GIFF Growth Identifi cation and Facilitation Framework GMM Generalized Method of Moments HRS Household Responsibility System IEF Index of Economic Freedom IMF International Monetary Fund LDC less developed countries NIE newly industrialized economies NSE New Structural Economics R&D research and development TCI Technology Choice Index TFP total factor productivity TVE township and village enterprises (China) UNU-WIDER United Nations University – World Institute for Development Economics Research XI

13

14 Introduction The quest for sustainable growth has been the most intriguing topic in the An Inquiry world for economists and policy makers since Adam Smith’s into the Nature and Causes of the Wealth of Nations was published in 1776. Measured by today’s living standards, all countries in the world were poor at the beginning of the 18th century. Their economies were pre- dominately based on agriculture. Growth of gross domestic product (GDP) per capita had lingered at around 0.05 percent a year for millennia. Only after the onset of the Industrial Revolution did per capita income growth in the now advanced countries accelerate, jumping to around 1 percent a year in the 19th century and doubling to about 2 percent in the 20th cen- tury. This was an unimaginable change. While it took about 1,400 years for world income to double before the 18th century, the same process took only about 70 years in the 19th century and only 35 years in the 20th cen- tury for the now advanced countries (Maddison 1995). Nevertheless, the acceleration of growth was largely limited to the United Kingdom, where the Industrial Revolution began, a few western European economies, and Britain’s “offshoots”: Australia, Canada, New Zealand, and the United States (Maddison 1982). The result was a great divergence in income levels as the ratio of the top few to the majority bottom-income countries increased from 8.7 in 1870 to 38 by 1960 (Pritchett 1997). 1

15 2 | New Structural Economics After World War II, most countries in the developing world gained economic and political independence and started their postwar or post- independence reconstruction. By the end of the 20th century, a small set of developing countries was able to achieve prolonged high growth, catching up with or signifi cantly narrowing their gap with the advanced indus- trial economies. Japan, in 1950 a developing country with a per capita income one-fi fth of the United States, reached 63 percent of U.S. income by 1970 and became the world’s second-largest economy. Japan’s rise was the result of an impressive annual growth performance of 9.6 percent dur- ing the 1950s and 1960s, driven by the transformation from an agrarian to an industrial economy and continuous upgrading in key manufacturing sectors. Using an outward-oriented, market-friendly development strat- egy, the Asian Tigers—Hong Kong SAR, China; the Republic of Korea; Singapore; and Taiwan, China—grew in excess of 7 percent annually between the early 1960s and the early 1990s, demonstrating that it is possible to maintain impressive growth rates and to close the gap with advanced economies. More recently, growth in several large economies, such as China, Brazil, and India, has taken off, turning them into new global growth poles (World Bank 2011). These high growth rates have cant reduction in poverty. Between 1981 and 2005, the led to a signifi percentage of people living below US$1.25 a day was halved, falling from 52 percent to 26 percent. This drop in poverty was nowhere as apparent as in my home country, China. In 1981 a staggering 84 percent of Chinese lived in poverty. By 2005 this proportion had fallen to 16 percent—well below the average for the developing world. Although the occurrence of high, sustained growth further diversifi ed in the 21th century to some Sub-Saharan African and Latin American countries, such growth still remains the exception rather than the rule. Most developing countries suffered from prolonged uninterrupted spells of anemic growth (Reddy and Minoui 2009). Between 1960 and 2009, only about one third of low-income countries reached at least middle-income status. Despite the rising weight of middle-income countries in supporting global growth, many of them have been stuck in the “middle-income trap.” Of the countries that were independent and had middle-income status in 1960, almost three-fourths remained middle-income or had regressed to low-income by 2009. The ones that made it to high-income status are

16 Introduction | 3 countries in Western Europe, Japan, the Asian Tigers, and two island econ- omies in Latin America (Barbados and Trinidad and Tobago). If we can learn from the failed development attempts by most developing countries and especially the few successes, explore the nature and determinants of economic growth, and provide policy makers with the tools to unleash their country’s growth potential, poverty could become within a generation or two a memory of the past. Sustained economic growth cannot happen without structural changes (Kuznets 1966). All countries that remain poor have failed to achieve structural transformation, that is, they have been unable to diversify away from agriculture and the production of traditional goods into manufactur- ing and other modern activities. In Sub-Saharan Africa, which constitutes the core of the development challenge today, agriculture continues to play a dominant role, accounting for 63 percent of the labor force. Its share of manufacturing in 2005 was lower than in 1965 (Lin 2011). Recent empiri- cal work confi rms that the bulk of the difference in growth between Asia and developing countries in Latin America and Africa can be explained by the contribution of structural change to overall labor productivity (McMillan and Rodrik 2011). Development economics fi rst became an independent subdiscipline of modern economics after World War II. Various schools of fi rst-generation development economists in fact emphasized the importance of structural change and saw structural differences as a result of market failures. Not surprisingly, they proposed to use government interventions to facilitate structural change through import substitution and gave priority to mod- ern advanced industries. It was a period when new protective devices such as quantitative restrictions on imports and exchange controls to manage the balance of payments were fi rst used on a large scale by most coun- tries. Using Keynesianism as the main intellectual foundation for their analyses, early development economists advocated a “dirigiste dogma” (Lal 1983), positing as the central tenant of their theories that develop- ing countries were irremediably different from industrial countries. Most developing countries and multilateral development institutions followed these policy recommendations. From Latin America to Europe, Asia, and Africa, results were disappointing, and the gap with the industrialized countries widened.

17 4 | New Structural Economics The failure of the government interventions inspired by the fi rst-wave development thinking generated a new wave, which highlighted govern- ment failures and adopted an astructural approach toward economic development that emphasized the essential function of markets in allocat- ing resources and providing incentives for economic development, ignored the structural differences among countries at different levels of develop- ment in their policy recommendations, and expected the structural change to happen spontaneously in a country’s development process. Keynesian macroeconomics was also challenged by the emergence of ation in the 1970s, the Latin American debt crisis, and the collapse stagfl of the socialist planning system in the 1980s. The rational expectations theory became the dominant intellectual framework for development and helped refute the structuralist theoretical foundation for the state’s role in using fi scal, monetary, and trade policy for economic development. The new development thinking emphasized getting the price right, creating a stable market environment, strengthening the institutions necessary for markets to function well (property rights, good governance, business environment, and the like), and building human capital (education and health) to supply the increasingly skilled labor required by advances in technology. Multilateral institutions and development agencies were the main advo- cates for this wave of thinking and infl uenced economic policies in devel- oping countries through their programs. They based much of their policy advice and conditionality on stabilization and structural adjustment pro- grams that refl ected the new dominant paradigm and promoted economic liberalization, privatization, and the implementation of rigorous stabiliza- tion programs. The results of these policies for growth and employment generation were at best controversial. Something strange and unexpected happened in the recent history of economic development: it was observed that developing countries that succeeded during the second half of the 20th century did not follow the dominant development thinking or the policy prescriptions of the fi rst and second wave. That puzzling fact convinced researchers to revisit some of the big assumptions underlying theories of economic development. As pointed out, countries that have led the world growth since the Indus- trial Revolution and developing countries that have successfully converged

18 Introduction | 5 with developed countries all experienced profound structural changes in the composition of employment and the relative contribution of primary, sec- ondary and tertiary sectors to aggregate growth. Drawing lessons from the intellectual advances, controversies, and disappointments of development economics, a third wave of development thinking, advanced by a small group of economists such as Dani Rodrik, Ricardo Hausmann, Andres Velasco, Philippe Aghion, Michael Spence, Ann Harrison, Célestin Monga, myself, and a few others is well under way. It aims at bringing structural change back to the core of development studies, and it emphasizes the important roles for the market and the state in the process of promoting economic development. These economists all agree that the market should be the basic mechanism for resource allocation, but that government must play an active role in coordinating investments for industrial upgrading and diversifi cation and in compensating for externalities generated by fi rst movers in the dynamic growth process. The “New Structural Economics” presented in this book is an attempt to set out this third wave of development thinking. Taking into account the lessons learned from the growth successes and failures of the last decades, it advances a neoclassical approach to study the determinants and dynam- ics of economic structure. It postulates that the economic structure of an economy is endogenous to its factor endowment structure and that sus- tained economic development is driven by changes in factor endowments and continuous technological innovation. The factor endowments in a country are given at any specifi c time and changeable over time. A country’s comparative advantages and thus its optimal industrial structure are determined by its factor endowments. Upgrading the industrial structure in a given country requires the upgrad- ing of the factor endowment structure from one that is relatively abundant in labor and natural resources to one that is relatively abundant in capital, the introduction of new technologies, and the corresponding improvement in infrastructure to facilitate economic operations. The new structural economics argues that the best way to upgrade a country’s endowment structure is to develop its industries at any specifi c time according to the comparative advantages determined by its given endowment structure at that time. The economy will be most competitive, the economic sur- plus will be the largest, and the capital accumulation and the upgrading

19 6 | New Structural Economics of factor endowment structure will be the fastest possible. For the pri- vate enterprises in a country to enter industries according to the country’s comparative advantages, relative factor prices must fully refl ect the rela- tive abundance of those factors, and those prices can be determined only through competition in a well-functioning market. Therefore, the market should be the basic institution of the economy. For the introduction of new technologies, developing countries can turn their backwardness into an advantage by borrowing or adapting technolo- gies that have already matured in richer economies. In contrast, advanced economies must produce at the global technology frontier and have to invest continuously in new R&D to achieve technological innovation. Hence developing countries have the potential to achieve a rate of techno- logical innovation several times higher than that of advanced countries. Upgrading the industrial structure as well as the corresponding improve- ment in infrastructure, however, entails coordination of investments and compensation for externalities generated by fi rst movers that cannot be internalized by private enterprises. Without this coordination and com- pensation, the process of economic development will slow. The govern- ment should therefore play an active role in facilitating structural change through mitigating the coordination and externality problem. Chapter I reviews the evolution of development thinking and presents the main arguments and extensions of New Structural Economics. This chapter also includes insightful comments on the framework from my col- leagues Anne Krueger, Dani Rodrik, and Joseph Stiglitz and my rejoinder to their comments. Chapter II shows how the New Structural Economics complements pre- vious thinking on development and growth. It compares the predictions derived from the New Structural Economics with the stylized facts of suc- cessful countries identifi ed by the Growth Report issued in 2008 by the Commission on Growth and Development and discusses the policy lessons that can be drawn from the New Structural Economics. The principle of comparative advantage and the role of the state in facilitating structural transformation, which are key aspects of the framework, are further dis- cussed in a subsequent debate between Ha-Joon Chang and myself. The Growth Identifi cation and Facilitation Framework (GIFF), which lays out a step-by-step approach for policy makers to facilitate structural

20 Introduction | 7 change based on the framework of the New Structural Economics, is presented in chapter III. It guides policy makers on how to identify new industries consistent with a country’s latent comparative advantage. It also presents information, coordination, and externality issues intrinsic to industrial upgrading and discusses government policies that can help overcome these constraints. Explaining why industrial policy has often failed in the past, the chapter also warns against government policies that are aimed at protecting selected fi rms and industries that defy a coun- comparative advantage. Dirk Willem te Velde, Suresh Tendulkar, try’s Alice Amsen, K. Y. Amoako, Howard Pack, and Wonhyuk Lim provide thought-provoking comments on the approach. The chapter concludes with a rejoinder. Chapter IV illustrates how to apply the GIFF in developing countries. Using the example of Nigeria, the chapter identifi es appropriate compara- tor countries and selects a wide range of industries in which Nigeria may have latent comparative advantage as the comparator countries may be losing theirs. The chapter argues that these industries, which include food processing, light manufacturing, suitcases, shoes, car parts, and petro- chemicals, may lend themselves to targeted interventions of the govern- ment. The paper also discusses binding constraints to growth in each of these industries’ value chains as well as mechanisms through which gov- ernance-related issues in the implementation of industrial policy could be addressed. Chapter V focuses on the question of fi nancial structure and devel- opment. Financial structure varies signifi cantly across countries and, within a country, at different levels of development. The chapter argues that the optimal fi nancial structure in an economy is endogenous to real demand for fi nancial services based on industrial structure, which in turn hinges on a country’s comparative advantages. Historically, the fi nancial depth rather than fi nan- nancial literature has argued that fi cial structure matters for economic development. This chapter pro- vides an overview of theoretical and empirical advances that support the notion that fi nancial structure is important for economic develop- ment and endogenous to its industrial structure. It also discusses the circumstances under which the actual fi nancial structure deviates from its optimal structure.

21 8 | New Structural Economics The New Structural Economics argues that countries that pursue a comparative-advantage-following development strategy perform better than other countries. In chapter VI, the book presents empirical evidence to support this notion. It shows that countries that follow their comparative advantage have higher growth, lower economic volatility, and less inequal- ity. It argues that the failure of most developing countries to converge with advanced economies can be explained largely by their governments’ inap- propriate development strategies. In the past, governments placed prior- ity on the development of certain capital-intensive industries rather than focusing their efforts on upgrading a country’s endowment structure and creating an enabling environment for the development of sectors aligned with a country’s comparative advantage. Chapter VII points out that as wages rise rapidly in dynamically grow- ing emerging market economies, such as China, India, Brazil, Indonesia and others, in the multipolar growth world of the 21st century, the labor- intensive industries in those emerging market economies will be losing comparative advantages and provide golden opportunities for other low- income countries to enter. China alone currently has 85 million manufac- turing jobs in labor-intensive industries. If low-income countries in Africa and other parts of the world are able to seize these jobs, they will be able to grow dynamically, reduce poverty, and improve living standards quickly. Lower-income countries should therefore turn their late-comer status to their advantage by identifying mature industries in carefully selected lead countries and facilitating the entry of their own private enterprises or for- eign direct investments from the comparator countries into those indus- tries. This chapter also summarizes key policy messages and provides con- cluding thoughts. As stated in the annual UNU-WIDER Lecture that I delivered in Maputo on May 4, 2011, I believe that every developing country, including those in Sub-Saharan Africa, can grow at 8 percent or more continuously for several decades, signifi cantly reducing poverty and becoming middle- or even high- income countries in the span of one or two generations, if its government has the right policy framework to facilitate the private sector’s development along the line of its comparative advantages and tap into the late-comer advantages (Lin 2011). I hope that the publication of this book will make a contribution toward the realization of that goal in the developing world.

22 Introduction | 9 References Modern Economic Growth. Kuznets, S. 1966. New Haven, CT: Yale University Press. Lal, Deepak. 1983. “The Poverty of ‘Development Economics.’” Institute of Economic Affairs, London. Lin, Justin Yifu. 2011. “From Flying Geese to Leading Dragons: New Opportuni- ties and Strategies for Structural Transformation in Developing Countries.” WIDER Annual Lecture 15, Helsinki: UNU-WIDER. (A shorter version of this Global Policy. paper is forthcoming in ) Maddison, Angus. 1982. Phases of Capitalist Development. Oxford, UK: Oxford University Press. ———. 1995. Monitoring the World Economy, 1820–1992. Paris: OECD. McMillan, Margaret, and Dani Rodrik. 2011. “Globalization, Structural Change and Productivity Growth,” Kennedy School of Government, Harvard University, Cambridge, MA., http://www.hks.harvard.edu/fs/drodrik/ Research%20papers/Globalization,%20Structural%20Change,%20and%20 Productivity%20Growth.pdf. Pritchett, Lant. 1997. “Divergence, Big Time.” Journal of Economic Perspectives 11 (3): 3–17. Reddy, Sanjay, and Camelia Minoiu. 2009. “Real Income Stagnation of Countries 1960–2001.” Journal of Development Studies 45 (1): 1–23. Spence, M. 2011. The Next Convergence: The Future of Economic Growth in a Multispeed World. New York: Farrar, Straus and Giroux. World Bank. 2011. Global Development Horizons—Multipolarity: The New Global Economy. Washington, DC: World Bank.

23

24 I New Structural Economics: A Framework for Rethinking Development

25

26 I New Structural Economics: A Framework for Rethinking *† Development Several decades from now, when economic historians look back on the story of the past hundred years, it is very likely that they will be intrigued by the mystery of diverging performances by various countries, especially during the second half of the twentieth century. On the one hand, they will be amazed by the rapid growth path followed by a small number of countries such as Brazil, Chile, China, Indonesia, India, Korea, Malaysia, Mauritius, Singapore, Thailand, and Vietnam, where the industrializa- tion process quickly transformed their subsistence, agrarian economies * Adapted from “New Structural Economics: A Framework for Rethinking Development,” by Justin Yifu Lin, originally published in The World Bank Research Observer (2011) 26 (2): 193–221, published by Oxford University Press on behalf of the International Bank for Reconstruction and Development / The World Bank. © 2011 The International Bank for Reconstruction and Development/ The World Bank. 13

27 14 | New Structural Economics and lifted several hundred million people out of poverty in the space of generation. On the other hand, they will be puzzled by the apparent one inability of many other countries, where more than one-sixth of humanity remained trapped in poverty, to generate sustainable growth. They will also notice that with the exception of a few successful economies, there was little economic convergence between rich and poor countries in spite of the many efforts made by developing countries and despite the assis- tance of many multilateral development agencies. Long-term sustainable and inclusive growth is the driving force for poverty reduction in developing countries, and for convergence with developed economies. The current global crisis, the most serious one since the Great Depression, calls for a rethinking of economic theories. It is therefore a good time for economists to reexamine development theories as well. This paper discusses the evolution of development thinking since the end of World War II and suggests a framework to enable developing countries to achieve sustainable growth, eliminate poverty, and narrow the income gap with the developed countries. The proposed framework, called a neoclassical approach to structure and change in the process of economic development, or new structural economics, is based on the fol- lowing ideas: First, an economy’s structure of factor endowments evolves from one level of development to another. Therefore, the industrial structure of a given economy will be different at different levels of development. Each industrial structure requires corresponding infrastructure (both tangible and intangible) to facilitate its operations and transactions. Second, each level of economic development is a point along the con- tinuum from a low-income agrarian economy to a high-income post- industrialized economy, not a dichotomy of two economic development levels (“poor” versus “rich” or “developing” versus “industrialized”). Industrial upgrading and infrastructure improvement targets in developing countries should not necessarily draw from those that exist in high-income countries. Third, at each given level of development, the market is the basic mechanism for effective resource allocation. However, economic devel- opment as a dynamic process entails structural changes, involving indus- trial upgrading and corresponding improvements in “hard” (tangible) and “soft” (intangible) infrastructure at each level. Such upgrading and

28 New Structural Economics: A Framework for Rethinking Development | 15 improvements require an inherent coordination, with large externalities to rms’ transaction costs and returns to capital investment. Thus, in addi- fi tion to an effective market mechanism, the government should play an active role in facilitating structural changes. The remainder of the paper is organized as follows: the next section examines the evolution of development thinking and offers a critical review of some of its main schools of thought. I then outline the basic principles and conceptual framework of the new structural economics, the function of the market, and the roles of a facilitating state. In the next section I highlight similarities and differences between old and new structural eco- nomics, and discuss some preliminary insights on major policy issues based on this new approach. A Short Review of Development Thinking and Experiences The process of sustainable per capita income increase and economic growth, characterized by continuous technological innovation and indus- trial upgrading, is a modern phenomenon. From Adam Smith to the early twentieth century, most economists believed that laissez-faire was the best vehicle for achieving sustainable growth in an economy. It was assumed that in thriving economies all decisions about resource allocation are made by economic agents interacting in markets free of government interven- tion. The price system determines not only what is produced and how but also for whom. Households and fi rms pursuing their own interests would be led, “as if by an invisible hand,” to do things that are in the interests of others and of society as a whole. Although the laissez-faire approach was challenged by Marxist economists and others, it became the domi- nant intellectual framework for the study of growth in all countries and remained so for a long time. It certainly provided many good insights on the process of economic development but it missed the importance of the process of continuous, fundamental technological changes and industrial upgrading, which distinguishes modern economic growth from premodern economic growth (Kuznets 1966). The study of economic development proceeds in two related but sep- arate tracks: growth theories and development theories. While some of the key ingredients of modern growth theory such as competitive behav- ior, equilibrium dynamics, the importance of physical capital and human

29 16 | New Structural Economics capital, the possibility of diminishing returns, and the impact of techno- logical progress can be found in the work of classical economists (Ramsey 1928; Schumpeter 1934), systematic modeling only started in the 1940s when some pioneers used primary factors to build generic models based on aggregate production functions. Harrod (1939) and Domar (1946) trig- gered extensive research along these lines. Following their initial work, the Solow-Swan model sparked the fi rst major wave of systematic growth analysis. The objective was to understand the mechanics of growth, iden- tify its determinants, and develop techniques of growth accounting, which would help explain changes in the momentum and role of economic policy. rst generation of growth researchers highlighted the centrality of That fi capital. One important prediction from these models was the idea of con- ditional convergence, derived from the assumption of diminishing returns to capital—poor economies with lower capital per worker (relative to their long-run or steady-state capital per worker) will grow faster. While that assumption allowed the model to maintain its key prediction of conditional convergence, it also seemed odd: technology, the main determinant of long- run growth, was kept outside of the model (Lin and Monga 2010). A new wave of growth modeling had to come up with a convincing theory of technological change. Endogenous growth theory, as it came to be known, maintained the assumption of nonrivalry because technology is indeed a very different type of factor from capital and labor—it can be used indefi nitely by others, at zero marginal cost (Romer 1987, 1990; Aghion and Howitt 1992). But it was important to take the next logical step and to understand better the public good characterization of technology and think of it as a partially excludable nonrival good. The new wave there- fore reclassifi ed technology not just as a public good but as a good that is subject to a certain level of private control. However, making it a partially excludable nonrival good and therefore giving it some degree of exclud- ability or appropriability was not suffi cient to ensure that incentives for its production and use were socially optimal. The move away from perfect competition was therefore necessary. It has yielded high methodological payoffs. While neoclassical models of growth took technology and factor accumulation as exogenous, endogenous growth models explain why tech- nology grows over time through new ideas and provide the microeconomic underpinnings for models of the technological frontier.

30 New Structural Economics: A Framework for Rethinking Development | 17 Another important question has been to understand how technological diffusion takes place across countries and generates or sustains growth— and why it does not take root in others. Various interesting possibilities have recently been explored in an attempt to answer that critical ques- tion (Jones 1998; Acemoglu, Johnson, and Robinson 2001; Glaeser and Shleifer 2002). Both on the theoretical and empirical fronts, progress has been made in our understanding of growth in recent decades. However, cant methodological diffi culties and chal- growth research still faces signifi lenges in identifying actionable policy levers to sustain and accelerate growth in specifi c countries. Intellectual progress has been even slower in the particular domain of development theories. It took a paper by Rosenstein-Rodan (1943) to bring development issues to the forefront of the discipline. The paper suggested that the virtuous circle of development depended essentially on the interaction between economies of scale at the level of individual fi rms and the size of the market. Specifi cally, it assumed that modern methods of production can be made more productive than traditional ones only if the market is large enough for their productivity edge to compensate for the necessity of paying higher wages. But the size of the market itself depends on the extent to which these modern techniques are adopted. Therefore, if the modernization process can be started on a very large scale, then the process of economic development will be self- reinforcing and self-sustaining. If not, countries will be indefi nitely trapped in poverty. Rosenstein-Rodan’s framework sparked a wave of similar ideas (Chang 1949; Lewis 1954; Myrdal 1957; Hirschman 1958) which came to be known as the structuralist approach to economic development. These early development theories held that the market encompassed insurmountable defects and that the state was a powerful supplementary means to accelerate the pace of economic development (Rosenstein-Rodan 1943; Nurkse 1953; Hirschman 1958). The slump of international trade in the Great Depres- sion led to export pessimism in the post-War period. In Latin America, for instance, political leaders and social elites were infl uenced strongly by the deterioration in the terms of trade, the economic diffi culty encountered during the Great Depression in the 1930s, and the thesis developed by Prebisch (1950) and Singer (1950). They believed that the decline in the terms of trade against the export of primary commodities was secular and

31 18 | New Structural Economics led to the transfer of income from resource-intensive developing countries to capital-intensive developed countries. They argued that the way for a developing country to avoid being exploited by developed countries was to develop domestic manufacturing industries through a process known as import substitution. Moreover, the emergence of previous colonies or semi- colonies as newly independent states in Asia and the Middle East, and later in Africa, was accompanied by strong nationalist sentiments. The results were disappointing in many cases. In many developing coun- tries, well-intended government interventions failed. This was the case across Latin American, African, and South Asian countries in the 1960s and 1970s when import substitution and protection were essential features of the development strategy. One of the main reasons for the failure of many former socialist and developing countries to achieve dynamic growth in their transitional processes was the fact that they attempted to defy the comparative advantage determined by their endowment structures and gave priority to development of capital-intensive heavy industries when capital in their economies was scarce. In order to implement such strate- gies, developing-country governments had to protect numerous nonviable enterprises in their priority sectors (Lin 2009a; Lin and Li 2009). By shielding unsustainable industries from import competition, devel- oping countries also imposed various types of other costs on their econo- mies. Protection typically led to: (i) an increase in the price of imports and import-substituting goods relative to the world price and distortions in incentives, pushing the economy to consume the wrong mix of goods from the point of view of economic effi ciency; (ii) the fragmentation of mar- kets, as the economy produced too many small-scale goods, which resulted again in loss of effi ciency; (iii) decreased competition from foreign fi rms and support for the monopoly power of domestic fi rms whose owners were politically well connected; and (iv) opportunities for rents and corruption, which raised input and transaction costs (Krueger 1974; Krugman 1993). As government-led economic development strategies based on the structuralist teachings failed in many countries, the free market approach appeared to triumph and infl uence development thinking. This trend was reinforced by a new revolution in macroeconomics. The prevailing Keynesian macroeconomics was challenged by the stagfl ation in the 1970s, the Latin American debt crisis, and the collapse of the socialist planning

32 New Structural Economics: A Framework for Rethinking Development | 19 system in the 1980s. The so-called rational expectations revolution emerged and refuted the structuralist theoretical foundation for the state’s role in using fi scal and monetary policy for economic development. The Latin American debt crisis began in 1982 when international fi nan- cial markets realized that the collapse of the Bretton Woods system had put some countries with unlimited access to foreign capital in a situation where they could not pay back their loans. The crisis was precipitated by a num- ber of interrelated exogenous shocks that toppled the Mexican and several other Latin American economies, which were already overburdened with a substantial percentage of the world’s outstanding debt (Cardoso and Helwege 1995). It prompted multilateral lending institutions and bilateral lenders—especially the United States—to call for a comprehensive set of reforms of Latin American economies and to advocate a set of free-market policies that followed the canons of the neoclassical paradigm, later known as the Washington Consensus (Williamson 1990). The Washington Consensus quickly came to be perceived as “a set of neoliberal policies that have been imposed on hapless countries by the Washington-based international fi nancial institutions and have led them to crisis and misery” (Williamson 2002). It promoted economic liberal- ization, privatization, and the implementation of rigorous stabilization programs. The results of these policies in terms of growth and employ- ment generation were at best controversial (Easterly, Loayza, and Montiel 1997; Easterly 2001). By the end of the 1990s and parallel to the dis- missal of structuralism and the prevalence of the free market approach, the development economics research community was witnessing the end of an era dominated by cross-country regressions, which attempted to identify growth determinants. That approach had been to focus on the independent and marginal effects of a multitude of growth determinants. This led to the linearization of complex theoretical models. Yet, the general view was that growth determinants interact with each other. To be successful, some pol- icy reforms must be implemented with other reforms. There was a general perception that the policy prescriptions stemming from such regressions did not produce tangible results. An alternative perspective on non-linearities was the Growth Diag- nostics or Decision Tree approach suggested by Hausmann, Rodrik, and Velasco (2005). They recognized the central role of structural change in

33 20 | New Structural Economics economic development and argued that there are “binding constraints” on growth in each country. These authors suggested that binding constraints can vary over time and across countries. They concluded that identifi ca- tion of the binding constraint was therefore key in practice. This frame- work highlighted pragmatically the inability of governments to reform everything and stressed the need to prioritize reforms, which should be done through the information revealed by shadow prices. It should be noted that the Growth Diagnostics approach is not operational unless one assumes away reform complementarities, which is the feature of linear growth regressions. The divergence in growth performance between developed and devel- oping countries, despite predictions of convergence from mainstream economic theory, has led to controversy. Some have concluded that the pol- icy prescriptions, or expectations about their effectiveness, or both, were wrong. Others have observed that growth researchers had paid limited attention to heterogeneity (the specifi c characteristics of each country). The suggestion that cross-country distribution may be multimodal (with the existence of “convergence clubs”) did not settle the debate about which new directions were needed for growth research. Instead, many basic ques- tions have come back on the agenda: Are development economists looking in the wrong place in their quest for the determinants of growth? Should the focus be on institutions (institutional outcomes), instead of or in addi- tion to policies? And, assuming that they are not refl ecting other factors, how can good institutional outcomes be generated? These unanswered questions were on the agenda for a long time. Start- ing in the 1980s, many development economists tried to understand better the causality of relationships and the various transmission chan- nels through which policies, institutional changes, or foreign aid affect growth. They were also the rationale for an increased focus of growth research on microbehavior issues at the household and fi rm levels, with two goals: (i) allowing for heterogeneity in the economy (across and within countries); and (ii) investigating how constraints to growth oper- ate at the microlevel. The growing disappointment and disillusionment with aid effectiveness also led to the quest for rigorous impact evaluation of development proj- ects and programs. This has generated a new approach to development

34 New Structural Economics: A Framework for Rethinking Development | 21 led by economists at the MIT Poverty Lab, whose goal is “to reduce pov- c evidence” through the erty by ensuring that policy is based on scientifi use of randomized control trials (RCT) or social experiments. Although RCT are good tools for understanding the effectiveness of some specifi c microprojects, they often do not start from a clear strategic assessment of how a particular method would fi t the knowledge gaps of highest priority (Ravallion 2009). All too often, research looks for topics “under the light.” The positive outcomes for policymaking are more often the occasional by-products of research than its objective from the outset. Recent microempirical studies may have indeed shed light on some important problems, such as the impact of the investment climate on fi rm performance or the impact of household behavior on produc- Rosenzweig and Wolpin 1985). But “there is a risk the bulk of tivity ( present-day research in development economics appears to be too nar- rowly focused and/or of too little generalizability to help much in the fi ght against poverty and to facilitate structural change and sustained growth” (World Bank 2010). The time has come to reexamine the state of development economics, to learn from past experiences and previous knowledge, and to offer new thinking and a new framework. Drawing lessons from past experience and from economic theories, the next section presents the key principles of a new structural economics, which is a neoclassical approach to economic 1 structure and dynamic change in the process of economic development. A Neoclassical Approach to Structure and Change The starting point for the analysis of economic development is an econ- omy’s endowments. Endowments are a given in an economy at any spe- cifi c time and are changeable over time. Following the tradition of classical economics, economists tend to think of a given country’s endowments as consisting only of its land (or natural resources), labor, and capital (both 2 These are in fact factor endowments, which fi rms physical and human). in an economy can use in production. It should be noted that the analysis of new structural economics focuses on the dynamics of the capital/labor ratio. This is because land is exogenously given in any realistic discussion of a country’s development and natural resources, such as mining resources,

35 22 | New Structural Economics exist underground in fi xed quantity and their discovery is often random. Conceptually, it is useful to add infrastructure as one more component in an economy’s endowments. Infrastructure includes hard (or tangible) infrastructure and soft (or intangible) infrastructure. Examples of hard infrastructure are highways, port facilities, airports, telecommunication systems, electricity grids, and other public utilities. Soft infrastructure con- sists of institutions, regulations, social capital, value systems, and other social, economic arrangements. Infrastructure affects the individual fi rm’s transaction costs and the marginal rate of return on investment. different Countries at different levels of development tend to have economic structures due to differences in their endowments. Factor endowments for countries at the early levels of development are typically characterized by a relative scarcity of capital and relative abundance of labor or resources. Their production activities tend to be labor intensive or resource intensive (mostly in subsistence agriculture, animal husbandry, fi shery, and the mining sector) and usually rely on conventional, mature technologies and produce “mature,” well-established products. Except for mining and plantations, their production has limited economies of scale. Their fi rm sizes are usually relatively small, with market transactions often informal, limited to local markets with familiar people. The hard and soft infrastructure required for facilitating that type of production and market transactions is limited and relatively simple and rudimentary. At the other extreme of the development spectrum, high-income coun- tries display a completely different endowment structure. The relatively abundant factor in their endowments is typically capital, not natural resources or labor. They tend to have comparative advantage in capital intensive industries with economies of scale in production. The various types of hard infrastructure (power, telecommunication, roads, port facili- ties, etc.) and soft infrastructure (regulatory and legal frameworks, cultural value systems, etc.) that are needed must comply with the necessities of national and global markets where business transactions are long distance and large in quantity and value. Economic development requires continuous introduction of new and better technology to an existing industry. Most people in low-income countries depend on agriculture for their livelihood. Improvements in agricultural technology are key to increasing farmers’ income and

36 New Structural Economics: A Framework for Rethinking Development | 23 reducing poverty. However, economic development also requires contin- uous diversifying and upgrading from existing industries to new, more capital-intensive ones. Without such a structural change, the scope for sustained increase in per capita income will be limited. Therefore, the discussion in this paper will focus mostly on issues related to industrial upgrading and diversifi cation. Developing countries have the advantage of backwardness in the upgrading process and a whole spectrum of industries with different levels of capital intensity available for them to choose. However, they must fi rst upgrade their factor endowment structure, which requires their stock of capital to grow more rapidly than the labor force (see Ju, Lin, and Wang 2009). When they move up the industrial ladder in the process of economic development, they also increase their scale of production—because of the indivisibility of capital equipment. Their fi rms become larger and need a bigger market, which in turn necessitates correspondent changes in power, transportation, fi nancial arrangements, and other soft infrastructure. The process of industrial upgrading and diversifi cation also increases the level of risk faced by fi rms. As fi rms move closer to the global technol- ogy frontier, it becomes increasingly diffi cult for them to borrow mature technology from advanced countries. They increasingly need to invent new technologies and products and thus face more risk. The idiosyncratic risk of a fi rm has three components based on risk sources: technological inno- vation, product innovation, and managerial capacity. At the early level of rms tend to use mature technologies to produce mature development, fi products for mature markets. At that level, the main source of risk is the managerial ability of fi rms’ owner-operators. At a higher level of develop- ment, fi rms often invent new technologies to produce new products for new markets. In addition to managerial capacity, such fi rms face risks arising from the maturity of technology and markets. Therefore, while technological innovation, product innovation, and managerial capacity all contribute to the overall level of risk associated with fi rms, their relative importance varies greatly from one industry to another and from one level of economic development to another. With changes in the size of fi rms, scope of the market, and nature of risk, along with the upgrading of the industrial structure, the require- ments for infrastructure services, both hard and soft, also change. If the

37 24 | New Structural Economics infrastructure is not improved simultaneously, the upgrading process in ciency, a phenomenon various industries may face the problem of x-ineffi discussed by Leibenstein (1957). Because the industrial structure in an economy at a specifi c time is endogenous to its given relative abundance of labor, capital, and natural resources at that time, the economy’s factor endowment will change with capital accumulation or population growth, pushing its industrial structure to deviate from the optimal determined by 3 its previous level. rms choose to enter industries and adopt technologies that are When fi consistent with the comparative advantage determined by changes in the 5 4 country’s factor endowments, As com- the economy is most competitive. rms grow, they claim larger domestic as well as petitive industries and fi international market shares and create the greatest possible economic sur- plus in the form of profi ts and salaries. Reinvested surpluses earn the high- est return possible as well, because the industrial structure is optimal for that endowment structure. Over time, this approach allows the economy to accumulate physical and human capital, upgrading the factor endowment rms structure as well as the industrial structure and making domestic fi more competitive over time in more capital- and skill-intensive products. Firms care about profi ts. For them spontaneously to enter industries and choose technologies consistent with the economy’s comparative advantage, the price system must refl ect the relative scarcity of factors in the country’s endowment. This only happens in an economy with competitive markets (Lin 2009a; Lin and Chang 2009). Therefore, a competitive market should be the economy’s fundamental mechanism for resource allocation at each level of its development. That kind of comparative-advantage-following approach in economic development may appear to be slow and frustrating in countries with major poverty challenges. In reality, it is the fastest way to accumulate capital and upgrade the endowment structure, and the upgrad- ing of industrial structure can be accelerated by better access to technol- ogy and industries already developed by and existing in more advanced countries. At each level in their development, fi rms in developing countries can acquire the technologies (and enter the industries) that are appropriate for their endowment structure, rather than having to reinvent the wheel (Gerschenkron 1962; Krugman 1979). This possibility to use off-the-shelf technology and to enter into existing industries is what has allowed some

38 New Structural Economics: A Framework for Rethinking Development | 25 of the East Asian newly industrialized economies to sustain annual GDP growth rates of 8 and even 10 percent. As a country climbs up the industrial and technological ladder, many rms becomes more other changes take place: the technology used by its fi sophisticated, and capital requirements increase, as well as the scale of production and the size of markets. Market transactions increasingly take place at arm’s length. A fl exible and smooth industrial and technologi- cal upgrading process therefore requires simultaneous improvements in nancial, and legal institutions, and in hard infrastructure so educational, fi rms in the newly upgraded industries can reduce transaction costs that fi and reach the production possibility frontier (Harrison and Rodríguez- Clare 2010). Clearly, individual fi rms cannot internalize all these changes cost effectively, and spontaneous coordination among many fi rms to meet these new challenges is often impossible. Change in infrastructure requires collective action or at least coordination between the provider of infrastruc- ture services and industrial fi rms. For this reason, it falls to the government either to introduce such changes or to coordinate them proactively. Successful industrial upgrading in responding to change in an econo- rms overcome issues my’s endowment structure requires that the pioneer fi of limited information regarding which new industries are the economy’s latent comparative advantages determined by the changing endowment structure. Valuable information externalities arise from the knowledge gained by pioneer fi rms in both success and failure. Therefore, in addition to playing a proactive role in the improvements of soft and hard infrastruc- tures, the government in a developing country, like that in a developed country, needs to compensate for the information externalities generated by pioneer fi rms (Rodrik 2004; Lin 2009a; Lin and Monga 2011; Harrison 6 and Rodríguez-Clare 2010). What Is “New” About the New Structural Economics? Like all learning ventures, economic development thinking is bound to be a continuous process of amalgamation and discovery, continuity, and reinvention. The existing stock of knowledge has been the result of many decades of work by thinkers from various backgrounds and disciplines and has come to light through several waves of theoretical and empirical

39 26 | New Structural Economics research. It is therefore only natural that the proposed new structural eco- nomics has some similarities to and differences from previous strands in the development economics literature. Its main value-added should be assessed on the new policy insights it provides and the pertinence of the research agenda ahead. Difference with Earlier Literature on Structural Change Earlier thinking on structural change in the context of economic develop- ment is mostly associated with Rostow (1990 [1960]) and Gerschenkron (1962). In trying to understand how economic development occurs and what strategies can be adopted to foster that process, the former suggested that countries can be placed in one of fi ve categories in terms of their level of growth: (i) traditional societies, characterized by subsistence economy, with output not traded or even recorded, the existence of barter, high levels of agriculture, and labor-intensive agriculture; (ii) societies with precondi- tions to growth, where there is an increase in capital use in agriculture, the development of mining industries, and some growth in savings and invest- ment; (iii) societies in take-off mode, with higher levels of investment and industrialization, accumulation of savings, and a decline in the share of the agricultural labor force; (iv) societies that drive to maturity and where wealth generation enables further investment in value adding industry and ed, and development—growth becomes self-sustaining, industry is diversifi more sophisticated technology is used; and (v) mass-consumption societies that achieve high output levels and where the services industry dominates the economy. Gerschenkron questioned Rostow’s proposition that all developing countries pass through a similar series of levels and its implication that it is possible to generalize the growth trajectory of different countries. For the new structural economics, economic development from a low level to a high level is a continuous spectrum, not a mechanical series of fi ve dis- tinguished levels. Although the change in an economy’s industrial structure refl ects the changes in that economy’s endowment structure, the develop- ment of industries in different countries with a similar endowment struc- ture can be achieved in different and nonlinear ways. This is especially true with the increased globalization of markets, the rapid development of new products, and constant technological change, as countries can exploit

40 New Structural Economics: A Framework for Rethinking Development | 27 opportunities that were not available in the past and specialize in industries that are likely to vary from one economy to another. The new structural economics also provides a framework for under- standing the endogeneity and exogeneity issues surrounding the key styl- ized facts of modern growth analysis that have been outlined by the Growth Commission (2008) and Jones and Romer (2009): an economy that fol- lows its comparative advantage in the development of its industries will be most competitive in domestic and world markets. As a result, the economy will generate potentially the largest income and surplus for savings. Capital investment will also have the largest possible return. Consequently, house- holds will have the highest savings propensity, resulting in an even faster upgrade of the country’s endowment structure (Lin and Monga 2010). Similarities to and Differences from Old Structural Economics In terms of similarities, the “new” and the “old” structural economics are both founded on structural differences between developed and develop- ing countries and acknowledge the active role of the state in facilitating the movement of the economy from a lower level of development to a higher one. However, there are profound differences between these two approaches regarding their targets and the modalities of state interven- tion. The old structural economics advocates development policies that go against an economy’s comparative advantage and advise governments in developing countries to develop advanced capital-intensive industries through direct administrative measures and price distortions. By contrast, the new structural economics stresses the central role of the market in resource allocation and advises the state to play a facilitating role to assist fi rms in the process of industrial upgrading by addressing externality and coordination issues. The differences between the two frameworks derive from their dissimi- lar views on the sources of structural rigidities: old structural economics assumes that the market failures that make the development of advanced capital-intensive industries diffi cult in developing countries are exoge- nously determined by structural rigidities due to the existence of monopo- lies, labor’s perverse response to price signals, and/or the immobility of factors. By contrast, the new structural economics posits that the failure to develop advanced capital-intensive industries in developing countries

41 28 | New Structural Economics is endogenously determined by their endowments. The relative scarcity in their capital endowment and/or the low level of soft and hard infra- structure in developing countries make the reallocations from the existing table for the industries to the advanced capital-intensive industries unprofi fi rms in a competitive market. Old structural economics assumes a dual and restrictive view of the world, with a binary classifi cation of only two possible categories of countries: “low-income, periphery countries” versus “high-income, core countries.” As a result, it views the differences in the industrial structure between developed and developing countries as expressing a dichotomy. Contrary to that vision, the new structural economics considers these differences as the refl ection of a whole spectrum that includes many dif- ferent levels of development. The new structural economics also rejects dependency theories. In an increasingly globalized world, it sees oppor- tunities for developing countries to counter negative historical trends by diversifying their economy and building industries that are consistent with their comparative advantage so as to accelerate growth and achieve convergence by exploiting the advantage of backwardness in an open, globalized world. Another major difference between the new and the old structural eco- nomics is the rationale for using key instruments of economic manage- ment. Old structural economics sees systematic government intervention in economic activities as the essential ingredient in the modernization objective. Among the key instruments used to move from “developing” countries to “industrialized” countries are generalized protectionism (such as government-imposed tariffs on imports to protect infant industries), rigid exchange-rate policies, fi nancial repression, and the creation of state- owned enterprises in most sectors. By contrast, the new structural economics recognizes that import sub- stitution is a natural phenomenon for a developing country climbing the industrial ladder in its development process, provided that it is consis- tent with the shift in comparative advantage that results from changes in its endowment structure. But it rejects conventional import- substitution strategies that rely on the use of fi scal policy or other distortions in low-income, labor- or resource-abundant economies to develop high cost, advanced capital-intensive industries that are not consistent with the country’s comparative advantage. It also stresses the idea that the industrial

42 New Structural Economics: A Framework for Rethinking Development | 29 upgrading process in a developing country should be consistent with the ects the accumula- change in the country’s comparative advantage that refl tion of human and physical capital and the change in its factor endow- ment structure—this ensures the viability of fi rms in new industries. The new structural economics concludes that the role of the state in industrial diversifi cation and upgrading should be limited to the provision of infor- mation about the new industries, the coordination of related investments rms in the same industries, the compensation of informa- across different fi rms, and the nurturing of new industries tion externalities for pioneer fi through incubation and encouragement of foreign direct investment (Lin 2009a; Lin and Chang 2009; Lin and Monga 2011). The state also needs to assume effectively its leadership role in the improvement of hard and soft infrastructure in order to reduce transaction costs on individual fi rms and so facilitate the economy’s industrial development process. New Structural Economics: Some Policy Insights The ultimate goal of development thinking is to provide policy advice that facilitates the quest for sustainable and inclusive economic and social progress in poor countries. Although specifi c policy measures to be derived from the new structural economics approach will require further research and depend very much on country context and circumstances, in this sec- tion I will make some conjectures about a few preliminary insights on various topics. Until Britain’s very high unemployment of the 1920s and the Fiscal Policy. Great Depression, economists generally held that the appropriate stance for fi scal policy was for governments to maintain balanced budgets. The severity of the early twentieth-century crises gave rise to the Keynesian idea of counter-cyclicality, which suggested that governments should use tax and expenditure policies to offset business cycles in the economy. By contrast, neoclassical economics offers doubts about the implicit assump- 7 tion behind the Keynesian model of a multiplier greater than one and its implication that governments are able to do something that the pri- vate sector has been unable to do: mobilize idle resources in the economy (unemployed labor and capital) at almost zero social cost, that is, with no corresponding decline in other parts of GDP (consumption, investment, and net exports). Instead, they warn against the possibility of the so-called

43 30 | New Structural Economics Ricardian equivalence trap and point to the fact that households tend to adjust their behavior for consumption or saving on the basis of expecta- scal policy (stimu- tions about the future. They suggest that expansionary fi lus packages) is perceived as immediate spending or tax cuts that will need to be repaid in the future. They conclude that the multiplier could be less than one in situations where the GDP is given and an increase in govern- ment spending does not lead to an equal rise in other parts of GDP. The neoclassical paradigm even suggests the possibility of some rare instances scal contrac- where multipliers are negative, pointing to situations where fi tions become expansionary (Francesco and Pagano 1991). scal From the viewpoint of new structural economics, the effects of fi policy may be different in developed and developing countries due to the differences in opportunities of using counter-cyclical expenditure for mak- ing productivity-enhanced investments. Physical infrastructure in general is a binding constraint for growth in developing countries, and govern- ments need to play a critical role in providing essential infrastructure to facilitate economic development. In such contexts, recessions are typically good times for making infrastructure investments, for three main reasons. First, such investments boost short-term demand and promote long-term 8 growth. Second, their investment cost is lower than in normal times. And third, the Ricardian equivalence trap can be avoided because the increase scal revenues can compensate for the cost of in future growth rates and fi these investments (Lin 2009b). If a developing country government follows the new structural econom- ics approach of facilitating the development of industries according to the country’s comparative advantage, its economy will be competitive and the fi scal position and the external account are likely to be sound, thanks to the likelihood of strong growth, good trade performance, and the lack of nonviable fi rms that the government has to subsidize. Under this scenario, the country will face fewer homegrown economic crises. If the economy is hit by external shocks such as the recent global crisis, the government will be in a good position to implement a counter-cyclical fi scal stimulus and invest in infrastructure and social projects. Such public investments can enhance the economy’s growth potential, reduce transaction costs on the private sector, increase the rate of return on private investment, and generate enough tax revenues in the future to liquidate the initial costs.

44 New Structural Economics: A Framework for Rethinking Development | 31 In addition to its different stance on fi scal stimulus, the new structural economics approach also offers a different strategy for managing natural resource wealth. In resource-abundant countries, it would recommend that an appropriate share of revenues from commodities be used to invest in human capital, infrastructure, social capital, and compensation for fi rst movers in new nonresource sectors so as to facilitate the structural trans- formation. To accomplish this with the greatest effect, these resources nance investment opportunities that remove binding constraints should fi on industrial diversifi cation and upgrading, especially in the infrastructure and education sectors. Microeconomic analyses show that even when fac- oor costs are comparable, ineffi tory fl ciencies in infrastructure can make it impossible for poor countries to compete on international markets. Freight and insurance costs in African countries are 250 percent of the global aver- 9 with road freight delays two to three times as long as in Asia. Lacking age, fi nancial resources and the appropriate policy frameworks, many of these countries are often unable to sustain much needed investment and main- tenance expenditures. In such contexts, the effective fi scal strategy would not be to keep natural resource revenues in sovereign funds and invest in foreign equity markets or projects but, rather, to use a substantial portion nancing domestic or regional projects that facilitate of the revenues for fi economic development and structural change—i.e., projects that stimulate the development of new manufacturing industries, diversify the economy, 10 provide jobs, and offer the potential of continuous upgrading. Monetary Policy. Old structural economics suggested that monetary policy should be under government control (not independent central banks) and directed at infl uencing interest rates and even sector credit allocation. But it also acknowledged that many other factors that infl uence the investment demand-schedule in developing countries are too powerful for monetary policy alone to achieve suffi cient levels of investment, channel resources in strategic sectors, and combat unemployment. Building on lessons from the rational expectations revolution, neoclas- sical economists doubted the idea that monetary policy could be used to support industrial development. They recommended that its main goal be price stability, and advocated the use of short-term interest rates by inde- pendent central banks to maintain the general level of prices (or to control

45 32 | New Structural Economics money supply growth), and not to stimulate economic activity and trigger ation. infl The new structural economics envisions the possibility of using inter- est rate policy in developing countries as a counter-cyclical tool and as an instrument to encourage infrastructure and industrial upgrading investments during recessions—measures that may contribute to productivity growth in the future. Monetary policy is often ineffective for stimulating investment and consumption in recessions and excess capacity situations in developed countries, especially when nominal interest rates hit the zero bound in a context of limited profi table investment opportunities, pessimistic expecta- tions, high unemployment rates, low confi dence about the future, and the likelihood of liquidity traps. It should be noted, however, that developing countries are less likely to encounter such liquidity traps. Even when faced with excess capacity in existing domestic industries, their scope for indus- trial upgrading and diversifi cation is large. Their fi rms have incentives to undertake productivity-enhancing, industrial- upgrading investments dur- ing recessions if interest rates are suffi ciently low. Furthermore, they tend to have many infrastructure bottlenecks. Lowering interest rates in such contexts would also encourage investments in infrastructure. The objective of monetary policy should be much broader than tradi- tionally conceived under neoclassical economics—in economic slumps, it should aim at encouraging investment that removes bottlenecks on growth. In practical terms, this implies not just that interest rates should be lowered in the slump, as would be the case in most circumstances under a stan- dard Taylor rule. It also implies that monetary authorities should resort to temporary interest rate subsidies, fl exible credit allocation rules, or similar time-bound devices, targeting infrastructure projects identifi ed by devel- opment banks as binding constraints, preferably in specifi c geographic locations where the payoff is the largest and where political economy con- straints can be more easily managed. There is ample consensus that fi nancial develop- Financial Development. ment is essential to sustaining economic growth. There is, however, much less agreement on the specifi c role it plays in that process. Starting with the observation that one of the major constraints facing developing countries was limited capital accumulation, old structural economics regarded the problems of the fi nancial sector in underdeveloped economies as resulting

46 New Structural Economics: A Framework for Rethinking Development | 33 from widespread market failures that could not be overcome by market 11 They recommended that governments adopt a hands-on forces alone. approach in that process, mobilize savings, and allocate credit to support the development of advanced capital-intensive industries. This very often led to fi nancial repression (McKinnon 1973; Shaw 1973). In some coun- tries, especially in Sub-Saharan Africa, the belief in soft-budget constraints cits in state-owned fi nancial institu- led governments to accumulate defi tions and created a pervasive business culture of self-repression not only for banks, but also for private enterprises (Monga 1997). Drawing conse- quences from such analyses, neoclassical economists advocated fi nancial liberalization. They contended that bureaucrats generally do not have the incentives or expertise to intervene effectively in credit allocation and pric- ing, and that a well-defi ned system of property rights, good contractual institutions, and competition would create the conditions for the emer- nancial system. They recommended that government gence of a sound fi exit from bank ownership and lift restrictions on the allocation of credit and the determination of interest rates (Caprio and Honohan 2001). While agreeing with the need to address the deleterious effects of fi nan- cial repression, the new structural economics would emphasize the fact that those distortions are often designed to protect nonviable fi rms in priority sectors in developing countries. It would then stress the importance of an appropriate sequencing of liberalization policies in domestic fi nance and foreign trade so as to achieve stability and dynamic growth simultaneously during transition. The new structural economics also posits that the optimal fi nancial structure at a given level of development may be determined by rms, and the usual the prevailing industrial structure, the average size of fi type of risk they face, all factors that are in turn endogenous to the econo- my’s factor endowments at that level. Observing that national policies fre- quently favor large banks and the equity market regardless of the structure of the economy, it would suggest that low-income countries choose small, local banks as the backbone of their fi nancial systems, instead of trying to replicate the fi countries. This nancial structure of advanced industrialized would allow small-scale fi rms in agriculture, industry, and the service sector to gain adequate fi nancial services. As industrial upgrading takes place and the economy relies increasingly on more capital-intensive industries, the fi nancial structure will change to give greater weight to large banks and sophisticated equity markets (Lin, Sun, and Jiang 2009).

47 34 | New Structural Economics Foreign Capital. In a world that they thought was characterized by the core-periphery relationship, old structural economists tended to view for- eign capital mainly as a tool in the hands of industrialized countries and rms to maintain harmful control over developing their multinational fi countries. They rejected the idea that free capital movements among coun- tries could deliver an effi cient allocation of resources and considered for- eign direct investment fl ows to poor countries as an instrument for foreign ownership and domination. They advocated tight restrictions on virtually all forms of international fi ows. nancial fl Neoclassical economic theory argues that international capital mobility serves several purposes: it allows countries with limited savings to attract fi nancing for productive domestic investment projects; it enables investors to diversify their portfolios; it spreads investment risk more broadly; and it promotes intertemporal trade—the trading of goods today for goods in the future (Eichengreen and others 1999). Therefore, the theory generally favors open or liberalized capital markets, with the expectation of more effi cient allocation of savings, increased possibilities for diversifi cation of investment risk, faster growth, and the dampening of business cycles. It should be noted, however, that some neoclassical economists also argue nancial markets in developing countries can be distorted that liberalized fi by incomplete information, large and volatile movements in and out of the system, and many other problems leading to suboptimal consequences that are damaging for general welfare. The new structural economics approach considers foreign direct investment to be a more favorable source of foreign capital for devel- oping countries than other capital fl ows because it is usually targeted toward industries consistent with a country’s comparative advantage. It is less prone to sudden reversals during panics than bank loans, debt fi nancing, and portfolio investment, and does not generate the same acute problems of fi nancial crises as do sharp reversals of debt and portfolio fl ows. In addition, direct investment generally brings technology, man- agement, access to markets, and social networking, which are often lack- ing in developing countries and are yet crucial for industrial upgrading. Thus, liberalizing inward direct investment should generally be an attrac- tive component of a broader development strategy. By contrast, portfolio investment that may move in and out quickly, in a large quantity, tends to

48 New Structural Economics: A Framework for Rethinking Development | 35 target speculative activities (mostly in equity markets or the housing sec- 12 uctuations. It should not be favored. The tor) and create bubbles and fl new structural economics approach may also shed new light on the puz- zle raised by Lucas (1990) about the fl ow of capital from capital scarce developing countries to capital abundant developed countries. With- out improvement of infrastructure and upgrading to new comparative advantage industries, the accumulation of capital in a developing country may encounter diminishing returns, causing lower returns to capital in developing countries, and justifying the subsequent outfl ow of capital to developed countries. Trade Policy. There have been various old structural economics approaches to external trade. But one constant feature is the belief that integration into the global economy is bound to maintain the existing world power struc- ture, with Western countries and their multinational corporations domi- nating poorer countries and exploiting their economies. In order to break the dependency trap, old structural economics thinkers have suggested that priority be given to import-substitution strategies, with developing econo- mies closed and protected until their modern industries can compete with advanced industrialized countries in world markets. A radically different view was adopted by economists in the 1980s. Observing that macroeconomic crises in developing countries almost always have an external dimension, they considered that their immediate cause was the lack of foreign exchange to service debts and purchase imports. They recommended trade liberalization and export promotion as a solu- tion to generate foreign exchange through export earnings. This was also consistent with the view that, in the long term, outward oriented develop- ment strategies are more effective than inward looking policies. This view was bolstered further by the argument that such a strategy would increase demand for unskilled labor and hence unskilled wages, as had happened in successful East Asian countries (Kanbur 2009). The analysis from the new structural economics would be consistent with the view from neoclassical economics that exports and imports are endogenous to the comparative advantage determined by a country’s endowment structure (they are essential features of the industrial upgrad- ing process and refl ect changes in comparative advantage). Globalization

49 36 | New Structural Economics offers a way for developing countries to exploit the advantages of backwardness and achieve a faster rate of innovation and structural possible for countries already on the global tech- transformation than is nology frontier. Openness is an essential channel for convergence. The new structural economics approach recognizes, however, that many developing countries start climbing the industrial ladder with the legacy of distortions from old structural economics strategies of import-substitution. It would therefore suggest a gradualist approach to trade liberalization. During transition, the state may consider providing some temporary protection to industries that are not consistent with a country’s comparative advantage, while liberalizing at the same time entry to other more competitive sectors that were controlled and repressed in the past. The dynamic growth in the newly liberalized sectors creates the conditions for reforming the old priority sectors. This pragmatic, dual-track approach may achieve the goal of growth without losers in the transition process (Naughton 1995; Lau, Qian, and Roland 2000; Subramanian and Roy 2003; Lin 2009a). Human Development. Old structural economics generally said little about the role of human development in economic growth. By contrast, neoclassical economics has shown that the continuing growth in per capita incomes of many countries during the nineteenth and twentieth centuries was mainly due to the expansion of scientifi c and technical knowledge that raised the productivity of labor and other inputs in pro- duction. Economic theory has demonstrated that growth is the result of synergies between new knowledge and human capital, which is why large increases in education and training have accompanied major advances in technological knowledge in all countries that have achieved signifi - cant economic growth. Education, training, and health, which are the most important investments in human capital, are considered to be the most important driving force for economic development (Becker 1975; Jones and Romer 2009). The new structural economics considers human capital to be one com- ponent of a country’s endowment. For economic agents, risks and uncer- tainty arise during the process of industrial upgrading and technological innovation that accompanies economic development. As various fi rms move up the industrial ladder to new, higher capital-intensity industries and get closer to the global industrial frontier, they face higher levels of risks.

50 New Structural Economics: A Framework for Rethinking Development | 37 Human capital increases workers’ ability to cope with risk and uncertainty (Schultz 1961) but its formation requires a long time. A person who loses the opportunity to receive education at a young age may not be able to compensate for that loss at a later age. In a dynamic growing economy, it is important to plan ahead and make human capital investments before the economy requires the set of skills associated with new industries and technologies. However, improvements in human capital should be com- mensurable with the accumulation of physical capital and the upgrading of industry in the economy. Otherwise, human capital will either become a binding constraint for economic development if it is under-supplied because of insuffi cient investment, or the country will have many frustrated highly educated youths if the industrial upgrading of the economy is not progress- ing fast enough to provide skilled jobs. A well-designed policy on human capital development should be an integral part of any country’s overall development strategy. The new struc- tural economics goes beyond the neoclassical generic prescription for edu- cation and suggests that development strategies include measures to invest in human capital that facilitates the upgrading of industries and prepares the economy to make full use of its resources. The key components of such strategies should follow Lucas’s (2002) suggestion to allow human capital to have both a quality and a quantity dimension. It should also include alternative policies for promoting skill formation that are targeted to dif- 13 with the government and the private sector ferent levels of the life cycle, working closely together to anticipate or respond to the skills needs in the 14 labor market. Singapore, one of the 13 high-growth economies that have been able to grow at more than 7 percent for periods of more than 25 years since World War II, provides a successful example of human capital devel- opment as a national strategy (Osman-Gani 2004), which goes beyond the schooling decision and recognizes that on-the-job training is an important component of aggregate human capital. Its human resource strategies have been continuously revised and adjusted in conjunction with other national strategic economic policies. Concluding Thoughts The new structural economics approach highlights the importance of endowments and differences in industrial structures at various levels of

51 38 | New Structural Economics development and the implications of distortions stemming from past, misguided, interventions by policymakers whose belief in old structural economics led them to over-estimate governments’ ability to correct market failures. It also points out the fact that policies advocated under the Washington Consensus often failed to take into consideration the structural differences between developed and developing countries and ignored the second-best nature of reforming various types of distortions in developing countries. The proposed new structural economics attempts to develop a general framework for understanding the causality behind the observed stylized facts of sustained growth. Specifi cally, the new structural economics pro- poses to: (i) develop an analytical framework that takes into account factor and infrastructure endowments, the levels of development, and the corresponding industrial, social, and economic structures of devel- oping countries; (ii) analyze the roles of the state and the market at each development level and the mechanics of the transition from one level to another; and (iii) focus on the causes of economic distortions and the government’s strategies for exit from the distortions. It is not an attempt to substitute another ideologically based policy framework for those that have dominated development thinking in past decades, yet showing little connection to the empirical realities of individual coun- tries. Rather, it is an approach that brings attention to the endowment structure and level of development of each country and suggests a path toward country-based research that is rigorous, innovative, and relevant to development policy. This framework stresses the need to understand better the implications of structural differences at various levels of a coun- try’s development—especially in terms of the appropriate institutions and policies, and the constraints and incentives for the private sector in the process of structural change. The current state of development economics and the severe impact of the global crisis on the economies of developing countries have gener- ated strong demand for a new framework for development thinking. The research agenda of the new structural economics should enrich research and enhance the understanding of the nature of economic development. This would help assist low- and middle-income countries in achieving dynamic, sustainable, and inclusive growth, and in eliminating poverty.

52 New Structural Economics: A Framework for Rethinking Development | 39 Notes The paper was presented as the Kuznets Lecture at the Economic Growth † Center, Yale University on March 1, 2011. The main arguments of this paper were fi rst presented at DEC’s fourth Lead Economists Meeting and at Lin’s fi rst anniversary at the Bank on June 2, 2009. A shorter version of the paper was presented at the conference on “Challenges and Strategies for Promoting Economic Growth,” organized by the Banco de México in Mexico City on October 19–20, 2009, and at public lectures at Cairo University on November 5, 2009, Korean Development Institute on November 17, 2009, OECD on December 8, 2009, UNU-WIDER on January 19, 2010, Stockholm Institute of Transitional Economics on January 21, 2010, National University of Manage- ment in Cambodia on September 8, 2010, Bank of Italy on April 26, 2011, and University of Dar es Salaam on April 29, 2011. Célestin Monga provided invaluable help in preparing this paper. The paper also benefi ts from discussions and comments from Gary Becker, Otaviano Canuto, Ha-Joon Chang, Luiz Pereira Da Silva, Augusto de la Torre, Christian Delvoie, Asli Demirgüç-Kunt, Shantayanan Devarajan, Hinh T. Dinh, Shah- rokh Fardoust, Ariel Fiszbein, Robert Fogel, Alan Gelb, Indermit S. Gill, Ann Harrison, James Heckman, Aart Kraay, Auguste Tano Kouame, Norman V. Loayza, Frank J. Lysy, Shiva S. Makki, William F. Maloney, Mustapha Kamel Nabli, Vikram Nehru, Howard Pack, Nadia Piffaretti, Mohammad Zia M. Qureshi, Martin Ravallion, Sergio Schmukler, Claudia Paz Sepúlveda, Luis Serven, and Harald Uhlig. I am also grateful for the editor and three referees for helpful comments and suggestions. I will refer to the early contributions by structuralist economists such as 1. Prebisch (1950) and Furtado (1964, 1970) and recent contributions by struc- turalist economists such as Taylor (1983, 1991, 2004) and Justman and Gurion (1991) as old structural economics. 2. The total endowments at a specifi c time—the economy’s total budgets at that time and the endowment structure, together with the households’ preferences and fi rms’ available production technologies—determine the relative factor and product prices in the economy. Total budgets and relative prices are two of the most fundamental parameters in economic analysis. Moreover, the endowments are given at any specifi c time and are changeable over time. These properties make endowments and the endowment structure the best starting point for analysis of economic development. Except in Heckscher- Ohlin trade theory, the economic profession has not given suffi cient attention to the implications of factor endowments and endowment structure. 3. The proposition that the industrial structure is endogenous to an economy’s endowment structure at each level of its development has been the subject

53 40 | New Structural Economics of extensive theoretical studies. For instance, Lin and Zhang (2009) develop an endogenous growth model that combines structural change with repeated product improvements to discuss the endogeneity of industrial structure, the appropriate technology, and economic growth in a less developed country (LDC) in a dynamic general-equilibrium framework. They use a two-sector model in which technological change in the traditional sector takes the form of horizontal innovation based on expanding variety as suggested in Romer (1990) while technological progress in the modern sector is accompanied by incessantly creating advanced capital-intensive industry to replace backward labor-intensive industry. This requires an intentional investment of resources by profi t-seeking fi rms or entrepreneurs (Grossman and Helpman 1994). The model shows that: (i) the optimal industrial structure in LDCs should not be the same as that in developed countries (DCs); (ii) the appropriate technol- ogy adopted in the modern sector in LDCs ought to be inside the technology frontier of the DCs; and (iii) a fi rm in an LDC that enters a capital-intensive, advanced industry (by DC standards) would be nonviable owing to the relative scarcity of capital in the LDC’s factor endowment. Ju, Lin, and Wang (2009) develop a dynamic general equilibrium model to show that industries will endogenously upgrade toward the more capital-intensive ones as the capital endowment becomes more abundant. The model features a continuous inverse- V-shaped pattern of industrial evolution driven by capital accumulation: As the capital endowment reaches a certain threshold, a new industry appears, prospers, then declines, and fi nally, disappears. While the industry is declining, a more capital-intensive industry appears and booms. Capital is mobile in an open economy. It is unlikely that the mobility of capital will equalize the cap- ital–labor ratio in high-income, capital-abundant countries and low-income, labor-abundant countries. This is because there are two main purposes for the capital to fl ow from a higher-income country to a lower-income country. The fi rst one is to exploit the lower-income country’s comparative advantage of abundant labor (or natural resources) so as to use the lower-income coun- try as its export base. For this purpose, the industry must be consistent with the recipient, lower-income country’s comparative advantage determined by its factor endowment, although the technology used by the foreign-invested fi rms may be somewhat more capital intensive than the indigenous fi rms. The second purpose of capital fl ow from a higher-income country is to get access to a lower-income country’s domestic markets. For this type of capital fl ow, the foreign-invested industries will be more capital intensive than the indigenous fi rms but only the types of production activities that are consistent with the host country’s comparative advantage, for example assembly of parts into fi nal products, will be located in the lower-income country. Therefore, the theoreti- cal insights derived from the assumption that the relative abundance of capital in a country is given at any specifi c time will hold even with capital mobility.

54 New Structural Economics: A Framework for Rethinking Development | 41 4. For nontradable goods and services, the nature of least-cost production tech- nology will also be endogenously determined by the endowment structure. That is, as capital becomes relatively abundant, the technology used to pro- duce nontradable goods and services will also become relatively capital inten- sive, just as happens in the tradable goods sector. For simplicity, the discussion in the paper will focus on the tradable sector. 5. Porter (1990) made the term “competitive advantage” popular. According to him, a nation will have competitive advantage in the global economy if the industries in the nation fulfi ll the following four conditions: (1) their indus- tries intensively use the nation’s abundant and relatively inexpensive factors of production; (2) their products have large domestic markets; (3) each industry forms a cluster; and (4) the domestic market for each industry is competitive. The fi rst condition in effect means that the industries should be the economy’s comparative advantage determined by the nation’s endowments. The third and the fourth conditions will hold only if the industries are consistent with the nation’s competitive advantage. Therefore, the four conditions can be reduced to two independent conditions: comparative advantage and domestic market size. Of these two independent conditions, comparative advantage is the most important because if an industry corresponds to the country’s comparative advantage, the industry’s product will have a global market. That is why many of the richest countries of the world are very small (Lin and Ren 2007). 6. Industries in advanced developed countries today are typically located on the global frontier and face uncertainty as to what the next frontier industries will be. This explains why government policy measures to support pioneer rms in such countries are usually in the form of general support to research fi in universities (which has externalities to private fi rms’ R&D), patents, preferential taxes for capital investments, mandates, defense contracts, and government procurement. Support in the form of preferential taxes, defense contracts, and government procurement are industry or product-specifi c. Government support to basic research also needs to be prioritized for certain types of potential industries or products because of budget constraints. How- ever, government support to pioneer fi rms in developing countries, especially low-income countries, often fails. One of the most important reasons is the rms in industries attempt by low-income countries governments to support fi that are inconsistent with the economy’s comparative advantages (Lin 2009a; Lin and Chang 2009). 7. Barro (2009) calls active fi scal policy of the Keynesian type “the extreme demand-side view” or the “new voodoo economics.” 8. Recent research suggests that economic returns on investment projects in devel- oping countries average 30–40 percent for telecommunications, more than 40 percent for electricity generation, and more than 200 percent for roads. In

55 42 | New Structural Economics Thailand, production loss due to power outages represented more than 50 percent of the total indirect costs of doing business in 2006. Firms often rely on their own generators to supplement the unreliable public electricity supply. In Pakistan, more than 60 percent of homes destroyed by fi re surveyed in 2002 owned a generator. The cost of maintaining a power generator is often high rms, which are impor- and burdensome, especially for small and medium-size fi tant sources of employment. Yet, while these costs must be privately borne, their benefi ts are felt across the economy. 9. This is percentage of cost (UNCTAD Statistical Database). The exploitation of natural resources can generate a large amount of revenues 10. but it is generally very capital intensive and provides limited job opportuni- ties. In a recent visit to Papua New Guinea, I observed that the Ok Tedi cop- per and gold mine in Tabubil generates almost 80 percent of the country’s export revenues and 40 percent of government revenues but provides only 2,000 jobs. A proposed liquefi ed natural gas project will double Papua New Guinea’s national income after its completion in 2012, but the project will only provide 8,000 jobs. The majority of Papua New Guinea’s 6.5 million population still live on subsistence agriculture. The contrast between the stan- dard of living of a few elite workers in modern mining and that of subsistence farmers is becoming a source of social tensions. A similar observation can be made about Botswana: the failure to diversify the economy from diamond mining and to generate employment opportunities may explain the widening disparity and deterioration of various human and social indicators, despite the diamond industry’s great success in sustaining Botswana’s growth miracle over the past 40 years. 11. Gerschenkron (1962) made a similar point, arguing that the private sector alone cannot effectively address the problems of access to fi nance in weak institutional environments. A sudden large infl ow of portfolio capital is most likely to be invested in specu- 12. lative sectors rather than in productive sectors. The reason is twofold: a large increase in investment in existing industries may encounter diminishing returns to capital, and the potential for quick and large industrial upgrading is limited by human capital, as well as soft and hard infrastructure constraints. Carneiro and Heckman (2003) have demonstrated the importance of both 13. cognitive and noncognitive skills that are formed early in life in accounting for gaps in schooling among social groups and other dimensions of socio- economic success. They have provided empirical evidence of a high return to early interventions and a low return to remedial or compensatory interven- tions later in life. 14. The list includes: Botswana; Brazil; China; Hong Kong SAR, China; Indone- sia; Japan; Korea; Malaysia; Malta; Oman; Singapore; Taiwan, China; and Thailand.

56 New Structural Economics: A Framework for Rethinking Development | 43 References Acemoglu, D., S. Johnson, and J.A. Robinson. 2001. “The Colonial Origins of Comparative Development: An Empirical Investigation.” American Economic Review 91: 1369–401. Aghion, P., and P. Howitt. 1992. “A Model of Growth through Creative Destruc- Econometrica 60(2): 323–51. tion.” Barro, R.J. 2009. “Government Spending Is No Free Lunch.” The Wall Street January 22. Journal, Becker, Gary S. 1975. Human Capital: A Theoretical and Empirical Analysis, 2nd ed. New York: Columbia University with Special Reference to Education. Press for NBER. Caprio, G., and P. Honohan. 2001. Finance for Growth: Policy Choices in a Vola- New York: World Bank and Oxford University Press. tile World. Latin America’s Economy. Cambridge, MA: Cardoso, E., and A. Helwege. 1995. MIT Press. Carneiro, P., and J.J. Heckman. 2003. “Human Capital Policy.” IZA Discussion Papers 821, Institute for the Study of Labor (IZA). Chang, P.K. 1949. Agriculture and Industrialization. Cambridge, MA: Harvard University Press. Domar, E. 1946. “Capital Expansion, Rate of Growth, and Employment.” Econometrica 14 (April): 137 – 47. Easterly, W. 2001. The Elusive Quest for Growth: Economists’ Adventures and Cambridge, MA: MIT Press. Misadventures in the Tropics. Easterly, W., N. Loayza, and P.J. Montiel. 1997. “Has Latin America’s Post- Reform Growth Been Disappointing?” World Bank Policy Research Paper 1708, World Bank, Washington, D.C., August. Eichengreen, B., M. Mussa, G. Dell’Ariccia, E. Detragiache, G.M. Milesi-Ferretti, and A. Tweedie. 1999. “Liberalizing Capital Movements: Some Analytical Issues.” Economic Issues no. 17. IMF, Washington, D.C. Francesco, G., and M. Pagano. 1991. “Can Severe Fiscal Contractions Be Expansionary?—Tales of Two Small European Countries.” In O.J. Blanchard and S. Fischer, NBER Macroeconomics Annual 1990. Cambridge, MA: MIT Press. Furtado, C. 1964. Development and Underdevelopment. Los Angeles: University of California Press. Economic Development of Latin America. London: Cambridge ———. 1970. University Press. Gerschenkron, A. 1962. Economic Backwardness in Historical Perspective: A Book of Essays. Cambridge, MA: Belknap Press of Harvard University Press. Glaeser, E., and A. Shleifer. 2002. “Legal Origins.” Quarterly Journal of Econom- ics 117 (November): 1193–229.

57 44 | New Structural Economics Grossman, G.M., and E. Helpman. 1994. “Endogenous Innovation in the Theory Journal of Economic Perspectives 8(1): 23–44. of Growth.” Growth Commission. 2008. “The Growth Report: Strategies for Sustained Growth and Inclusive Development.” Washington, D.C. Harrison, A., and A. Rodríguez-Clare. 2010. “Trade, Foreign Investment, and Handbook of Industrial Policy for Developing Countries.” In D. Rodrik (ed.), Economic Growth, Vol. 5. Amsterdam, The Netherlands: North-Holland, p. 4039–213. The Economic Journal Harrod, R.F. 1939. “An Essay in Dynamic Theory.” 49(193): 14–33. Hausmann, R., D. Rodrik, and A. Velasco. 2005. “Growth Diagnostics.” In J. Stiglitz and N. Serra (eds.), The Washington Consensus Reconsid- ered: Towards a New Global Governance. Oxford: Oxford University Press. Hirschman, A.O. 1958. New Haven, The Strategy of Economic Development. CT: Yale University Press. Introduction to Economic Growth. Jones, C.I. 1998. New York: W.W. Norton. Jones, C.I., and P.M. Romer. 2009. “The New Kaldor Facts: Ideas, Institutions, Population, and Human Capital.” NBER Working Paper Series 15094. Ju, J., J.Y. Lin, and Y. Wang. 2009. “Endowment Structures, Industrial Dynamics, and Economic Growth.” Policy Research Working Papers Series 5055, World Bank, Washington, D.C. Justman, M., and B. Gurion. 1991. “Structuralist Perspective on the Role of Tech- nology in Economic Growth and Development.” 19(9): World Development 1167–83. Kanbur, R. 2009. “The Crisis, Economic Development Thinking, and Protecting the Poor.” Presentation to the World Bank’s Executive Board, July. Krueger, A. 1974. “The Political Economy of Rent-Seeking Society.” American 64(3): 291–303. Economic Review Krugman, P. 1979. “A Model of Innovation, Technology Transfer, and the World Distribution of Income.” Journal of Political Economy 87(2): 253–66. ———. 1993. “Protection in Developing Countries.” In R. Dornbusch (ed.), Policymaking in the Open Economy: Concepts and Case Studies in Economic Performance. New York: Oxford University Press, 127–48. Kuznets, S. 1966. Modern Economic Growth: Rate, Structure and Spread. New Haven, CT: Yale University Press. Lau, L.J., Y. Qian, and G. Roland. 2000. “Reform without Losers: An Interpre- tation of China’s Dual-track Approach to Transition.” Journal of Political Economy 108(1): 120–43. Leibenstein, H. 1957. Economic Backwardness and Economic Growth: Studies in the Theory of Economic Development. New York: Wiley. Lewis, W.A. 1954. “Economic Development with Unlimited Supplies of Labor.” The Manchester School, May.

58 New Structural Economics: A Framework for Rethinking Development | 45 Lin, J.Y. 2009a. Economic Development and Transition: Thought, Strategy, and Cambridge: Cambridge University Press. Viability. ———. 2009b. “Beyond Keynesianism.” Harvard International Review 31(2): 14–17. Lin, J.Y., and H. Chang. 2009. “DPR Debate: Should Industrial Policy in Devel- oping Countries Conform to Comparative Advantage or Defy It?” Develop- 27(5): 483–502. (Reprinted as the debate in chapter II ment Policy Review of this volume.) Lin, J.Y., and F. Li. 2009. “Development Strategy, Viability, and Economic Distor- tions in Developing Countries.” Policy Research Working Paper 4906, World Bank, Washington, D.C., April. Lin, J.Y., and C. Monga. 2010. “The Growth Report and New Structural Eco- nomics.” Policy Research Working Papers Series 5336, World Bank, Washing- ton, D.C. (Reprinted in chapter II of this volume.) cation and Facilitation: The Role of ———. 2011. “DPR Debate: Growth Identifi Development Policy Review the State in the Dynamics of Structural Change.” 29(3): 259–310. (Reprinted in chapter III of this volume.) Lin, J.Y., and R. Ren. 2007. “East Asian Miracle Debate Revisited” (in Chinese). Jingji Yanjiu (Economic Research Journal) 42(8): 4–12. Lin, J.Y., and P. Zhang. 2009. “Industrial Structure, Appropriate Technology and Economic Growth in Less Developed Countries.” Policy Research Working Paper 4906, World Bank, Washington, D.C., April. Lin, J.Y., X. Sun, and Y. Jiang. 2009. “Towards a Theory of Optimal Finan- cial Structure.” Policy Research Working Papers Series 5038, World Bank, Washington, D.C. Lucas Jr., R.E. 1990. “Why Doesn’t Capital Flow from Rich to Poor Countries?” American Economic Review 80(2): 92–96. Lectures on Economic Growth. ———. 2002. Cambridge, MA: Harvard Univer- sity Press. Money and Capital in Economic Development. McKinnon, R.I. 1973. Washington, D.C: Brookings Institution. Monga, C. 1997. L’argent des autres—Banques et petites entreprises en Afrique: le cas du Cameroun. Paris: LDGJ-Montchretien. Myrdal, G. 1957. Economic Theory and Under-developed Regions. London: Duckworth. Growing Out of Plan: Chinese Economic Reform Naughton, B. 1995. 1978–1993. Cambridge: Cambridge University Press. Nurkse, R. 1953. Problems of Capital Formation in Underdeveloped Countries. New York: Oxford University Press. Osman-Gani, A.M. 2004. “Human Capital Development in Singapore: An Analysis of National Policy Perspectives.” Advances in Developing Human Resources 6(3): 276–87. Porter, M.E. 1990. The Competitive Advantage of Nations. New York: Free Press.

59 46 | New Structural Economics Prebisch, R. 1950. The Economic Development of Latin America and its Princi- New York: United Nations. Reprinted in Economic Bulletin for pal Problems. Latin America 7(1): 1–22. Ramsey, F.P. 1928. “A Mathematical Theory of Saving.” Economic Journal 38 (152): 543–59. Ravallion, M. 2009. “Evaluation in the Practice of Development.” The World 24(1): 29–53. Bank Research Observer Rodrik, D. 2004. “Industrial Policy for the Twenty-First Century.” Cambridge, MA. [http://ksghome. harvard.edu/~drodrik/unidosep.pdf]. Romer, P.M. 1987. “Growth Based on Increasing Returns Due to Specialization.” 77(2): 56–62. American Economic Review ———. 1990. “Endogenous Technological Change.” Journal of Political Econ- 98(5, Part 2): omy The Problem of Development: A Conference of the Insti- tute for the Study of Free Enterprise Systems, October, pp. S71–S102. Rosenstein-Rodan, P. 1943. “Problems of Industrialization of Eastern and South- Economic Journal 111(210–11, June–September): 202 –11. eastern Europe.” c Experience, Household Rosenzweig, M.R., and K.I. Wolpin. 1985. “Scientifi Structure and Intergenerational Transfers: Farm Family Land and Labor Arrangements in Developing Countries.” Quarterly Journal of Economics 100, Supplement. Rostow, W.W. 1990 [1960]. T he Stages of Economic Growth: A Non-Communist Manifesto, 3rd edn. New York: Cambridge University Press. American Economic Schultz, T.W. 1961. “Investments in Human Capital.” 51 (1): 1–17. Review Schumpeter, J., 1934. The Theory of Economic Development. Cambridge, MA: Harvard University Press. Shaw, E. 1973. New York: Financial Deepening in Economic Development. Oxford University Press. Singer, H. 1950. “The Distribution of Gains between Investing and Borrowing Countries.” American Economic Review 40(May): 473–85. Subramanian, A., and D. Roy. 2003. “Who Can Explain the Mauritian Miracle? Mede, Romer, Sachs, or Rodrik?” In. D. Rodrik (ed.), In Search of Prosperity: Analytic Narratives on Economic Growth. Princeton: Princeton University Press, 205–43. Taylor, L. 1983. Structuralist Macroeconomics: Applicable Models for the Third World. New York: Basic Books.. ———. 1991. Income Distribution, Infl ation and Growth: Lectures on Structur- alist Macroeconomic Theory. Cambridge, MA: MIT Press. ———. 2004. Reconstructing Macroeconomics: Structuralist Proposals and Cri- tiques of the Mainstream. Cambridge, MA: Harvard University Press. UNCTAD Statistical Database. http://www.unctad.org/templates/page .asp?intItemID=2364&lang=1.

60 New Structural Economics: A Framework for Rethinking Development | 47 Williamson, J. 1990. “What Washington Means by Policy Reform.” In J. Williamson (ed.), Latin American Adjustment: How Much Has Happened? Washington, D.C.: Institute for International Economics. ———. 2002. “Did the Washington Consensus Fail?” [http://www.peterson institute.org/publications/papers/paper.cfm?ResearchID=488]. World Bank. 2005. Economic Growth in the 1990s: Learning from a Decade of Reform. Washington, D.C. ———. 2010. “Research for Development: A World Bank Perspective on Future Directions for Research.” Policy Research Working Paper 5437, Washington, D.C.

61 I Comments *† Anne Krueger eld, there has been a search Ever since development economics became a fi for “the” key to development. Physical capital accumulation, human capital, industrial development, institutional quality, social capital, and a variety of other factors have been the focus at one time or another. As each became the focal point, there was a parallel explicit or implied role of government. If I understand Justin Lin correctly, he is saying that the “new struc- tural economics” (NSE) accepts that earlier thought ignored comparative advantage, which should be market determined, but that growth requires improvements in ‘hard’ (tangible) and ‘soft’ (intangible) infrastructure at each stage. Such upgrading and improvements require coordination and inhere with large externalities to fi rms’ transaction costs and returns to capital investment. Thus, in addition to an effective market mecha- nism, the government should play an active role in facilitating structural change (p. 28). He seems also to believe that growth depends almost entirely on industry growth and believes that constant “upgrading” or moving up the value added chain is the central challenge. He says that “the laissez- faire approach . . . missed the importance of the process of continuous, * Adapted from “Comments on “New Structural Economics’ by Justin Lin,” by Anne Krueger, originally published in The World Bank Research Observer (2011) 26 (2): 222–26, published by Oxford University Press on behalf of the International Bank for Reconstruction and Development/ The World Bank. © 2011 The International Bank for Reconstruction and Development / The World Bank. 48

62 Comments on “New Structural Economics: A Framework for Rethinking Development” | 49 fundamental technological changes and industrial upgrading, which distinguishes modern economic growth from premodern economic growth” (p. 15). It is questionable whether such changes and upgrading must take place early in the development process. In many countries, unskilled labor has moved to unskilled-labor-intensive industries, with expansion of those industries’ outputs for a period during which more and more workers acquired acquaintance with modern factory techniques, and exports of the unskilled-labor-intensive goods increased. Only later in the development process did upgrading become a major part of industrial growth once there cant absorption of rural labor, and much of it happened had been signifi rms in response to rising real wages, lower capital costs, and in existing fi learning through exposure to the international market. However, in most countries rural labor could be absorbed only as agri- cultural productivity increased; Lin’s NSE seems to equate growth with industrial expansion, ignoring the importance of increased productivity of the large fraction of the labor force (and of land) in rural areas. Fail- ure to invest in agricultural research and development and in rural health and education has been a major weakness of many countries’ develop- ment strategies. While strides have been made in reducing discrimination against agriculture, the NSE as exposited by Lin would appear to support the industrial and urban bias that has itself constituted a very large distor- tion in some countries. It will come as no surprise that I agree that the market should be used to determine comparative advantage, and that governments have respon- sibilities for insuring an appropriate incentive framework and provision of infrastructure (both hard and, as he terms it, “soft”). But there is nothing new in that. What purports to be the “new” part is the assertion that coordination and upgrading of infrastructure should in some way be related to particular industries. It is at this point where a question arises: most economists would accept the view that cost–benefi t analysis should be used in the choice of infrastructure projects. If “exter- nalities” and “coordination” are important, are they important for specifi c industries or for the entire industrial economy? If the former, how are those industries to be identifi ed, and how would the externalities be estimated in cost–benefi t analysis? Or would they? If infrastructure is seen to be industry-specifi c, it is not clear what it is. As with the possible existence

63 50 | New Structural Economics of infant industries, it is one thing to believe that there are such industries (perhaps) and quite another to identify ahead of time which they are. And even if such industries exist and are identifi ed, questions arise as to the incentives that would be appropriate for the government to foster these rm-specifi c treatment? Tariffs? Subsidies to industries. (Would they be fi fi rms or industries? Each has huge problems.) And if it is more “conven- tional,” what is new? If infrastructure is specifi c to industry (or a group of industries), the same questions must be addressed. Some hints are given as to what Lin has in mind: “successful industrial upgrading in responding to change in an economy’s endowment structure requires that the pioneer fi rms overcome issues of limited information regarding which new industries are the economy’s latent comparative advantages determined by the changing endowment structure. Valuable information externalities arise from the knowledge gained by pioneer rms in both success and failure. Therefore, in addition to playing a fi proactive role in the improvements of soft and hard infrastructures, the government in a developing county, like that in a developed country, needs to compensate for the information externalities generated by pio- neer fi rms” (p. 25). Here, the infant industry concerns arise again. How can these externali- ties be forecast? As Baldwin (1969) pointed out, there are major diffi culties cation of such externalities. with this argument, quite aside from the identifi And fi rms producing unskilled-labor-intensive goods and exporting them have usually learned of the opportunities provided by the international market and chosen to upgrade as their experience has increased. Learning does not seem to have been a major issue for fi rms in South Korea, Taiwan, and elsewhere. Another hint as to what Lin has in mind comes from his advocacy of coordination of infrastructure investments. According to him, “Change in infrastructure requires collective action or at least coordination between the provider of infrastructure services and industrial fi rms. For this reason, it falls to the government either to introduce such changes or to coordinate them proactively.”(p. 25) How this would be carried out is unclear; Lin insists that infrastructure must be upgraded with growth as long as it is consistent with the evolving future direction of comparative advantage, but does not elaborate on how that future direction should be identifi ed.

64 Comments on “New Structural Economics: A Framework for Rethinking Development” | 51 Involving individual fi rms and industries in decisions as to infrastructure rms’ investments would appear to offer far too much scope for individual fi uence over these investments. and industries’ infl Although it is certainly true that not everything can be done at once, focus on selected areas for large investments at the neglect of the rest of the economy is a highly questionable strategy. Why it would be preferable to allocate scarce capital so that some activities have excellent infrastruc- ture while others must manage with seriously defi cient infrastructure is not clear: without further evidence, it would appear to be a distortion. Further, questions can also be raised as to why “soft infrastructure,” such as the “business environment” (which consists of such things as the commercial code, the structure of taxes and subsidies, regulations, and so on), cannot be economy wide. And the criteria by which there would be designation of a given area, or the types of industries that would be eligible, as the recipi- ent of special treatment are not discussed. What the hard infrastructure is that does not consist of items such as roads and ports, and is industry specifi c, is not discussed. But all of this hinges on the proposition that decisionmakers in the pub- lic sector can ascertain the appropriate rate of “upgrading” and the extent of the supposed externalities. This raises a host of issues. There is, fi rst, the consideration that even if one could know which activities would have rms enter, comparative advantage, that advantage often develops as small fi some of which are successful and grow larger. Any strategy of “upgrad- ing” would inevitably favor larger, established fi rms, and hence encounter the same sorts of problems as did the older import-substitution strategy which, as Lin recognizes, failed. “Picking winners” as industries is diffi cult; it cannot be fi rm specifi c or the usual problems of corruption and cronyism arise. And yet supporting an industry or industries as an undifferentiated cult: are textiles an industry? Or is synthetic fi ber an industry? entity is diffi Or is nylon an industry? And, of course, the breakdown could go further. And as capital and skills per person accumulate, how is it to be decided where the industrial park or export processing zone should be? And which fi rms should be eligible to enter it? Another strand of Lin’s argument pertains to the role of distortions. He appears to be saying that countries that earlier adopted import-substitu- tion strategies have distorted industrial structures that should affect policy.

65 52 | New Structural Economics In particular, he says: “many developing countries start climbing the indus- trial ladder with the legacy of distortions from old structural strategies of import-substitution. [The new structural economics] would therefore sug- gest a gradualist approach to trade liberalization. During transition, the state may consider some temporary protection to industries that are not consistent with a country’s comparative advantage, while liberalizing at the same time entry to other more competitive sectors that were controlled and repressed in the past” (p. 36). Here, as elsewhere, little guidance is given as to how much protection industries would be provided with; how long that protection would last; how industries to be protected would be chosen; and so on. But even more important, one can imagine the political pressures for greater protection for longer periods. Protection of some industries is disprotection of oth- ers, as is well known, so reform efforts would clearly be dampened. Even worse, a major challenge for liberalizing reform is for it to be credible that the altered policies are not reversible. Lin’s prescription would greatly increase the challenge of creating credibility, and a slower transition would be a longer period during which growth was slow and political pressures opposing liberalization at all were mounting. In all, there is much in Lin’s analysis with which most would agree, but focus on governmentally led identifi cation of industries with “latent com- parative advantage” and industry-specifi c provision of infrastructure is not convincing. Lin calls for much research. A fi rst task should be to show that there are industry (or industry-cluster) externalities, how they could be iden- tifi ed and measured ex ante, and what sorts of government support would improve potential welfare and growth prospects without generating the same sorts of rent-seeking opportunities as import substitution policies did. Until that research is undertaken, the NSE will, it is to be feared, be taken as a license for governments to support specifi c industries (and worse yet, perhaps even fi rms), in ways that may be no more conducive to growth than were the old, failed, import-substitution policies. Note † Anne Krueger is Professor of International Economics in the School of Advanced International Studies (SAIS) at Johns Hopkins University and a Senior Fellow at the Stanford Center for International Development.

66 Comments on “New Structural Economics: A Framework for Rethinking Development” | 53 Reference Journal Baldwin, Robert E. 1969. “The Case against Infant Industry Protection.” of Political Economy 77(3): 295–305. *† Dani Rodrik Justin Lin wants to make structuralist economics respectable again, and I applaud him for that. He wants to marry structuralism with neoclassical economic reasoning, and I applaud this idea too. So he has two cheers from me. I withhold my third cheer so I can quibble with some of what he writes. The central insight of structuralism is that developing countries are qualitatively different from developed ones. They are not just radially shrunk versions of rich countries. In order to understand the challenges of under-development, you have to understand how the structure of employ- ment and production—in particular the large gaps between the social marginal products of labor in traditional versus modern activities—is determined and how the obstacles that block structural transformation can be overcome. The central insight of neoclassical economics is that people respond to incentives. We need to understand the incentives of, say, teachers to show up for work and impart valuable skills to their students or of entrepreneurs to invest in new economic activities if we are going to have useful things to say to governments about what they ought to do. (And of course, let’s not cials must have the incentive to do the economi- forget that government offi cally “correct” things, too.) If we put these two sets of ideas together, we can have a useful devel- opment economics, one that does not dismiss the tools of contemporary economic analysis and yet is sensitive to the specifi c circumstances of devel- oping economies. This is the kind of development economics that is appro- presume priately nuanced in its take on government intervention. It doesn’t * Adapted from “Comments on “New Structural Economics’ by Justin Lin,” by Dani Rodrik, originally published in The World Bank Research Observer (2011) 26 (2): 227–29, published by Oxford University Press on behalf of the International Bank for Reconstruction and Development/ The World Bank. © 2011 The International Bank for Reconstruction and Development / The World Bank.

67 54 | New Structural Economics omniscience or altruism on the part of governments. It has a healthy respect for the power and effectiveness of markets. But it does not blithely assume that development is an automatic process that takes care of itself as long as government stays out of the picture. So as Lin rightly emphasizes, the state has a useful role to play in pro- moting industrial diversifi cation and upgrading. He lists among desirable functions the provision of information about new industries, the coordina- tion of investments across fi rms and industries, the internalization of infor- mational externalities, and the incubation of new industries through the encouragement of foreign direct investment. Policies of this kind may be uous in advanced economies, but they are essential unnecessary or superfl if poor countries are to progress. To distinguish his brand of structuralist development economics from old-style structuralism, Lin writes that a key difference is that the old school advocated policies that go against an economy’s comparative advantage. The new approach, by contrast, “stresses the central role of the market . . . and advises the state to play a facilitating role to assist fi rms in the process of industrial upgrading by addressing externality and coordination issues.” Lin argues that government policies should “follow” comparative advan- tage, rather than “defy” it. Here is where I quibble with Lin’s argument. It seems to me that Lin wants to argue both for and against comparative advantage at the same time, and I cannot quite see how this can be done. If one believes that externality and coordination problems need to be addressed, as Lin appar- ently does, one must believe that such problems are preventing fi rms from investing appropriately. One must believe that markets are sending entre- preneurs the wrong signals—invest here, not there—and that allocating resources according to comparative advantage, as revealed by market prices, would be socially suboptimal. Comparative advantage has practical meaning for fi rms only insofar as it gets refl ected in prices. So when Lin asks governments to step in to address market failures and recommends the type of policies I have listed above—the coordination of investments, the incubation of new industries, etc.—he too is asking them to defy comparative advantage as revealed in market prices. In this respect, there is less difference between what the old school said and what the new school is saying.

68 Comments on “New Structural Economics: A Framework for Rethinking Development” | 55 Lin doesn’t want governments to employ “conventional” import sub- stitution strategies to build capital-intensive industries which “are not consistent with the country’s comparative advantage.” But isn’t building industries that defy comparative advantage what Japan and South Korea did, in their time? Isn’t it what China has been doing, and quite success- fully, for some time now? According to my calculations, the export bundle of China is that of a country between three and six times richer. If China, with its huge surplus of agricultural labor, were to specialize in the type of products that its factor endowments recommend, would it now be export- ing the advanced products that it is? Some people draw a distinction between static and dynamic compara- tive advantage in this context, but I don’t think that is the relevant dis- tinction. Market failures drive a wedge between market prices and social marginal valuations, and distort the relative costs that signal comparative advantage. Whether these distortions are introduced into intertemporal relative prices or today’s relative prices is largely secondary. The policies that Lin recommends are meant to offset such market distortions, and their intended effect is to induce fi rms to make choices that defy comparative advantage. I suspect that my difference with Lin is mainly methodological—and perhaps even just terminological—and may have little practical import. What Lin probably has in mind is that today’s industrial policies need to have a softer touch than that which structuralists of old tended to recom- mend. They must be more respectful of markets and incentives; they must show greater awareness of the potential of government failures; and they must focus specifi cally on market failures rather than vague shortcomings of the private sector. I would agree with all this. But a deeper question relates to the policy implications one draws from all this. In principle, market failures need to be addressed with appropri- ately targeted policies. So if the problem is one of information spillovers, the fi rst-best is to subsidize the information generating process. If the problem is lack of coordination, the fi rst-best is for the government to bring the parties together and coordinate their investments. In practice, though, the relevant market failures cannot be always closely identifi ed and the directly targeted remedies may not be available. The practical reality is that the type of policies structuralism calls for—whether of the

69 56 | New Structural Economics traditional or the contemporary type—have to be applied in a second- best setting. And in such a setting, nothing is all that straightforward anymore. Presumably this is the reason why Lin recommends, for example, a gradual approach to trade liberalization. Such an approach is, at best, a second-best remedy to some loosely specifi ed market failures that either cannot be precisely identifi ed ex ante or cannot be fully treated with fi rst- best Pigovian interventions. But how different is this from the old struc- turalist approach? Didn’t most structuralists also view protection as a temporary expedient, to be done away with once the requisite industrial capabilities were built? To repeat, my differences with Justin Lin are second order, and they are swamped by our areas of agreement. My quibbles are a bit like the internal doctrinal debates waged among communists—does the revolution require the intensifi cation of the class struggle, or can that stage be skipped?—when much of the rest of the world is on a different wavelength altogether. As a fellow traveler, I am greatly encouraged by what Justin Lin is try- ing to do. It is high time that the common sense exhibited in his approach reclaimed its mantle in development economics. Note † Dani Rodrik is professor of international political economy at Harvard University’s John F. Kennedy School of Government. *† Joseph E. Stiglitz Twelve years ago, when I was chief economist of the World Bank, I sug- gested that the major challenge to development economics was learning the lessons of the previous several decades: a small group of countries, mostly in Asia, but a few in other regions, had had phenomenal success, beyond anything that had been anticipated by economists; while many other countries had experienced slow growth, or even worse, stagnation * Adapted from “Rethinking Development Economics,” by Joseph Stiglitz, originally published in The World Bank Research Observer (2011) 26 (2): 230–36, published by Oxford University Press on behalf of the International Bank for Reconstruction and Development / The World Bank. © 2011 The International Bank for Reconstruction and Development / The World Bank.

70 Comments on “New Structural Economics: A Framework for Rethinking Development” | 57 and decline—inconsistent with the standard models in economics which predicted convergence. The successful countries had followed policies that were markedly different from those of the Washington Consensus, though they shared some elements in common; those policies had not brought high growth, stability, or poverty reduction. Shortly after I left the World Bank, the crisis in Argentina—which had been held up as the poster child of the country that had followed Washington Consensus policies—reinforced the doubts about that strategy. The global fi nancial crisis, too, has cast doubt over the neoclassical paradigm in advanced industrial countries, and rightly so. Much of devel- opment economics had been viewed as asking how developing countries could successfully transition toward the kinds of market-oriented policy frameworks that came to be called “American style capitalism.” The debate was not about the goal, but the path to that goal, with some advocating “shock therapy,” while others focused on pacing and sequencing—a more gradualist tack. The global fi nancial crisis has now raised questions about that model even for developed countries. In this short essay, I want to argue that the long-term experiences in growth and stability of both developed and less developed countries, as well as the deeper theoretical understanding of the strengths and limitations of market economies, provide support for a “new structural” approach to development—an approach similar in some ways to that advocated by Justin Lin in his paper, but markedly different in others. This approach sees the limitations of markets as being greater than he suggests—even well func- tioning market economies are, on their own, neither effi cient nor stable. The only period in the history of modern capitalism when there has not been repeated fi nancial crises was the short period after the Great Depression when the major countries around the world adopted, and enforced, strong fi nancial regulations. Interestingly this was also a period of rapid growth and a period in which the fruits of that growth were widely shared. But government not only has a restraining role; it has a constructive and catalytic role—in promoting entrepreneurship, providing the social and physical infrastructure, ensuring access to education and fi nance, and supporting technology and innovation. The perspective that I am putting forward differs not only in its view of the effi ciency and stability of unfettered markets, but also in what it

71 58 | New Structural Economics sees as the primary driver of economic growth. Since Solow’s pioneering work more than a half-century ago (Solow 1957), it has been recognized that the major source of increases in per capita income are advances in 1 technology. The argument that improvements in knowledge are a primary source of growth is even more compelling for developing countries. As the World Development Report for 1998–99 emphasized, what separates developing and developed countries is not just a gap in resources, but a disparity in knowledge. There are well understood limits to the pace with which coun- tries can accumulate capital, but the limitations on the speed with which the gap in knowledge can be closed are less clear. But the view that creating a learning society, focusing on absorbing and adapting, and eventually producing knowledge, provides markedly differ- ent perspectives on development strategies than those provided by the neo- classical model. That model centered attention on increasing capital and the effi cient allocation of resources. Since the appropriate sectoral structure of the economy naturally depends on the resource endowment, there will be a natural evolution of the economy’s structure over time. Markets allo- cate resources effi ciently, enabling the structure to change as the (endog- enous) endowments change. A government’s main role, in this view, is not to put impediments in the market. The standard market failures approach criticized these conclusions by focusing on a variety of market imperfections: For instance, imperfections in capital markets meant that fi nance was often not available for new enterprises that were required as part of this sectoral adjustment. Indi- viduals on their own couldn’t fi nance their education. There are pervasive externalities—not only environmental externalities but also those associ- ated with systemic risk, so evidenced in the current crisis. Research over the past 20 years has explored the consequences of market failures like imperfect capital markets, traced these imperfections back to problems of imperfect and asymmetric information, and proposed a set of remedies, which in some countries, in some periods, have worked remarkably well. Good fi nancial regulations in countries like India protected them against the ravages of the global fi nancial crisis. But the perspective of the “learning society”—or, as Greenwald and I call it, the “infant economy”—adds a new dimension to the analysis

72 Comments on “New Structural Economics: A Framework for Rethinking Development” | 59 (Greenwald and Stiglitz 2006). Knowledge is different from an ordinary commodity. The accumulation of knowledge is inherently associated with externalities—knowledge spillovers. Knowledge itself is a public good. If the accumulation, absorption, adaptation, production, and transfer of knowledge are at the center of successful development, then there is no presumption that markets, on their own, will lead to successful outcomes. Indeed there is a presumption that they will not. The “new structuralist approach” advocated by Justin Lin is perfectly aligned with this perspective. Lin provides guidance as to how governments should direct the economy; he emphasizes that they should strive to shape the economy in a way that is consistent with its comparative advantage. The problem is that some of the most important elements of comparative advantage are endogenous. Switzerland’s comparative advantage in watch- making has little to do with its geography. Standard Heckscher-Ohlin theory (emphasizing that trade in goods was a substitute for movement in factors) was formulated in a period before globalization allowed the kinds of fl ows of capital that occur today. With fully mobile capital, outside of agriculture, natural resource endowments need not provide the basis for explaining patterns of production and 2 specialization. In short, there is no reason for countries to need to limit ned. themselves to patterns dictated by endowments, as conventionally defi More important is the “endowment” of knowledge and entrepreneurship. A major focus of policy should be on how to enhance and shape those endowments. Even if a government would like to avoid addressing these issues, it cannot; for what the government does (or does not do) has consequences, positive and negative, for the development of the “learning society.” This is obviously so for investments in infrastructure, technology, and education; but also for fi nancial, trade, intellectual property rights and competition policies. At the center of creating a learning society is the identifying of sectors that are more amenable to learning, with benefi ts not captured by fi rms themselves, so that there will be underinvestment in learning. Elsewhere Greenwald and I have argued that an implication of this is the encour- agement of the industrial sector, which typically has large spillovers. This approach provides an interpretation of the success of Asia’s export-led

73 60 | New Structural Economics growth. Had Korea allowed market forces on their own to prevail, it would ciency not have embarked on its amazing development successes. Static effi entailed that Korea produced rice; indeed the country might today have been among the most effi cient rice farmers—but it would still be a poor country. As Arrow pointed out (1962), one learns by doing (and one learns how to learn by learning [Stiglitz 1987]). This discussion highlights the fundamental difference with neoclassical approaches emphasizing short-run effi ciency. The fundamental trade-offs ciency should be familiar from the debate between static and dynamic effi over patent laws. 3 A major concern with these industrial policies concerns implementa- tion—do developing countries have the requisite capacities? We need to put this question in context. There is probably no country that has grown successfully without an important role, not just in restraining and creating markets, but also in promoting such industrial policies, from the countries of East Asia today to the advanced industrial countries, not just during their developmental stages, but even today. The task is to adopt policies and practices—to create institutions like an effective civil service—that enhance the quality of the public sector. The successful countries did so. Policies that either intentionally or unintentionally weaken the state are not likely to do so. Economic policies have to refl ect the capacity of the state to implement them. One of the arguments in favor of exchange rate policies that encour- age export industries is that they are broad based: the government does not have to pick particular “strategic” sectors to support. As always, there are trade-offs: effi ciency might be enhanced if the sectors with the largest externalities could be targeted. There are other broad-based policies, such as a development-oriented intellectual-property regime, and investment and fi nancial policies that encourage transfer of technology and the promotion of local entrepreneur- ship, that can help promote a learning and innovation society (Hausmann and Rodrik 2003; Stiglitz 2004; Emran and Stiglitz 2009; Hoff 2010). Some forms of fi nancial and capital market liberalization may be counter- productive. Interventions will never be perfect, nor need they be to effect an improve- 4 ment in economic performance. The choice is not between an imperfect

74 Comments on “New Structural Economics: A Framework for Rethinking Development” | 61 government and a perfect market. It is between imperfect governments and imperfect markets, each of which has to serve as a check on the other; they need to be seen as complementary, and we need to seek a balance between the two—a balance which is not just a matter of assigning certain tasks to one, and others to the other, but rather designing systems where they interact effectively. While I have been discussing the economics of development, that sub- ject cannot be separated from broader aspects of societal transformation (Stiglitz 1998), as Hirschman emphasized in his writings (1958, 1982). Race and caste are social constructs that effectively inhibit the human development of large parts of the population in many parts of the world. The study of how these constructs get formed, and how they change, is thus a central part of developmental studies (Hoff and Stiglitz 2010). In this article, I have emphasized the creation of a learning society. The eco- nomics of doing so entails policies that change sectoral composition. But at the root of success is the education system and how it inculcates attitudes toward change and skills of learning. Other policies (for example legal sys- tems, gender-based microcredit schemes, affi rmative action programs) can also play an important role. Before concluding, I want to make two further remarks. The fi rst concerns the relationship between growth and poverty reduction. While growth may be necessary for sustained poverty reduction, it is not suffi cient. Not all development policies are pro-poor; some are anti- nancial and capital market liberalization have, at least poor. Policies like fi in some countries, contributed to greater instability, and a consequence of 5 Contractionary monetary and fi scal poli- that instability is more poverty. cies in response to crises exacerbate the downturns, leading to higher unem- ployment and a higher incidence of poverty. Policies to promote a learning economy too can either be pro- or anti-poor, but the most successful poli- cies will necessarily be broad-based, engendering a transformation of the learning capacities of all citizens, and will therefore be pro-poor. The second comment relates to the broader objectives of development, which should be sustainable improvements in the well-being of the citizens 6 of the country, and the metrics we use to assess success. Our metrics don’t typically capture the increase in the wealth of a country that is a result of the learning strategies advocated here. It is only gradually, over time, that the benefi ts are realized and recognized.

75 62 | New Structural Economics The aftermath of the global fi nancial crisis should be an exciting time for economists, including development economists, since it dramatically revealed fl aws in the reigning paradigm. This paradigm has had enormous infl uence in development economics, though that infl uence was already waning, because its prescriptions had failed. Fortunately there are alter- native frameworks already available—a plethora of ideas that should provide the basis for new understandings of why a few countries have succeeded so well and some have failed so miserably. Out of this under- standing, perhaps we will be able to mold new policy frameworks that will provide the basis of a new era of growth—growth that will both be sustainable and enhance the well-being of most citizens in the poorest countries of the world. Notes † Joseph Stiglitz is a professor of fi nance and business at Columbia University and chair of the university’s Committee on Global Thought. This article was originally a paper prepared for a World Bank symposium, based on Justin Yifu Lin’s paper, “New Structural Economics” (chapter I in this book). The perspec- tive taken here is based on joint work with Bruce Greenwald (2006; forthcom- ing). Stiglitz is indebted to Eamon Kirchen-Allen for research assistance. 1. Even before Solow, Schumpeter had argued that the strength of a market economy resided in its ability to promote innovation and invention; and, shortly after Solow’s work, there developed a large literature on endogenous growth, associated with names like Arrow, Shell, Nordhaus, Atkinson, Das- gupta, Uzawa, Kennedy, Fellner, and Stiglitz, followed on in the 1980s and 1990s by the work of Romer. (See for example Atkinson and Stiglitz, 1969; Dasgupta and Stiglitz, 1980a and 1980b; Fellner, 1961; Kennedy, 1964; Nor- dhaus, 1969a and 1969b, Romer, 1994; Shell, 1966 and 1967; Uzawa, 1965.) The earlier work on endogenous (sometimes referred to as induced) innovation addressed not only the rate of innovation but its direction. For a discussion of more recent contributions in this line of research, see Stiglitz (2006). 2. Indeed, the work of Krugman has emphasized that today most trade is not related in fact to differences in factor endowments. 3. I use the term broadly to embrace any policy attempting to affect the direction of the economy. 4. Indeed, if all projects were successful, it suggests that the government is under- taking too little risk. 5. As I have also noted, such policies may have an adverse effect in enhancing domestic learning capacities.

76 Comments on “New Structural Economics: A Framework for Rethinking Development” | 63 6. The International Commission on the Measurement of Economic Performance ect either sustain- and Social Progress emphasized the failures of GDP to refl ability or well-being (Fitoussi, Sen, and Stiglitz 2010). GDP per capita does not say anything about how well most citizens are doing; it can be going up even though most citizens’ incomes are declining (as has been happening in the United States). GDP focuses on production in the country, not on incomes earned by those in the country, and takes no account of environmental degrada- tion or resource depletion, or, more broadly, of sustainability. The United States and Argentina both provide examples of countries whose growth appeared to be good—but both were based on unsustainable debts, used to fi nance con- sumption booms, not investment. References Arrow, Kenneth J. 1962. “The Economic Implications of Learning by Doing.” 29:155 –73. Review of Economic Studies Atkinson, A.B., and J.E. Stiglitz. 1969. “A New View of Technological Change.” The Economic Journal 79(315): 573–8. Dasgupta, P., and J.E. Stiglitz. 1980a. “Industrial Structure and the Nature of The Economic Journal Innovative Activity.” 90(358): 266–93. ———. 1980b. “Uncertainty, Market Structure and the Speed of R&D.” Bell Journal of Economics 11(1): 1–28. Emran, S., and J.E. Stiglitz. 2009. “Financial Liberalization, Financial Restraint, and Entrepreneurial Development.” Working paper, Institute for International Economic Policy Working Paper Series Elliott School of International Affairs The George Washington University, January (www2.gsb.columbia.edu/faculty/ jstiglitz/download/papers/2009_Financial_Liberalization.pdf). The Fellner, W. 1961. “Two Propositions in the Theory of Induced Innovations.” Economic Journal 71(282): 305–8. Fitoussi, J., A. Sen, and J.E. Stiglitz. 2010. Mismeasuring Our Lives: Why GDP Doesn’t Add Up. New York: The New Press. (The Report of the Commis- sion in the Measurement of Economic Performance and Social Progress, also known as the Sarkhozy Commission.) Greenwald, B., and J.E. Stiglitz. 2006. “Helping Infant Economies Grow: Foun- dations of Trade Policies for Developing Countries.” American Economic Review: AEA Papers and Proceedings 96(2): 141–6. ———. Forthcoming. Creating a Learning Society: A New Paradigm For Devel- opment and Social Progress. New York: Columbia University Press. Hausmann, R., and D. Rodrik. 2003. “Economic Development as Self-Discovery.” Journal of Development Economics 72(2): 603–33. Hirschman, A.O. 1958. The Strategy of Economic Development. New Haven, CT: Yale University Press.

77 64 | New Structural Economics ———. 1982. “The Rise and Decline of Development Economics.” Gersovitz, and W.A. Lewis, eds., The Theory and Experience of In M. Economic Development. London: Allen and Unwin: 372–90. Hoff, K. 2010. “Dysfunctional Finance: Positive Shocks and Negative Outcomes.” Policy Research Working Paper 5183, The World Bank Development Research Group Macroeconomics and Growth Team, January. Hoff, K., and J.E. Stiglitz. 2010. “Equilibrium Fictions: A Cognitive Approach American Economic Review 100(2): 141–6. to Societal Rigidity.” Kennedy, C. 1964. “Induced Bias in Innovation and the Theory of Distribution.” 74(295): 541–7. Economic Journal Nordhaus, W.D. 1969a. “An Economic Theory of Technological Change.” American Economic Association Papers and Proceedings 59: 18–28. ———. 1969b. Invention, Growth and Welfare: A Theoretical Treatment Cambridge, MA: MIT Press. of Technological Change, The Journal of Romer, P. 1994. “The Origins of Endogenous Growth.” Economic Perspectives 8(1): 3–22. Shell, K. 1966. “Toward a Theory of Inventive Activity and Capital Accumulation.” American Economic Association Papers and Proceedings 56: 62–8. Essays on the Theory of Optimal Economic Growth. ———. ed. 1967, Cambridge, MA: MIT Press. Solow, Robert M. 1957. “Technical Change and the Aggregate Production Review of Economics and Statistics Function.” 39(3): 312–20. Stiglitz, J.E. 1987. “Learning to Learn, Localized Learning and Technological Progress.” In P. Dasgupta and P. Stoneman, eds., Economic Policy and Technological Performance. Cambridge, New York: Cambridge University Press: 125–53. ———. 1998. “Towards a New Paradigm for Development: Strategies, Policies and Processes.” The 9th Raul Prebisch Lecture delivered at the Palais des Nations, Geneva, October 19, UNCTAD. Also Chapter 2 in Ha-Joon Chang, ed., The Rebel Within. London: Wimbledon Publishing Company, 2001: 57–93. ———. 2004. “Towards a Pro-Development and Balanced Intellectual Prop- erty Regime.” Keynote address presented at the Ministerial Conference on Intellectual Property for Least Developed Countries, World Intellectual Property Organization, Seoul, October 25. http://www2.gsb.columbia .edu/faculty/jstiglitz/download/2004_TOWARDS_A_PRO_DEVELOPMENT. htm.

78 Comments on “New Structural Economics: A Framework for Rethinking Development” | 65 ———. 2006. “Samuelson and the Factor Bias of Technological Change.” In M. Szenberg, L. Ramrattan, and A.A. Gottesman, eds., Samuelsonian Economics and the Twenty-First Century. New York: Oxford University Press: 235–51. Uzawa, H. 1965. “Optimum Technical Change in an Aggregate Model of Eco- nomic Growth.” International Economic Review 6(1): 18–31.

79 I I Rejoinder Development Thinking 3.0: The Road Ahead Development economics appeared after World War II with the intention of helping developing countries industrialize their economies, reduce pov- erty, and narrow their income gaps with advanced countries. However, the developing countries that followed its recommendations to formulate their development policies failed to achieve the intended goals. In a paper on New Structural Economics: A Framework for Rethinking Development published in the most recent issue of the World Bank Research Observer [and reprinted as chapter I in this book], I took up the challenge of syn- thesizing half a century of various approaches proposed by development economics, and suggested a way forward. I am very fortunate and honored that my paper was critically discussed in the same issue of the journal by Anne Krueger, Dani Rodrik, and Joseph Stiglitz, who are among the best minds and most respected experts in the profession—two of them happen to be my predecessors as Chief Economist at the World Bank. I basically argue that early researchers who launched development eco- nomics as a sub-discipline of modern economics focused on market fail- ures and advocated old structuralist, state-led development policies. These policies did not properly account for comparative advantage and failed to create competitive industries. In reaction, a second wave of develop- ment thinking inspired by neo-liberalism focused on government failures and recommended Washington Consensus–type policies that also failed to deliver sustainable, inclusive growth and poverty reduction in developing countries. 66

80 Rejoinder to Comments on Chapter I | 67 Looking at the economic history of all successful economies since the Industrial Revolution, I have suggested a general framework for engag- ing in a third phase of development thinking that focuses on structural change, driven by changes in endowment structure and comparative advantages. That framework, encapsulated in the idea of New Structural Economics (NSE), would help the state play a proactive facilitating role in structural transformation. It would also require policy makers to be more disciplined in designing and implementing strategies around the function of the market. As could be expected, Anne Krueger, Joseph Stiglitz, Dani Rodrik, and I all agree on the importance and need to reignite the debate on development nancial and economic recipes—especially in light of the current global fi crisis. But we also have subtle and important differences of ideas on the true lessons from economic history and economic theory. My biggest divergence of views is with Anne Krueger, who questions whether the fundamental technological changes and industrial upgrading— which I consider to be at the heart of, and crucial to, the mechanics of growth—must take place early in the economic development. She contends that “only later in the development process did upgrading become a major part of industrial growth once there had been signifi cant absorption of rural labor, and much of it happened in existing fi rms in response to rising real wages, lower capital costs, and learning through exposure to the inter- national market” (p. 49). My view of economic development is slightly dif- ferent: The migration of unskilled rural labor to unskilled labor-intensive industries is a form of structural change that may not occur spontaneously. I believe that proactive action must be taken by policy makers to manage the demand for labor: Indeed it is necessary for the government to facilitate the growth of existing and emerging unskilled labor-intensive industries along the line argued in the NSE. Without such action, many rural out-migrants will be unemployed as has been the case in Africa, Latin America and many other developing countries. On the supply side of the labor market, the government also needs to provide basic education and training to enhance the rural out-migrants’ ability to adapt to the new working environment and requirements in the industrial sector. Moreover, successful catching-up countries may start their upgrading process long before their rural surplus labor is exhausted. One example is China. With 39.1 percent of China’s

81 68 | New Structural Economics labor force still working in the primary sector in 2009, that issue is hotly debated in academic circles: some economists wonder whether China has reached the so-called Lewis turning point and depleted its labor surplus. However, the quick and continuous upgrading of China’s industries is still going on, exemplifi ed by the quality and varieties of China’s exports to the U.S. market. A similar situation was observed in 1980 when 34 percent of Korea’s labor force was still in agriculture. However, Korea had already entered not only industries such as consumer electronics, but also ship- building, automaking, and memory chips by that time. I fully agree with Anne that agricultural productivity needs to be improved in parallel with industrialization (p. 48). But again, in order to improve agricultural productivity and increase farm income, the govern- ment must play a proactive role in making new agricultural technology available, providing extension services, improving irrigation, and expand- ing market channels. The government also needs to create conditions to facilitate the diversifi cation of agriculture into new, higher-value-added cash crops. While Anne agrees that “the market should be used to determine com- parative advantage, and that governments have responsibilities for insuring an appropriate incentive framework and provision of infrastructure, both hard and soft,” she objects, most notably, to government intervention aim- ing at fostering the development of specifi c industries and doubts “why it would be preferable to allocate scarce capital so that some activities have excellent infrastructure while others must manage with seriously defi cient infrastructure” (p. 51). In fact, identifi cation of new industries and prioritization of govern- ment’s limited resources to facilitate the development of those industries are both essential for successful growth strategies in developing countries. Why? Because the infrastructure improvements required are often indus- try specifi c. One simply has to look at the list of recent success stories in African countries to understand the necessity for identifi cation: tex- tiles in Mauritius, apparel in Lesotho, cotton in Burkina Faso, cut fl owers in Ethiopia, mango in Mali and gorilla tourism in Rwanda all required that governments provide different types of infrastructure. The refrigera- tion facilities needed at the airport and regular fl ights to ship Ethiopia’s cut fl owers to the auctions in Europe are obviously quite different from

82 Rejoinder to Comments on Chapter I | 69 the improvements required at the port facilities for textile exports in Mauritius. Similarly, the type of infrastructure needed for the garment industry in Lesotho is distinct from the one needed for mango production and export in Mali or for attracting gorilla tourism in Rwanda. Because scal resources and implementation capacity are limited, the government fi in each of those countries had to prioritize and decide which particular infrastructure they should improve or where to optimally locate the public services to make those success stories happen. Deng Xiaoping explained that pragmatic wisdom at the beginning of China’s transition to a market economy when he advocated allowing a few regions and people to get rich fi rst so as to achieve common prosperity for all people in the nation. The dynamic growth in those regions and industries would increase fi scal rev- enues, giving the government more resources to improve infrastructure for other regions in the nation later. Identifi cation of new sectors or lines of business and prioritization of infrastructure investment are also necessary because to be competitive in the globalized world, a new industry not only must align with the coun- try’s comparative advantage so that its factor costs of production can be at the lowest possible level but it also must have the lowest possible transaction-related costs. Why? Suppose a country’s infrastructure and business environment are good and industrial upgrading and diversifi cation happen spontaneously. Without government coordination, fi rms may enter into too many different industries that are all consistent with the country’s comparative advantage. As a result, most industries may not form clusters that are suffi ciently large and will not be competitive in the domestic and international markets. A few clusters may emerge eventually after many failures. Such a “trial and error” process is likely to be long and costly, reducing the individual fi rms’ expected returns and incentives to upgrade or diversify to new industries, and slowing down the country’s economic development. It is therefore imperative for a facilitating state in a developing country to identify and select new industries that are consistent with com- parative advantage, use its limited resources to improve infrastructure for a limited number of carefully selected industries, provide adequate incentives for fi rst movers, and coordinate private fi rms’ related investments in those industries so that clusters can be formed successfully and quickly. Whether the government plays the identifi cation and facilitation role may explain

83 70 | New Structural Economics why some developing countries can grow at 8 percent or more for several decades while most others fail to have a similar performance. I agree with Anne that cost-benefi t analysis is indeed an excellent tool that should be used for evaluating the potential merits of every single project (p. 49). Such an analysis sheds light on the valid- infrastructure ity of competing alternatives and can help make better public investment decisions. It forces policy makers to provide quantitative data to back up qualitative arguments and is therefore an invaluable technique for increas- ing social welfare. But it is microeconomic by nature. Without the iden- cation of potentially successful industries and their likely location and tifi needed infrastructure, policy makers are confronted with too many pos- sible feasible projects that all need careful cost-benefi t analysis. Moreover, ts and costs that for every public investment project, there are many benefi are intangible and therefore diffi cult to value. It is also well known that the results of that analysis can be very sensitive to the choice of the discount rate, and that the information used to determine future benefi ts and costs is limited by current knowledge. In her discussion of infant industries, Anne observes that fi rms that pro- duce and export unskilled-labor-intensive goods have usually learned from the opportunities arising from the dynamics of international market. She rms notes that “learning does not seem to have been a major issue for fi in South Korea, Taiwan, and elsewhere” (p. 50). Learning may not be an issue if it is a by-product of the fi rms’ business activities, but if it is not such rms may not have the incentives a spontaneous element of their activities, fi to invest in it. A low-income country should have comparative advantage in the production of many unskilled, labor-intensive manufacturing prod- ucts that it still imports. Such product market information should be avail- able freely to any entrepreneur in the country. But production information about where to buy the equipment and intermediate inputs to manufacture rm to these imported products, and knowledge about how to operate a fi produce them, are relatively costly to obtain for most entrepreneurs in low- income countries. Furthermore, the coordination of related investment in infrastructure, access to fi nance for investment and operation, or the avail- ability of foreign exchange for importing equipment for developing the new industry may still be serious issues for private fi rms even if learning about product market and production information is not a problem.

84 Rejoinder to Comments on Chapter I | 71 Anne’s skepticism of all industry-specifi c interventions—a skepticism widely shared in the mainstream economic profession and Washington- based development institutions—results from the pervasive failures of government’s attempts to pick winners in the past. Those failures were mostly due to the misguided attempts by many governments to develop industries that were inconsistent with their countries’ comparative advan- tages. Firms in those industries were not viable in open, competitive mar- kets, and their investment and survival depended on heavy government protection, large subsidies, and direct resource allocations through mea- sures such as monopoly rent, high tariffs, quota restrictions, and subsi- dized credits. The large rents embedded in those measures created many distortions and easily became the targets of political capture. All this cre- ated diffi cult governance problems. The likelihood of these problems aris- ing is much reduced when the government facilitates the development of new industries that are consistent with the country’s changing comparative advantage determined by the change in its endowment structure, as sug- gested in the NSE. Anne also worries that the identifi cation of any new industry for upgrad- ing “would inevitably favor larger, established fi rms and hence encounter the same sorts of problems as did the older import-substitution strategy” (p. 51). Her worry is valid for the old structuralist import-substitution strategy because the industries favored went against the comparative advantages of the countries that adopted it. Such industries were too capi- tal intensive and only a few rich and politically well-connected fi rms could enter them. However, if the identifi ed new industries are consistent with the rms will country’s comparative advantages, capital intensive or not, many fi be able to enter and contest the dominance of large fi rms, as exemplifi ed by the auto industry in Japan in the 1960s, the textile industry in Mauritius and electronics in Taiwan, China, in the 1970s, and the garment industry in Bangladesh and salmon-farming in Chile in the 1980s. The type of government incentives for the fi rst movers advocated in the NSE is limited to compensating for the externalities generated by the fi rst movers rather than supporting nonviable fi rms, as in the case of old struc- turalist import-substitution strategy. Therefore, tax holidays for fi rst mov- ers for a few years, and preferential access to credit and foreign exchange (in countries where lack of access is a binding constraint) would be enough.

85 72 | New Structural Economics Finally, Anne questions the uncertainty surrounding the scope, depth, and length of government protection and points out the risk of political capture and rent-seeking in situations where a government adopts a dual- track approach during its transition from a heavily distorted economy to a well-functioning market economy. She argues that “a major challenge for liberalizing reform is for it to be credible that the altered policies are not reversible. Lin’s prescription would greatly increase the challenge of creating credibility, and a slower transition would be a longer period dur- ing which growth was slow and political pressures opposing liberalization at all were mounting” (p. 52). The credibility argument was used to sup- port the shock therapy in the transitions of East Europe and the former rms were Soviet Union in the early 1990s. However, even though those fi privatized, governments in transition economies were very often forced to provide other disguised and less effi cient forms of subsidies and protec- tion to ward off large unemployment and subsequent social and politi- cal instability. As a result, most transition economies encountered the awkward situation of “shock without therapy.” Instead of a “J-curve” recovery as promised by the proponents of shock therapy, those econo- mies encountered an “L-curve” growth path (a prolonged sluggish growth after a sharp decline in the GDP) during their transition. By contrast, good performers such as China, Vietnam, Laos, Slovenia, and Uzbekistan have reformed their distorted economies by adopting a more pragmatic, dual- track approach, which consists of progressively phasing out government support to “nonviable” fi rms in priority sectors and at the same time lib- eralizing the entry of formerly repressed private enterprises, joint ventures, and foreign direct investment in sectors aligned with their comparative advantages. The lesson is clear: for any developing country confronted with severe distortions and poor growth performance, the best way to gain confi dence and credibility in its liberalization reforms is to achieve stability and dynamic growth in the transition process. The comments by Joe Stiglitz and Dani Rodrik on my paper reveal more differences of emphasis and style than divergence on substance. I agree with Dani’s assessment that our difference “is mainly methodological— and perhaps even just terminological—and may have little practical import” (p. 55). However, there are a few differences that are worth highlighting.

86 Rejoinder to Comments on Chapter I | 73 Beyond the traditional need for regulation, Joe sees a catalytic role for governments “in promoting entrepreneurship, providing the social and physical infrastructure, ensuring access to education and fi nance, and sup- porting technology and innovation” (p. 57). He strongly challenges the belief in the effi ciency and stability of unfettered markets, and stresses the need for advances in technology as a key condition for increases in per capita income. Consequently, he favors public action for the creation of a “learning society” (p. 58). I agree with Joe on the importance of learning. However, the content and mechanism of learning may be different for countries at different lev- els of development. Developing countries that are still at the early phase of their development generally do not have the necessary human and physical capital to leapfrog into capital-intensive, high-tech industries. The more effective route for their learning and development is to exploit the advantages of backwardness and upgrade and diversify into new indus- tries according to the changing comparative advantages determined by the changes in their endowment structure. The subsequent dynamics of growth, accumulation of human and physical capital, and industrial and technological upgrading eventually open up possibilities to enter and mas- ter capital- and knowledge-intensive industries at the global frontier. The need to generate new knowledge through indigenous innovations in an economy increases with its economic development and the narrowing of the knowledge gap (the distance to the global technology/industrial fron- tier). Therefore, the learning and the enhancement of human capital should be commensurate with the level of economic development. Otherwise the attempt to create a “learning society” by increasing education alone may not correspond to the emergence of new, dynamic sectors consistent with the comparative advantage refl ected in its endowment structure. Should this happen, the educated young people will not fi nd suitable employment opportunities, causing a waste of scarce human and educational resources and, most likely, social tensions, as has happened in North African and many other developing countries. Joe points out that “some of the most important elements of comparative advantage are endogenous” and contends that “Switzerland’s comparative advantage in watch-making has little to do with its geography” (p. 59). The fact is that watch-making was a new industry in the 16th century.

87 74 | New Structural Economics Switzerland goldsmiths started making watches in 1541 and formed the 1 rst watch-makers’ guild in 1601. According to Maddison’s (2010) esti- fi mations, Switzerland’s per capita income in 1600 was 750 measured in 1990’s international dollars, which was 77 percent of Britain’s in the same year. Therefore, Switzerland was one of the “high-income” countries in the world at that time. To continue its income growth, it had to upgrade its industries to some new higher-value-added industries. While Switzerland’s comparative advantage in watching-making has little to do with its geography, as Joe pointed out, geography may be a crucial reason for its leadership in this industry since the 16th century. Watches are small, light, and high value added, with potential for continu- ous technology improvements. Such an industry is particularly suitable for a landlocked country like Switzerland. This may explain why Switzerland has kept its watch-making industry by maintaining technology leadership through continuous innovations since the 16th century but gave up other industries, such as garments, textiles, and footwear, which fl ourished in Switzerland in its early history. Joe may be a bit too optimistic when he suggests that full capital mobil- ity in a globalized world allows countries to free themselves from patterns dictated by endowments, as conventionally defi ned. He postulates that “with fully mobile capital, outside of agriculture, natural resource endow- ments need not provide the basis for explaining patterns of production and specialization” (p. 59). However, short-term capital fl ows are too volatile to be a reliable source for long-term productive investments in developing countries. We observed that during the East Asian fi nancial crisis of the late 1990s. By contrast, foreign direct investments are more reliable because they are motivated by the search for profi ts. They mostly go to tradable sectors or production activities which are consistent with a host country’s comparative advantage so as to use that location as an export base, or to enter the host country’s domestic market—except when they are driven by occasional cases of privatization of large non-tradable sectors such as utili- ties and telecommunication. Because of his optimism about the mobility of capital, Joe highlights the importance of knowledge and entrepreneurship endowment. The importance of knowledge and entrepreneurship cannot be overemphasized. They are indeed driving forces for industrial upgrad- ing and diversifi cation in a dynamically growing economy. Nevertheless,

88 Rejoinder to Comments on Chapter I | 75 as discussed above, the type of new knowledge that is useful for a coun- try’s development depends on the needs of new industries that align with the country’s comparative advantage. An entrepreneur’s investment in an open, competitive market is unlikely to be successful if he or she invests in industries that do not align with the country’s comparative advantage. Joe proposes the undervaluation of the exchange rate as a broad-based policy for encouraging the upgrading of tradable industries (p. 60). This is a delicate issue: it may help exports but it always makes imports of equip- ment more expensive, which is an obstacle for industrial upgrading and cation (because fi rms need new capital equipment from abroad to diversifi upgrade or diversify into new industries). Therefore such a policy may help existing industries’ exports but may not be conducive to long-term growth. Successful developing countries seem to have adopted a policy of undervaluation of real exchange rate if the Balassa-Samuelson theorem is used as a reference. However, the explanation may be the following: these countries typically converge from a dual economy with large surplus labor to a modern economy with a unifi ed national labor market. At some point that theorem does not apply: before the depletion of surplus labor, the wage rate in the tradable and non-tradable sectors will not increase, which is a required mechanism for real appreciation in the theorem. What then looks like undervaluation may actually be an equilibrium exchange rate. Dani’s quibble with my approach seems to be related to his assump- tion that coordination and externality issues only exist in situations where markets send entrepreneurs the wrong signals. He therefore suggests that I may be arguing “both for and against comparative advantage” (p. 55). This deserves a clarifi cation: Comparative advantage is determined by fac- tor endowment. If an industry is consistent with a country’s comparative advantage, the factor cost of production will be lower than otherwise. But for that industry to be competitive in its domestic and international mar- kets, transaction-related costs should also be reduced to their lowest pos- sible level. Yet, individual fi rms cannot internalize the reduction of many of the transaction-related costs arising from issues such as provision of infrastructure, logistics, fi nance, educated labor and so forth. Without gov- ernment coordination and facilitation to reduce such costs and compensate for the externalities generated by the fi rst mover, these industries are likely to simply remain as the latent comparative advantage of the economy. An

89 76 | New Structural Economics illustration of this problem is the fact that low-income countries typically have comparative advantages in most unskilled, labor-intensive industries but few of them are able to be competitive in those industries—precisely because governments fail to effectively play their facilitating role. There- fore, the answer to Dani’s objection lies in the distinction between a coun- try’s latent comparative advantage, which determines the factor costs of production, and its actual comparative advantage (or, in Michael Porter’s term, competitive advantage), which requires in addition the reduction of transaction-related costs. My recommendation that governments step into the economic process and address market failures should therefore not be misunderstood as an attempt to defy an economy’s “natural” or “inexorable” comparative advantage as revealed in market prices but as a way of opening the black box of business competitiveness, converting an economy’s potential into reality, and igniting the march of domestic fi rms toward market success. The differences between Dani’s and my understanding of the govern- ment’s role arise to a large extent from our diverging interpretations of experiences in successful countries such as Japan, Korea and China. He regards the successful catching up in Japan and South Korea as evidence for the need to defy a country’s comparative advantage (p. 55). When Japan embarked upon its industrialization path in the early years of the Meiji period (1868–1912), it was an agrarian society in which farming, forestry and fi shing employed more than 70 percent of the working population and represented over 60 percent of national output. Throughout the Meiji, Taisho (1912–26) and pre-war Showa (1926–36) periods, the top exports were raw silk yarn, tea, and marine products. The main market for these commodities was the United States. Historians remind us that along with the opening up of Japanese ports, demand for these primary commodities quickly ballooned and domestic producers greatly profi ted from it. Silk in particular brought wealth to rural areas and generated much coveted for- eign exchange. Also, mining, which had continued from the previous Edo Period, was largely requisitioned by the government and later sold off to the private sector to become one of Japan’s principal industries. The suc- cess of these sectors allowed Japan’s per capita income to increase 40 per- cent from $737 in 1870 to $1,012 in 1890 and again to $2,026 at the 2 onset of Great Depression in 1929 (Maddison 2010). From the point of

90 Rejoinder to Comments on Chapter I | 77 view of NSE, that success contributed to capital accumulation and changed Japan’s endowment structure and comparative advantage. In the words of Japanese historian Kenichi Ohno, “the industrialization of the Meiji Period light industrial revolution, which made its way from importing to was a domestic production and then onto exporting. Within this transition, cot- ton production played a central role. The iron and steel, shipbuilding and chemical industries, as well as the manufacture of electrical machinery and appliances were in their infancy and the country was still in the process of learning by imitating the West... By late Meiji, private-sector production in the areas of shipbuilding, railway carriages and machine instruments 3 had slowly emerged.” Japan’s industrialization proceeded in a fl ying-geese pattern, moving step by step from simple, labor-intensive manufacturing goods to more capital- and technology-intensive manufacturing goods (Akamatsu 1962). Korea also adopted a realistic approach to industrial upgrading and adjusted its strategy to enter industries that were consistent with its latent (and evolving) comparative advantage. In the 1960s Korea developed and exported labor-intensive products such as garments, plywood, and wigs. With capital accumulation and a change in its endowment structure due to success, Korea upgraded to more capital-intensive sectors such as the automotive sector. But at the initial stage, domestic manufacturers concen- trated mostly on assembly of imported parts, which was labor-intensive and in line with their comparative advantage at the time. Similarly, in electronics, the focus was initially on household appliances, such as TVs, washing machines and refrigerators before the country moved to the pro- duction of memory chips, the least technologically complex segment of the information industry. Korea’s technological ascent has been rapid, at a pace commensurate to changes in underlying comparative advantage. Such changes refl ected rapid accumulation of physical and human capi- tal resulting from the dynamic growth, which could only occur because the country’s main industrial sectors remained consistent with its existing comparative advantage. Similarly, Dani’s observation that China has been defying its comparative advantage successfully, with its export bundle resembling that of a country between three and six times richer, neglects the fact that these are mostly processed products. China only provides value-added in labor-intensive

91 78 | New Structural Economics assembly and accessories. Empirical research (Wang and Wei 2010) shows that China’s exports are consistent with China’s comparative advantage. Dani also questions the differences between my recommendation for gradual trade liberalization and the old structuralist policies. The lat- ter approach advocated protection and subsidies to build new industries that were not aligned with comparative advantage, whereas the dual- track, gradual approach for trade liberalization that I recommend advises governments in transition economies to provide temporary protection or subsidies to old industries that were not viable in an open, competi- tive market but were established under the misguided old structuralist strategy. The pragmatic dual-track approach helps a transition economy avoid unnecessary and costly economic and social disruption, and eventu- ally leads to a system of market-based prices and resource allocation as explained in my response to Anne’s comments. Summing up, it appears that Anne’s questions about the practicality of my framework arise mostly from interrogations on how to identify new industries that are consistent with a country’s latent comparative advan- rst tage, and how to administer the coordination and incentives for the fi movers. Joe’s and Dani’s advocacy of broad-based interventions such as undervalued real exchange rates to support the trade sector but reluctance to embrace the idea of sector-specifi c policies are also related to the puzzle about how to identify industries aligned with latent comparative advantage. entitled “Growth Those questions are addressed in a companion paper Identifi cation and Facilitation,” co-authored with Célestin Monga and published in Development Policy Review (chapter III of this book). Based on economic analysis and historical experiences, the growth identifi cation and facilitation framework that we propose suggests that policy makers with similar in fast-growing countries identify dynamic tradable industries endowment structures, and with a per capita income about double their own. If domestic private fi rms in these sectors are already present, pol- icy makers should identify them and remove constraints on those fi rms’ technological upgrading or on entry by other fi rms. In industries where no domestic fi rms are present, policy makers could aim to attract foreign direct investment from the countries being emulated or organize programs for incubating new fi rms. The government should also pay attention to the development by private enterprises of new and competitive products,

92 Rejoinder to Comments on Chapter I | 79 and support the scaling up of successful private-sector innovations in new industries. In countries with a poor business environment, special economic zones or industrial parks can facilitate fi rm entry, foreign direct investment, and the formation of industrial clusters. Finally, the government might help pioneering fi rms in the new industries by offering tax incentives for a lim- ited period, co-fi nancing investments, or providing access to land or for- eign exchange. I am grateful to Anne, Joe, Dani, and many others who have provided comments and constructive criticism of my paper. Despite our differences, there seems to be an emerging consensus on the need to reconcile lessons from the fi rst two major waves of development thinking (structuralism and neo-liberalism) into a new synthesis that recognizes and defi nes the proper roles of state and markets. The road ahead towards that third wave (which might be termed “Development Thinking 3.0”) will obviously involve healthy and useful intellectual disagreements. Because, as Confucius once said, “Real knowledge is to know the extent of one’s ignorance.” Notes 1. http://www.fhs.ch/en/history.php. 2. Dollars are 1990 dollars adjusted for Geary-Khamis purchasing power parity. Translated excerpts from the book Tojokoku no Globalization: Jiritsuteki 3. Hatten wa Kanoka (Globalization of Developing Countries: Is Autonomous Development Possible?) by Toyo Keizai Shimposha (2000), quoted by Japan’s National Graduate Institute for Policy Studies (GRIPS). See http://www.grips .ac.jp/forum-e/pdf_e01/eastasia/ch5.pdf. References Akamatsu, K. 1962. “A Historical Pattern of Economic Growth in Developing Countries,” (Tokyo), supplement issue no. 1: The Developing Economies 3–25. Maddison, A. 2010. “Historical Statistics of the World Economy: 1–2008 AD” (www.ggdc.net/maddison/Historical_Statistics/vertical-fi le_02-2010.xls). Wang, Z., and S. Wei. 2010. “What Accounts for the Rising Sophistication of China’s Exports,” in “China’s Growing Role in World Trade,” eds. R, Feenstra and S. Wei. Chicago: University of Chicago Press.

93

94 II The Growth Report and New Structural Economics

95

96 II The Growth Report and New Structural Economics † with Célestin Monga 1. Introduction Economists have always been confl icted between the need to understand the dynamics of business cycles, and the study of long-term growth—both of which are important for human welfare. The world economy has just 1 nancial and economic crisis, which has justifi ed the experienced a severe fi intellectual focus on stabilization policies, especially given the role that coordinated and decisive monetary and fi scal policies have played in pre- venting the global recession from becoming a worldwide depression. But the persistence of poverty in many parts of the world and the potential long-term impact of the crisis on global poverty reduction also highlight the importance of policies that are conducive to sustainable and inclusive 2 growth. Economic growth is indeed the main source of divergences in liv- ing standards across countries and regions of the world. As Barro and Sala- i-Martin (1995) observe, “if we can learn about government policy options 83

97 84 | New Structural Economics that have even small effects on long-term growth rates, we can contribute much more to improvements in standards of living than has been provided by the entire history of macroeconomic analysis of countercyclical policy ne tuning.” and fi In fact, economic growth may be the single most important issue con- fronting economists today. The differences in output per worker and national income across countries are still puzzling. According to cal- culations by Maddison (2001), world population rose 22-fold over the past millennium. Per capita income increased 13-fold, world GDP nearly 300-fold. This contrasts sharply with the preceding millennium, when world population grew by only a sixth, and there was no advance in per capita income. Measured in today’s living standards, all countries in the world were poor in the beginning of the 18th century. Sustained growth in income per capita only picked up after 1820: per capita income rose more than eightfold. A well-known fact confi rmed by the recent crisis is the observation that countries that have sustained high rates of growth have also performed well despite the global meltdown. Their dynamic performance has made them more resilient. With strong external balance sheets and ample room for fi scal maneuver before the crisis, they were able to implement coun- tercyclical policies to combat external shocks. “A crisis is a terrible thing to waste,” said Paul Romer, one of the preeminent theorists of growth. Despite its heavy human, fi nancial and economic cost, the recent recession provides a unique opportunity to refl ect on the knowledge from several decades of growth research, draw policy lessons from the experience of successful countries, and explore new approaches going forward. Looking at the data, one may be surprised to note that the recession has obscured the broader economic narrative of our time, which is the remark- able economic performance of many poor countries, especially in the past ten years. Leaving aside the United States, which ranks third, the four most populous countries of the world (Brazil, China, India, Indonesia), have made great strides, averaging annual growth rates well over 6 percent a year. That is a vast improvement in the standards of living for more than 40 percent of the world’s population. The same trends are in place in many other South American countries (Chile, Colombia, Peru) and in some African countries (Botswana, Mauritius, Tunisia, Ghana).

98 The Growth Report and New Structural Economics | 85 To be sure, poverty reduction is still a very challenging development ghting poverty is not issue. In an increasingly globalized world where fi only a moral responsibility but also a strategy for confronting some of the major problems (diseases, malnutrition, insecurity and violence) that ignore boundaries and contribute to global insecurity, thinking about new ways of generating and sustaining growth is a crucial task for economists. It is therefore essential to continue searching for new ideas on the mechan- ics of wealth creation. Over the last 50 years, much progress has been 3 made, most recently with the work of the Growth Commission Report. But beyond a consensus on broad principles and the rejection of one-size- fi ts-all approaches, economists still face signifi cant challenges in identifying actionable policy levers that are directly relevant to specifi c countries. This paper reassesses the evolution of knowledge on growth and sug- gests a new structural approach to the analysis. Section 2 offers a brief, critical review of lessons learned from growth research and examines the remaining challenges—especially from the policy standpoint. Section 3 highlights the important recent contribution of the Growth Commission Report and the identifi cation of stylized facts associated with sustained and inclusive growth. Section 4 provides a consistent framework for under- standing its key fi ndings through the lenses of new structural economics. Section 5 offers some concluding thoughts. 2. The Quest for Growth: An Unfi nished Journey Economic historians who have examined the evolution of growth per- formance throughout history tend to divide it into three distinct periods: The fi rst one, which spanned most of human history up to the middle of the 18th century, was marked by static living standards, despite popula- tion growth—the so-called Malthusian conditions. The second one, which lasted from about 1750 to the 1820s, was characterized by some improve- ment in living standards, and changes in demographic trends (higher fertil- ity rates and lower mortality rates). The third epoch, observed initially in England at the end of the fi rst quarter of the 19th century, has been that of modern economic growth (Cameron, 1993). Deciphering the mystery of modern economic growth and explaining convergence and divergence have been major topics of research, especially since the 1950s. While much

99 86 | New Structural Economics progress has been achieved on theoretical and empirical grounds, much remains to be understood on the policy front. Growth Analysis in Historical Perspective c factors that have sustained it and The analysis of growth—and the specifi accompanied the structural changes associated with it—became a major topic of interest for thinkers in general and economists in particular in the early 18th century. David Hume, whom Rostow claims to be “the fi rst modern economist” (1990: 18) placed economic analysis at the center of his analysis of the human condition. He also offered economic concepts that are considered to “form a reasonably coherent and consistent the- ory of the dynamics of growth.” Classical economists who followed in his footsteps—such as Adam Smith, Alfred Marshall, David Ricardo and Allyn Young—were also obsessed with economic growth. Perhaps because of their fascination with the idea of human progress celebrated during the Enlightenment, they explored the determinants of economic development, and the role that policymakers could play in fostering prosperity. Their pioneering work highlighted important notions such as factor accumula- tion, factor substitution, technical change, or specialization, which are at the core of modern growth theory. But growth analysis slowed down after the Great Depression, as the intellectual focus shifted from long-run to short-run issues. In fact, with the notable exception of the pioneering work of Robert Solow, for much of the 20th century and certainly through the 1960s and 1970s, macroecono- mists tended to study business cycles issues that characterized the post-War period. As they tried to better understand stabilization policies—monetary and fi scal measures to avoid disruptive and costly infl ation—few resources were devoted to the analysis of the long-run determinants of growth. Things changed in the 1980s when many prominent researchers focused their attention on differences in economic performance among countries. Surveys of economic growth and levels of performance in different parts of the world economy show that growth has indeed been uneven across countries and regions: between 1900 and 2001, per capita GDP in Western Europe increased by a factor of 6.65 (6.7 in Western offshoots), compared 4 to 5.2 in Latin America, 4.2 in Eastern Europe, and only 2.5 in Africa. The number of people living in high-growth environments or in countries with

100 The Growth Report and New Structural Economics | 87 OECD per capita income levels has increased in the past 30 years by a fac- tor of four, from 1 billion to about 4 billion (Growth Commission, 2008). Following the initial work by Harrod and Domar, the Solow-Swan model sparked the fi rst major wave of systematic growth analysis. The objective was to understand the mechanics of growth, identify its determi- nants, and to develop techniques of growth accounting, which would help explain changes in the momentum and the role of economic policy. That fi rst generation of growth researchers highlighted the centrality of capi- tal. Their models featured neoclassical forms of production functions with cations that relied on constant returns to scale, diminishing returns specifi and some elasticity of substitution between inputs. In order to present a general equilibrium model of the economy, these researchers adopted a constant saving rate rule. This was a crude assumption but a major step forward in tool building, as it offered a clear demonstration that general equilibrium theory could be applied convincingly to real world issues. One important prediction from these models was the idea of conditional con- vergence, derived from the assumption of diminishing returns to capital— poor economies with lower capital per worker (relative to their long-run or 5 steady state capital per worker) will grow faster. The major strength of that line of growth research was the explicit introduction of technology—in addition to capital and labor—in the theo- retical and empirical analysis. But the limited toolkit available at the time created a major shortcoming to that approach: technology was presented as an exogenously given public good. The major prediction of the model based on the assumption of diminishing returns to capital was the idea that per capita growth will cease in the absence of continuous improvements in technology. While that assumption allowed the model to maintain its key prediction of conditional convergence, it also seemed odd: technology, the 6 main determinant of long-run growth, was kept outside of the model. A new wave of growth modeling had to come up with a convincing the- ory of technological change—one that frees up the neoclassical model from the exogeneity of the main determinant of long-term growth. A fi rst step was to design a theory of continuous growth fuelled by non-diminishing returns to investment on a broad class of physical and human capital. The process could go on indefi nitely if returns do not diminish as economies grow (Romer 1986). A second, more effective approach was to move away

101 88 | New Structural Economics from the straightjacket of perfect competition, and incorporate imperfect competition and R&D theories in growth modeling—the rationale here being that such bold methodological moves helped explain why the econ- omy would not run out of new ideas, and growth rates could be kept posi- tive in the long run (Romer 1987, 1990; Aghion and Howitt, 1992). Endogenous growth theory, as it came to be known, maintained the assumption of nonrivalry because technology is indeed a very different type of factor from capital and labor—because it can be used indefi nitely by others, at zero marginal cost. But it was important to take the next logical step and to better understand the public good characterization of technol- ogy, and think of it as a partially excludable nonrival good. The new wave therefore reclassifi ed technology not just as a public good but as a good that is subject to a certain level of private control. By making it a partially excludable nonrival good and therefore giving it some degree of exclud- ability or appropriability, it was possible to ensure that incentives mat- ter for its production and use. The move away from perfect competition was therefore necessary. It has yielded high methodological payoffs. While neoclassical models of growth took technology and factor accumulation as exogenous, endogenous growth models explain why technology grows over time through new ideas, and provide the microeconomic underpin- nings for models of the technological frontier. Another important question has been to understand how technological diffusion takes place across countries and generates or sustains growth— and why it does not take root in others. Various interesting possibilities have recently been explored in an attempt to answer that critical question: one option has been to add an avenue for technology transfer as a new component to the endogenous growth model, that is, “endogenizing” the mechanism by which different countries achieve the ability to use various intermediate capital goods (Jones, 1998). Another popular route is to try to identify the fundamental determinants of growth through political economy models. Contrary to previous waves of growth modeling, this line of research focuses not on the proximate determinants of growth but on the impact on growth of such factors as institutions or the qual- ity of governance (Acemoglu and Robinson, 2001; Glaeser and Shleifer, 2002). Several other approaches to growth research have yielded various insights to the mystery of modern economic growth (Barro and Sala-i- Martin 2003; Jones 1998).

102 The Growth Report and New Structural Economics | 89 Challenges of Explaining Convergence—and Divergence On both the theoretical and empirical fronts, progress has been made in our understanding of growth in recent decades. On the theoretical front, the analysis of endogenous technical innovation and increasing returns to scale has provided economists with a rich general framework for capturing the broad picture and the mechanics of economic growth. From Solow’s work, we know the importance of the role of capital accumulation (both physical and human) and technical change in the growth process. From con- 7 and many others, we also learned tributions by Becker, Heckman, Lucas about the importance of human capital through diffusion of new knowl- edge or on-the-job learning, often stimulated by trade, and the so-called college wage premium. From work by North (1981), with supporting the- oretical and empirical analyses exemplifi ed by the works of Acemoglu and Robinson (2001), Greif (1993), and Glaeser and Shleifer (2002), we have learned that growth is in large part driven by innovation and institutions that have evolved in countries where innovative activity is promoted and conditions are in place for change to take place. From Romer and the endog- enous growth theorists, we have understood the need to change the focus of growth theory from accumulation to knowledge creation and innovation. In sum, we know quite a lot about some of the basic ingredients of growth. On the empirical side, the availability of standardized data sets— especially the Penn World tables—has stimulated interest in cross- country work that highlights systematic differences between high-growth and low-growth countries with regard to: (i) Initial conditions such as productivity, human capital, demographic structure, infrastructure, fi nancial development, or inequality; (ii) Policy variables of various sorts such as trade openness, macroeconomic stability, levels and composi- tion of public spending, taxation, or regulation; and (iii) Institutional variables such as general governance indicators, administrative capacity, rule of law, protection of property rights, or corruption. cant methodological diffi cul- However, growth research still faces signifi ties, and challenges in identifying actionable policy levers to sustain and 8 c countries. accelerate growth in specifi Deaton (2009) expresses the general sentiment of despair among econo- mists when he notes that “empiricists and theorists seem further apart now than at any period in the last quarter century. Yet reintegration is hardly an

103 90 | New Structural Economics option because without it there is no chance of long-term scientifi c prog- ress.” Despite many decades of theoretical advances and the development of new techniques to help policymakers in developing countries identify systematically constraints to growth, the intellectual and policy agenda ahead is indeed still daunting. Contrary to the prediction of most neoclassical models, convergence among world economies has been a limited phenomenon (Pritchett 1997). In 2008, GDP per capita in the United States (the world’s richest coun- try) was three times higher than per capita income in neighboring Mexico, 16 times higher than the per capita income in India, and 145 times the per capita income of the Democratic Republic of Congo. That gap is still wid- ening. In most of the past century, incomes in developing countries have fallen far behind those in developed countries, both proportionately and 9 absolutely. Yet, empirical observation reveals that divergence between industrial- ized and developing countries is not inexorable: in the past two centu- ries, some countries have been able to catch up with the most advanced economies (most notably Germany, France, and the USA in the late 19th century, and the Nordic countries, Japan, and the 13 economies analyzed in the Growth Commission Report in the 20th century). After the Industrial Revolution began in England in the mid-eighteenth cen- tury, experiments conducted in laboratories became the major source of technological invention and innovation (Lin, 1995). This was espe- cially true for those macro-inventions that consisted of radical new ideas and involved large, discrete, novel changes, as defi ned by Mokyr (1990). For developed countries, such inventions were essential to technologi- cal advances. With investment in research and development, innovation became endogenous (Romer, 1986; Lucas, 1988). Industrial structures were upgraded continuously and productivity increased. As a result, developed countries began to take off and the divergence between the North and the South appeared (Baumol, 1994). Historical evidence suggests that the growth process followed a simi- lar pattern in developing economies such as the four East Asian dragons (Korea, Singapore, Taiwan, Hong Kong), which converged to the income levels of advanced western countries in the second half of the 20th cen- tury. The same process subsequently allowed countries as diverse as China,

104 The Growth Report and New Structural Economics | 91 Vietnam, Botswana, or Mauritius to achieve rapid and sustained growth in the 1980s and 1990s (Lin, 2003, 2009; Rodrik, 2005). Except for that select group, most developing countries have failed to achieve their eco- nomic growth ambitions since World War II. In fact, many have encoun- tered frequent crises despite efforts from their governments and assistance from international development agencies. Yet, their experiences highlight the need to understand how developing countries can create the conditions for facilitating the fl ow of technologies and unleash growth, even in the context of sub-optimal microeconomic policies, weak institutions, and the absence of full-fl edged private property rights. The failures of growth research to predict divergence on a large scale indi- cates that the proposed theories did not capture the fundamental factor(s) that determines whether or not a developing country will converge. Some researchers have recently argued that the evolution of economic perfor- mance of nations is determined by conditional convergence—the idea that countries converge when all other macroeconomic variables that proxy for differences in steady-state characteristics are held constant—or to put it differently, the distribution of world income reveals the existence 10 of convergence clubs among countries. But the puzzle of diverging per- formances may be more easily sorted out through comparative analysis based on in-depth country studies and historical experience: the key ingre- dients for convergence of successful economies seem to lie in their abil- ity to change their endowment structure, increase the pace of adoption of new ideas, speed up the process of industrial upgrading, and improve institutions simultaneously. Understanding and replicating the economic strategies and policies that allowed latecomers to catch up with the most advanced economies is still a major challenge for economists and policy- makers around the world. New Directions in Applied Growth Research The disappointments of growth research—most notably from the perspective of policymakers seeking specifi c action plans to generate prosperity—have led to a reassessment of the validity and usefulness of existing knowledge, and to the development of radically new approaches. An important study by the World Bank (2005) that focused on lessons of the 1990s highlighted the complexity of economic growth and recognized that it is not amenable

105 92 | New Structural Economics to simple formulas. The report also noted that the reforms carried out in - many developing countries in the 1990s focused too narrowly on the effi cient use of resources, not on the expansion of capacity and growth. While they enabled better use of existing capacity, thereby establishing the basis cient incentives for for sustained long-run growth, they did not provide suffi 11 expanding that capacity. The report concluded that there is no unique, universal set of rules to guide policymakers. It recommended less reliance on simple formulas and the elusive search for “best practices,” and greater reliance on deeper economic analysis to identify each country’s one or two most binding constraints on growth. That line of research is exemplifi ed by the Growth Diagnostics frame- work, which aims at identifying the one or two most binding constraints on any developing economy, and then focusing on lifting those. The main rationale is to ensure that economic reforms are contingent on the eco- nomic environment. “Presented with a laundry list of needed reforms, poli- cymakers have either tried to fi x all of the problems at once or started with reforms that were not crucial to their country’s growth potential. And, more often than not, reforms have gotten in each other’s way, with reform in one area creating unanticipated distortions in another area. By focusing on the one area that represents the biggest hurdle to growth, countries will be more likely to achieve success from their reform efforts” (Hausmann, Rodrik and Velasco, 2008). The proposed approach offers a decision tree methodology to help identify the relevant binding constraints for each country. While it does not specifi cally identify the political costs and ben- efi ts of various reform strategies, its focus on alternative hypotheses can help clarify the options available to policymakers for responding to politi- cal constraints. “We are concerned mainly with short-run constraints. In this sense, our focus is on igniting growth and identifying constraints that tomorrow’s inevitably emerge as an economy expands, not on anticipating constraints on growth” (Hausmann, Rodrik and Velasco, 2008). A key lesson from that approach is the notion that different countries (or even the same country at different points in time) require different pol- icy choices to facilitate growth, and that the ‘big principles’ that growth requires—sound money, property rights, openness, free markets—can take many forms and that achieving them requires country-specifi c context and

106 The Growth Report and New Structural Economics | 93 information. In particular, these principles need not take any one precise institutional or policy form. Each country is assumed to have some bind- ing constraints to its growth potential, and failure to identify and remove them would impede economic performance, even if every other produc- tion factor is satisfactory. The Growth Diagnostics approach is certainly an important advance in growth analysis. However, its model does not fully 12 fl esh out the notion of “binding constraint.” nitions are The variable defi deliberately left quite imprecise, which makes it challenging to operational- ize them. Another infl uential new approach is the one adopted by researchers at the MIT Poverty Lab, who suggest that the quest for growth be re-centered on assessing the impact of a development project or program (against explicit counterfactual outcomes). Starting with the idea that credible impact evalu- ations are needed to ensure that the most effective programs are scaled up at the national or international levels, they design randomized control ts trials (RCTs) or social experiments that can be used to leverage the benefi of knowing which programs work and which do not (Dufl o and Kremer, 2003). Their approach is based on the notion that the standard aggregate growth paradigm relies, to a large extent and mistakenly, on the assump- tion of a rational representative agent. Stressing heterogeneity in country circumstances and among micro agents, this new wave of research attempts to explicitly account for the heterogeneity of individual households and 13 It has produced some useful fi rms in development analysis and policy. c micro projects. tools for understanding the effectiveness of some specifi But even assuming that they can actually transfer lessons from localized 14 development experiences to different geographic or cultural areas, RCTs still fall short in providing useful overall guidance to policymakers con- fronted with the design of development strategies. While these new approaches to growth research have shed light on important questions, they have not provided suffi cient guidance on how policymakers could foster the process of industrial upgrading and struc- tural change. It would be desirable to complement them with structural analyses of the determinants of growth—specifi cally the identifi cation of factors that would allow poor economies to move from one stage of devel- opment to another.

107 94 | New Structural Economics 3. The Unique Contribution of the Growth Report Despite intellectual progress, some of the key questions on the growth research agenda today remain the same as those that confronted previ- ous generations of researchers: If growth is driven in large part by inno- vation, why are some countries successful at innovating and adapting to change, while others are not? What are the forces that drive convergence and what are the factors that stifl e material progress? What are the condi- tions for the kind of structural change that allow low-income countries to become middle-income and then high-income economies? What are the most important determinants of growth (initial conditions, institutions, and policies)? What is the appropriate role for governments and markets in the growth dynamics? Faced with the diffi culty of providing clear answers to such pressing questions and the impossibility of deriving actionable policy recommen- dations from growth analyses, some growth researchers have found it useful to avoid searching for robust determinants of growth, and to look instead for the stylized facts that can guide economic policy in develop- ing countries. This approach goes back several decades, most notably to Kaldor’s (1961) six characteristics of 20th century growth, derived from United States and United Kingdom macroeconomic data: (i) sustained rate of increase in labor productivity; (ii) sustained rate of increase in capital per worker; (iii) stable real interest rate or return on capital; (iv) stable ratio of capital to output; (v) stable shares of capital and labor as fractions of national income; and (vi) a wide variation in the rate of growth of fast growing economies, of the order of 2-5 percent. More recently, Jones and Romer (2009) have identifi ed a different set of stylized facts: (i) increases in the extent of the market—via globalization and urbanization; (ii) acceleration of the pace of growth over time, from virtually zero to relatively rapid rates; (iii) variation in the rate of growth of GDP per capita, which increases with the distance from the technol- ogy frontier; (iv) large income and total factor productivity differences; (v) increases in human capital per worker; and (vi) long-run stability of relative wages. The Growth Report: Strategies for Sustained Growth and Inclusive Development , a landmark study issued in 2008 by the Commission on

108 The Growth Report and New Structural Economics | 95 Growth and Development, followed a similar approach but took it to a ndings of several other empirical studies initi- new level. It built on the fi ated by the World Bank during the past two decades to reassess the past theories of economic growth and poverty reduction, and rethink its policy 15 advice to developing countries. Launched in April 2006, the Commission brought together 22 leading practitioners from government, business and policymaking arenas, mostly from the developing world. It was chaired by Nobel Laureate Michael Spence and Danny Leipziger, a World Bank Vice- President. Over a period of two years the Commission sought to “gather the best understanding there is about the policies and strategies that under- lie rapid and sustained economic growth and poverty reduction.” The Commission was established to take stock of the state of theoreti- cal and empirical knowledge on economic growth with a view to drawing implications for policy, and avoiding the trap of purely theoretical exer- cises. It provides the following motivation for its work: (i) the sense that poverty cannot be reduced in isolation of economic growth, and that this link has been missing in many development strategies; (ii) increasing evi- dence that the economic and social forces underlying rapid and sustained growth are much less well understood than generally thought—economic advice to developing countries has been given with more confi dence that justifi ed by the state of knowledge; (iii) realization that the accumulation of highly relevant (both successful and unsuccessful) growth experiences over the past 20 years provides a unique source of learning; and (iv) grow- ing awareness that, except for China and India, and other rapidly growing economies in East Asia, developing countries need to accelerate their rates of growth signifi cantly for their incomes to catch up with income levels in industrialized countries, and for the world to achieve a better balance in the distribution of wealth and opportunity. The uniqueness of the Commission lies not only in its very diverse composition but also in the way it has reexamined growth analysis. Its approach has been to “try to assimilate and digest the cumulative experience of growth and development as well as careful and thought- ful policy analysis in a wide spectrum of fi elds. We then seek to share this understanding with political leaders and policymakers in developing countries, including the next generation of leaders; with an international community of advisors; and with investors, policymakers and leaders in

109 96 | New Structural Economics advanced countries and international institutions who share the same 16 (Growth Commission, 2008, p. x). goals” The Report starts with the observation that “fast, sustained growth does not happen spontaneously. It requires long-term commitment by a country’s political leaders, a commitment pursued with patience, persever- ance, and pragmatism” (Growth Commission, 2008, p. 2). It then identi- 17 fi es some of the distinctive characteristics of 13 high-growth economies that have been able to grow at more than 7 percent for periods of more than 25 years since World War II. At that pace of expansion, an economy 18 almost doubles in size every decade. The Report then asks how other developing countries can emulate them. Observing that each country has specifi c characteristics and historical experiences that must be refl ected in its growth strategy, it does not attempt to provide a generic formula for policymakers to apply. However, it offers a framework that can help poli- cymakers design a growth strategy. While it does not lay out a full set of answers, it suggests the right questions to be addressed. The conclusion is an optimistic one: rapid, sustained growth is not a miracle confi ned to certain parts of the world. It can be achieved by all developing countries. More important than the list of “growth ingredi- ents,” which includes a wide range of policy prescriptions whose validity depends on specifi c contexts and conditions, the Report lists “fi ve striking points of resemblance” among all highly successful countries: • Openness to the global economy . During their periods of fast growth, all the successful economies made the most of the global economy. They did so in at least two ways: fi rst, they imported ideas, technology and know- how from the rest of the world—a world that has become more open and more tightly integrated since the end of World War II. Second, they exploited global demand, which provided an almost infi nite market for their goods. In sum, successful economies “all imported what the rest of world knew, and exported what it wanted.” The unsuccessful countries did the opposite. The lesson here is clear: in order to achieve sustained and dynamic growth, a developing country must: (i) rely on its com- parative advantage (that is, export what the rest of the world needs and upgrade its industries step by step at a pace consistent with the change in its endowment structure so as to make its economy competitive); and (ii) tap the potential of advantage of backwardness (imported ideas,

110 The Growth Report and New Structural Economics | 97 technology and know-how from the rest of the world in the process of its industrial upgrading). • Macroeconomic stability . The second stylized fact of high-growth coun- tries is their maintenance of stable macroeconomic environments. Dur- ing their most successful periods, all 13 countries avoided the kind of unpredictability in fi scal and monetary policies that damage private sec- tor investment. While growth was sometimes accompanied by moderate infl ation in some of them (Korea in the 1970s, China in the mid-1990s), budget defi cits or even high ratios of debt-to-GDP, the situation never got out of control. High saving and investment rates . Another characteristic of high-growth • countries is their willingness to forgo current consumption in pursuit of higher levels of incomes in the future. High saving rates were matched by high investment rates. The fact that countries such as Singapore or Malaysia adopted mandatory saving schemes have led some research- ers to stress the importance of deliberate saving policies as the main cause for these high saving and investment rates (Montiel and Serven, 2008). In fact, the main explanation may be the ability of these coun- tries to produce large economic surplus and to generate rates of return on investment that were high enough to provide strong incentives to save. In the 1970s, Southeast Asia and Latin America had similar sav- ings rates. Twenty years later, the Asian rate was about 20 percentage points higher. Market allocation . The Report notes that the 20th century saw many • experiments with alternatives to a market system. They all failed to help developing countries achieve sustained growth. While successful coun- tries may differ in the intensity and strength of their property rights systems, they all adopted a well-functioning market mechanism that provided adequate price signals, transparent decision-making and good incentives. Their governments also did not resist the market forces in the reallocation of capital and labor from sector to sector, industry to industry. • Leadership and governance . Sustained growth that can help over- come poverty is typically a multi-decade process, which only takes place in a stable and functional investment environment. It requires political leadership and effective, pragmatic and sometimes activist governments.

111 98 | New Structural Economics The Growth Commission Report also identifi es a series of “bad ideas” to be avoided by policymakers in their search for growth. The non- exhaustive list includes: subsidizing energy; relying on the civil service to scal defi cits by cutting expenditures on deal with joblessness; reducing fi infrastructure investment; providing open-ended protection to domestic rms; imposing price controls to stem infl ation; banning exports for long fi periods of time; resisting urbanization and measuring educational prog- ress through infrastructure; ignoring environmental issues as an “unaf- fordable luxury”; adopting regulation of the banking system; or allowing the exchange rate to appreciate excessively. Summing up, it can be said that the Report represents a major step forward as it provides a practical approach to help policymakers today understand the economic dynamics of catching up, and to identify the pre- cise (and probably country-specifi c) mechanics of creating the appropriate infrastructures, incentive systems, and institutions to facilitate and sus- tain the evolving growth process. It also offers a new challenge to growth researchers, who must come up with a conceptual framework for making sense of its main fi ndings. 4. A New Structural Analysis of the Growth Report The stylized facts identifi ed by the Growth Commission Report can be either endogenous or exogenous variables to the growth process. In order to disentangle causes and effects, and prioritize public policies, it is useful to go beyond the mere association that these stylized facts suggest, and refl ect on the dynamics of possible causal relationships. As Zellner (1979) pointed out, this requires some generally acceptable economic theory. The new structural economics approach provides such a framework. Principles of New Structural Economics The new structural economics framework (Lin, 2010) is based on the analysis of the growth process in modern times and across continents. It starts with the observation that the main feature of modern economic development is continuous technological innovation and structural change. The optimal industrial structure in an economy, that is, the indus- trial structure that will make the economy most competitive domestically

112 The Growth Report and New Structural Economics | 99 and internationally at any specifi c time, is endogenous to its comparative advantage, which in turn is determined by the given endowment struc- 19 ture of the economy at that time. Economies that try to grow simply by adding more and more physical capital or labor to the existing industries eventually run into diminishing returns; and economies that try to deviate from their comparative advantage are likely to perform poorly. Because the optimal industrial structure at any given time is endogenous to the existing factor endowments, a country trying to move up the ladder rst change its endowment structure. of technological development must fi With capital accumulation, the economy’s factor endowment structure evolves, pushing its industrial structure to deviate from the optimal deter- mined by its previous level. Firms then need to upgrade their industries and technologies accordingly in order to maintain market competitiveness. If the economy follows its comparative advantage in the development of its industries, its industries will be most competitive in domestic and world markets. As a result, they will gain the largest possible market share and generate potentially the largest surplus. Capital investment will also have the largest possible return. Consequently, households will have the high- est savings propensity, resulting in an even faster upgrade of the country’s endowment structure. A developing country that follows its comparative advantage to develop its industries can also benefi t from the advantage of backwardness in the upgrading process and grow faster than advanced countries. Enterprises in developing countries can benefi t from the industrial and technological gap with developed countries by acquiring industrial and technological inno- vations that are consistent with their new comparative advantage through learning and borrowing from developed countries. The main question then is how to ensure that the economy grows in a manner that is consistent with its comparative advantage. The goal of most fi rms everywhere is profi t maximization, which is, ceteris pari- bus, a function of relative prices of factor inputs. The criterion they use to select their industries and technology is typically the relative prices of capital, labor and natural resources. Therefore, the precondition for fi rms to follow the comparative advantage of the economy in their choice of technologies and industries is to have a relative price system which can refl ect the relative scarcity of these production factors in the endowment

113 100 | New Structural Economics structure. Such a relative price system exists only in a competitive market system. In developing countries where this is not usually not the case, it is necessary that government action be taken to improve various market institutions so as to create and protect effective competition in the prod- uct and factor markets. In the process of industrial upgrading, fi rms need to have information about production technologies and product markets. If information is not freely available, each fi rm will need to invest resources to search for rms in developing countries, it, collect it, and analyze it. For individual fi industrial upgrading is therefore a high-reward, high-risk process. First movers who attempt to enter new industries can either fail—because they target the wrong industries—or succeed—because the industry is consis- tent with the country’s new comparative advantage. In the case of success, their experience offers valuable and free information to other prospective entrants. They will not have monopoly rent because of competition from new entry. Moreover, these fi rst movers often need to devote resources to train workers on the new business processes and techniques, who may be then hired by competitors. First movers generate demand for new activities and human capital which may not have existed otherwise. Even in situa- tions where they fail, their bad experience also provides useful knowledge to other fi rms. Yet, they must bear the costs of failure. In other words, the rst movers’ investments is usually much larger than social value of the fi their private value and there is an asymmetry between the fi rst movers’ gain from success and the cost of failure. Successful industrial upgrading in an economy also requires new types of fi nancial, legal, and other “soft” (or intangible) and “hard” (or tangible) infrastructure to facilitate production and market transactions and allow the economy to reach its production possibility frontier. The improvement of the hard and soft infrastructure requires coordination beyond individual fi rms’ decisions. Economic development is therefore a dynamic process marked with externalities and requiring coordination. While the market is a neces- sary basic mechanism for effective resource allocation at each given stage of development, governments must play a proactive, facilitating role for an economy to move from one stage to another. They must intervene to allow markets to function properly. They can do so by (i) provid- ing information about new industries that are consistent with the new

114 The Growth Report and New Structural Economics | 101 comparative advantage determined by change in the economy’s endow- ment structure; (ii) coordinating investments in related industries and the required improvements in infrastructure; (iii) subsidizing activities with externalities in the process of industrial upgrading and structural change; and (iv) catalyzing the development of new industries by incubation or by attracting foreign direct investment to overcome the defi cits in social capital and other intangible constraints. In sum, the new structural economics framework is three-pronged: it includes an understanding of a country’s comparative advantage defi ned as the evolving potential of its endowment structure; reliance on the market as the optimal resource allocation mechanism at any given stage of devel- opment; and the recognition of a facilitating role of the state in the process of industrial upgrading. It helps explain the economic performance of the most successful developing countries. Key Findings of the Growth Commission: A New Structural Analysis The new structural economics provides a framework for understanding the endogeneity and exogeneity issues surrounding the fi ve stylized facts of the Growth Commission Report: (i) exploiting the world economy through openness; (ii) maintaining macroeconomic stability; (iii) keeping high rates of saving and investment; (v) using markets to allocate resources; and (v) having committed, credible, and capable governments. The fi rst three stylized facts are logical outcomes of a country following its comparative advantage determined by its factor endowments in each stage of develop- ment. The fourth stylized fact, the market mechanism, is the precondition for a country to follow its comparative advantage. The last stylized fact, a committed, credible, and capable government, is a prediction as well as a consequence of following comparative advantage. First, if a country follows its comparative advantage in its develop- 20 it will have an open economy, and produce whatever is ment strategy, consistent with its existing endowment structure and export to the inter- 21 national market, while importing whatever goods and services are not in its comparative advantage. Its trade dependency ratio will be endog- enous to its comparative advantage and will be larger than would be the case otherwise. Its economy will become competitive and its endowment

115 102 | New Structural Economics structure and industrial structure will be upgraded at the fastest pace possible. In the industrial upgrading process, the country will be able to tap into the advantage of backwardness by borrowing technologies and industries from advanced countries. The country will achieve a much faster rate of growth than the advanced countries, as its innovation cost will be smaller than that of countries already on the global technology frontier. Its economy will therefore achieve convergence with high-income countries. From that perspective, exploiting the world economy through stylized fact 1 ) is a result of the growth strategy that facilitates openness ( industrial upgrading according to the comparative advantage determined by the country’s endowment structure. stylized fact 2 ) is also a consequence of a Macroeconomic stability ( country following comparative advantage in its development strategy. If a country does so, its economy will be competitive. Its industries will be via- ble in an open, competitive market (Lin 2009). The upgrading of industries will mainly rely on its own capital accumulation process. The government will have a strong fi scal position, for several reasons: fi rst, it will reap the benefi ts of dynamic growth; second, there will be no need for subsidizing non-viable fi rms; and third, the economy will generate more job opportu- nities and less unemployment. The country will also be much less exposed to homegrown crises due to uncompetitive industries, currency mismatch, or fi scal crises. Because of its external competitiveness and limited reliance ows for growth, the country is also likely to have strong on capital infl external accounts. Therefore, the government will be in a strong position to adopt countercyclical measures if there are shocks to the economy from global crises. Recording high rates of saving and investment ( stylized fact 3 ) is another logical result of the new structural economics approach of developing industries that are consistent with comparative advantage. Such a strategy allows a developing economy to be most competitive and produce the larg- est possible economic surplus (profi ts). This yields the highest savings for the economy. Competitive industries also imply high return on investment, which in turn provides additional incentives to save and invest. Moreover, good public investments can enhance the economy’s growth potential, reduce transaction costs on the private sector, increase the rate of return

116 The Growth Report and New Structural Economics | 103 on private investment, and generate enough tax revenues in the future to liquidate the initial costs. Adopting a market system to allocate resources ( ) is a stylized fact 4 necessary condition for an economy to follow comparative advantage in its development. Most fi rms are set up to pursue profi ts. They will follow the economy’s comparative advantage in their decisions regarding the adop- ect the rela- tion of technology and entry into industries if relative prices refl tive scarcity of each factor in the endowment structure. This only happens in an economy with competitive markets (Lin 2009; Lin and Chang 2009). Therefore, a competitive market is the economy’s optimal mechanism for resource allocation at each stage of its development. Building committed, credible, and capable governments ( stylized fact 5 ), that is, creating a facilitating state, is also a condition for an economy to adopt a comparative-advantage-following strategy in its development pro- cess. For a developing economy to upgrade from one industrial structure to another, the government needs to play a facilitating role in improving soft and hard infrastructures and in overcoming the information, coordi- nation and externality issues. Therefore, a committed, credible and capable government is a precondition for sustainable growth. But capable states can also be seen as a consequence of that strategy: if the government’s goal is to facilitate a development process that is consistent with the country’s comparative advantage, its intervention will be implemented more easily and more successfully, which will strengthen its credibility. So a committed, credible and capable state can also be viewed as the outcome of the coun- try’s following its comparative advantage in its development. Beyond those stylized facts, the Growth Commission Report also iden- tifi ed “bad ideas” to be avoided by policymakers in developing countries. While the Report prudently offers the caveat that there are situations and circumstances that may justify limited or temporary resort to some of the policies listed under that category, it notes that “the overwhelming weight of evidence suggests that such policies involve large costs and their stated objectives—which are often admirable—are usually much better served through other means” (p. 68). These “bad ideas” include costly or unsustainable policy decisions such as subsidizing energy, relying on the civil service to deal with joblessness, providing open-ended protection,

117 104 | New Structural Economics reducing fi cits by cutting expenditures on infrastructure invest- scal defi ment, or allowing the exchange rate to appreciate excessively. Policy recommendations derived from the new structural economics approach would help developing country governments avoid such “bad ideas.” Energy subsidies for instance are adopted in most countries to sup- port nonviable fi rms (political economy rationale), or to help the poor (equity rationale). Large, costly and unsustainable government subsidies in developing countries arise from the fact that development strategies deviate substantially from their optimal industrial structure. If a country follows its comparative advantage in its development strategy, few of its state-owned or private enterprises will be nonviable, and there will be rms. Its economy will achieve dynamic no need to provide subsidies to fi growth, which would allow poverty to be reduced rapidly. There will be little need to subsidize the poor through price distortions. By growing fast, the economy will create many job opportunities. Viable private fi rms offer the best insurance against joblessness, so there will be no need to use public employment as a tool to deal with joblessness. Moreover, the government will not have to use open-ended protection to support or subsidize nonvi- able fi rms. Thanks to the country’s good economic performance, the government’s fi cation for the scal position is likely to be strong and there will be no justifi kind of erratic budget policies (expenditure cuts, public investment delays, payment arrears, salary freezes, etc.) that are often caused by large fi s- cal defi cits. Likewise, a government that implements a development strat- egy consistent with the country’s comparative advantage will not have to resort to an overvalued exchange rate as a means for subsidizing nonviable fi rms that are created in the framework of comparative-advantage-defying, import-substitution policies. 5. Conclusion The quest for economic growth has preoccupied economists and policy- th makers since at least the 18 century. Much progress has been achieved over the past 50 years, most notably on theoretical and empirical grounds. On the theoretical front, the analysis of endogenous technical innova- tion and increasing returns to scale has provided economists with a rich

118 The Growth Report and New Structural Economics | 105 general framework for capturing the broad picture and the mechanics of economic growth. On the empirical side, the availability of standardized data sets such as the Penn World Tables has stimulated interest in cross- country work that highlights systematic differences between high-growth and low-growth countries with regard to initial conditions, and policy and institutional variables. Yet, despite progress, policymakers around the world—especially in developing countries, still face diffi c actionable culty in identifying specifi policy levers that can help ignite and sustain the type of dynamic growth rates that are necessary to reduce poverty. In recent years, growth research- ers have responded to their concerns by trying to address various new cation of challenges: the lack of convergence among countries; the identifi robust determinants of economic performance; the design of the support- ing institutions for innovation and technological change, which are widely acknowledged to be the foundations for structural change and prosperity; and the identifi cation of binding constraints to growth, the evaluation of successful development programs through randomized control trials, with the goal of scaling them up whenever possible. By adopting a radically different approach to growth analysis, the Growth Report has made an important contribution to knowledge. It has identifi ed fi ve stylized facts (openness, macroeconomic stability, high rates of saving and investment, market mechanism, committed, credible and capable government) that can guide policymaking in developing countries. But in doing so, the Report has not disentangled causes and consequences. The new structural economics framework proposed in Lin (2010) helps explain the endogeneity and exogeneity issues surrounding these fi ve styl- ized facts. A central proposition that runs through this paper is that, devel- oping countries that implement economic policies in contradiction with their comparative advantage tend to perform poorly and suffer macro- economic instability. They do not exploit the benefi ts of globalization to the fullest. Typical features of such strategies are large budget defi cits due to government support of nonviable fi rms, infl ationary policies caused by excessive consumption, fi nancial repression, and over-valued exchange rates in the context of low productivity. By contrast, countries that adopt comparative-advantage-following strategies are typically in the position to

119 106 | New Structural Economics achieve dynamic growth. They rely on the market as the key mechanism for allocating resources at any given stage of development, and they have credible and capable governments. As a consequence of following their comparative advantage, they have an open economy, achieve macroeco- nomic stability, and record high rates of saving and investment. Notes † Célestin Monga, a native of Cameroon, is a Senior Advisor to the World Bank Senior Vice President and Chief Economist. In his 13-year career in the World Bank, he has held positions in both operations and in the research department. He has also served on the Board of Directors of MIT’s Sloan School of Man- agement (Sloan Fellows) and taught at Boston University and the University of Bordeaux (France). The losses precipitated by the fi nancial crisis have been enormous. Total 1. capitalization of world stock markets halved in 2008—about $32 trillion of wealth. The losses in household wealth during 2008 were about $11 tril- lion in the United States ($8.5 trillion in fi nancial assets and $2.5 trillion in housing assets) and were estimated at $1.5 trillion in the United Kingdom ($0.6 trillion in fi nancial assets and $0.9 trillion in housing assets). Losses of such magnitude have signifi cant wealth effects on consumption and sav- ings. Industrial production fell sharply in many developed and emerging countries and for the fi rst time since 1929, world trade contracted in 2009. Data sources: Global Stability Reports; IMF Survey Magazine, June 24, 2009. 2. There were 1.4 billion people living under the international poverty line of c growth pro- $1.25 a day before the global crisis. Applying the country-specifi jections to survey-based data and aggregating, World Bank experts calculate that the crisis will add 50 million people to the 2009 count of the number of people living below $1.25 a day and 57 million to the count of the number of people living under $2 a day. Given current growth projections for 2010, there will be a further impact on poverty in that year, with the cumulative impacts rising to an extra 64 million people living under $1.25 a day and 76 million more under $2 a day by 2010. 3. The report was released in 2008 and titled The Growth Report: Strategies for Sustained Growth and Inclusive Development. The Commission was consti- tuted of 20 experienced policymakers and two Nobel prize-winning econo- mists, Michael Spence and Robert Solow. Its work has been supported by the Governments of Australia, Sweden, the Netherlands, and the United Kingdom, the William and Flora Hewlett Foundation, and the World Bank Group.

120 The Growth Report and New Structural Economics | 107 4. The World Economy: Historical Statistics , avail- Maddison (2007). See also able at http://www.ggdc.net/maddison/. Western offshoots, a term used in Maddison (2001), include Australia, Canada, New Zealand, and the United States. 5. Conditional convergence is a key property in Solow-Swan models. It is con- ditional because in these models, the steady-state levels of capital and output per worker depend on characteristics that vary across economies: saving rate, population growth rate, and the position of the production function. Many recent empirical studies have suggested that many other sources of cross- country variations such as government policies or the initial stock of human capital should be included in the analysis. 6. The Cass (1965) and Koopmans (1965) versions of the neoclassical model, which built on Ramsey’s analysis of consumer optimization, attempted to pro- vide an endogenous determination of saving rates. While this extension helped preserve conditional convergence, it did not solve the problem of long-run growth being determined by exogenous technological progress. 7. See, in particular, Becker (1992); Heckman (2006); Lucas (2004). 8. This is the case not only in development economics but also in various sub- disciplines of macroeconomics. Following the 2008-09 global crisis, a heated debate erupted among economists over the pertinence of the dominant models and their policy prescriptions. See for instance Blanchfl ower (2009), Krugman (2009), or Stiglitz (2009). For an assessment of controversies in development economics over methodological and policy issues, see Deaton (2009) and Ravallion (2009). 9. From 1870 to 1990, the ratio of per capita incomes between the richest and the poorest countries increased by roughly a factor of fi ve. See Pritchett (1997). That is the view expressed by Barro and Sala-i-Martin (1992); and Baumol 10. (1986). Prescott (1999) is even more optimistic and expresses the view that continued divergence is not an option, and that the world distribution of income will eventually converge. 11. As Zagha et al. (2006) note, “whereas reforms can help achieve effi ciency gains, they will not put the economy on a sustained growth path unless they also strengthen production incentives and address market or government failures that undercut efforts to accumulate capital and boost productivity.” Pritchett (2006) suggests that economists abandon the quest for a single growth theory, and focus instead on developing a collection of growth and transition theories tailored to countries’ particular circumstances. 12. The methodology proposed for the identifi cation of the binding constraints to growth relies on shadow prices. Even in countries where data on shadow prices are widely available, it is not clear that this would accurately identify

121 108 | New Structural Economics areas in which progress is most needed in each country. For example, one could imagine a simple model of growth for a low income country where tech- nology and human capital are complementary. In such a country, the returns to education and technology adoption would both be low due to low levels of human capital and technology. An exclusive focus on shadow prices and an ignorance of cross-country comparison of levels would then suggest no need to improve education levels and encourage technology adoption. See Banerjee and Dufl o (2005). Bourguignon (2006) offers a compelling theo- 13. retical framework for making the same case. 14. Critics of RCTs point to the fact that they often do not start from a clear stra- tegic assessment of how a particular method would fi t the knowledge gaps of highest priority. See Ravallion (2009). These previous studies include, among others, the 15. (1993), East Asian Miracle the (2005), and the World Development Report on Agri- Growth in the 1990’s culture for Development (World Bank 2007). 16. rst, it The way the Commission organized its work was also quite unusual: fi defi ned themes and issues deemed important for growth and development. Then, it invited world renowned academics, practitioners and experts to author papers exploring the state of knowledge in these themes and issues; those were reviewed and discussed at workshops. A working group which interacted with academics and commissioners, reviewed and commented on papers throughout the process. The working group also supported the Chair- nal report by reviewing interim drafts and providing man in drafting the fi comments. 17. The list includes: Botswana, Brazil, China, Hong Kong (China), Indonesia, Japan, Korea, Malaysia, Malta, Oman, Singapore, Taiwan (China), and Thailand. 18. Because growth rates of this magnitude for such long periods were unheard of before the latter part of the 20th century, the authors acknowledge that their work could have been called a report on “economic miracles,” except that they believe the term is a misnomer: unlike miracles, sustained high growth can be explained and repeated. 19. competitive advantage refers to a situation where domestic A country’s industries fulfi ll the following four conditions: (i) They intensively use the nation’s abundant and relatively inexpensive factors of production; (ii) Their products have large domestic markets; (iii) Each industry forms a cluster; and (iv) domestic market for each industry is competitive (Porter 1990). A coun- try’s comparative advantage is the situation in which it produces a good or ser- vice at a lower opportunity cost than that of its competitors. Such a condition is based on the country’s possession of comparative advantage in that product or service determined by its endowment structure at any given time (Lin 2010).

122 The Growth Report and New Structural Economics | 109 The fi rst condition for competitive advantage listed by Porter supposes that the industries should be the economy’s comparative advantage determined by the nations’ endowments. The third and the fourth conditions will hold only if the industries are consistent with the nation’s competitive advantage. Therefore, the four conditions can be reduced to two independent conditions: the com- parative advantage and domestic market size. Between these two independent conditions, the comparative advantage is the most important because if an industry corresponds to the country’s comparative advantage, the industry’s product will have a global market. That is why many of the richest countries of the world are very small (Lin and Ren 2007). 20. We defi ne the development strategy here in the same way as Rodrik (2005), referring to policies and institutional arrangements adopted by the government in a developing country for achieving economic convergence with the living standards prevailing in advanced countries. 21. Exportable manufacturing goods are of particular importance, as they allow late-comers in the industrialization process to position themselves in industries where they have lower wages and other competitive advantages than more advanced economies. References Acemoglu, D. and J.A. Robinson, 2001. “The Colonial Origins of Comparative Development: An Empirical Investigation.” American Economic Review , September. Aghion, P. and P. Howitt, 1992. “A Model of Growth through Creative Destruction.” Econometrica , vol. 60, no. 2, March, pp. 323–351. Banerjee, A. and E. Dufl o, 2005. “Growth Theory through the Lens of Develop- Handbook of ment Economics.” In Philippe Aghion & Steven Durlauf (eds.), , vol. 1, chapter 7. Economic Growth Economic Growth , Cambridge, MIT Barro, R. J., and X. Sala-i-Martin, 1995. Press. Barro, R. J. and X. Sala-i-Martin, 1992. “Convergence.” Journal of Political Economy , vol. 100, no. 2. Baumol, W., 1994. “Multivariate Growth Patterns: Contagion and Common Forces as Possible Sources of Convergence.” In W. Baumol, R. Nelson and E. Wolf (eds.), Convergence of Productivity, Cross-National Studies and Historical Evidence , New York: Oxford University Press. Baumol, W., 1986. “Productivity Growth, Convergence, and Welfare: What the Long-Run Data Show.” American Economic Review , vol. 76, December, pp. 1072–85. Becker, G.S., 1992. “Education, Labor Force Quality, and the Economy: The Adam Smith Address.” Business Economics , vol. 27, no. 1, January, pp. 7–12.

123 110 | New Structural Economics Blanchfl ower, D., 2009. “The Future of Monetary Policy.” Speech at Cardiff University, March 24. Bourguignon, F., 2006. “Economic Growth: Heterogeneity and Firm-Level Disag- gregation.” PREM Lecture, Washington, D.C., World Bank, April. , 2d ed., Oxford, Cameron, R., 1993. A Concise Economic History of the World Oxford University Press. Cass, D., 1965. “Optimum Growth in an Aggregative Model of Capital Review of Economic Studies Accumulation.” , no. 32, July, pp. 233–240. Instruments of Development: Randomization in the Deaton, A., 2009. Tropics, and the Search for the Elusive Keys to Economic Development , Princeton, N.J., Princeton University Center for Health and Wellbeing, January. Dufl o, E. and M. Kremer, 2003. “Use of Randomization in the Evaluation of Development Effectiveness.” Paper prepared for the World Bank Operations Evaluation Department, July. Quarterly Journal of Glaeser, E. and A. Shleifer, 2002. “Legal Origins.” Economics , vol. 117, November, pp. 1193–1229. Growth Commission, 2008. The Growth Report: Strategies for Sustained Growth and Inclusive Development , Washington D.C. Hausmann, R., D. Rodrik, and A. Velasco, 2008. “Growth Diagnostics.” In N. Serra and J.E. Stiglitz (eds.), The Washington Consensus Reconsidered: Towards a New Global Governance , New York, Oxford University Press, pp. 324–354. Heckman, J.J., 2006. “Skill Formation and the Economics of Investing in Disadvantaged Children.” , vol. 312 (5782), June, pp. 1900–1902. Science Jones, C.I. and P.M. Romer, 2009. The New Kaldor Facts: Ideas, Institutions, Population, and Human Capital , working paper no. 15094, NBER, June. Jones, C. I., 1998. , New York, W.W. Introduction to Economic Growth Norton. Kaldor, N., 1961. “Capital Accumulation and Economic Growth.” In: F.A. Lutz and D.C. Hague (eds.), The Theory of Capital , New York, St. Martin Press, pp. 177–222. Koopmans, T.C., 1965. “On the Concept of Optimal Economic Growth.” In: , North The Econometric Approach to Development Planning, Amsterdam Holland. Krugman, P., 2009. “How Did Economists Get it so Wrong?” New York Times Magazine , September 2. Lin, J.Y., 2010. New Structural Economics: A Framework for Rethinking Development , Policy Research Working Papers, no. 5197, Washington D.C., World Bank. Lin, J.Y. 2009. Economic Development and Transition: Thought, Strategy, and Viability , Cambridge, Cambridge University Press.

124 The Growth Report and New Structural Economics | 111 Lin, J.Y. 2003. “Development Strategy, Viability and Economic Convergence.” , Vol. 53 (2), pp. 277–308. (A Economic Development and Cultural Change version of this article appears in chapter VI of this volume.) Lin, J. Y., 1995. “The Needham Puzzle: Why the Industrial Revolution Did Not Economic Development and Cultural Change Originate in China.” , Vol. 41 (2), 269–92. Lin, J.Y., and H. Chang, 2009. “DPR Debate: Should Industrial Policy in Developing Countries Conform to Comparative Advantage or Defy It?” Development Policy Review , Vol. 27, No. 5, pp. 483–502. (Reprinted in chapter II of this volume.) Lin, J.Y. and Ren, R., 2007. “East Asian Miracle Debate Revisited” (in Chinese). Jingji Yanjiu (Economic Research Journal), Vol. 42 (8), 4–12. Lucas, R.E., 2004. , Cambridge, MA., Harvard Lectures on Economic Growth University Press. Journal Lucas, R. E., Jr, 1988. “On the Mechanism of Economic Development.” , Vol. 22, 3–42. of Monetary Economics Maddison, A., 2007. Contours of the World Economy 1–2030 AD: Essays in Oxford University Press. Macro-Economic History. The World Economy: A Millennial Perspective , Paris, Maddison, A., 2001. OECD. Mokyr, J., 1990. The Lever of Riches: Technological Creativity and Economic Progress , New York and Oxford: Oxford University Press. Montiel, P. and L. Serven, 2008. “Real Exchange Rates, Saving, and Growth: Is There a Link?” Background Paper, Commission on Growth and Development, Washington D.C. North, D., 1981. . New York: Norton. Structure and Change in Economic History Porter, M. E., 1990. The Competitive Advantage of Nations , New York: Free Press. Prescott, E., 1999. “Interview with Edward Prescott.” In B. Snowdown and H. Conversations with Economists: Interpreting Macroeconomics , Vane (eds.), Edward Elgar. Pritchett, L., 2006. “The Quest Continues.” Finance and Development , vol.43, no. 1. Pritchett, L., 1997. “Divergence, Big Time.” , Journal of Economic Perspectives vol. 11. No. 3, summer, pp. 3–17. Ravallion, M., 2009. “Evaluation in the Practice of Development.” The World Bank Research Observer , vol. 24, no. 1, February, pp. 29–53. Rodrik, D., 2005. “Growth Strategies.” In Philippe Aghion & Steven Durlauf (ed.), Handbook of Economic Growth , vol. 1, chapter 14, pages 967–1014, Elsevier. Romer, P. M., 1986. “Increasing Returns and Long-Run Growth.” Journal of Political Economy , vol. 95, no. 5, October, pp. 1002–1037.

125 112 | New Structural Economics Romer, P. M., 1987. “Growth Based on Increasing reurns Due to Specialization.” , vol. 77, no. 2, May, pp. 56–62. American Economic Review Journal of Political Romer, P. M., 1990. “Endogenous Technological Change.” Economy , vol. 98, no. 5, October, part II, S71-S102. Theorists of Economic Growth from David Hume to the Rostow, W.W., 1990. , New York, Oxford Univer- Present—With a Perspective on the Next Century sity Press. Stiglitz, J., 2009. Freefall: America, Free Markets, and the Sinking of the World Economy , New York, W. W. Norton & Co. World Bank, 2007. World Development Report 2008: Agriculture for Develop- ment, Washington D.C. World Bank, 2005. Economic Growth in the 1990s: Learning from a Decade of Reform , Washington D.C. Zagha, R., I. Gill, and G. Nankani, 2006. “Rethinking Growth,” Finance and Development , vol. 43, no. 1. Zellner, A., 1979. “Causality and Econometrics, Policy and Policymaking,” Carnegie-Rochester Conference Series on Public Policy , vol. 10, pp. 9–54.

126 II Debate Should Industrial Policy in Developing Countries Conform to Comparative * Advantage or Defy It? A Debate between Justin Lin † and Ha-Joon Chang Justin Lin Introduction: Growth and Industrial Upgrading At a time when cyclical turbulence threatens to distract us from the longer-run goal of promoting sustained growth and development, I wel- come the chance to launch a discussion on this crucial topic with my friend Ha-Joon Chang. The Nobel laureate Robert Lucas (1988) has commented that ‘Once one starts to think about them [questions of eco- nomic growth], it is hard to think about anything else’. What he had in mind was the remarkable sustained growth in productivity and living Adapted from “Should Industrial Policy in Developing Countries Conform to Comparative Advan- * , tage or Defy It? A Debate Between Justin Lin and Ha-Joon Chang,” Development Policy Review Ha-Joon 27 (5), August 2009 (DOI: 10.111/J.1467-7679.2009.00456.X). © 2009 Justin Lin, Chang. © 2009 Overseas Development Institute. Reprinted with the permission of John Wiley and Sons / Blackwell Publishing. 113

127 114 | New Structural Economics standards that has characterised especially the countries of East Asia in recent decades, compared with the stagnation that, at least at that time, affl icted much of the rest of the developing world. To Professor Lucas’ comment, I would add that, once you start thinking about growth, it is hard not to focus on the continuous industrial and tech- nological upgrading that is characteristic of sustained economic growth. In theory, as has long been recognised, poor countries should be able to take advantage of their backwardness, by importing modern technology and institutions developed elsewhere. But while some countries have done this well, many others have been far less successful at industrial upgrading and therefore at poverty reduction. What is it that makes it possible in one or two generations for a country to go from exporting wigs and plywood to competing in the most technologically advanced sectors? The answer is not simply ‘a dynamic private sector’, though that is the ultimate driver. Historical examples make it clear that the answer must include effective government policies to catalyse private-sector growth. Governments have adopted a variety of measures to promote industrialisa- tion and technological upgrading, with a wide variety of results. Used well, the unique powers available to governments can be wielded to initiate and support long-run sustained improvements in factors and productivity. Our central task as development economists is to learn from these historical examples, as well as from economic theory and empirics, so that we can help today’s poorer countries to map out and follow a sustained growth path. In this essay, I shall argue that industrial upgrading and technological advance are best promoted by what I call a facilitating state—a state that facilitates the private sector’s ability to exploit the country’s areas of com- parative advantage. As I shall explain, the key is to make use of the coun- try’s current comparative advantage—not in the factors of production that it may have someday, but in the factors of production that it has now. The Case for a State Role: Market Failures That Block Innovation First, however, it is necessary to justify why the state needs to take the lead in development, because the facilitating-state approach requires government to do much more than a pure laissez-faire approach would allow. Developing economies are ridden with market failures, which can- not be ignored simply because we fear government failure. One such

128 Debate on Industrial Policy in Developing Countries | 115 market failure is caused by important information externalities . Economic innovations—whether they succeed or fail—yield information about prof- table market opportunities. But because much of this itable and unprofi information is available not only to the innovators themselves but also to competitors and potential imitators, who do not bear any of the costs of the innovation, it will tend to be undersupplied by the market. Government subsidies are one possible mechanism for encouraging innovation and off- rst-mover disadvantage. setting this fi co-ordination problems . Develop- A second market failure is caused by ing countries lag behind more developed countries, not only in technol- ogy and industrial structure, but also in human capital, infrastructure and institutions. For a country to climb up the industrial and technological lad- der, a host of other changes also need to take place: technologies become more complicated, capital requirements increase, the scale of production increases, the size of markets increases, and market exchanges increasingly take place at arm’s length. A fl exible and smooth industrial and technologi- cal upgrading therefore requires simultaneous improvements in education, fi nancial and legal institutions, and infrastructure. Individual fi rms clearly cannot internalise all these changes cost-effectively, and co-ordination among many fi rms to achieve these changes will often be impossible. For this reason, it falls to government either to introduce such changes itself or 1 to co-ordinate them. In these cases, the positive externalities of fi rm entry and experimenta- tion and needs for co-ordination can justify government intervention, and do so in a way that is perfectly compatible with neoclassical economic theory. It is true that the force of this argument is lessened by the high risk of government failures, but fear of poor governance does not absolve us of responsibility for trying to design effective strategies for facilitating development. Another Nobel laureate, W. Arthur Lewis (1955), correctly pointed out that ‘[N]o country has made economic progress without positive stimulus from intelligent governments’, even as he warned of the ‘mischief done to economic life by governments’. A half-century later, it remains true that there are few if any examples of governments that have succeeded with a purely laissez-faire approach that does not try to come to grips with market failures, and far more examples of rapid growth in coun- tries whose governments have led effectively. Therefore, it is incumbent

129 116 | New Structural Economics upon policy-makers and researchers to identify the most effective ways of promoting the productivity growth and change in industrial structure necessary for development. The Facilitating State: Helping the Private Sector Exploit Comparative Advantage In summary, these severe market failures can provide a rationale for government intervention to kick-start growth. But what kind of interven- tion? The key to answering that question is recognising that the optimal endogenous to the country’s endowment structure— industrial structure is in terms of its relative abundance of labour and skills, capital, and natu- rst upgrading ral resources. Upgrading the industrial structure requires fi the endowment structure, or else the resulting industrial structure will become a drag on development. Therefore the government’s role is to make sure that the economy is well launched on this endogenous process of upgrading. Let me explain this. The role of the facilitating state is to encourage the emergence of fi rms, industries, and sectors that, once launched, will make effective use of the country’s current comparative advantage. In many poor countries, that will mean focusing on labour- and/or resource-intensive types of production activities and services. Even with the increased inter- ows of recent decades, low-cost capital remains rela- national capital fl tively scarce, whereas labour and resources are relatively abundant and less costly. Focusing on labour- and resource-intensive production activities rms to be competitive in domestic and interna- allows poor countries’ fi tional markets. The facilitating state provides the necessary co-ordination to remove the barriers to the emergence of these fi rms and their related industries, and gives them a helping nudge to overcome externalities, but then is able to let them grow and advance organically because of their comparative advantage. As the competitive industries and fi rms grow, they will claim larger mar- ket share and create the greatest possible economic surplus, in the form of profi ts and salaries. When the surplus is reinvested, it earns the highest return possible as well, because the industrial structure is optimal for that endowment structure. Over time, this strategy allows the economy to accu- mulate physical and human capital, upgrading the endowment structure as

130 Debate on Industrial Policy in Developing Countries | 117 well as the industrial structure and making domestic fi rms more competi- tive over time in more capital- and skill-intensive products. While this approach sounds gradual— comparative-advantage-following and hence unsatisfying, when we consider the enormity of the poverty challenge—in fact progress is accelerated by the availability of technology and industries already developed by and existing in more advanced coun- tries. Firms in developing countries can at each stage in their development acquire the technologies and enter into industries appropriate for their endowment structure, rather than having to do frontier innovation them- selves. This ability to use off-the-shelf technology and to enter into existing industries is what has made possible the sustained annual GDP growth rates of 8 and even 10% achieved by some of the East Asian NIEs. The State as Midwife, not Permanent Nursemaid Too often, developing-country policy-makers have tried to take a short cut in this endogenous process of industrial and technological upgrading. They have fi xed their sights and their policies on an ideal industrial struc- ture that they associate with modernisation, but that structure is of course usually capital- and skill-intensive and is characteristic of a higher-income country than their own. As I have argued in my Marshall Lectures (Lin, 2009), industrial strategies of the often newly-independent developing countries in the 1950s and 1960s were informed by incorrect perceptions of the binding constraints on development. These countries adopted devel- opment strategies that placed a priority on capital-intensive heavy indus- tries, that is, industries that made intensive use of a factor that they largely lacked, and that neglected to use many of the factors that they had in great abundance, such as unskilled labour and natural resources. In effect, these policy-makers took the optimal industrial structure as something that they could impose exogenously, rather than something that results from the characteristics of the economy and changes over time. This approach can be thought of as comparative-advantage-defying , and it has high costs, both fi nancially and in terms of governance quality. To implement this strategy, governments have to provide substantial pro- tection and subsidisation to fi rms that are not viable without government subsidies and protection and cannot quickly become internationally com- petitive. Such fi rms cannot generate any real surplus for society. Without a

131 118 | New Structural Economics continuous fl nance improvements ow of surplus, it will be far harder to fi in the factors of production—notably, capital and skilled labour—that are in turn necessary to make a more advanced industrial structure viable over the medium term. By distorting market signals and shifting resources from competitive to noncompetitive sectors, high levels of protection and subsi- dies slow the country’s accumulation of physical and human capital. They also encourage fi rms to divert their energies from productive entrepreneur- ship into rent-seeking, which corrupts institutions and further slows capi- tal accumulation. Suppose the government tries to protect and subsidise the growth of capital-intensive industries, or other industries in which it has no compara- tive advantage. In that case, the accumulation of capital and the upgrading of endowment structure are retarded, slowing the upgrading of its optimal technology/industrial structure. Rather than serving as midwife to healthy new industries, it is likely to fi nd itself becoming a long-run nursemaid to sickly infant industries that never mature. The culture of rent-seeking that is likely to emerge will calcify the web of protection even more and make later reforms more diffi cult. Comparative vs. Competitive Advantage Putting domestic fi rms in a position to exploit the country’s compara- tive advantage may sound sensible but old-fashioned. How does exploit- ing comparative advantage compare with the promotion of ‘competitive advantage’, a strategy popularised by Michael Porter (1990) over the past two decades? In that literature, the four key sources of competitive advan- tage are: • sectors/industries that make good use of factors that are abundant domestically; • large domestic markets, to enable fi rms to achieve scale; • industrial clusters; and vibrant domestic competition, to encourage effi ciency and productivity • growth. But these requirements can be simplifi ed, in my view. First, consider domestic competition: if a country’s strategy defi es comparative advantage, it will generally be unable to enforce competition, because non-viable fi rms

132 Debate on Industrial Policy in Developing Countries | 119 will need to be protected. Industrial clusters will also be hard to build and sustain, because, unless the government gives subsidies and protec- tion, fi rms will not enter into this industry. However, the government will not be able to give subsidies and protection to many fi rms in an industry at the same time so as to form an industrial cluster. And if the country follows its comparative advantage, large domestic markets become unnec- rms should be able to compete on glo- essary, because the industries and fi bal markets. Thus these four requirements boil down largely to a single prescription: exploit your comparative advantage. Closing Notes I am happy to launch this exchange with my friend and colleague Ha-Joon. economic We both care deeply about understanding the roots of rapid growth and poverty reduction, and we have both thought carefully about the East Asian growth successes of the past two generations. There will doubtless be differences in the conclusions we reach on trade and indus- trial policy, but it is illustrative that neither of us questions the importance of a major state role in promoting economic development. Perhaps this is because in the countries we know most intimately—China and South Korea—a crucial ingredient in growth was a capable and largely devel- opmentally oriented state. The issue is identifying the key role played by the state in those countries and other rapid developers. My reading of these cases is that, while they took proactive steps to accelerate industrial upgrading, their success was spurred primarily by a state that made pos- sible the effective exploitation of comparative advantage at each stage of development. Notes † rst in an occasional series of DPR Debates, designed to illumi- This is the fi nate specifi c issues of international development policy. Each debate will bring together two well-known researchers or practitioners, giving them the opportu- nity, over three rounds, to test and challenge each other’s ideas. The debates are intended to be robust but accessible, rooted in rigorous research but useful to Development Policy Review . the wide readership of Ha-Joon Chang is a reader in the Political Economy of Development, Faculty of Economics, University of Cambridge. He is the author, inter alia , of Kick- ing Away the Ladder: Development Strategy in Historical Perspective (Anthem

133 120 | New Structural Economics Press, 2002), and Bad Samaritans: Rich Nations, Poor Policies, and the Threat (Random House, 2007). to the Developing World 1. Note that this is a different argument from the co-ordination role often pro- posed in the past for developing-country governments. That ‘big push’ line of rm’s viability depends on inputs from argument stressed that if each potential fi another fi rms may emerge. rm that does not yet exist, none of the potential fi In this case, the government can theoretically move the economy to a higher- welfare equilibrium with a big push that leads to the concurrent emergence of rms (see Rosenstein-Rodan, 1961; and Murphy et upstream and downstream fi al., 1989). But changing global conditions have made the traditional big-push argument less compelling. The reduction in transportation and information costs in recent decades has led to global production networks in which many countries, both developed and developing, produce only certain parts of a fi nal product according to each country’s comparative advantage. References Theory of Economic Growth . London: Allen & Unwin. Lewis, W. Arthur (1955) Lin, Justin Yifu (2009) Economic Development and Transition: Thought, Strategy, and Viability. Marshall Lectures, 2007/8. Cambridge: Cambridge University Press. Lucas, Robert E, Jr. (1988) ‘On the Mechanics of Economic Development’, 22 (1): 3–42. Journal of Monetary Economics Murphy, Kevin M., Shleifer, Andrei and Vishny, Robert W. (1989) ‘Industrializa- Journal of Political Economy 97(5): 1003–26. tion and the Big Push’, The Competitive Advantage of Nations . London: Free Porter, Michael E. (1990) Press. Rosenstein-Rodan, P. (1961) ‘Notes on the Theory of the “Big Push’, in H. S. Ellis and H. C. Wallich (eds), Economic Development for Latin America. New York: St Martin’s Press. Ha-Joon Chang It is a pleasure to debate this issue with Justin Lin, whose intellectual inter- ests are exceptionally wide-ranging and whose theoretical position, while fi rmly grounded in neoclassical economics, is never dogmatic. In his opening essay, Justin acknowledges the importance of industrial upgrading for economic growth and development. This is a point that is often missed by today’s development mainstream, which emphasises static allocative effi ciency; so Justin’s emphasis on industrial upgrading is really welcome.

134 Debate on Industrial Policy in Developing Countries | 121 On top of that, Justin also acknowledges the positive role that state intervention can play in promoting industrial upgrading, given important market failures that exist in the supply of new technological knowledge, such as the externalities generated by innovators experimenting with new things and the co-ordination failures across different input markets (for example, education, fi nance, legal institutions, and infrastructure). Justin also rightly warns against the possibility of government failure, but goes on to note that ‘there are few if any examples of governments that have succeeded with a purely laissez-faire approach that does not try to come to grips with market failures, and far more examples of rapid growth in countries whose governments have led effectively’. Up to this point, we are on the same platform. However, there are some important differences in our views. Our main difference is that, whereas Justin believes that state intervention, while important, should be basically about facilitating the exploitation of a country’s comparative advantage, I believe that comparative advantage, while important, is no more than the base line, and that a country needs to defy its comparative advantage in order to upgrade its industry. The concept of comparative advantage, fi rst invented by David Ricardo, is one of the few concepts in economics that is more than common sense (the others include Keynes’ notion of effective demand and Schumpeter’s concept of innovation). The beauty of this concept is that it shows how even a country with no absolute international cost advantage in any indus- try may benefi t from international trade by specialising in industries at which it is least bad. Indeed, it was the brilliance of Ricardo’s concept that rst drew me into economics. And as a guide to fi nding out the best way to fi maximise a country’s current consumption opportunities, given its current endowments , we cannot do better than that. As is well known, this theory, especially in the Heckscher-Ohlin- Samuelson version that Justin uses, is based on some stringent assump- tions. Of course, all theories have assumptions and therefore the fact that there are some stringent assumptions in itself cannot be a point of criti- cism. However, we still need to ask whether the particular assumptions made by a model are appropriate for the particular questions we happen to be asking. My contention is that, while the assumptions made by the HOS theory may be acceptable when we are interested in short-term

135 122 | New Structural Economics allocative effi nd out whether a country ciency (i.e., when we want to fi ciency), they are is exploiting its given resources with the maximum effi not acceptable if we are interested in medium-term adjustment and long- term development. First, let us look at the issue of medium-term adjustment. One of the key assumptions of the HOS theory is the assumption of perfect factor mobility (within each country). When this is assumed, no one loses out from changes in trade pattern caused by external shocks. So, if a steel mill shuts down because, say, the government reduces tariffs on steel, the resources employed in the industry (the workers, the buildings, the blast furnaces) will be employed (at the same or higher levels of pro- ductivity and thus higher returns) by another industry that has become relatively more profi table, say, the computer industry. No one loses from the process. However, in reality, factors of production are usually fi xed in their physical qualities. Blast furnaces from a bankrupt steel mill cannot be re-moulded into a machine making computers. Steel workers do not have the right skills for the computer industry: unless they are retrained, they will remain unemployed; at best, they will end up working in low-skill jobs, where their existing skills are totally wasted. In other words, even if the country as a whole benefi ts from trade liberalisation (which is not always the case even in the short run), the owners of factors of production that have low or no mobility are going to lose from it, unless there is delib- erate compensation. This is why trade liberalisation has produced so many ‘losers’, despite the prediction of HOS theory. This is a more serious problem in developing countries, where the com- pensation mechanism is weak, if not non-existent. In developed countries, the welfare state works as a mechanism partially to compensate losers from the trade-adjustment process through unemployment benefi t, guar- antees of health care and education, and even guarantees of a minimum income. In some countries, such as Sweden and other Scandinavian coun- tries, there are also highly effective re-training schemes for unemployed workers. In most developing countries, however, such mechanisms are very weak and often virtually non-existent. As a result, the victims of trade adjustment in these countries are not even partially compensated for the sacrifi ce that they have made for the rest of society.

136 Debate on Industrial Policy in Developing Countries | 123 If the assumption of perfect factor mobility makes HOS inadequate for the analysis of medium-term adjustment, its assumption about technol- ogy makes it particularly unsuited to the analysis of long-term economic development. The assumption in the HOS model is that there is only one best tech- nology for producing a particular product and, more importantly, that all countries have the same ability to use that technology. So, in the HOS the- ory, if Ecuador should not be producing BMWs, it is not because it cannot do it, but because doing it has too high an opportunity cost, as producing BMWs will use too much of its scarce factor of production—capital. However, this is assuming away the very thing that makes some countries developed and others not—namely, their differential abilities to develop and use technologies, or what is known as ‘technological capabilities’. In the end, the rich countries are rich and the poor countries are poor because the former can use, and develop, technologies that the latter cannot use, let alone develop. Moreover, the nature of the process of acquiring higher technological capabilities is such that a country trying to catch up with a more techno- logically advanced country needs to set up and protect industries in which it does not have comparative advantage. Why should that be the case? Can the country not wait until it accumulates enough physical and human capi- tal before it enters a more advanced industry that uses physical and human capital more intensively? Unfortunately, it cannot be done quite like that. Factor accumulation does not happen as an abstract process. There is no such thing as general ‘capital’ or ‘labour’ that a country can accumulate and that it can deploy wherever necessary. Capital is accumulated in concrete forms, such as machine tools for the car parts industry, blast furnaces, or textile machines. This means that, even if a country has the right capital-labour ratio for the automobile industry, it cannot enter the industry if its capital has been accumulated in the form of, say, textile machines. Likewise, even if a coun- try accumulates more human capital to justify its entry into the automobile industry, it cannot start making cars if all its engineers and workers were trained for the textile industry. Most (although not all) technological capabilities are accumulated through concrete production experiences, and at that in the forms of

137 124 | New Structural Economics ‘collective knowledge’ embodied in organisational routines and institu- tional memories. Even if a country has all the right machines, engineers, and workers (which is not possible anyway, as I have just explained), they rm overnight still cannot be combined into an internationally competitive fi because they actually need to be put through a (potentially very lengthy) learning process before they can acquire all the necessary technological capabilities. This is why Japan had to protect its car industry with high tariffs for nearly four decades, provide a lot of direct and indirect subsidies, and vir- tually ban foreign direct investment in the industry before it could become competitive in the world market. It is for the same reason that the electron- ics subsidiary of the Nokia group had to be cross-subsidised by its sister companies for 17 years before it made any profi t. History is full of exam- ples of this kind, from eighteenth-century Britain to late twentieth-century Korea. Of course, Justin is absolutely right in saying that deviating too much from one’s comparative advantages is to be avoided. Comparative advan- tage does offer a useful guideline in telling us how much the country is sac- rifi cing by protecting its infant industries. The more you deviate from your comparative advantage, the more you pay in order to acquire capabilities in new industries. However, this does not mean that a country should conform to its com- parative advantage, as Justin puts it. As I have argued, given the nature of the process of factor accumulation and technological capability-building, it is simply not possible for a backward economy to accumulate capabili- ties in new industries without defying comparative advantage and actually entering the industry before it has the ‘right’ factor endowments. Given this, a good neoclassical economist may be tempted to argue that a country should do a cost-benefi t analysis before deciding to enter a new industry, weighing the costs of technological upgrading against the expected future returns, using comparative advantage as the measuring rod. However, this is a logical but ultimately misleading way of looking at the process. The problem is that it is very diffi cult to predict how long the acquisition of the necessary technological capabilities is going to take and how much ‘return’ it will bring in the end. So it is not as if Nokia entered the electronics industry in 1960 because it could clearly calculate that it

138 Debate on Industrial Policy in Developing Countries | 125 would need to invest such and such amount in developing the electron- ics industry (through cross-subsidies) for exactly 17 years but then would reap huge future returns of such and such amount. Nokia probably did not t in electronics. It prob- think that it would take 17 years to make a profi ably did not know how large the eventual return was going to be. That is the nature of entrepreneurial decision-making in a world with bounded rationality and fundamental uncertainty. In other words, unless you actu- ally enter the industry and develop it, it is impossible to know how long it will take for the country to acquire the necessary technological capabilities to become internationally competitive. At the most general level, Justin and I share the same policy conclu- sions. We agree that industrial upgrading is necessary for economic devel- opment. We agree that it will not happen purely through market forces and will need government intervention. We also agree that the govern- ment should not push the economy too far away from its current structure too quickly. However, there are some important differences between the two of us. In the theory of neoclassical comparative advantage that Justin uses, the issue of limited factor mobility is neglected, resulting in the systemic under- estimation of the costs of trade liberalisation and hence the need for good redistribution mechanisms. More importantly, technological capabilities are missing from the theory, when they are really what distinguishes devel- oped countries from developing ones. Once we realise that a lot of tech- nological capabilities are acquired in an industry-specifi c manner through actual production experiences, we begin to see that it is by defi nition nec- essary to defy comparative advantage if a country is going to enter new industries and upgrade its industrial structure. And the length and the strength of such protection can be very large, as the examples of Toyota, Nokia, and countless other examples of successful infant-industry protec- tion show, and also inherently diffi cult to predict. Justin Lin Ha-Joon summarises well our key areas of agreement: government has a role to play in promoting technological and industrial upgrading, but there are risks in deviating too far from a country’s comparative advantage. Our

139 126 | New Structural Economics differences lie in how to defi ne ‘too far’—how to interpret trade models and historical evidence, and how to promote technological learning cost- effectively. Do Adjustment Costs and Technological Differences Really Undermine the Theory of Comparative Advantage? Ha-Joon argues that, because of imperfect factor mobility (in effect, adjustment costs) and simplifi ed assumptions about technology, argu- ments against infant-industry protection that are based on standard trade models (such as Baldwin, 1969) do not provide good guidance for policy. Clearly, there are frictions in labour-market adjustment to changes in c. industrial competitiveness, and physical capital is often industry-specifi Workers cannot move costlessly from one industry to another, or from one region to another, and many developing-country governments do little to compensate the losers. But adjustment costs can easily be incor- porated into standard trade models, without undermining the basic theory of comparative advantage (Mussa, 1978). Moreover, when a country loses comparative advantage in the existing industry, the indus- try-specifi c capital can be relocated in the form of foreign direct invest- ment to other countries, in what has been called a fl ying-geese pattern of economic development in East Asia and many other parts of the world (Akamatsu, 1962). Ha-Joon’s second point is that the Heckscher-Ohlin-Samuelson model incorrectly assumes that the same technology is available to producers in all countries. Yet the theory of comparative advantage does not hinge on having identical technology. Ricardo’s original model of comparative advantage recognised that England and Portugal had different technologies for producing wine and cloth, for example. Moreover, theoretical models are intended to be simplifi cations; in empirical trade models, richer and poorer countries are routinely recognised to be using different technologies. Thanks to the dramatic reduction in information and transportation costs, countries at different stages of development could even concentrate on different segments of the same industry, each using different technologies and producing different products according to comparative advantages. Take the information industry as an example: high-income countries, like the US, specialise in product/technology development; middle-income

140 Debate on Industrial Policy in Developing Countries | 127 countries, like Malaysia, concentrate on the fabrication of chips; and lower-middle-income countries, like China, focus on the production of spare parts and the assembly of fi nal products. Ha-Joon correctly observes that in reality trade liberalisation has pro- duced many losers in the past two decades. But this is because those coun- tries started with many industries that were inconsistent with their areas of comparative advantage, as a result of comparative-advantage-defying (CAD) strategies that their governments had adopted in the past. Remov- ing protection in a shock-therapy manner caused the collapse of nonviable rms. However, if, in the liberalisation process, the government liberalises fi the entry to sectors in which the country has comparative advantage, and phases out protections to the CAD industries gradually, as argued in my Marshall Lectures (Lin, 2009), the country can obtain a Pareto improve- ment by achieving stability and dynamic growth simultaneously in the proc- ess. Indeed, this is how China has managed its transition from a planned to a market economy. What Do We Learn about Technological Upgrading from the Success Stories? Underlying Ha-Joon’s line of argument is research that he and others have done on some of the most rapid industrialisers. Here, I will comment on the case of Korea with a brief note about his Nokia example as well. On the one hand, it is hard to argue that an active industrial and trade policy substantially hindered growth in the Republic of Korea. The coun- try did protect certain sectors with high trade barriers, and in some cases took an aggressive approach to industrial upgrading into capital-intensive industries. And over the past 40 years, Korea has achieved remarkable GDP growth rates, and has performed impressively on industrial upgrad- ing, into such industries as automobiles and semiconductors. Yet we should not overstate the extent to which Korea pushed ahead of its comparative advantage. In the automotive sector, for example, early in its growth period, Korean manufacturers concentrated mostly on the assembly of imported parts—which was labour-intensive and in line with their comparative advantage at the time. Similarly, in electronics, the focus was initially on household appliances, such as TVs, washing machines, and refrigerators, and then moved on to memory chips, the least technologically

141 128 | New Structural Economics complex segment of the information industry. Korea’s technological ascent has been rapid, but then so has its accumulation of physical and human capital, due to the conformity of Korea’s main industrial sectors to the exist- ing comparative advantages, and hence its changes in underlying compara- tive advantage. Equally important, the Korean government had a record of managing the protected sectors in ways that kept them subject to market discipline, which made large-scale deviation from the economy’s comparative advan- tage impossible. Industries benefi ting from protection and subsidisation were required to prove on export markets that their competitiveness was increasing over time. In addition, the government worked hard to make sure that Korean manufacturers could access intermediate inputs at world prices, for example through duty-drawback and exemption schemes and export-processing zones. So the government clearly recognised that com- parative advantage mattered, and that successful technological upgrading depended on fi rms being infl uenced by world prices for both inputs and outputs. The evidence indicates that Korea’s government served as a facili- tating state, as argued in my opening contribution. Let me add a footnote on the Nokia example, which I would interpret differently from Ha-Joon. Nokia’s technological upgrading—from timber company to footwear, to manufacturing for Philips and then manufacturer of own-brand household electronics, and fi nally to mobile-phone power- house—took place roughly in line with the growth of Finland’s stocks of physical and human capital. The Finnish government helped in ways that were far-sighted, but that I would interpret as consistent with the facili- tating role in a comparative-advantage-following strategy. It promoted R&D and competition in the mobile-phone industry in the 1970s, creating and building on a pan-Nordic mobile network (Ali-Yrkkö and Hermans, 2004). The learning-by-doing that Nokia gained was invaluable, but the core element of this strategy was not high levels of protection of the domestic market. Nokia apparently cross-subsidised the development of its mobile-phone division through profi ts in other areas. However, Finland’s per capita income in 1970, measured in 1990s’ purchasing power parity, had already reached 9,600 international dollars, which was at a level close to Germany’s 10,800 dollars in the same year (Maddison, 2006). Nokia’s decision is wholly consistent with a model of technological/ industrial

142 Debate on Industrial Policy in Developing Countries | 129 upgrading by a profi rm in an open, competitive, t-maximising private fi high-income country. Are Dynamic Comparative Advantage and Infant-Industry Protection Sound Foundations for Industrial Policy? Finally, we should turn to the question of Ha-Joon’s theoretical founda- tion for using trade policy as a tool for promoting industrial upgrading. His argument is based on the idea of dynamic comparative advantage and infant-industry protection. Nevertheless, if industrial upgrading pro- ceeds step by step in conjunction with changes in comparative advantage, learning costs are lower than if the country attempts a big leap. As an anal- ogy, think of mathematics learning. Typically, a student starts by study- ing algebra, then proceeds through calculus to real analysis. If instead he started with real analysis, even though he might eventually master it, the learning costs would most likely be much higher than otherwise. Similarly, if a fi rm begins by manufacturing bicycles, then learns to make motor- cycles, and eventually moves into making automobiles, the total learning costs will probably be much lower than if it starts with the daunting task of mastering the effi cient production of automobiles. When a government chooses to provide protection or incentives to fi rms in sectors that may be viable only in twenty or more years, it will inevitably have to draw resources from fi rms in areas of current comparative advan- tage. This will reduce the surpluses they earn, and will therefore slow capi- tal accumulation and the upgrading of the country’s endowment structure and comparative advantage, making the infant industry stay as an infant much longer than otherwise (Baldwin, 1969; Saure, 2007). Furthermore, excessive protection risks institutionalising a culture of rent-seeking. Given how important the quality of institutions and gover- nance is to development, the indirect effects of protection through poor governance may be even more damaging than the direct effects. References Ali-Yrkkö, Jyrki and Hermans, Raine (2004) ‘Nokia: A Giant in the Finnish Innovation System’, in Gerd Schienstock (ed.), Embracing the Knowledge Economy: The Dynamic Transformation of the Finnish Innovation System . Cheltenham: Edward Elgar Publishing.

143 130 | New Structural Economics Akamatsu, Kaname (1962) ‘A Historical Pattern of Economic Growth in Devel- The Developing Economies, oping Countries’, Preliminary Issue No.1: 3–25. Baldwin, Robert E. (1969) ‘The Case Against Infant-Industry Tariff Protection’, Journal of Political Economy 77 (3): 295–305. Lin, Justin Yifu (2009) Economic Development and Transition: Thought, Strategy and Viability . Cambridge: Cambridge University Press. The World Economy . Paris: OECD. Maddison, Angus (2006) Mussa, Michael (1978) ‘Dynamic Adjustment in the Heckscher-Ohlin-Samuelson Model’, 86 (5): 775–91. Journal of Political Economy Journal of Saure, Philip (2007) ‘Revisiting the Infant Industry Argument’, 84 (1): 104–17. Development Economics Ha-Joon Chang Even though we come from different theoretical traditions, Justin and I agree on the broad framework for the analysis of industrial upgrading. To be sure, we have our differences. While we may both be of the view that comparative advantage is an important principle, I see it as only a ‘base line’, whereas Justin thinks it should be stuck to very closely, if not per- fectly. We agree on the importance of adjustment costs and technological learning, but we differ in how important we think they are and we analyse them in different ways. However, these are differences whose clarifi cation actually helps us think through some of the fi ner points and advances our knowledge, rather than those that lead to unproductive bickering. First, on adjustment costs. Justin is right in saying that these costs can be (and occasionally have been) incorporated into mainstream trade models. But my question is: if adjustments costs are important, why have they been so much neglected in practice by mainstream economists, who keep rec- ommending trade liberalisation with only perfunctory, if any, attention to adjustment costs? It is not enough to say that adjustment costs can be incor- porated into mainstream models. Intellectual leaders in the mainstream camp, like Justin, should encourage people actually to do it and then fully apply the results in designing trade-policy reform. The same applies to the assumption of identical technology. If it is better not to assume identical technology (as Justin implicitly acknowledges), why do mainstream econo- mists keep using the HOS version of comparative advantage rather than

144 Debate on Industrial Policy in Developing Countries | 131 the Ricardian version, in which differences in technology determine the comparative advantages of different nations? As for Justin’s point that even activity-specifi c assets do not need to lose their value entirely in the adjustment process because they can be shifted to another country, I thank him for reminding me of this important point. However, this mainly applies to physical assets and then only to a limited extent. Not all physical assets can be shipped abroad and many of them need complementary assets and skills if they are to realise their full produc- c skills (or human capital, tive potential. Moreover, workers with specifi if you like) cannot move to the ‘next-goose’ country, except for a limited number of technicians who may be called upon to advise the factories in the new host countries. For the workers, it is cold comfort to learn that the physical assets they used to work with may preserve some of their value by moving to another country. To make things worse, the workers usually have fewer and less diversifi ed assets (even including their own human capital) than the owners of physical assets, so they are less capable of cop- ing with the consequences of the adjustment, even if they are subject to the same magnitude of shocks (in proportional terms) as the capitalists. Thus seen, Justin’s ‘fl ying geese’ point does not lessen the need to incor- porate adjustment costs into trade policy design. If anything, it actually highlights the need to better design compensation schemes for the workers with specifi c skills (for example, subsidised re-training programmes). Justin argues that trade liberalisation in the last two decades has pro- duced many losers ‘because those countries started with many industries that were inconsistent with their areas of comparative advantage’ because of wrong policies in the past. This may often (although not always) have been the case, but it does not justify the way trade liberalisation has been conducted in the last two decades. If we know that a country has deviated ‘too much’ from its comparative advantage, the prudent course of action will be not to try to liberalise trade too much too quickly, as otherwise the adjustment costs will be very high. Two wrongs do not make a right. This naturally leads me to Justin’s second point—the challenge of decid- ing how much to deviate from comparative advantage. Using the Korean and Finnish examples, he argues that these countries succeeded because they did not deviate from their comparative advantages too much. He is

145 132 | New Structural Economics right in saying that Korea’s move along the ‘ladder’ of international divi- sion of labour has often been carried out in small, if rapid, steps. Although I do not fully agree with this characterisation (for example, the moves into industries like steel and shipbuilding were big leaps, with virtually no ‘intermediate’ steps), I also agree that making excessive leaps can result in excessive learning costs. Thus seen, we could suppose some kind of inverted-U-shaped relation- ship between an economy’s deviation from comparative advantage and its cient in the short run, but growth rate. If it deviates too little, it may be effi its long-term growth is slowed down, as it is not upgrading. Up to a point, therefore, increasing deviation from comparative advantage will accelerate growth. After a point, negative effects of protection (for example, exces- sive learning costs, rent-seeking) may overwhelm the acceleration in pro- ductivity growth that the ‘infant’ industries generate, resulting in negative growth overall. I think Justin would probably agree with the above way of seeing things. However, there is one big disagreement between the two of us in applying this idea. It is the question of ‘how much (deviation from comparative advantage) is too much?’ (or where is the apex in the inverted-U curve?) Using the Finnish example, Justin says that Nokia was justifi ed in mov- ing into the electronics industry, as Finland was already a pretty rich coun- try, with per capita income (in international dollars) only 13% lower than that of Germany in 1970 ($9,577 vs. $10,839). However, the relevant year is not 1970 but 1960, which is when the electronics subsidiary of Nokia was set up, and in that year the income gap with Germany was much 1 greater, at 23% ($7,705 vs. $6,230). Anyway, these fi gures are purchas- ing power parity (PPP) fi ate a poorer country’s gures, which tend to infl income. PPP fi gures are preferable if we are interested in measuring com- parative living standards, but if we are interested in comparative advantage in international trade, current dollar fi gures, rather than PPP fi gures, are better fi gures to use. 2 In 1960, If we use current dollars, the picture becomes quite different. the per capita income of Finland was only 41% that of the US, the fron- tier country in electronics and overall ($1,172 vs. $2,881). This does not look like the case of a country sticking closely to comparative advantage. If Finland’s decision regarding Nokia does not look ‘wrong’ enough, how

146 Debate on Industrial Policy in Developing Countries | 133 about Japan? In 1961, the per capita income of Japan was a mere 19% that of the US ($563 vs. $2,934), but Japan was then protecting and promoting all sorts of ‘wrong’ industries—automobiles, steel, shipbuilding, and so on. For an even more dramatic example, take the case of South Korea. Its (then) state-owned steel mill, POSCO, which had been set up in 1968, started production in 1972, when its per capita income was a mere 5.5% 3 that of the US ($322 vs. $5,838). To make it worse, in the same year, South Korea decided to deviate even further from its comparative advantage by launching its ambitious Heavy and Chemical Industrialisation programme, which promoted shipbuilding, (home-designed) automobiles, machinery, and many other ‘wrong’ industries. Even as late as 1983, when Samsung decided to design its own semiconductors, Korea’s income was only 14% that of the US ($2,118 vs. $15,008). Does this sound like a ‘comparative- advantage-conforming’ strategy, as Justin calls it? A further diffi culty with Justin’s argument is that in all these examples of defi ance of comparative advantage, the market gave Finland, Japan, and Korea unambiguous signals that they should not promote those industries; all the companies in those industries ran losses or earned profi ts on paper only because they were subsidised by profi table companies in the same business group and/or by the government (directly through subsidies and indirectly through protection and entry restrictions). But if Justin thinks Nokia’s experience is ‘consistent with a model of technological/industrial upgrading by a profi rm in an open, competitive, t-maximising private fi high-income country’, is he saying that market signals are not to be taken seriously? Within the neoclassical framework, how else are we to judge whether or not a country is following its comparative advantage, except by looking at profi ts and losses made by the relevant companies? I think that, deep down, Justin and I actually agree. We agree that coun- tries should deviate from comparative advantage to upgrade their economy, although Justin thinks this deviation should be fairly small and I think it can be big. However, because Justin is too faithful to neoclassical econom- ics, he has to say that a country with an income level that is only 5% of the frontier country moving into one of the most capital-intensive industries (Korea and steel) is consistent with the theory of comparative advantage. Once Justin frees himself from the shackles of neoclassical economics, our debate will be more like two carpenters having a friendly disagreement

147 134 | New Structural Economics over what kind of hinges and door handles to use for a new cabinet that they are building together, on whose basic design they agree. Notes All the PPP income fi gures are from Maddison (2006: Tables 1-c for Europe, 1. 2-c for the USA, 5-c for South Korea). 2. All the current dollar income fi gures are from http://www.nationmaster.com/ red/graph/eco_gdp_percapeconomy-gdp-per-capita, which draws on the World Bank and the CIA data. 3. Even in PPP terms, its income was only 16% that of the US ($2,561 vs. $15,944). Justin Lin I’ve enjoyed this extended exchange, which has given us a chance to highlight our differences, while recognising our points of agreement. In response to Ha-Joon’s latest submission, it is useful to focus on two points: the dynamic nature of industrial upgrading, and the role of government in promoting it. Industrial Upgrading as a Dynamic Process First, let me reiterate that innovation is necessary for industrial upgrad- ing and development, and that government has a role in supporting that innovation for the positive externalities innovation brings to an economy’s development. It is hard work to climb technological ladders, to use a meta- phor employed by Ha-Joon and others. The developed countries that are at the technology frontiers recognise this. They provide considerable public support to fi rms in their frontier industries—directly by giving a patent to a new invention and sometimes also through defence contracts; and indi- rectly through supporting basic research at universities, which ultimately spills over into product development and benefi ts fi rms and industries at the technological frontier. As inside-the-frontier innovations in developing countries involve similar risk and externalities, public support can be desir- able and justifi able in that context too. Well-thought-out subsidisation is not only consistent with the role of a facilitating state, but is even implied. However, as pointed out in my fi rst essay, the subsidies to compensate for

148 Debate on Industrial Policy in Developing Countries | 135 an innovative fi rm’s externality will be small compared with those that rms in industries that go against would be required to protect non-viable fi an economy’s comparative advantage. Second, industrial upgrading in an economy is a continuous process. Although government needs to help solve externality and co-ordination problems for the pioneer fi rms, their upgrading is based on the fact that the economy has successfully exploited its existing comparative advantages and its endowment structure, as well as comparative advantage shifting. When the Korean government started its world-class state-owned Pohang Iron and Steel company in 1968, to use Ha-Joon’s example, that invest- ment was built upon the success of development in garments, plywood, wigs, footwear, and other labour-intensive industries. With the success of those labour-intensive industries, Korea accumulated capital and the capi- tal intensity of its endowment structure increased. From the perspective of the comparative-advantage-following strategy, the upgrading of a few fi rms into more capital-intensive industries became a necessity. The ‘fl ying geese’ metaphor is useful in the domestic context as well as the international one: when an economy follows its comparative advan- tage in economic development, its endowment structure and comparative advantage change dynamically. Some fi rms need to play the role of a ‘lead goose’ so as to pioneer the upgrading into new industries. This appears to be one area of difference between Ha-Joon and me: I see the lead goose as a small but important leading wedge in a dynamic process, whereas he sees it as a more quantitatively signifi cant part of the economy making larger discrete technological leaps. The quantitative difference can cause a quali- tative difference. When the lead goose is a small wedge in the dynamic pro- cess, the nature of the economy is consistent with its comparative advan- tage. Unlike the upgrading in the comparative-advantage-defying strategy discussed in my fi rst essay, the subsidies to the lead goose can derive mostly from intra-fi rm profi ts obtained in the operations of other products in competitive markets, as in the case of Samsung and Nokia. Third, the global technological frontier is continually being pushed out- ward. Industries such as steel production and shipbuilding were among the most advanced industries globally in the nineteenth century, but by the mid- twentieth century they no longer held this leading-edge position. Compared with new industries, such as aviation, information, and heavy chemicals,

149 136 | New Structural Economics their technologies had become mature. Investments in these mature indus- tries required a large amount of capital, compared with traditional labour- intensive industries, but their capital intensities were much lower than in the new emergent industries. It is therefore not surprising that, with some government support for overcoming the diffi culty of mobilising a large amount of capital in an economy with an underdeveloped fi nancial sector, these industries are viable in countries that have achieved or are approach- ing lower-middle-income status. When Korea established Pohang Iron and Steel, its per capita income in dollar terms was just 5.5% that of the US, as pointed out by Ha-Joon. I would also like to mention that China had become the largest producer of steel in the world by 2000, at a time when 1 its per capita income in dollar terms was only about 2.5% of the US level. Korea and China were able to succeed in the steel industry at a relatively low income level because steel had become a mature and relatively low capital-intensity industry in the global industrial spectrum. A related point is that, within industries, some segments are more acces- sible to developing countries than others. Manufacturing includes various stages—product R&D, design, production of complex parts, production of simpler parts, and assembly—and they all have different factor require- ments and are consistent with different patterns of comparative advan- tage. Countries therefore scale the ladder of technological sophistication and capital intensity within industries dynamically in a fl ying-geese pattern as well. Samsung’s entry in 1983 into the development of the 64-kilobit dynamic random access memory (DRAM) chip, which was relatively low- tech on the microchip spectrum at that time and was produced with the proprietary technology from Micron of the United States and Sharp of Japan, was built on some 15 years of successful operations in consumer electronics. It is worth noting that, in spite of the success of its entry into microchips in 1983, Samsung, on the one hand, has not entered the more complicated and advanced CPU chips and, on the other hand, has main- tained its successful operations in consumer electronics. Facilitating Comparative Advantage, with Equal Parts Vision and Realism To sum up my argument in this exchange, I reiterate that the comparative- advantage-following approach is dynamic in nature and the state should

150 Debate on Industrial Policy in Developing Countries | 137 play a facilitating role in that process. This means that economic develop- ment in a country should exploit pragmatically the existing opportunities embedded in the country’s areas of comparative advantage, while recogn- ising the potential for industrial upgrading when those areas of compara- tive advantage have been exploited. Industrial upgrading is an innovation involving risks and externalities, whether in developed or developing coun- tries, and thus requires the government to play a facilitating role. Gov- ernments in developing countries can play that role through the channels of information, co-ordination and compensation for externalities, as dis- rst essay. cussed in my fi Ha-Joon’s rhetorical jibe notwithstanding, neoclassical economics is simply a useful tool in all this, not a constraint. It is fl exible enough to model the externalities, dynamics, and co-ordination failures that give the government a role to play, while also providing the metrics to judge whether government is supporting industries that take the economy too far from its areas of comparative advantage. Without the former, developing coun- tries may lack the wisdom to seize opportunities to develop competitive industries and lay the foundation for sustainable industrial upgrading and development. But without the latter, as the historical record emphasises, governments can make any number of costly mistakes, most notably by funding large-scale, unrealistic and unsustainable comparative-advantage- defying projects and industries. By facilitating industrial upgrading where domestic fi rms will be able to survive and thrive, government can intervene in ways that yield the greatest social returns. Notes 1. Here I use Ha-Joon’s method of comparison based on market exchange rates, but PPP incomes are the more appropriate basis for comparison, in my view. Although market exchange rates govern international trade, PPP fi gures are better indicators of the level of development and capacity of an economy, and are therefore more relevant for discussions of industrial upgrading. Ha-Joon Chang As the exchange shows, Justin and I agree on many things. Both of us rec- ognise that ‘climbing up the ladder’ is a hard slog that involves more than ‘getting the prices right’. It requires, inter alia , intelligent industrial policy,

151 138 | New Structural Economics organisation building, and efforts to accumulate technological capabili- ties through R&D, training and production experiences. We agree that, in climbing up the ladder, a country can skip some rungs with the help of industrial policy, but that it can slip, fall, and even be destroyed, if it tries to jump too many rungs. The principle of comparative advantage, Justin says and I agree, can tell us what a country’s ‘natural’ climbing ability is and thus help us to see how much risk it is taking in trying to skip a certain number of rungs. However, we have some important differences. exible enough to allow Justin emphasises that neoclassical economics is fl us to deal with all the complex issues arising during the development pro- cess. I think it is not enough. I agree that neoclassical economics is a lot more fl exible than is usually recognised by many of its critics and that it can justify most types of state intervention, even of pretty ‘unorthodox’ kinds. After all, in the 1930s, the famous Marxist Oskar Lange tried to justify socialist planning with a neoclassical general equilibrium model. However, the rational-choice, individualistic foundation of neoclassical economics limits its ability to analyse the uncertain and collective nature of the technological learning process, which is at the heart of economic development. I have emphasised the importance of bounded rational- ity, fundamental uncertainty (and not just calculable risk), and collective knowledge in the development process. This means that the industrial upgrading process will be messy. It will not be possible for a country to follow market signals closely and enter an industry when its factor endow- ments are right, as will happen with the smooth comparative-advantage- conforming strategy that Justin advocates. In the real world, fi rms with uncertain prospects need to be created, protected, subsidised, and nur- tured, possibly for decades, if industrial upgrading is to be achieved. In practical terms, my difference with Justin lies primarily in the extent to which we think the defi ance of comparative advantage is advisable. While Justin believes that the skipping of the rungs in climbing the ladder should be very small (‘comparative-advantage-conforming’ in his words), I believe that it can be, and sometimes has to be, large (‘comparative-advantage- defying’ in his words). There is, of course, a chance that such an attempt

152 Debate on Industrial Policy in Developing Countries | 139 may not succeed, but that is the nature of any venture into new activities, whether purely private or assisted by the state. Justin is right in pointing out that Korea’s forays into industries like steel, shipbuilding, and microchips were not as dramatic as they may have looked at fi rst sight. By the time Korea entered them, steel and shipbuilding were technologically mature, although I am not sure whether that neces- sarily means lower capital intensity, as Justin assumes; technological matu- rity will increase capital intensity by leading to a greater embodiment of technologies in capital goods, while it may reduce capital intensity by low- ering the relative prices of the relevant capital goods. Even in microchips, the segment that Korea entered, namely, the DRAM chip, was (and still is) technologically the easiest. However, all these still do not mean that Korea’s entry into these indus- tries was comparative-advantage-conforming. First of all, technologically mature or not, the fact remains that industries like steel were still way too capital-intensive for Korea at the time (or, for that matter, today’s China). More interestingly, Korea’s success in steel was owed especially to the fact that it reaped the maximum scale economy by deliberately going for the most up-to-date and capital-intensive technology available (bought from New Nippon Steel). Most importantly, the market clearly signalled that these were ‘wrong’ industries to enter, by making the producers run losses or forcing the gov- ernment or the relevant business groups to manufacture ‘artifi cial’ profi ts by protecting and subsidising them. I do not think any version of neoclas- sical economic theory can justify protecting an industry for four decades (for example, Japanese and Korean cars) or cross-subsidising a loss-making subsidiary for 17 years (Nokia). I have learned a lot from this exchange with Justin. We come from dif- ferent intellectual traditions, but we have conducted a cordial and very productive debate that bears no bitterness or petty point-scoring. I wish there could be more exchanges like this in the pages of Development Policy Review and elsewhere.

153

154 III Growth Identification and Facilitation: The Role of the State in the Dynamics of Structural Change

155

156 III Growth Identification and Facilitation: The Role of the State in the Dynamics of Structural Change *† with Célestin Monga 1. Introduction The recent global crisis, the most serious since the Great Depression, has forced economists and policy-makers to rethink their approaches to macroeconomic management. For developing countries, in the midst of a * Adapted from “DPR Debate: Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change,” Development Policy Review, 29 (3), May 2011 (DOI: 10.1111/ j.1467-7679.2011.00534.x). © 2011 Lin, J., Monga, C., te Velde, D. W., Tendulkar, S. D., Amsden, A., Amoako, K. Y., Pack, H., and Lim, W. © 2011 Overseas Development Institute. Reprinted with the permission of John Wiley and Sons / Blackwell Publishing. 143

157 144 | New Structural Economics fi nancial and economic turmoil not of their own making, the road ahead is likely to be rocky. Because of the sluggish recovery in high-income coun- tries and the heavy cost of the crisis, they will have to confront a more dif- cult global environment for their exports and fi nancing conditions. Yet, in fi order to continue tackling the enormous challenge of poverty and achieve convergence, they must return to the pre-crisis path of dynamic growth. How to promote economic growth has been a main topic for economic The Wealth of Nations discourse since the publication of Adam Smith’s in 1776. Market mechanisms have proved essential for valuing the basic ingredients for production and providing the right price signals and the appropriate incentive system for the effi cient allocation of resources. How- ever, modern economic growth—a fairly recent phenomenon (Maddison, 2001)—is a process of continuous technological innovation, industrial upgrading and diversifi cation, and improvements in the various types of infrastructure and institutional arrangements that constitute the context for business development and wealth creation (Kuznets, 1966). Historical evidence shows that all countries that have successfully transformed from agrarian to modern advanced economies—both the old industrial powers of Western Europe and North America, and the newly industrialised economies of East Asia—have had governments that played a pro-active role in assisting individual fi rms in overcoming the inevitable co-ordination and externality problems. In fact, the governments in high- income countries today continue to do so. However, the sad fact is that almost every government in the developing world has attempted, at some point, to play that facilitating role, but most have failed. In this article, we argue that these pervasive failures are mostly due to government inability to come up with good criteria for identifying industries appropriate for a given country’s endowment structure and level of development. In fact, government propensity to target industries that are too ambitious and not aligned with a country’s comparative advantage largely explains why their 1 By contrast, attempts to ‘pick winners’ have resulted in ‘picking losers’. spontaneously or intentionally, the governments in successful develop- ing countries have typically targeted mature industries in countries with an endowment structure similar to and a level of development not much more advanced than theirs. The main lesson is straightforward: to facili- tate industrial upgrading and diversifi cation, government policy must be

158 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 145 anchored in industries with latent comparative advantage so that, once the new industries are established, they can quickly become competitive domestically and internationally. This article broadens the scope of analysis of industrial policy by introducing an important distinction between two types of government interventions. First are those that facilitate structural change by aiming to provide information, compensate for externalities, and co-ordinate 2 improvements in the ‘hard’ and ‘soft’ infrastructure that are needed for the private sector to grow in a manner consistent with the dynamic change in the economy’s comparative advantage. Second are those whose objec- tive is to protect selected fi rms and industries that are in defi ance of the comparative advantage determined by the existing endowment structure— either in new sectors that are too advanced or in old sectors that have lost comparative advantage. The remainder of the article is structured as follows: Section 2 explains the importance of well-functioning markets and the rationale for a facili- tating state in the process of dynamic economic growth. Section 3 briefl y reviews some important lessons from early industrial development strate- gies around the world and analyses the role of the state in the process of structural change in today’s advanced economies. It also examines similar attempts by developing-country governments to adopt policy interventions to facilitate industrial upgrading and economic diversifi cation, and analy- ses the reasons for their success or failure. Building on the foundations of new structural economics (Lin, 2010), Section 4 provides a framework for formulating industrial policy based on a new approach entitled ‘growth identifi cation and facilitation’. Section 5 offers some concluding thoughts. 2. Structural Change, Effi cient Markets and a Facilitating State Economists have long been intrigued by the mystery of modern economic growth, typically observed through the seemingly divergent evolution of the change in per capita gross domestic product among countries. Since taking off sometime around 1820 (Maddison, 2001), the world growth rate has risen more or less steadily, peaking during a ‘golden age’ (1950–73) when it averaged almost 3% per year. But such progress has been uneven across regions, countries, and time. Sustained growth has led to improved

159 146 | New Structural Economics living standards, fi rst in Western Europe, North America and Japan, and more recently in newly industrialised (NIEs) and other emerging market economies. Cross-country income distribution that initially widened (with the proportional gap between the richest and poorest countries growing more than fi vefold from 1870 to 1990) (Pritchett, 1997) has slowed in recent decades among groups of countries. With the narrowing of the top end of the distribution, there seem to be ‘convergence clubs’ among nations (Evans, 1996). Still, many of the poorest countries, especially in Africa, are excluded from the convergence process. Modern growth theory has attempted to explain the diverging paths followed. Despite differences in approach and methodology, there is wide consensus that the variation of living standards across countries and time ects differences in the rate of capital accumulation and produc- mostly refl tivity growth. Empirical studies carried out from the perspective of develop- ment accounting show that, among these two broad factors, ‘productivity differences among countries are the dominant explanation for income dif- ferences. Similarly, differences in productivity growth are the most impor- tant explanation for differences in income growth rates among countries’ (Howitt and Weill, 2010: 43-4). Over the long term, productivity growth is 3 associated with technological and structural change, namely, to reduce the costs of producing the same outputs with better knowledge and to relocate 4 resources from lower value-added to higher value-added industries. It can therefore be said that continuous technological innovation, indus- trial upgrading, economic diversifi cation and an acceleration of income growth are the main features of modern economic growth (Kuznets, 1966; 5 Maddison, 2006). Each country at any specifi c time possesses given fac- tor endowments consisting of land (natural resources), labour and capital (both physical and human), which are the total budgets that the country can allocate to primary, secondary and tertiary industries to produce goods and services. These are changeable over time, and conceptually it is useful 6 Both to add both ‘hard’ and ‘soft’ infrastructure to the mix (Lin, 2010). types are essential to the competitiveness of domestic fi rms because they affect transaction costs and the marginal rate of return on investment. At any given point in time, ceteris paribus, the structure of a country’s endowment, that is, the relative abundances of factors that the country pos- sesses, determines the relative factor prices and thus the optimal industrial

160 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 147 structure (Ju et al., 2009). A low-income country with abundant labour or natural resources and scarce capital will have comparative advantage and be competitive in labour- or resource-intensive industries. Similarly, a high-income country with abundant capital and scarce labour will have comparative advantage and be competitive in capital-intensive industries. The optimal industrial structure in a country, which will make it most com- petitive, is therefore endogenously determined by its endowment structure. For a developing country to reach the advanced countries’ income level, it needs to upgrade its industrial structure relative to their capital-intensity. However, to achieve that, it must fi rst close its endowment gap with that of the advanced countries, and the strategy to get there is to follow its com- parative advantage at each stage of its development. When fi rms choose to enter industries and adopt technologies consistent with that country’s comparative advantage, the economy is most competitive. These fi rms will claim the largest possible market shares and create the greatest possible economic surplus in the form of profi ts and salaries. Because of the com- petitiveness, re-invested surpluses earn the highest return, which allows the economy to accumulate even more physical and human capital over time. This dynamic can lead to a virtuous circle: it can upgrade the country’s factor-endowment structure as well as the industrial structure, and also make domestic fi rms more competitive in more capital- and skill-intensive products over time. A fi rm’s objective is to maximise profi t, not to exploit the economy’s comparative advantage. It will follow the comparative advantage in choos- ing its industry and technology in the development process only if the rela- tive factor prices refl ect the relative abundances of factors in the economy (Lin, 2009; Lin and Chang, 2009). Relative factor prices of such nature will exist only in a competitive market system. An effi cient market mecha- nism is therefore a required institution for the economy to follow its com- parative advantage in the process of dynamic development. However, in spite of the importance of the market mechanism, for the following information, co-ordination, and externality reasons, it is also desirable for the government to play a pro-active role in facilitating indus- trial upgrading and diversifi cation in the development process. First, the decision to upgrade or diversify is never an obvious choice. A pioneer fi rm may fail due to the lack of complementary inputs or adequate

161 148 | New Structural Economics infrastructure for the new industry, or the targeted industry may simply not be consistent with the economy’s comparative advantage. Industrial upgrading and diversifi cation are therefore likely to be a costly trial-and- error exercise of discovery, even with the advantage of backwardness (Hausmann and Rodrik, 2003). In order to be successful in a competi- tive market, fi rms in a developing country need information about which industries within the global industrial frontier align with the country’s latent comparative advantage. Information has the same properties as public goods. The costs of col- lecting and processing information are substantial; however, the marginal rm to share the information is almost zero, cost of allowing one more fi once the information is generated. Therefore, the government can play a facilitating role by investing in information collection and processing and making information about the relevant new industries freely avail- able to fi rms. In addition, the choice of a new industry may also shape the economy’s future growth potential in a path-dependent way through c human and social capital. The government the accumulation of specifi is better than individual private fi rms at analysing information about this and making that information available to the public. Second, technological innovation and industrial diversifi cation and upgrading are typically accompanied by changes in capital and skills requirements for fi rms, as well as changes in their market scope and infra- structure needs due to the evolving nature of production embodied in the process. In other words, industrial upgrading and diversifi cation are typi- cally accompanied by changes in hard and soft infrastructure requirements. For example, with the change from agrarian production to manufacturing and from simple to advanced manufacturing in the development process, the scale of production and market scope become increasingly large, and with them the demand for transportation and power. Individual fi rms are not capable of internalising these provisions or deploying the kind of co- ordination efforts among fi rms in different sectors needed to meet those 7 increasing demands. Even if some large single companies were willing to fi nance a national road or a power network, co-ordination through the public sector would be needed to ensure consistency, effi ciency and preven- tion of natural monopolies when the national economy grows. In addition to the hard infrastructure, in a low-income country fi rms in small-scale,

162 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 149 labour-intensive agriculture and manufacturing need only an unskilled nancial and marketing system. But labour force and an unsophisticated fi high-skilled when the economy expands into modern manufacturing, labour and large funds for lump-sum investments in equipment, work- ing capital and/or export fi nancing are needed, as well as new market- ing arrangements. However, individual fi rms are usually not capable of internalising the needed changes in soft infrastructure. Here again, there is a need for the state to provide or co-ordinate some of those changes in different sectors of the economy so as to facilitate the individual fi rms’ 8 upgrading and diversifi cation. Third, innovation, which underlies the industrial upgrading and diver- cation process, is by nature a very risky endeavour. Even when gov- sifi rms with the ernments are willing and capable of helping by providing fi necessary information and co-ordination, success is not guaranteed. Firms can fail because the targeted industry is too ambitious, or the market too small, or the co-ordination inadequate. But even such cases of failure offer useful information to other fi rms, indicating that the targeted industries are inappropriate and should be re-examined. First-mover fi rms therefore pay the cost of failure and produce valuable information for other fi rms. And when they succeed, their experience also provides information externalities to other fi rms: their success proves that the new industry is aligned with the rms economy’s new comparative advantage, thus prompting many new fi 9 to enter the industry. rms eliminates the possible rents The subsequent large entry of new fi that the fi rm, rst mover may enjoy. From the perspective of an individual fi the incentive to be a pioneer fi rm is repressed because of the asymmetry between the high cost of failure and the limited advantage of success. Unless there is compensation for the information externalities that the pio- neer fi rms will have the incentive to be the fi rst movers rm creates, few fi and thus the process of industrial upgrading and diversifi cation as well as economic growth will be impeded (Aghion, 2009; Romer, 1990). In a developed country with global-frontier industries, a successful fi rst mover can in general be rewarded with a patent and enjoys the rent created by a period of monopoly for its innovation. For a developing country, its new industry is most likely to be a matured industry located within the global industrial frontier. So the fi rst mover will not be able to obtain a patent

163 150 | New Structural Economics for its entry into a new industry. Some form of direct support from gov- rms that are willing to take the risk to move to new ernment to pioneer fi 10 able. industries is therefore justifi Compared with developed countries whose industries are located on the global frontier and their industrial upgrading and diversifi cation rely on their own generation of new knowledge through a process of trial and error, developing countries in the catching-up process move within the global industrial frontier and have the advantage of backwardness. In other words, they can rely on borrowing the existing technology and industrial ideas from the advanced countries. This method of acquiring innovation has a lower cost and is less risky than the one used by fi rms in developed 11 countries (Krugman, 1979). Therefore, in a developing country commit- ted to the market system, if fi rms know how to tap into the potential of the advantage of backwardness and the government pro-actively provides information, co-ordination, and externality compensation in the process of industrial upgrading and diversifi cation, the country can grow much faster than a developed country and achieve the goal of converging with high- income countries (Lin, 2009). After all, this was the case for Britain before the eighteenth century; for Germany, France and the United States in the nineteenth century; and the Nordic countries, Japan, Korea, Taiwan-China, Singapore, Malaysia and other East Asian economies in the twentieth cen- tury (Amsden, 1989; Chang, 2003; Gerschenkron, 1962; Wade, 1990). 3. Picking Winners or Losers: Lessons from Experience There is wide consensus among economic historians on the important role played by the state in facilitating structural change and helping sustain it across time and across developed countries. However, except for a few successful cases post-World War II, the governments in most developing countries have failed to play that desirable role. It is therefore essential to briefl y review historical and contemporary experiences of state interven- tion, to draw lessons from the many failures and few successes. 3.1 The Role of the State in Structural Change in Advanced Economies There is ample historical evidence that today’s most advanced economies relied heavily on government intervention to ignite and facilitate their

164 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 151 take-off and catch-up processes, which allowed them to build strong industrial bases and sustain the momentum of their growth over long periods. In his well-known survey of trade and industrial policies lead- ing to early economic transformations in the Western world, List (1841) documented various policy instruments through which governments pro- tected domestic industries or even intervened to support the development of specifi c industries—many of which became successful and provided the 12 bedrock for national industrial development. Likewise, Chang (2003) has reviewed economic developments during the period when most of the currently advanced economies went through their industrial revolutions (between the end of the Napoleonic Wars in 1815 and the beginning of World War I in 1914). He has documented various patterns of state intervention that have allowed these countries to successfully implement their catch-up strategies. Contrary to conventional wisdom that often attributes the Western industrial successes to laissez- faire and free-market policies, the historical evidence shows that the use of industrial, trade, and technology policies was the main ingredient for their successful structural transformation. This ranged from the frequent use of import duties or even import bans for infant-industry protection to industrial promotion through monopoly grants and cheap supplies from government factories, various subsidies, public-private partnerships, and direct state investment, especially in Britain and the US (Trebilcok, 1981). All European countries trying to catch up with Britain devoted efforts to technology policy. Up to the middle of the fi rst Industrial Revolution, the main channel for technological transfer was the movement of skilled work- ers who embodied new knowledge. Latecomers to the industrialisation process, such as France, attempted to acquire them on a large scale from Britain, but the British government banned the emigration of skilled work- 13 When new technologies ers for more than a century, starting in 1719. became embodied in machines, they too were put under government con- trol: various laws were adopted throughout the eighteenth and nineteenth centuries to ban the export of ‘tools and utensils’. In all advanced economies, the government supported the acquisi- tion of foreign technology, ‘sometimes by legal means such as fi nancing study tours and apprenticeships, and sometimes through illegal measures, which included support for industrial espionage, smuggling of contraband machinery, and refusal to acknowledge foreign patents’ (Chang, 2003: 18).

165 152 | New Structural Economics In Germany (Prussia), for instance, Frederick the Great annexed the industrial province of Silesia and promoted the steel and linen industries. Advanced technologies such as iron-puddling, the coke furnace or the steam engine were subsequently imported from more successful countries (Kindleberger, 1978). Government intervention took many forms in the early experiences of industrialisation. In Japan, the government created many factories (‘pilot plants’) in shipbuilding, mining, textiles, etc., most of which were subse- quently sold off to the private sector at very low prices and further sub- sidised. This helped launch the process of industrialisation and diversifi - 14 cation. Even when government-run enterprises performed poorly, there were many cases of failures that generated a burgeoning private sector. 15 This was most notably the case in Japan during the Meiji Restoration when a vibrant textile industry emerged from the failure of the poorly rms were successful because managed state-owned enterprise. Private fi they learned the skill and management from the state-owned fi rms, and introduced various process innovations to replace expensive equipment with inexpensive labour, which was Japan’s comparative advantage at the 16 time (Otsuka et al., 1988). Developed-country governments continue to adopt various measures to support industrial upgrading and diversifi cation, even though these policies may not be announced under the formal label of ‘industrial policy’. Besides patent systems, which are industry-neutral, other such measures typically include support for basic research, mandates, allocation of defence con- tracts and large public procurements. Local governments also often pro- vide all kinds of incentives to private fi rms to attract them to particular geographic areas and induce new investments. The application of all these measures needs to identify specifi c industries or products and amounts to ‘picking winners’. A prime example is the US, where the government has constantly offered strong incentives to private businesses and academic institutions for discovering new ideas that are valuable for sustaining growth, as well as making such ideas non-rival—besides building infrastructure in key eco- nomic sectors such as transportation and providing fi nancing to education and training in order to build the country’s skills base in various industries. This is routinely done through subsidies for research and development,

166 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 153 and through the granting of patents and copyrights. The Advanced Tech- nology Program, for instance, launched in 1990, has been instrumental in the research and development of promising high-risk technologies. Gov- ernment subsidies can also be found in areas such as defence, energy, trans- portation and home construction. 17 The ongoing debate over the need for a US industrial policy has not changed the hard facts about the important role played by the federal and state governments in industrial development in recent decades. Their interventions include the allocation of large amounts of public funding to defence-related procurements and R&D spending, which have large spill- over effects throughout the economy (Shapiro and Taylor, 1990). In fact, the share of the federal government in total R&D spending, which was only 16% in 1930, has remained between 50 and 66% during the post-World War II years (Owen, 1966; Mowery and Rosenberg, 1993). As Chang observes, ‘industries such as computers, aerospace and the internet, where the U.S.A. still maintains an international edge despite the decline in its overall technological leadership, would not have existed without defence- related R&D funding by the country’s federal government’. Government support is also critical in other important segments of the economy such as the health industry: public funding to the National Institutes of Health, which in turn support a large fraction of R&D by biotechnological fi rms, has been essential in helping the US maintain its lead in that industry. The same is true in Europe where discussions of active industrial policy 18 have been taking place since the end of World War II. In fact, many of Europe’s most remarkable industrial successes (space programme Ari- ane, aircraft manufacturer Airbus, etc.) were achieved in the context of intergovernmental co-operation, with decisive political support from the European Union. Since the early 1990s, the European Commission has issued several policy papers on the subject, including the 1994 report An Industrial Competitiveness Policy for the European Union , which set the stage for more determined government interventions. Other offi cial strategy documents have focused on the risk of de-industrialisation, the regulatory burden, the impact of enlargement of the EU on the competi- tiveness of European companies and their location, etc. In the context of the review of the Lisbon Strategy in March 2005, EU Member States set the objective of ‘creating a solid industrial base’, and reiterated the

167 154 | New Structural Economics increasing importance attached to R&D and innovation in all forms, as 19 well as information and communication technologies. France has always favoured government-sponsored economic pro- grammes in which the public and private sectors co-ordinate their efforts to develop new technologies and industries. The French government often provides fi nancial support and capital to the private sector by direct subsi- 20 dies, tax credits, or government-run developmental banks. In Britain, the government, which defi nes itself as ‘a market shaper’, has recently released a new industrial policy aimed at: supporting enterprise and entrepreneurial activity, including the access to fi nance required for starting and grow- ing fi rms; fostering knowledge creation and its application; helping people develop the skills and capabilities to fi nd work and build the businesses and industries of the future; investing in the infrastructure required to sup- port a modern low-carbon economy; ensuring open and competitive mar- kets to drive innovation and rising productivity; and building on industrial strengths where Britain has particular expertise or might gain a compara- tive advantage, and where government action can have an impact (British Government, 2009). Another interesting case is that of Finland, a late but successful state-led industrialisation. According to Jäntti and Vartiainen (2009), the economic policy that achieved this objective was a mix of heavy government inter- vention and private incentives. Government intervention aimed at a fast build-up of industrial capital in order to ensure a solid manufacturing base. The main features of the country’s growth regime were: a high rate of capi- tal accumulation, which often required the use of administrative rationing of credit through interest-rate controls as well as a policy of selective loan approvals for capital-equipment investment; and a high rate of investment in targeted areas of manufacturing, the paper and pulp and metalworking industries in particular. State enterprises were established in the basic metal and chemical-fertiliser industries, and in the energy sector. As late as in the 1980s, state-owned enterprises accounted for 18% of the country’s total industry value-added (Kosonen, 1992). Almost all developing countries have tried to replicate the earlier mod- els of state-led structural change, especially after World War II. From the planned economies of Eastern Europe and Asia to left-leaning or even liberal regimes in Latin America, Asia, Africa and throughout the Arab

168 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 155 world, many governments have adopted various policy measures to pro- mote industrial development and industrial upgrading (Chenery, 1961). While there have been a few successes in East Asia, most of these attempts have failed to deliver the expected results (Krueger and Tuncer, 1982; Lal, 1994; Pack and Saggi, 2006). Nevertheless, the governments in developing countries will continue to attempt to play the facilitating role. It is there- fore all the more important to understand better why some countries have been able to succeed while most others have failed, so that it is possible to advise the governments to do the right things and avoid the mistakes (Rodrik, 2009). 3.2 The Recipe for Success—or Failure There are two main reasons for the controversies and confusion about industrial policy in developing countries. First, economists have tended to focus their attention on the failed policies implemented and not on the objectives and the broader strategic choices made in the successful cases. Second, too often very different types of government interventions are lumped together in regression analyses, with little consideration specifi - cally as to which ones may have attempted to facilitate the emergence of industries that are consistent with latent comparative advantage. Summing up the research fi ndings on how to achieve sustained growth through structural transformation and the diffusion of ideas and accumu- lation of knowledge, Romer notes that ‘the challenge is to fi nd better forms of government intervention, ones that have better economic effects and pose fewer political and institutional risks’ (1990: 66). He also points out that ‘the temptation for economists, however, has always been to duck the complicated political and institutional issues that this kind of analysis raises and instead to work backward from a desired policy conclusion to a simple economic model that supports it’. In fact, the real challenge for economists and policy-makers in any country may be instead to identify the new industries that are consistent with the economy’s comparative advantage, which evolves as the endowment structure changes. A common feature of the industrial upgrading and diversifi cation strate- gies adopted by successful countries (the most advanced ones and the East Asian NIEs in the post-War period) was the fact that they targeted mature industries in countries not too far advanced compared with their own

169 156 | New Structural Economics levels of per capita income. This may have been the single most important cause for their success. Throughout human history, it appears that pioneer countries have always played (and often unwillingly) the role of an ‘eco- nomic compass’ for latecomers. Going back to the sixteenth century, the Netherlands played that role for Britain, which in turn served as a model and target for the US, Germany, and France in the late nineteenth and early twentieth centuries and for Japan in the mid-twentieth century. Likewise, Japan was imitated by Korea, Taiwan-China, Hong Kong-China, and Sin- gapore in the 1960s and 1970s. Mauritius picked Hong Kong-China as its ‘compass’ in its catch-up strategy in the 1970s. China chose Korea, Taiwan-China, and Hong Kong-China in the 1980s. Two main lessons can be drawn from these successful cases of state-led structural-change strategies. First, it appears that the government imple- mented policies to facilitate the development of new industries in a way that was consistent with the country’s latent comparative advantage as determined by its endowment structure. Therefore, fi rms, once established with government support in information, co-ordination, and sometimes 21 Second and even limited subsidies, have turned out to be competitive. more important, to ensure that they would tap into their latent and evolv- ing comparative advantage, the government targeted mature industries in countries that were, on average, about 100% higher than their own level 22 When Britain of per capita income, measured in purchasing power parity. applied industrial policies to catch up with the Netherlands in the sixteenth and seventeenth centuries, its per capita income was about 70% that of the Netherlands. When Germany, France, and the US used industrial policy to catch up with Britain in the nineteenth century, their per capita incomes were about 60 to 75% that of Britain. Similarly, when Japan’s industrial policy targeted the US automobile industry in the 1960s, its per capita income was about 40% that of the US. When Korea and Taiwan-China adopted industrial policies to facilitate their industrial upgrading in the 1960s and 1970s, they targeted industries in Japan instead of the US, and for a good reason: their per capita incomes were about 35% that of Japan 23 and only about 10% that of the US at the time. Looking closely at the elements of successful catch-up strategies, it appears that the specifi cs of policy interventions depended on the particular binding constraints for these new industries and on country circumstances.

170 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 157 But while the interventions were often different, the patterns of industrial development were similar across countries. They all started from labour- intensive industries, such as garments, textiles, toys and electronics, in the early stage of development and proceeded to move up the industrial ladder 24 step by step to more capital-intensive industries. The East Asian NIEs, for instance, exploited the fact that their endowment structures were similar to Japan’s to follow its development in a fl ying-geese pattern (Akamatsu, 1962; Kim, 1988). This was possible because the per capita income gaps 25 with their target-country were not large (Ito, 1980). The story of Korea is a particularly good illustration of this strategy. The government took a pro-active approach to industrial upgrading, and adjusted its strategy to enter industries that were consistent with the coun- try’s latent (and evolving) comparative advantage. In the automotive sec- tor, for example, early in Korea’s growth period, domestic manufacturers concentrated mostly on assembly of imported parts, which was labour- intensive and in line with their comparative advantage at the time. Simi- larly, in electronics, the focus was initially on household appliances, such as TVs, washing machines and refrigerators, and then moved to memory chips, the least technologically complex segment of the information indus- try. Korea’s technological ascent has been rapid, as has its accumulation of physical and human capital due to the conformity of its main industrial sectors with the existing comparative advantage and, hence, its changes 26 As a result, Korea has achieved in underlying comparative advantage. remarkable GDP growth rates in the past forty years and has performed impressively in industrial upgrading into such industries as automobiles and semiconductors. Developing countries in other regions of the world pursued the same path with excellent results. Chile, one of the Pacifi c Rim countries, success- fully targeted industries that were consistent with its comparative advan- tage determined by its natural endowment, as well as industries that were already mature in more advanced countries. While free-market reforms introduced in the early 1970s brought many benefi ts to the country, they were slowly accompanied by market failures (Diaz-Alejandro, 1985). In recognition of these problems, the government has supported private- sector growth through a number of policy instruments, including the pro- vision of agricultural public goods by a state institution (Servicio Agricola

171 158 | New Structural Economics Granadero); guarantees for loans to small enterprises; a semi-public entre- preneurial institution (Fundacion Chile) responsible for the development of the salmon industry; the ‘simplify drawback’ mechanism, which provided subsidies to new exports; the various programmes of the national develop- ment agency (Corporacion de Fomento de la Produccion, CORFO); and the National Council on Innovation for Competitiveness. In recent years, the country has experienced ‘a burst of export discov- eries of new comparative advantages’ (Agosin et al., 2008) and dynamic growth. Key to this success has been the diversifi cation of Chile’s tradi- tional resource-based industries of mining, forestry, fi shing and agriculture, coupled with a strong drive to increase exports. The initial dependence on copper has been gradually reduced in favour of aluminum smelting. Forestry products have been expanded into salmon aquaculture and agri- culture into wine production, as well as freezing and canning fruits and vegetables. Manufacturing has been less successful but many foreign fi rms have chosen to locate in Chile as it offers a secure platform from which to supply other markets across South America. Mauritius, one of the most successful African economies, took off in the 1970s by targeting labour-intensive industries such as textiles and gar- ments. These industries were mature in Hong Kong, its ‘compass econ- omy’. Both economies share the same endowment structure and the per capita income in Mauritius was about half that in Hong Kong-China in 27 The Mauritius Industrial Development Authority (MIDA) and the 1970s. Export Processing Zones Development Authority were created by the gov- ernment to attract Hong Kong-China’s investment in its export processing zone. The vision was to position Mauritius as a world-class export hub on the Hong Kong-China model. Together, they have contributed to the country’s emergence as an economic powerhouse. By contrast, many countries designed and implemented catch-up strate- gies that were too ambitious for establishing the ‘commanding heights’, given their level of development. Historical examples of such mistakes go back to countries such as Hungary or Russia, which tried to replicate industries in place in Britain in the late nineteenth century (Gerschenkron, 1962). While GDP statistics are scarce for individual countries, purchasing power parity estimates by Maddison (2006) indicate that their per capita

172 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 159 GDP represented 25% and 30% that of Britain in 1900. Such a large gap 28 made any attempt by the former to develop British industries unrealistic. Most developing countries fell into the same trap after World War II. They often targeted advanced industries in advanced economies when their per capita incomes represented only a very small fraction of that of high- income countries. After gaining their independence from colonial powers, many countries considered the development of advanced heavy industries as a key symbol of their freedom, a sign of strength, and a political state- ment of their reputation on the international scene. Across Latin America, Africa and South Asia, some of these newly independent countries were run by political leaders with leftist inclinations who chose to follow the prevail- ing Stalinist model of state-led industrialisation through the development of advanced heavy industries, regardless of their political denominations. State resources were used in the industrialisation push, with resources directly allocated to various investments, and large public enterprises set up in almost every sector of the economy—all deemed strategic for the survival and modernisation of the nation. Under the ‘macroeconomics of national- ism’ (Monga, 2006), the criteria for designing industrial policies and select- ing specifi c sectors for government intervention were mostly political. In parallel to political aspirations for heavy-industry development, there was an obsession with ‘market failure’ in academic circles—especially uential economists and policy-makers in Latin America where many infl (Albert Hirschman, Raul Prebisch, Roberto Campos and Celso Furtado, among others) argued that industrialisation and growth could not take place spontaneously in developing countries because of structural rigidities 29 and co-ordination problems. They recommended that government sup- port be provided to the manufacturing industry for these countries to catch up with developed countries, regardless of the large income gap between the two. Too often, such industrial policy defi ed the prevailing comparative advantage of many poor countries where factor endowments were charac- terised by the abundance of labour. By implementing the capital-intensive heavy industry-oriented development strategy, they were not able to build fi rms capable of surviving in open, competitive markets. Because of their high capital needs and their structurally high production costs, these public

173 160 | New Structural Economics enterprises were not viable. Even when they were well managed, they t in an undistorted and competi- could not earn a socially acceptable profi tive market. A good example is Egypt’s industrialisation programme in the 1950s, which featured heavy industries such as iron, steel and chemicals. The country’s per capita income represented about 5% that of the US, the world’s most important steel producer at the time. Unless the government continuously provided costly subsidies and/or protection, Egyptian fi rms could not attract private investment. The limited fi scal resource capaci- ties of the state made such large-scale protection and subsidies unsustain- able. In such situations, governments have had to resort to administrative measures—granting market monopolies to fi rms in the so-called priority sectors, suppressing interest rates, over-valuing domestic currencies, and controlling the prices of raw materials—in order to reduce the costs of investment and continuous operation of their non-viable public enterprises (Lin, 2009). These various experiments provide valuable lessons for economic policy. They highlight conditions under which industrial policies can suc- ceed or fail. Failures occur when countries target industries that are too advanced, far beyond their latent comparative advantage. In such cir- rms cannot be viable in open, com- cumstances, government-supported fi petitive markets. Their survival depends on heavy protection and large subsidies through various means such as high tariffs, quota restrictions and subsidised credit. The large rents embedded in these measures easily become the targets of political capture and create diffi cult governance 30 problems (Lin, 2010). 4. A Framework for Growth Identifi cation and Facilitation The historical and contemporary evidence showing that governments always play an important role in facilitating industrial upgrading and diversifi cation in all successful countries may not be enough to validate an idea that has been mired in controversy for so long. Many economists who agree with the general notion that government intervention is an indispensable ingredient of structural transformation have maintained their opposition to industrial policy because of the lack of a general framework that can be used to guide policy-making. As Charles Schultze,

174 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 161 chairman of the Council of Economic Advisers under US President Jimmy Carter, once put it: rst problem for the government in carrying out an industrial policy is that we The fi actually know precious little about identifying, before the fact, a ‘winning’ industrial structure. There is not a set of economic criteria that determine what gives different countries preeminence in particular lines of business. Nor is it at all clear what the substantive criteria would be for deciding which older industries to protect or restruc- ture. (Schultze, 1983) It is therefore useful to draw on the theories of comparative advantage and backwardness advantage as well as the successful and failed expe- riences of industrial policies discussed in Section 3 to codify some basic principles that can guide the formation of successful industrial policy. The fi rst step is to identify new industries in which a country may have latent comparative advantage, and the second is to remove the constraints that impede the emergence of industries with such advantage and create the conditions to allow them to become the country’s actual comparative advantage. Here, we propose a six-step process: 31 • First, the government in a developing country can identify the list 32 and services that have been produced for about of tradeable goods 20 years in dynamically growing countries with similar endowment structures and a per capita income that is about 100% higher than 33 their own. • Second, among the industries in that list, the government may give pri- rms have already entered ority to those which some domestic private fi 34 spontaneously, and try to identify: (i) the obstacles that are preventing rms from upgrading the quality of their products; or (ii) the bar- these fi 35 This rms. riers that limit entry to those industries by other private fi could be done through the combination of various methods such as value-chain analysis or the Growth Diagnostic Framework suggested by Hausmann et al. (2008). The government can then implement policy to remove these binding constraints and use randomised controlled experi- ments to test the effects of this so as to ensure the effectiveness of scaling up these policies at the national level (Dufl o, 2004). • Third, some of those industries in the list may be completely new to domestic fi rms. In such cases, the government could adopt specifi c mea- sures to encourage fi rms in the higher-income countries identifi ed in the

175 162 | New Structural Economics fi rst step to invest in these industries, so as to take advantage of the lower labour costs. The government may also set up incubation programmes 36 to catalyse the entry of private domestic fi rms into these industries. ed on the list of potential Fourth, in addition to the industries identifi • opportunities for tradeable goods and services in step 1, developing- country governments should pay close attention to successful self- discoveries by private enterprises and provide support to scale up 37 these industries. Fifth, in developing countries with poor infrastructure and an unfriendly • business environment, the government can invest in industrial parks or export processing zones and make the necessary improvements to attract rms and/or foreign fi domestic private fi rms that may be willing to invest in the targeted industries. Improvements in infrastructure and the busi- ness environment can reduce transaction costs and facilitate industrial development. However, because of budget and capacity constraints, most governments will not be able to make the desirable improvements for the whole economy within a reasonable timeframe. Focusing on improv- ing the infrastructure and business environment in industrial parks or 38 export processing zones is, therefore, a more manageable alternative. Industrial parks and export processing zones also have the benefi ts of encouraging industrial clustering. • Sixth, the government may also provide incentives to domestic pioneer rms or foreign investors working within the list of industries identifi ed fi in step 1 in order to compensate for the non-rival public knowledge created by their investments. These incentives should be limited both in time and in fi nancial cost. They may take the form of a corporate 39 income-tax holiday for a limited number of years, direct credits to co- 40 to import key nance investments, or priority access to foreign reserves fi equipment. The incentives should not and need not be in the form of monopoly rent, high tariffs, or other distortions. The risk of rent-seeking 41 rms in step 4 that For fi and political capture can therefore be avoided. discovered new industries successfully by themselves, the government may award them special recognition for their contribution to the coun- 42 try’s economic development. The industries identifi ed through the above process should be consistent with the country’s latent comparative advantage. Once the pioneer fi rms

176 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 163 come in successfully, many other fi rms will enter these industries as well. The government’s facilitating role is mainly restricted to provision of infor- mation, co-ordination of hard and soft infrastructure improvement, and compensation for externalities. Government facilitation through the above approach is likely to help developing countries tap into the potential of the advantage of backwardness and realise a dynamic and sustained growth. 4.1 Possible Ways of Identifying Binding Constraints The facilitation of industrial growth has been the subject of a rich body of research and several approaches have recently been suggested by various 43 authors. While these are all likely to yield useful results, none of them focuses specifi cally on the identifi cation of industries in which a developing country may have latent comparative advantage. The intellectual legacy of the failure of industrial policies based on development strategies that were inconsistent with comparative advantage has certainly led many econo- mists to conclude that it may be impossible for any government to ‘pick winners’ successfully. In the absence of a framework for industrial identifi cation, the existing literature has been limited to exploring ways of improving the business environment and infrastructure, which indeed affect fi rms’ operations and transaction costs. There is a robust empirical knowledge based on quanti- tative data on fi rm performance and perceptions-based data on the severity rms in the developing world. of a number of potential constraints facing fi It points out that in most of sub-Saharan Africa, for instance, fi rms tend to consider many areas of the investment climate major obstacles to business development and the adoption of more sophisticated technology. Finance and access to land seem to be areas of particular concern to smaller fi rms; larger fi rms tend to perceive labour regulations and the availability of skilled labour as the main constraints to their activity; fi rms across the board are concerned about corruption and infrastructure—especially net- work utilities such as electricity, telecommunications, transportation and water (Gelb et al., 2007). Despite their usefulness, investment-climate surveys, which try to cap- ture the policy and institutional environment within which fi rms oper- ate, can be misused or misinterpreted. Just as individual perceptions of well-being are subjective and do not necessarily correlate with objective

177 164 | New Structural Economics measures such as income or consumption, fi rms’ perceptions of binding constraints to their development often differ from actual determinants of performance. This limitation is due to the very nature of the investment-cli- mate data and the way they are often used. In a typical survey, the manag- ers of a sample of fi rms are asked to rate each dimension of the investment climate (such as ‘infrastructure’, ‘access to fi nancing’, ‘corruption’, etc.) on a scale of 1 to 4, corresponding to the degree to which it is an obstacle to 44 rm performance. High mean reported values for particular dimensions fi of the investment climate are then interpreted as evidence of the severity of obstacles to growth. However, this may not be the case. Despite their intimate knowledge of their business processes and operating environment, fi rms may not fully recognise the true origin of their main problems and mistakenly identify as a constraint something which is in fact a symptom of another less obvious problem. Because of these shortcomings, investment-climate constraints are increasingly complemented by the World Bank’s ‘Doing Business’ indi- cators, which are based on expert surveys (not just fi rm-level perceptions) and provide a more comparable cross-country perspective across a detailed range of regulation. The problem remains, as survey results often vary depending on whether respondents are asked to rate their most important constraints, or to rank them. While ranking appears to be favoured by researchers who have examined different methodologies, since it forces stronger expression and relationships (Alvin and Krosnick, 1985), it may not be entirely reliable: rms or experts asked to rank constraints may not have a good basis for fi determining whether their top-ranked constraint is serious or not. Ranking without a solid and meaningful benchmark against which local fi rms can rate the severity of a particular constraint may not provide useful informa- tion. In addition, there are instances where picking any single quantitative rms often face several constraints simul- criterion could be misleading, as fi taneously. Ranking all of them as important may not be very helpful for policy-making. In order to account for the major role of fi rm heterogeneity in growth analysis, one must go beyond extracting means of investment- climate variables from fi rm-level surveys. Careful econometric modelling of fi rm performance is therefore needed to identify which particular vari- able has the biggest effect on growth. In other words, the policy variables

178 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 165 with the greatest economic impact can be quite different from those with 45 the highest perceived values. Investment-climate surveys have two more limitations. They do not provide information about industries that do not yet exist, but in which a country has latent comparative advantage. And the existing industries that are surveyed may not be consistent with the country’s comparative advan- tage, either because they are too advanced (as a legacy of a development strategy that defi ed comparative advantage), or because they have become fundamentally uncompetitive (as a result of a general wage increase that accompanied the country’s development). These two additional limitations make it highly desirable for investment-climate surveys to cover only a sample of fi rms that meet the criteria of viability, and can represent the economy’s true potential. Another important problem with the recognition of obstacles to growth is the fact that many other constraints to business development are endog- enous to the industries that might be targeted by a developing country. Good examples are specifi c types of human capital, fi nancing instruments, or infrastructure that may be needed only by fi rms moving to specifi c industries. Identifying and removing them may require the use of several complementary analytical tools. One useful tool is the Growth Diagnos- tics Framework suggested by Hausmann et al. (2008). It is based on the observation that, when presented with a laundry list of needed reforms, policy-makers either struggle to try to solve all of the problems at once or start with reforms that are not critical to their country’s growth poten- tial. Because reforms in one area may create unanticipated distortions in another, focusing on the one that represents the biggest hurdle to growth is the most promising avenue to success. Countries should therefore fi gure out the one or two most binding constraints on their economies and focus on lifting those. The Growth Diagnostics approach provides a decision-tree methodol- ogy to help identify the relevant binding constraints for any given country. It starts with a taxonomy of possible causes of low growth in developing countries, which generally suffer from either a high cost of fi nance (due either to low economic and social returns or to a large gap between social and private returns), or low private return on investment. The main step in the diagnostic analysis is to fi gure out which of these conditions more

179 166 | New Structural Economics accurately characterises the economy in question. The use of this frame- work highlights the fact that, in some countries, the growth strategy should identify the reasons for the low returns on investment, while it must explain why domestic savings do not rise to exploit large returns on investment in other countries. While the Growth Diagnostics Framework attempts to take the policy discussion of growth forward, its focus and the specifi cation of its model remain quite macroeconomic. This is understandable; after all, growth is a macroeconomic concept and taking the analysis to a sector level would raise issues of sector interactions and trade-offs. Moreover, the Growth Diagnostic Framework is also imprecise in its links to the institutions that facilitate the growth process. The methodology proposed for the identifi cation of the binding constraints to growth is not always straightforward. Even if data on shadow prices were widely avail- able, it is not obvious that this would accurately identify areas in which progress is most needed in each country. For example, one could imagine a simple model of growth for a low-income country where technology and human capital are complementary. In such a country, the returns to edu- cation and technology adoption would both be low due to low levels of human capital and technology. An exclusive focus on shadow prices and an ignorance of cross-country comparison of levels would then suggest no need to improve education levels and encourage technology adoption. In fact, even in situations where the Growth Diagnostics approach leads to relative certainty about the binding constraints to growth in any given country, there is still a wide range of policy options available to choose from. It is therefore necessary for policy-makers to rely not just on one approach but to use several different macro and micro tools to identify binding constraints to growth. Microeconomic analyses of growth show that differentiated fi rm dynamics drive a good part of aggregate productiv- ity growth and capital accumulation. Establishing a diagnostic at the aggre- gate level requires a good knowledge of what happens at the micro level. In particular, monitoring the entry and exit of fi rms and the policy variables that affect them is essential to understanding overall gains in productivity in economies subject to strong structural changes (Bourguignon, 2006). One must take account of heterogeneity in country circumstances and among micro agents. This can more effectively be done through country- specifi c analyses.

180 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 167 Finally, even if one could identify relevant binding constraints to indus- trial development in industries with comparative advantage and induce improvements in a country’s business environment, the crucial issues of rst movers and of co-ordination would remain externality encountered by fi unresolved. Despite the removal of the constraints, a country may then fi nd its industrial upgrading and diversifi cation process stalled. It is therefore necessary that the Growth Diagnostics Framework and other methods of targeting obstacles to industrial upgrading be used in conjunction with the cation and facilitation approach. growth identifi 5. Conclusion The current crisis has infl icted heavy costs on economies around the world. Unemployment is at record levels in many countries, fi scal fragility is a leg- acy of the crisis in many countries, and capacity-utilisation rates in indus- try remain substantially below pre-crisis levels. Many developing countries have the potential to grow faster than developed countries and are now confronted with the challenge of fi nding new sources of growth in the con- text of a world of multi-polar growth (Zoellick, 2010). In this regard, the role of developing-country governments in inducing and accompanying structural change (industrial upgrading and economic diversifi cation) to promote growth, employment and poverty reduction must regain centre stage. Indeed, historical evidence and economic theory suggest that while markets are indispensable mechanisms for allocating resources to the most productive sectors and industries, government intervention—through the provision of information, co-ordination of hard and soft infrastructure improvement, and compensation for externalities—is equally indispens- able for helping economies move from one stage of development to another (Lin, 2010). Because of the many failures observed throughout the world in the post- World War II period, industrial policy has raised serious doubts among economists and policy-makers. Taking into consideration O’Brien and Keyder’s recommendation that ‘countries should (if possible) be studied in terms of some unique capacity for development at different stages of their history’ (1978: 15), this article has examined the mechanics of structural change in today’s advanced economies and the reason for success in a few

181 168 | New Structural Economics developing countries in East Asia and elsewhere, as well as suggesting a framework for government intervention in the economy. The article has argued that the failure of industrial policy is most likely cation to arise from mistakes made by policy-makers in the growth-identifi process. Industrial policies implemented by governments in developed and developing countries usually fall into one of two broad categories: (i) they attempt to facilitate the development of new industries that are either too advanced and thus far from the comparative advantage of the economy, or too old and have lost comparative advantage; or (ii) they try to facili- tate the development of new industries that are consistent with the latent comparative advantage of the economy. Only the latter type of policy is likely to succeed. High-performing developed and developing countries are those where governments were able to play an active role in the indus- trial upgrading and diversifi cation process by helping fi rms take advantage of market opportunities. They have generally done so by overcoming the information, co-ordination, and externality issues, and by providing ade- quate hard and soft infrastructure to private agents. It is expected that the growth identifi cation and facilitation approach put forward in the article can help governments in developing countries identify the right industries in their attempts to facilitate structural transformation in the development of their countries. Notes † Célestin Monga, a native of Cameroon, is a Senior Advisor to the World Bank Senior Vice President and Chief Economist. In his 13-year career in the World Bank, he has held positions in both operations and in the research department. He has also served on the Board of Directors of MIT’s Sloan School of Man- agement (Sloan Fellows) and taught at Boston University and the University of Bordeaux (France). 1. To protect jobs, governments in both developed and developing countries may also support old, declining industries, which have already lost their compara- tive advantages. Such policies will fail as well. 2. Examples of hard infrastructure are highways, port facilities, airports, tele- communication systems, electricity grids and other public utilities. Soft infrastructure consists of institutions, regulations, social capital, value sys- tems, and other social and economic arrangements. For further discussion on their impacts on economic development, see Lin (2010).

182 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 169 3. ned here as knowledge (intangible intellectual capital) of Technology is defi nal utility. It differs from human or physi- how to transform basic inputs into fi cal capital by its non-rival nature. Effi ciency is the way technology is used— with the goal of optimality, especially in the allocation of resources. In the growth literature, structural change has not received as much atten- 4. tion as technological change because of the use of a one-sector model, which is incapable of handling issues related to structural change, in the standard growth accounting and regression research. 5. Maddison (2006) estimated that, in Western Europe, the annual per capita income growth rate before the 18th century was about 0.05%, accelerated to about 1% in the 18th and 19th centuries, and reached 2% in the 20th century. The required time for doubling per capita income thus reduced from 1400 years before the 18th century to 70 years in the 18th and 19th centuries, and further to 35 years in the 20th century. 6. The difference between factors of production and infrastructure is that the supply and demand of the former are determined individually by households and fi rms, whereas the latter in most cases is supplied by the community or governments by collective actions. For example, the application of chemical fertilisers in rice and wheat require 7. modern semi-dwarf varieties to avoid the lodging problem, and the use of modern seeds often requires timely irrigation. Individual farmers will not be able to do this by themselves. There is also a need for access to credits beyond individual farmers’ capacity. Similarly, the diversifi cation from farm to non-farm industries or from small-scale traditional to modern industries also requires the provision of many new inputs and improvements in hard and soft rm’s decision. infrastructure, which cannot be internalised in any individual fi The success of Ecuador’s cut fl owers export in the 1980s is a good example. 8. The fact that Ecuador had latent comparative advantages in producing and exporting cut fl owers to the US market was known in the 1970s. But the indus- try did not expand and exports did not take off until the government helped arrange regular fl ights and investment in cooling facilities near airports in the 1980s (Harrison and Rodríguez-Clare, 2010). A similar story applies to Ethio- pia’s cut fl owers export to the European market. In issues related to the pro- vision of skilled labour, Germany’s dual system of vocational education and training has been a major factor in the country’s economic success over the past six decades. 9. In a recent fi eld study in Zambia, we found that a local entrepreneur success- fully started the production of corrugated roofi ng sheets. Within a year, more than 20 fi rms had joined in. 10. Precisely because of such positive information externalities, governments in developed countries also provide various forms of targeted support to fi rms

183 170 | New Structural Economics engaged in innovation, such as funding of basic research, preferential taxes, mandates, defence contracts, and procurement policies. 11. The possibility of borrowing existing knowledge does not mean that develop- ing countries need not engage in indigenous innovation. To be successful, they need to undertake a process of innovation that makes the borrowed technol- ogy suitable to local conditions, and also to carry out product innovation in sectors where they are already world leaders, or not too far behind the world leader. For further discussion, see Lin and Ren (2007). 12. List’s book covers the rise of economic powerhouses in a variety of con- texts, from Italian cities such as Venice to Hanseatic cities such as Hamburg or Lübeck, and countries such as the Netherlands, England, Spain, Portugal, France, Germany and the United States. 13. The ban lasted until 1825. See Landes (1969). 14. For a theoretical exposition, see Jones et al. (1990) and World Bank (1995). 15. The Meiji period (1868-1912) marked the beginning of an era of major politi- cal, economic, and social change which, according to conventional wisdom, brought about the modernisation and Westernisation of Japan. See Beasley (1972). 16. A common reason for the failure of state-owned enterprises is the govern- ment’s attempt to use them as a vehicle to develop industries or adopt tech- nologies inconsistent with the country’s comparative advantage (Lin and Tan, 1999). Such attempts create a policy burden for state-owned fi rms which the state is compelled to provide with subsidies and protection. Information asym- metry prevents governments from knowing exactly what level of subsidies and protection would be adequate and state-owned fi rms use the policy burden as an excuse to ask for more subsidies and protection, which gives rise to the problem of soft-budget constraint (Kornai, 1986). 17. During the 1984 presidential campaign, Democratic candidate Walter Mondale argued that the economic policies of the country were ‘destroying industry— not building it’, and that federal aid should be directed to ‘those communities and regions hit hardest by economic change’ (quoted by McKenzie, 2007). Economists Bluestone and Harrison (1982) argued that the ongoing process of deindustrialisation amounted to a ‘wide-spread, systematic disinvestment in the nation’s productive capacity’. Pointing to the post-War economic success of Japan, which he credited to industrial policies orchestrated by its Ministry of International Trade and Industry (MITI), Thurow (1980) worried that, if left alone, ‘our economy and our institutions will not provide jobs for everyone who wants to work’, and that ‘we have a moral responsibility to guarantee full employment’. He observed that ‘major investment decisions have become too important to be left to the private market alone . . . Japan Inc. needs to be met with U.S.A. Inc’. Others recommended various measures such as the cre- ation of national and regional economic development banks similar to Herbert

184 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 171 Hoover’s Reconstruction Finance Corporation, which would use subsidies and federal loan guarantees to slow the contraction of declining industries and speed the development of emerging industries; the launch of ‘Tripartite coun- cils’ at the national, regional, and fi rm levels, which would be composed of representatives from management, labour and government and would seek consensus on how capital investment should be allocated. While often con- ceding on protectionist proposals, other economists and political leaders have maintained strong opposition to any coherent industrial-policy programmes. 18. The European Coal and Steel Community (ECSC) was created in 1951 and the European Atomic Energy Community (EURATOM) in 1957. 19. In October 2005, the European Commission announced seven new horizontal initiatives in order to: ‘(1) consolidate the EU’s legal framework in the area of intellectual property, (2) take into account the links between the issues of competitiveness and environmental protection, (3) adapt the trade policy with a view to developing the competitiveness of European industry, (4) simplify the law governing certain industrial sectors (i.e. construction, food industry), (5) remedy the shortage of skilled labour in certain sectors (i.e. new technolo- gies, textiles), (6) anticipate and support the structural changes in industry, by taking this objective into consideration in other EU policies (structural funds, in particular), and (7), adopt an integrated European approach to industrial research and innovation.’ 20. Several proposals are currently under consideration to stimulate innovation and growth. The recently issued Juppé-Roccard report by two former Prime Ministers (a socialist and a conservative) recommends that France raises 35 billion euros (US$52 bn) through public borrowing to be spent on universities and research (providing them with endowments and incentives to merge or become independent and private), the green economy and high-tech to pro- pel growth. Among the projects are plans to expand the high-speed Internet, develop green cities, and support innovative small businesses and France’s cutting-edge aerospace and nuclear industries. Of the 35 bn euros to be raised, 13 bn will come from the reimbursed bailout packages given to French banks, with the remaining 20 bn to be raised on the fi nancial markets. 21. The idea of a dynamic comparative advantage is often used to justify industrial policy and government support to fi rms (Redding, 1999). In our analysis, however, the argument is valid only if the government’s sup- port is limited to overcoming information and co-ordination costs and the externalities associated with the pioneer status of fi rst-movers. The targeted industry should be consistent with the comparative advantage of the economy and the fi rms in the new industry should be viable, other- wise they will collapse once the government support is removed. If the tar- geted industry is outside the country’s comparative advantage, the required open-ended support to the subsidised fi rms will crowd out the resources

185 172 | New Structural Economics available to other fi rms that operate in industries consistent with the com- parative advantage. This will obviously slow down economic growth and capital accumulation, and it will take more time for the economy to reach the stage targeted by the dynamic-advantage policy than an economy that follows a comparative-advantage strategy (Lin and Zhang, 2007). 22. For the purposes of this article, the use of per capita income measured in pur- chasing power parity is better than that of the market-exchange rate because, in cross-country comparisons, the former refl ects the level of development and the cost of production better. For a discussion of industrial policies in these countries, see Chang (2003); and 23. for the estimations of per capita income for the above countries, see Maddison (2006). 24. Countries in similar stages of development may specialise in different indus- tries. However, the level of capital intensity in their industries will be simi- lar. For example, in recent years, China is achieving dynamic growth by specialising in the labour-intensive manufacturing industries, such as elec- tronics, toys and textiles, whereas India’s growth has relied on specialis- ing in call centres, programming, and business process services, which are labour-intensive activities within the information industry. 25. In a similar spirit, Hausmann and Klinger (2006) recently investigated the evolution of a country’s level of sophistication in exports and found that this process was easier when the move was to ‘nearby’ products in the product c inputs such as space. This is because every industry requires highly specifi knowledge, physical assets, intermediate inputs, labour skills, infrastructure, property rights, regulatory requirements, or other public goods. Established industries somehow have sorted out the many potential failures involved in assuring the presence of all of these inputs. The barriers preventing the emer- gence of new industries are less binding for nearby industries, which only require slight adaptations of existing inputs. 26. For the debate on the conformity of Korea’s industrial upgrading with its evolv- ing comparative advantage, see the exchange between Lin and Chang (2009). 27. According to Maddison (2006), Hong Kong’s per capita income in 1970 mea- sured in 1990 international dollars was 5,695, whereas that of Mauritius was 2,945. 28. As discussed earlier, a similar policy was pursued successfully in Germany, France and the US at the same time. Their per capita incomes ranged from 60% to 75% that of Britain. 29. The new fi eld of development economics was regarded as covering underdevel- opment because ‘conventional economics’ did not apply (Hirschman, 1982). Early trade and development theories and policy prescriptions were based on

186 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 173 some widely accepted stylised facts and premises about developing countries (Krueger, 1997). These included: (i) developing economies’ production structures were oriented heavily towards primary-commodity production; (ii) if develop- ing countries adopted policies of free trade, their comparative advantage would forever lie in primary-commodity production; (iii) the global income elasticity and the price elasticity of demand for primary commodities were low; (iv) capi- tal accumulation was crucial for growth and, in the early stage of development, it could occur only with the importation of capital goods. Based on these facts and premises, it was a straight step to believing that the process of development was industrialisation, which consisted primarily of the substitution of domestic production of manufactured goods for imports (Chenery, 1958). 30. The other reason for the failure of industrial policy in developing countries is that the policy targets industries that have already lost comparative advantage, but governments want to protect them for socio political reasons (such as pro- viding employment, often in urban areas). The government refers to both the central and local governments. The process 31. discussed here can also be used by multilateral development agencies and non- governmental organisations to promote industrial upgrading and diversifi ca- tion in developing countries. 32. The tradeable goods refer to manufactured products, agricultural products, and fi shery as well other natural-resources products. Because of the ascen- dance and dominance of international production networks in manufacturing industries, the manufactured goods here refer not only to the fi nal products but nal products in manufacturing industries. also to intermediate inputs of fi As discussed in Section 3, this is the most important principle for a developing 33. country to reap the advantage of backwardness in its industrial upgrading and diversifi cation. This is because, for a dynamically growing economy, its wage rate is increasing rapidly and is likely to start losing comparative advantage in the industries that it has produced for many years. Therefore, the industries will become the latent comparative advantage of countries with a similar endow- ment structure but with a lower wage. The principle also means that when a country grows beyond the income level of 50% of the most advanced country, it will become increasingly diffi cult to identify industries that are likely to be its latent comparative advantage. The country’s industries will locate increasingly close to the global frontier and its industries’ upgrading and diversifi cation will increasingly rely on indigenous innovations. Therefore, the government’s poli- cies to support industrial upgrading and diversifi cation will increasingly resem- ble those of the advanced countries. The chance of those policies failing to achieve the intended goal will also increase. As for low-income countries with per capita income measured at about $1,000 in purchasing power parity (PPP) terms currently, in addition to identifying matured tradeable goods in countries

187 174 | New Structural Economics at about $2,000 currently, it may also identify tradeable goods produced in countries that had similar per capita income levels 20 or so years ago and have been growing dynamically since then. In particular, China, Vietnam and India had a similar or even lower income levels 30 years ago than most of today’s poor sub-Saharan countries. Therefore, for today’s poor countries, they may identify the list of goods and services produced in China, Vietnam and India 20 years ago as references. They may also review their imports and identify the list of simple manufacturing goods, which are labour-intensive, have limited economies of scale, and require only small investments, as the targets of their cation. The idea put forward is similar to industrial upgrading and diversifi that of monkeys jumping to nearby trees (Hausmann and Klinger, 2006), but the step proposed here is much easier to implement than the product-space analysis proposed by them. 34. This is because every industry requires some highly specifi c inputs such as knowledge, physical assets, intermediate inputs, labour skills, and so on. The existence of some private fi rms in the industry indicates that the economy at least partially possesses these crucial inputs. Chile has produced wine for a long time. Its recent success in the wine industry 35. is a good example. The change from a negligible wine exporter to the world’s fi fth-largest exporter in the 1970s benefi tted greatly from the government’s programmes to disseminate foreign technology to local farmers and vineyards through Grupos de Transferencia Tecnológica and to promote Chilean wine abroad through Export Promotion Offi ce, ProChile (Benavente, 2006). Lessons from successful Asian countries can be of relevance here. When local 36. rms had no historical knowledge in a particular industry, the state often Asian fi attracted foreign direct investment and/or promoted joint ventures. After the transition to a market economy in the 1980s, China, for instance, pro-actively invited direct investment from Hong Kong-China, Taiwan-China, Korea and Japan—a promotion policy which helped the local economy to get started in various industries. Bangladesh’s vibrant garment industry also started with direct investment from Daewoo, a Korean manufacturer, in the 1970s. After a few years, enough knowledge transfer had taken place and the direct invest- ment became a sort of ‘incubation’, local garment plants mushroomed, and most of them could be traced back to that fi rm (Mottaleb and rst Korean fi Sonobe, 2009; Rhee, 1990; Rhee and Belot, 1990). The booming cut-fl ower export business in Ecuador from the 1980s on also started with three compa- nies founded by Colombia’s fl ower growers (Sawers, 2005). The government can also set up an industrial park to incubate new industries. Taiwan-China’s Hsingchu Science-based Industrial Park for the development of electronic and IT industries (Mathews, 2006) and the Fundacion Chile’s demonstration of commercial salmon farming (Katz, 2006) are two successful examples of the government’s incubation of new industries.

188 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 175 37. India’s information industry is a good example. Indian professionals in Sili- con Valley helped Indian companies to take advantage of expanding oppor- tunities for outsourced IT work in the 1980s. Once the potential of software exports was demonstrated, the Indian government helped build a high-speed data-communications infrastructure that allowed overseas Indians to return home and set up offshore sites for US clients. The Indian software industy has grown more than 30% p.a. for 20 years, with 2008 exports close to $60 billion (Bhatnagar, 2006). Ethiopia’s success in cut fl owers exports is another cation of cut fl owers export and the example. Before the government’s identifi provision of supports in its industrial policy in the 1990s, a local pirate fi rm had exported cut fl owers to the European market for over 10 years. Asparagus in Peru is also a good example. The possibility of growing asparagus, a foreign crop, was discovered by Peruvian farmers in the 1950s. However, the industry and exports did not take off in earnest until 1985 when USAID provided a grant for a farmers’ association to obtain advice from a specialist from the University of California, Davis, who had recently invented the UC-157 variety suitable for the US market, and from another expert who showed members of the association’s experimental station how to set up seedbeds for large-scale production and package the products for export. The state also supported co operative institutions such as the Peruvian Asparagus Institute and Frío Aereo Asociación Civil for engaging in research, technology transfer, market studies, export drives, and quality promotion, and invested in the freezing and packing plants that handled 80% of fresh asparagus exports. With these inter- ventions, Peru has overtaken China and become the largest asparagus exporter in the world (O’Brien and Rodriguez, 2004). In addition to infrastructure, many African countries, for instance, also face the 38. constraint of rigid labour regulation. To overcome this, Mauritius has allowed employment to be fl exible in the export process zone, while maintaining the existing regulation for the domestic economy (Mistry and Treebhoohun, 2009). The measure commonly used in China to attract FDI is to exempt from corpo- 39. rate income tax for the fi rst two years and reduce the tax by half for a further three years. 40. Direct credits and access to foreign reserves are desirable measures in countries with fi nancial depressions and foreign-exchange control. 41. The likelihood of capture is proportional to the magnitude of protection and subsidies. If the targeted industries are consistent with the country’s inherent comparative advantages, and the protection and subsidies are used to compensate the pioneer fi rms for their positive information externalities, the magnitude of protection and subsidies should be small, and the elites will not have the incentives to use their political capital to capture the small rent. In addition, once the pioneering fi rms are successful, many new fi rms will enter and the market will become competitive, which will further reduce

189 176 | New Structural Economics the danger of capture by elites. Alternatively, if the government’s goal is to support the development of industries that go against the country’s com- parative advantages, the fi rms in the targeted industries will not be viable in competitive markets and the required subsidies and protections will be large, and are likely to become the target of rent-seeking and political cap- ture (Lin, 2009). ex-post reward idea to Professor Shang-jin Wei. 42. We owe this 43. See, for example, Di Maio (2008) and Agosin et al. (2009). Ayyagari et al. (2008) present the mean reported values for a number of invest- 44. ment-climate variables in a sample of over 6,000 fi rms in 80 countries. In the overall sample, taxes and regulation, political instability, infl nanc- ation and fi ing are reported as being the greatest obstacles to fi rm growth. Bourguignon (2006) observes: ‘“Extracting means” is the way I would char- 45. acterize the Investment Climate Assessment exercises that the Bank is now carrying out. Like the “Doing Business” indicators, these are undoubtedly useful. However, what they give us is essentially new and better right-hand side variables in cross-country regressions, not necessarily better data for country-specifi c analysis. The goal should be to use investment climate sur- veys to measure the sensitivity of fi rms of different types to investment climate variables, as another way of determining exactly which variable corresponds to a major obstacle to growth.’ References Some Thoughts on Industrial Policy and Growth Aghion, P. (2009) . Working Paper No. 2009–09. Paris: OFCE-Sciences Po. Agosin, M., Larraín, C., and Grau, N. (2009) ‘Industrial Policy in Chile’, Work- ing Papers wp294, University of Chile, Department of Economics. Akamatsu, K. (1962) ‘A Historical Pattern of Economic Growth in Developing Countries,’ The Development Economies, Preliminary Issue No. 1: 3–25. Alvin, D. F. and Krosnick, J. A. (1985) ‘The Measurement of Values in Surveys: A Comparison of Ratings and Rankings’, Public Opinion Quarterly 49 (4): 535–52. Asia’s Next Giant . New York and Oxford: Oxford Amsden, A. H. (1989) University Press. Ayyagari, M., Demirgüç-Kunt, A. and Maksimovic, V. (2008) ‘How Well Do Institutional Theories Explain Firms’ Perceptions of Property Rights?’, Review of Financial Studies 21 (4): 1833–71. Bhatnagar, S. (2006) ‘India’s Software Industry’, in Vandana Chandra (ed.), Technology, Adaptation, and Exports: How Some Developing Countries Got It Right . Washington, DC: World Bank.

190 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 177 Beasley, W. G. (1972) . Stanford, CA: Stanford University The Meiji Restoration Press. Benavente, Jose Miguel (2006) ‘Wine Production in Chile’, in Vandana Chandra (ed.), Technology, Adaptation, and Exports: How Some Developing Countries . Washington, DC: World Bank. Got It Right Bluestone, B. and Harrison, B. (1982) The Deindustrialization of America: Plant Closings, Community Abandonment, and the Dismantling of Basic Industry. New York: Basic Books. Bourguignon, F. (2006) Economic Growth: Heterogeneity and Firm-Level Disaggregation . PREM Lecture. Washington, DC: World Bank. British Government (2009) ‘Going for Growth: Our Future Prosperity’. London (www.bis.gov.uk/wp-content/uploads/2010/01/GoingForGrowth.pdf). Chang, H.-J. (2003) Kicking Away the Ladder: Development Strategy in Historical Perspective . London, Anthem Press. Chenery, H. B. (1961) ‘Comparative Advantage and Development Policy’, American Economic Review 51 (1): 18–51. Chenery, H. B. (1958) ‘The Role of Industrialization in Development The Economics of Programmes’ in A. N. Agarwala and S. P. Singh (eds.), Underdevelopment . Bombay: Oxford University Press. Diaz-Alejandro, C. (1985) ‘Good-Bye Financial Repression, Hello Financial Crash’, Journal of Development Economics 19: 1–24. Di Maio, M. (2008) Industrial Policies in Developing Countries: History and Perspectives. Working Paper No. 48-2008. Italy: Macerata University, Department of Finance and Economic Sciences. o, E. (2004) ‘Scaling Up and Evaluation’ in: F. Bourguignon and B. Pleskovic Dufl (eds), Annual World Bank Conference on Development Economics 2004. Washington, DC: World Bank. Evans, P. (1996) ‘Using Cross-country Income Differences’ in P. Aghion and S. N. Durlauf (eds), , vol. 1. Amsterdam: North- Handbook of Economic Growth Holland. Gelb, A. et al. (2007) What Matters to African Firms? The Relevance of Perception Data. Policy Research Working Paper No. 4446. Washington, DC: World Bank. Gerschenkron, A. (1962) Economic Backwardness in Historical Perspective: A Book of Essays . Cambridge, MA: Belknap Press of Harvard University Press. Harrison, A. and Rodríguez-Clare, A. (2010) ‘Trade, Foreign Investment, and Industrial Policy for Developing Countries’, Handbook of Development Economics 5: 4039 –213. Hausmann, R. and Klinger, B. (2006) Structural Transformation and Patterns of Comparative Advantage in the Product Space. Working Paper No. 128. Cambridge, MA: Harvard University Center for International Development. Hausmann, R. and Rodrik, D. (2003) ‘Economic Development as Self-Discovery’, Journal of Development Economics 72 (December).

191 178 | New Structural Economics Hausmann, R., Rodrik, D. and Velasco, A. (2008) ‘Growth Diagnostics’ in The Washington Consensus Reconsidered: N. Serra and J. E. Stiglitz (eds.), Towards a New Global Governance , New York: Oxford University Press. Hirschman, A. O. (1982) ‘The Rise and Decline of Development Economics’ in The Theory and Experience of Economic M. Gersovitz and W. A. Lewis (eds), . London: Allen and Unwin. Development Howitt, P. and Weil, D. (2010) ‘Economic Growth’ in S. N. Durlauf and L. E. New York: Macmillan Palgrave. Blume (eds), Economic Growth. Journal of Economic Theory Ito, T. (1980) ‘Disequilibrium Growth Theory’, 23: 380 –409. The Finnish Developmental State and its Jäntti, M. and Vartiainen, J. (2009) Research Paper No. 2009/35. Helsinki: United Nations Growth Regime. University. Jones, L. et al. (1990) t Methodology . Selling Public Enterprises: A Cost-Benefi Cambridge, MA: MIT Press. Ju, J., Lin, J. Y. and Wang, Y. (2009) Endowment Structures, Industrial Dynamics and Economic Growth . Policy Research Working Paper No. 5055. Washington, DC: World Bank. Katz, J. (2006) ‘Salmon Farming in Chile’, in Vandana Chandra (ed.), Technol- ogy, Adaptation, and Exports: How Some Developing Countries Got It Right . Washington, DC: World Bank. Kim, Y. H. (1988) Higashi Ajia Kogyoka to Sekai Shihonshugi ( Industrialisation of East Asia and World Capitalism ). Tokyo: Toyo Keizai Shimpo-sha. Kindleberger, C. (1978) ‘Germany’s Overtaking of England, 1806–1914’ in Economic Response: Comparative Studies in Trade, Finance, and Growth . Cambridge, MA: Harvard University Press. Kornai, J. (1986) ‘The Soft Budget Constraint’. Kyklos 39 (1): 3–30. Social Kosonen, K. (1992) ‘Economic Growth’ in J. Pekkarinen et al. (eds.), Corporatism. Oxford: Clarendon Press. Krueger, A. O. (1997) ‘Trade Policy and Economic Development: How We Learn’, American Economic Review 87 (1): 1–22. Krueger, A. O. and Tuncer, B. (1982) ‘An Empirical Test of the Infant Industry Argument’, American Economic Review 72: 1142–152. Krugman, P. (1979) ‘A Model of Innovation, Technology Transfer, and the World Journal of Political Economy 87 (2): 253–66. Distribution of Income’, Kuznets, S. (1966) Modern Economic Growth: Rate, Structure and Spread . New Haven, CT and London: Yale University Press. Lal, D. (1994) Against Dirigisme: The Case for Unshackling Economic Markets . San Francisco: International Center for Economic Growth, ICS Press. Landes, D. (1969) The Unbound Prometheus: Technological Change and Indus- trial Development in Western Europe from 1750 to the Present. Cambridge: Cambridge University Press.

192 Growth Identifi cation and Facilitation: The Role of the State in the Dynamics of Structural Change | 179 Lin, J. Y. (2010) New Structural Economics: A Framework for Rethinking Development . Policy Research Working Paper No. 5197. Washington, DC: World Bank. Economic Development and Transition: Thought, Strategy, and Lin, J. Y. (2009) . Cambridge: Cambridge University Press. Viability Lin, J. Y. and Chang, H.-J. (2009) ‘DPR Debate: Should Industrial Policy in Developing Countries Conform to Comparative Advantage or Defy It?’, Development Policy Review 27 (5): 483–502. Lin, J. Y. and Ren, R. (2007) ‘East Asian Miracle Debate Revisited’ (in Chinese), ( Economic Research Journal ) 42 (8): 4–12. Jingji Yanjiu Lin, J. Y. and Tan, G. (1999) ‘Policy Burdens, Accountability, and the Soft American Economic Review: Papers and Proceedings Budget Constraint’, 89 (2): 426–31. Lin, J. Y. and Zhang, Pengfei (2007) Development Strategy, Optimal Industrial Structure and Economic Growth in Less Developed Countries . CID Working Paper No. 19. Cambridge, MA: Harvard University Center for International Development. List, F. (1841 [1930]) ( The Das Nationale System der Politischen Ökonomie National System of Political Economy ). Vol. 6. Schriften, Reden, Briefe . A. Sommer (ed.). Berlin: Reinmar Hobbing. Maddison, A. (2006) The World Economy . Paris: Organisation for Economic operation and Development. Co The World Economy: A Millennial Perspective . Paris: Maddison, A. (2001) OECD. Mathews, J. A. (2006) ‘Electronics in Taiwan: A Case of Technological Learning’ Technology, Adaptation, and Exports: How Some in Vandana Chandra (ed.), . Washington, DC: World Bank. Developing Countries Got It Right McKenzie, R. B. (2007) ‘Industrial Policy’ in D. R. Henderson (ed.), The Concise Encyclopedia of Economics. Liberty Fund (www.econlib.org/library/Enc1/ IndustrialPolicy.html). Mistry, P. S and Treebhoohun, N. (2009) The Export of Tradeable Services in Mauritius: A Commonwealth Case Study in Economic Transformation . London: Commonwealth Secretariat. Monga, C. (2006) ‘Commodities, Mercedes-Benz, and Adjustment: An Episode in West African History’ in E. K. Akyeampong (ed.), Themes in West Africa’s History . Oxford: James Currey. Mottaleb, K. A. and Sonobe, T. (2009) ‘Inquiry into the Rapid Growth of the Garment Industry in Bangladesh’. Tokyo: Foundation for Advanced Studies on International Development (mimeo). Mowery, D. and Rosenberg, N. (1993) ‘The US National Innovation System’ in R. Nelson (ed.), National Innovation Systems: A Comparative Analysis . Oxford: Oxford University Press.

193 180 | New Structural Economics O’Brien, T. M. and Rodriguez, A. D. (2004) ‘Improving Competitiveness and Market Access for Agricultural Exports through the Development and Appli- cation of Food Safety and Quality Standards: the Example of Peruvian Aspara- gus’. San Jose: Agricultural Health and Food Safety Program, Inter-American Institute for Co operation on Agriculture. O’Brien, P. and Keyder, C. (1978) Economic Growth in Britain and France 1789–1914: Two Paths to the Twentieth Century. London: George Allen and Unwin. Otsuka, K., Ranis, G. and Saxonhouse, G. (1988) Comparative Technology Choice in Development: The Indian and Japanese Cotton Textile Industries . London: Macmillan Press. Owen, G. (1966) Industry in the USA . London: Penguin Books. Pack, H. and Saggi, K. (2006) ‘Is There a Case for Industrial Policy? A Critical Survey’, 21 (2): 267–97. World Bank Research Observer Pritchett, L. (1997) ‘Divergence, Big Time’, Journal of Economic Perspectives 11 (3): 3–17. Redding, S. (1999) ‘Dynamic Comparative Advantage and the Welfare Effects of Oxford Economic Papers 51 (1): 15–39. Trade’, Rhee, Y. W. (1990) ‘The Catalyst Model of Development: Lessons from Bangladesh’s Success with Garment Exports’, 18 (2): World Development 333–46. Rhee, Y. W. and Belot, T. (1990) Export Catalysts in Low-income Countries . World Bank Discussion Papers No. 72. Washington, DC: World Bank. Rodrik, D. (2009) ‘Industrial Policy: Don’t Ask Why, Ask How’, Middle East 1 (1): 1–29. Development Journal Journal of Political Romer, P. M. (1990) ‘Endogenous Technological Change’, Economy 98 (5): part II, S71–S102. Sawers, L. (2005) ‘Nontraditional or New Traditional Exports: Ecuador’s Flower Latin American Research Review 40 (3): 40–66. Boom’, Brookings Review October: Schultze, C. (1983) ‘Industrial Policy: A Dissent’, 3–12. Shapiro, H. and Taylor, L. (1990) ‘The State and Industrial Change’, World Development 18 (6). Thurow, L. (1980) The Zero-Sum Society: Distribution and the Possibilities for Change. New York: Basic Books. Trebilcok, C. (1981) The Industrialization of Continental Powers, 1780–1914 . London: Longman. Wade, R. (1990) Governing the Market. Princeton, NJ: Princeton University Press. World Bank (1995) Bureaucrats in Business: The Economics and Politics of Government Ownership . New York: Oxford University Press. Zoellick, R. B. (2010) The End of the Third World? Modernizing Multilateralism for a Multipolar Growth. Washington, DC: World Bank.

194 III Comments and Rejoinder* † Introduction, by Dirk Willem te Velde Assessing the role of the state in the dynamics of structural change, as Lin and Monga do, is not new. However, the novel and noteworthy contribu- tion of their article is that it provides a practical procedure to identify and facilitate growth through a six-step procedure. This approach complements existing approaches such as the growth diagnostics approach (Hausmann et al., 2005), the competitiveness approach (Porter and Schwab, 2008), the investment-climate survey approach (World Bank, 2005), or the capability approach (Cantore et al., 2011): Step 1: Governments should select dynamically growing countries with • a similar endowment structure and with about 100% higher per capita income than their own average. They must then identify tradeable indus- tries that have grown well in those countries for the previous 20 years. rms are already present in those indus- • Step 2: If some private domestic fi tries, they should identify constraints to technological upgrading or fur- rm entry, and take action to remove such constraints. ther fi rms are present, policy-makers Step 3: In industries where no domestic fi • may try to attract foreign direct investment (FDI) from countries listed rm-incubation programmes. in step 1, or organize new fi cation and Facilitation: The Role of the State in the Adapted from “DPR Debate: Growth Identifi * 29 (3), May 2011 (DOI: 10.1111/ Development Policy Review, Dynamics of Structural Change,” j.1467-7679.2011.00534.x). © 2011 Lin, J., Monga, C., te Velde, D. W., Tendulkar, S. D., Amsden, A., Amoako, K. Y., Pack, H., and Lim, W. © 2011 Overseas Development Institute. Reprinted with the permission of John Wiley and Sons / Blackwell Publishing. 181

195 182 | New Structural Economics • ed in step 1, the government Step 4: In addition to the industries identifi should also pay attention to spontaneous self-discovery by private enter- prises and support the scaling up of the successful private innovations in new industries. • Step 5: In countries with poor infrastructure and a bad business environ- ment, special economic zones or industrial parks may be used to over- come barriers to fi rm entry and FDI and encourage the formation of industrial clusters. Step 6: The government should be willing to compensate pioneer fi rms • ed above with tax incentives for a limited period, in the industries identifi cofi nancing for investments, or access to foreign exchange. Without reviewing the article in detail, one can pose some obvious questions regarding the framework. For example, the fi rst step of the new framework requires a country to identify sectors in which it has a com- parative advantage on the basis of goods and services that have been pro- duced for 20 years in similar countries. But what happens if the current circumstances have changed so fundamentally that a comparison with the past is less informative (for example, the rise of emerging power, new com- munications technology including fi ber optic cables, new production pro- cesses, new global rules and institutions, climate change)? What if demand patterns have shifted so fundamentally (for example, the rise of the middle classes in China and India, global fi nancial crisis) that different products are now more successful compared with those in the past? What if there are measurement issues, for example information communication technol- ogy (ICT) service exports may be quite diffi cult to measure but this might be just the export in which small landlocked countries with access to a good quality fi ber optic cable have a comparative advantage? What if com- parator countries are actually very different geographically or institution- ciently ally? So one could have doubts that step 1 would actually be suffi informative under all these circumstances. Or with respect to step 2, the argument is made for government support (to remove binding constraints to growth), but it is not clear how a country knows which policy or instrument works best in which case (a comment which can also be made on the growth diagnostics literature, for example). So even if the right industry and constraints are identifi ed, the wrong policy instrument might still lead to an unintended outcome. This links to a wider

196 Comments and Rejoinder on Growth Identifi cation and Facilitation | 183 point which is underemphasized in the article, namely, that the conditions under which policies are implemented (government capacities, political incentives, the nature of state-business relations) can also be crucial for the success of industrial policies as the need for policies to follow the compara- tive advantage of a country (which is implicit in steps 1–6). However, these issues aside, the growth identifi cation and facilitation framework is presented as an alternative to the existing frameworks for analysis. We asked fi ve distinguished experts to comment. Comments from Experts Suresh Tendulkar, professor of economics, retired, at the Delhi School of Economics, University of Delhi, comments on the distinction between the roles of the state in facilitating growth and in identifying sources of growth. Tendulkar accepts that there is an important role for growth facilitation, but is less certain about the role of the state in growth identi- fi cation. He asks how the over-enthusiastic government can be restrained from taking on much more than it can effectively handle, and refers to the South Asian context. He also cautions that time-bound incentives are not straightforward. Alice Amsden of the Massachusetts Institute of Technology makes three points. First, she argues that industrial policies in countries ranging from the Middle East’s energy belt to the Asian manufacturing corridor and the BRIC (Brazil, Russia, India and China) economies have been more success- ful in practice than is portrayed by Lin and Monga. Secondly, she suggests that the authors’ two-track approach (identifi cation and facilitation) is bet- ter than Michael Porter’s value-chain approach and Ricardo Hausmann’s ‘jumping-monkey’ model, because it involves building up business knowl- edge which is a more complete approach than is embodied in the other concepts. Thirdly, she argues that Lin and Monga’s model can be enhanced by using industrial policy to invest overseas and attract skills. K. Y. Amoako, founder and president of the African Center for Eco- nomic Transformation (ACET) in Accra, argues that Lin and Monga’s approach is a practical and useful starting guide, suggesting that such a pragmatic approach is to be welcomed especially because it comes from the World Bank which traditionally has not believed in a proactive indus- trial policy. However, Amoako also argues that the article pays too much

197 184 | New Structural Economics attention to supporting products that follow a comparative advantage and too little attention to the acquisition of new technological capabilities and learning. Successful industrialisation has not always been based on com- petitive markets, and African countries have not always succeeded, despite following their comparative advantage. Howard Pack, Department of Business and Public Policy at the Whar- ton School, University of Pennsylvania, suggests that the algorithm involved in selecting industries on the basis of a selection in richer compa- rable countries and then following the country’s comparative advantage is problematic. This is in part because the economic structure of the richer country could be the result of distorting policies, and in part because a formidable set of policies is required for successful policies which go beyond the mere identifi cation of potential products. Governments should address a whole list of issues which are likely to go beyond the capacity of any government. Wonhyuk Lim, director of policy research at the Center for Interna- tional Development, Korea Development Institute, agrees that policy advice based on the ideas of comparative advantage, self-discovery and the facilitating state will help policy-makers in the early stage of develop- ment, but argues that more needs to be done to move a country beyond the middle-income trap. South Korea defi ed its comparative advantage by moving into heavy and chemical industries by means of building specifi c skills, fi lling specifi c gaps in the value chain, relying on a select set of busi- ness groups and strategic choice. Rejoinder Lin and Monga close the debate with a rejoinder. They answer many of these points directly, for example by arguing that growth identifi cation and facilitation go hand in hand implicitly. They also agree with many of the constructive comments made, for example those by Professor Amsden on the importance of gaining experience in managing business organiza- tions, or on the way the model can be enhanced by using industrial policy to invest overseas. But there are also some points where the reader is invited to make up his or her own mind. For example, Lin and Monga repeat time and again that their model focuses on ‘development of industries consistent with latent

198 Comments and Rejoinder on Growth Identifi cation and Facilitation | 185 comparative advantages’. It is unlikely that many will object to this being indeed their intent, but they might object to whether it can be achieved in practice. Or, when Lin and Monga say that it ‘simply points to the neces- sity to set clear, transparent and rigorous criteria that mitigate ... support [for] uncompetitive industries’, few would doubt that that is important, but many would rather question the ability of governments to implement these criteria successfully, as Tendulkar emphasized. A further discussion, which has not been completely closed, is to what extent countries should follow a path consistent with their static com- parative advantages or whether they should create dynamic comparative advantages (as implied by Amoako). Another debate left somewhat open concerns knowledge (as brought up by Pack). Government offi cials are unlikely to know enough to be able to support industries in the way that is intended and to target the develop- ment of industries consistent with latent comparative advantages. When do they know enough? Conclusion The article by Lin and Monga introduces a useful and practical 6-step plan for governments to facilitate growth which seems a credible alter- native to the existing frameworks (growth diagnostic framework, com- petiveness analysis, capability analysis, investment-climate analysis). In doing so, it reinvigorates the debate on the appropriate role of growth policy in development, which has once again become a topical issue in development economics. Most comments on Lin and Monga’s contribution to the literature agree with the importance attached to the role of the state in growth facilitation. Many also value the practical policy advice embodied in the approach. However, there is some disagreement about the capacity of the state to deliver on growth identifi cation. Furthermore, the two-track approach in Lin and Monga relies on coun- tries following their comparative advantage, and several comments suggest that countries actually need to defy their comparative advantage involving a more complex set of policies than that suggested by the framework. In conclusion, most agree that the approach could establish useful ingre- dients in successful industrial development and comes close to a recipe for

199 186 | New Structural Economics growth. But there are doubts about the ability of the cooks to use the recipe and turn the ingredients into a fully cooked meal. Nonetheless, it is hoped that the application of the six-step procedure will produce relevant infor- mation that can help countries to grow faster. Indeed, apart from questions on what is the binding constraint in a country, developing-country policy- makers often ask: how did other countries achieve in the past what we would like to achieve now? Note Dirk Willem te Velde is head of Investment and Growth and Trade Programmes † at the Overseas Development Institute, London. References Cantore, N., Ellis, K., Massa, I., and te Velde, D. W. (2011) ‘Managing Change and Cultivating Opportunity: The Case for a Capability Index Measuring Countries’ Ability to Manage Change Effectively’. KPMG-ODI report. London: KPMG and Overseas Development Institute. Hausmann, R., Rodrik, D., and Velasco, A. (2005) Growth Diagnostics. Cambridge, MA: Harvard University Center for International Development (www.hks.harvard.edu/fs/rhausma/new/growthdiag.pdf). Porter, M. and Schwab, K. (2008) The Global Competitiveness Report 2008– Geneva: World Economic Forum. 2009. World Bank (2005) World Development Report on the Investment Climate. New York: Oxford University Press for the World Bank. † Suresh D. Tendulkar Lin and Monga rightly admit that economists have been intrigued by the mystery of economic growth. Gone are the days of self-assurance and confi dence of Walter Rostow in the 1950s who provided a predict- able and certain roadmap of growth for every underdeveloped country. A large number of economic theoretical analytical models fl ourished shortly thereafter. While Rostow did not remain uncontested in his time (Kuznets and Gerschenkron readily come to mind), economists have been much more circumspect since then, despite or possibly because of the wealth of data becoming available. And rightly so. The growth brigade started with 1.5 to 2.0% a year in per capita terms in the eighteenth century with the Netherlands and Britain. It was joined by the United States, Germany

200 Comments and Rejoinder on Growth Identifi cation and Facilitation | 187 and France in the nineteenth century, and the bar was raised signifi cantly by the latecomers: Japan in the 1950s, Hong Kong, Singapore, South Korea and Taiwan in the 1960s, China, Thailand, Malaysia and Indo- nesia in the 1970s, India in the 1980s, and other emerging economies since then. With the two most populous countries, China and India, join- ing the brigade, the population-weighted inter-country inequality in per capita GDP has shown a welcome decline. The number of countries in the brigade, however, remains countable. While we are reasonably cer- tain about common ex-post descriptive features associated with rapid growth, very few countries have managed to grow rapidly (say, 3% or higher in per capita annual terms) in a sustained fashion over more than two decades. Nor do we know nor can we predict with certainty what triggers spurts of sustained growth in any given country. Undeterred, Lin and Monga embark on the ambitious and admirably persuasive enterprise of setting a two-fold agenda for government inter- vention in growth facilitation (provision of hard and soft infrastructure) and growth identifi cation (continuing technological upgrading and diver- sifi cation through anticipatory industrial policy) for developing countries striving for rapid economic growth for poverty reduction. The objective is indeed laudable beyond doubt. What is also on their side is the lessons they seek to draw from well-documented ex-post analyses of the successes as well as failures of state interventions during the pre- as well as post- World War II period. While my heart wants their enterprise to succeed, my head remains uncomfortable. Let me therefore express the sources of my discomfort with introspective comments which are indeed coloured by my South Asian, especially Indian, lenses, while recognizing that the authors have an East and South-East Asian, including Chinese, perspective. Less contentious and more readily acceptable is the important role of the government in growth facilitation, that is, the provision of hard (adequate networks of road, rail and air transport and communications, electricity grids and other public utilities) and soft (basic governance including competitive market institutions, fi nancial system and regula- tion, basic health, and primary and secondary education services includ- ing vocational training) infrastructure. Because of externalities and their public good character, this is indeed the core legitimate domain of the government. The adequacy of physical facilities and the cost-effectiveness

201 188 | New Structural Economics of operation and delivery of corresponding public services cut down the transaction costs of exchange for the private sector and impart competi- tive edge to the economic structure. cult, uncertain and hence contentious is the role of growth More diffi identifi cation. Lin and Monga’s excellent historical analysis brings out many failures and fewer successes, more over-enthusiastic though often well-intended policy excesses, more heavy-handed than ‘right’ mixes of non-intrusive interventions, persistence of once successful policies being indiscriminately extended beyond their effective time span, indiscriminate extension of the public sector well beyond the minimum core and discre- ing the dynamism of functioning markets and tionary controls often stifl leading to rampant rent-seeking activities. While one may readily concede that some ex-ante judgments regarding the choice of industries consistent with endowment structure and potential comparative advantage may go wrong, the discipline of time-bound withdrawal of concessions in the face of ineffectiveness and even timely exit from some patently unsuccessful policies including subsidies and tariff protection is diffi cult to obtain. The question ironically becomes: how do you control the over-enthusiastic government from taking on much more than it can effectively handle? In cation and consequent ex-ante nurturing of my assessment, growth identifi picked winners and keeping them under a tight, time-bound leash are a much more diffi cult and risky enterprise based on the South Asian experi- ence in this context. This does not rule out occasional, lucky successes, but this has to be ascertained by the experience of actual experiential judgment and confi dence about the existence of warranted discipline. Note † Suresh Tendulkar is Professor of Economics (Retd.), Delhi School of Econom- ics, University of Delhi, India. † Alice Amsden In their critical essay on the role of the state, Justin Lin and Célestin Monga concentrate on the concept of comparative advantage as the clue to slaying the dragon of underdevelopment. Are they moving ahead or simply stand- ing still?

202 Comments and Rejoinder on Growth Identifi cation and Facilitation | 189 Comparative advantage can be construed deductively or inductively, as an abstract theory or as something that bubbles up from below. Economists mostly conceive of it deductively. Lin and Monga argue that pervasive fail- ures in developing countries are mostly due to the inability of governments to come up with good criteria for identifying industries that are appropri- ate for their country’s endowment structure and level of development, i.e., their ‘latent’ comparative advantage. But the two vast de-colonised regions with the most successful industrial policies, fast GDP growth and steeply falling poverty, the Far East and Middle East, followed their comparative advantage by nosing around their neighbours; if an industry grew next door, this was taken as de facto proof of its comparative advantage—what more concrete evidence can there be? If an export-processing zone works, if a national oil company raises domestic supply and tax revenues more than an international oil company, other countries try to follow suit, and having a blueprint to follow (imitation has been overwhelmingly South- South) makes it easier for them to succeed. Two great regional role models have evolved, whose meaning of ‘indus- trial policy’, unlike a theory’s, meanders, depending on who has joined and exogenous jolts. (The World Trade Organization’s restrictions on sub- sidies, I would argue, have driven the industrial policies of ‘emerging’ and eld’ ‘emerged’ economies underground, creating a sort of ‘level playing fi of subterfuges.) The OPEC development role model (as distinct from the OPEC price cartel), which started in Iran and Saudi Arabia with Mexico’s 1938 oil nationalisations as example, employs millions of workers from as far away as Bangladesh and approximates the ‘labour-scarce, resource- rich economy’ described in the 1950s by Hla Myint. The Far East role model, running along the lines of the labour-surplus economy analyzed by W. Arthur Lewis, formed around post-War Japan, which was neither developed nor underdeveloped at the time and which targeted industries according to simpler criteria than Lin’s or Monga’s: government support went to industries with dense linkages and high productivity growth rates internationally, fi rst silk and cotton textiles (see Amsden and Suzumura, 2001). Large countries like the BRICs have industrial policies that strad- dle Asia’s manufacturing corridor and the Middle East’s energy belt, so in fact industrial policy has probably been more successful than the premise of Lin and Monga’s article, that ‘most have failed’.

203 190 | New Structural Economics Industrial policy has failed in the peasant export economy, the third pro- totype of Lewis, Myint and other classical economists, but it is questionable if this is a viable economic formation, given hyper-fast population increases (the 30 countries in 2005-10 with the highest estimated population growth rates include 24 smallholder-type economies, 23 from Africa). Landless- ness, unemployment and underemployment are high, but labour costs are not low enough and manufacturing experience is not deep enough to com- pete against labour-surplus economies like India. What, besides population planning, is the best industrial policy to help peasant economies, many of which have newly discovered energy and mineral resources such as Sudan, Angola, Cameroon and Ghana: the OPEC development role model, with nearby Nigeria teaching what not to do, or Lin and Monga’s ‘important distinction’ between two types of government intervention? The latter differentiate policies that facilitate structural change by overcoming infor- mation, and coordination and externality issues’ from those that aim at protecting certain selected fi rms and industries that defy the comparative advantage determined by the existing endowment structure (once called dynamic comparative advantage). Their distinction seems sensible but vague, at least for the vast energy and mining sector, the great hope of the defunct peasant economic confi guration. Developing countries without unlimited labour supplies, rich natural resources or a credible role model close by, such as Colombia, Morocco, Nicaragua and Nepal, are in need of advice about how to ‘pick winners’. Lin and Monga’s criteria face competition from those of Michael Porter (the value chain) and Ricardo Hausmann (the jumping monkey). I think Lin and Monga’s two-track approach is better than theirs because, if I understand its broad implications, comparative advantage boils down to having ‘knowledge of a business’, an empirical construct that rests on a roadmap of where an industry is going, production engineering skills, and project execution capabilities to get an investment up and running. ( Taiwan’s electronics fi rms invested in producing CD-ROMs, despite falling world prices, once government R&D had skirted restrictive patents and there was a sense that Japan would graduate to producing DVDs.) By contrast, where a monkey jumps or where a country situates itself on a value chain largely involves decisions determined by the narrow criterion of factor proportions.

204 Comments and Rejoinder on Growth Identifi cation and Facilitation | 191 The challenge Lin and Monga face is how to accelerate the growth of professionally managed business organizations and their unique skills. Business knowledge depends on experience, which I would say is the critical missing element in economies deprived of East Asia’s pre-War manufacturing culture (which was fortuitously strengthened by Japan’s regional war preparations). Experience can be understood using a learn- ing curve, except that learning is not repetitive. Experience depends on gaining tacit and undocumented knowledge of multiple activities that may change simultaneously—a harder task than gaining ‘information’ (a fact). How can an industrial policy hasten the acquisition of experience? Two possibilities that I think would move Lin and Monga’s argu- ment forward would be to follow what the role models of East Asia and the Middle East are doing, and use industrial policy to: (i) invest overseas (outward foreign direct investment); and (ii) reverse the fl ow of talented brain drain (and create a level playing fi eld for local talent), which may change a small country’s comparative advantage overnight. When Malaysia’s government reformed its industrial policy towards its Malay population—instead of subsidising Malay-owned enterprises in Malaysia, it began acquiring foreign companies and giving Malays equity in them—it developed skills to choose a specifi c overseas asset to buy, the same ‘knowledge of a business’ it needs to target a successful investment at home. Similarly with SABIC, Saudi Arabia’s state-owned petrochemical company, which acquired General Electric’s chemi- cal business in China, the feedstock costs of both petrochemicals and chemicals draw on intelligence of the supply of and demand for oil. An outward FDI can thus have positive spillovers on domestic income, employment, income distribution (as in Malaysia’s case) and picking winners. Reversing talented brain drain by creating economic opportunities at home is a costly challenge, but one with potentially high returns, because embedded in returnees’ experience is an inductive clue as to the specifi c industries a government should support. Globalism’s imperfections, moreover, have created a willingness on the part of some professionals to return home, opportunities permitting. Morris Chang, a top executive in Texas Instruments, went back to Taiwan to run its new state-owned semi- conductor company because he claimed that at TI he had hit a ‘ yellow

205 192 | New Structural Economics glass ceiling’. West African executives in Unilever talk about a ‘black glass ceiling’. Industrial policy is inherently nationalistic, and the role of government is to nurture nationalism of a productive type. Note † Alice Amsden is Barton L. Weller Professor of Political Economy, Massachu- setts Institute of Technology (MIT), Cambridge, MA. Reference Amsden, A. H. and Suzumura, K. (2001) ‘An Interview with Miyohei Shinohara: Journal of the Japanese and Nonconformism in Japanese Economic Thought’, 15: 341–60. International Economies † K. Y. Amoako Lin and Monga’s article emphasizes the importance for sustained growth of technological innovation, industrial upgrading and diversifi cation, and improvements in infrastructure and institutions. It notes that, while the market mechanism is essential for effi cient resource allocation, it may not be suffi cient to enable fi rms to overcome the problems of information, co-ordination and externality that often stand in the way of achieving the above-listed requirements for sustained growth in developing countries. It then points out that the historical evidence shows that governments in almost all the successful countries (i.e., the industrialized countries and the recent East Asian success stories) played and are continuing to play pro-ac- tive roles in helping fi rms in their economies to overcome these problems. It further states that governments in almost all developing countries have also tried to intervene in their economies for similar reasons, but most have failed. The article’s central thesis is that the failures were due to the fact that the governments intervened by trying to defy their economies’ exist- ing comparative advantage, meaning that they tried to promote products that did not refl ect their relative factor endowments, particularly of capital and labour. To address this, it proposes a process that policy-makers in developing countries may follow to pick industries or products to promote or facilitate. The authors argue that, for a developing country to diversify its exports, the government should ‘... identify the list of tradeable goods and services

206 Comments and Rejoinder on Growth Identifi cation and Facilitation | 193 that have been produced for about 20 years in dynamically growing coun- tries with similar endowment structures and per capita income that is about 100% higher than their own’, and then remove the binding constraints or take necessary measures to facilitate their export development, including attraction of Foreign Direct Investment (FDI). The government should also look for successful discoveries by domestic enterprises and provide appro- priate support. The article provides examples of the type of facilitation or support that may be given. Overall, I fi nd this proposal a very practical and useful starting guide for a government keen on diversifying and upgrading its country’s exports. It is to be welcomed that an article that looks at the role for the state in late-industrialisation in such a pragmatic manner emanates from the World Bank, which spent a great part of the 1980s and 1990s denying any positive or pro-active role for the state in industrialisation and promoting liberalization and privatization programmes to support that view. With the demonstrably superior economic performance of the East Asian countries (for example, Korea, Taiwan, Singapore and Hong Kong) in recent times, in the fi rst three of which the governments pursued active industrial poli- cies, the Bank has had to revise its position, starting with the East Asian Miracle Study (World Bank, 1993). It is to be hoped that this article by Lin and Monga will help move the Bank further along on the line of pragma- tism. As they write, paraphrasing Rodrik (2009), “. . . instead of advising the governments in developing countries to give up playing the facilitating role, it is ‘more important to understand better why some countries have been able to succeed while most others have failed, so that it is possible to advise the governments to do the right things and avoid the mistakes’”. While I fully share the authors’ view that the state has a positive role to play in facilitating industrialisation and commend their proposal, I would have liked to see a little more fl exibility in the way they use comparative advantage (i.e., relative factor proportions) to review the cases of success and failure in diversifi cation and industrial upgrading. The apparent use of the Hecksher-Olin-Samuelson (HOS) framework to interpret industrial policy appears to me to be too confi ning theoretically, and also does not seem to adequately explain the country experiences. Conceptually, the idea that a country should try to focus on products for which it has the required factors in relative abundance makes eminent sense. However, this assumes competitive markets, both internationally

207 194 | New Structural Economics and domestically, which may not be the case. It is also very static and may not take account of the likely demand, price, and prospects of technologi- cal change and learning of products on the world market. The fact that a country’s factor endowments enable it to produce product A more cheaply today relative to product B does not necessarily mean that, in the medium to long term, it is better off producing A instead of B, if in fact B has bet- ter demand and technological change and learning prospects. Admittedly, producing A today is more likely to raise national income and, presum- ably, thereby savings, which would enable the country to augment its capi- tal. However, if the country is aiming to ‘catch up’ in industrialisation, it would at some point have to defy its existing comparative advantage and take non-marginal steps away from its current productive structure (i.e. try to produce B). Certainly, this would be a more risky move, but it would be taking high risks for potential high rewards. For me, therefore, the policy questions are twofold: (i) what should be the mix of the A and B products in industrial policy at any given time, and how should they change over time? And (ii) having chosen a particular mix, especially a mix including B products, what are the complementary policies that minimize the risks and enhance the chances of success? The latter question brings to the fore the issue of the acquisition of technologies and technological capabilities. In the HOS framework, which seems to inform Lin and Monga’s article, this issue is side-stepped by the assumption that technologies are equally acces- sible and can be effi ciently operated by all producers. This assumption is clearly problematic. In fact, to my mind, the core development prob- lem that industrial policy should address is precisely the access, effi cient deployment, absorption and adaptation of technology (Lall, 2003, 2004). To meet this challenge takes more than focusing on existing comparative advantage as determined by existing relative capital-labour ratios. In fact, if each country diversifi ed and upgraded its industries only by try- ing to break into markets that countries ahead of it on the ‘industrial ladder’ were becoming less competitive due to rising labour costs, then one would expect the industrial rankings of countries to be rather static over time. There would be hardly any instances of ‘catch up’ or overtaking; the US and Germany would not have overtaken Britain in industrialisation; Japan would not have become dominant in automobile exports; nor would Korea have become among the most effi cient steel producers. My reading of the

208 Comments and Rejoinder on Growth Identifi cation and Facilitation | 195 experiences of Japan and the East Asian countries is that the governments promoted a mixture of A and B types of industries, with the mix changing over time, but supported the industries with a host of fi scal, exchange-rate, trade and credit instruments. They also built strong institutions, pursued active technology and FDI policies, aggressively developed skills, and pro- actively engaged in industrial restructuring. One cannot be sure that what arose from all these government interventions could always be character- ized as competitive markets that enabled fi rms to develop according to com- parative advantage (see Johnson, 1982; Amsden, 1989; Wade, 1990; Evans, 1995; World Bank 1993, and Chang, 2006). It should also be noted that in Africa many of the industries established during the import-substitution era failed, although a large number of them were engaged in manufactures of textiles and other simple consumer goods that refl ected the comparative advantage of the countries in terms of relative factor endowments. What can be taken from all this is that following comparative advan- tage is very important, but is only one of a whole host of policies, institu- tions, capacities and arrangements that have to be deployed together in order to increase the chances of success of industrial policy. And for a country that wants to accelerate its industrial catching-up, it may be nec- essary for it to defy its current comparative advantage to some extent and promote a carefully selected small sub-set of products that are ‘high-tech’ (i.e., from the point of view of the country’s current production structure). This would require a government that is capable, organized, disciplined and prepared to work closely with the private sector, and yet be able to subject it to rigorous performance criteria. The above are just questions of nuance. As has already been noted, I do think the article makes a valuable contribution by providing a practical and sensible way for countries to initiate industrial policy. At the African Center for Economic Transformation (ACET), we are engaged precisely in exploring ways for African countries to transform their economies through, among other things, sensible industrial policy. We therefore welcome this contribution from Lin and Monga. Note † K. Y. Amaoko is Founder and President, African Center for Economic Trans- formation (ACET), Accra, Ghana.

209 196 | New Structural Economics References Asia’s Next Giant. New York and Oxford: Oxford Amsden, A. H. (1989) University Press. Chang, H.-J. (2006) The East Asian Development Experience: The Miracle, the Crisis and the Future. London: Zed Books, and Penang, Malaysia: Third World Network. Evans, P. (1995) Embedded Autonomy – States and Industrial Transformation. Princeton, NJ: Princeton University Press. MITI and the Japanese Miracle: The Growth of Industrial Johnson, C. (1982) Stanford, CA: Stanford University Press. Policy, 1925–1975. Lall, S. (2004) ‘Selective Industrial and Trade Policies in Developing Countries: Theoretical and Empirical Issues’ in C. Soludo, O. Ogbu and Ha-Joon Chang (eds.), The Politics of Trade and Industrial Policy in Africa – Forced Consensus? Trenton, NJ and Asmara, Eritrea: Africa World Press, Inc. Lall, S. (2003) ‘Reinventing Industrial Strategy: The Role of Government Policy in Building Industrial Competitiveness’. Paper prepared for the Intergovernmental Group on Monetary Affairs and Development (G-24). Rodrik, D. (2009) ‘Industrial Policy: Don’t Ask Why, Ask How’, Middle East Development Journal 1 (1): 1–29. Governing the Market. Wade, R. (1990) Princeton, NJ: Princeton University Press. World Bank (1993) The East Asian Miracle: Economic Growth and Public Policy. New York: Oxford University Press. † Howard Pack Lin and Monga have covered an enormous range of issues in this intrigu- ing article. They correctly argue for a rethinking of whether a more activist policy is necessary to stimulate manufacturing development in the least industrialised economies, especially in sub-Saharan Africa which has a rel- atively low level of manufacturing share of GDP and also surprisingly little small-scale manufacturing. It is worth noting that the same is true of the Arab economies (Noland and Pack, 2007). Given the growth in population and labour force and the need to fi nd new sources of employment, the issue is of great importance. The main contribution of the article is a reiteration of the need to conform with comparative advantage in development, an argument made cogently in Lin’s Marshall lectures, while seeking to trans- form the economy to more advanced activities. The new argument is an algorithm for identifying successful sectors; this seems to be deeply infl u- enced by East Asian experience. Phrased perhaps too baldly, the algorithm

210 Comments and Rejoinder on Growth Identifi cation and Facilitation | 197 suggests looking at the industries in nations that are more advanced but not too much so—for Korea and Taiwan targeting Japan’s structure as Japan was ‘only’ three times richer. This algorithm is problematic. First, the structure of the richer nation may not be economically optimal even for that country, but is itself the result of distorting policies. Some of Japan’s industrial development between 1868 and 1941 refl ected a felt urgency to develop a serious mili- tary potential which did allow the Japanese to deploy battleships in the ected Russo-Japanese War of 1905. Japan’s metallurgical capabilities, refl partly in post-1950 industrial development, built upon the skills devel- oped in the 1930s that contributed to Japan’s initial success in World War II. Similarly, the USSR in the 1920s and ’30s emphasized heavy industry in an attempt to build up military capacity but also as a perceived path to industrial success. India in the early 1950s emulated the Soviet path which deeply infl uenced some Indians such as Mahalanobis, then chair of the Planning Commission. India violated the Lin-Monga dictum of pursuing labour-intensive industry, but that experience, replicated in many other nations pursuing import-substituting industrialisation, does suggest one of the perils of emulating more ‘advanced’ nations; the body politic may be tempted to throw aside strict economic rationality and pursue technologi- cally advanced and capital-intensive sectors, steel in the 1950s and ’60s and high technology today. Once started down the road of emulation, technocrats may not be able to rein in their bosses. Korea’s economists have shown the considerable cost to the heavy and chemical industry pro- gramme of the 1970s and ’80s and it is possible that the interim costs were suffi ciently large that the protection failed to satisfy the Mill-Bastable test (Yoo, 1990). Moreover, the industrial policy of these nations was embedded within a macroeconomic framework that was conducive to growth, including (World Bank, 1993): • exceptionally high saving and investment rates which continued for four decades, leading to the high growth rates of the capital-labour ratios; • a rapid increase in education measured in years, but also high achieve- ment in science and mathematics measured in international tests and growing tertiary education enrolments in science and engineering;

211 198 | New Structural Economics • an enormous expansion of infrastructure including transportation, ports and roads that was not sectorally targeted; • an emphasis on transferring technology from the rest of the world, whether in the form of technology licensing, FDI, the use of foreign consultants, or in some cases reverse engineering; • the use of export growth as the sine qua non for the continuation of aid rms, government programmes benefi ting fi rms being contingent on to fi rms to increase their productivity and was the exporting. This forced fi impetus for the demand for importing more advanced technologies. But export growth was also abetted by macro policies that limited domes- tic absorption and maintained a relatively constant real exchange rate that allowed potential exporters to gauge potential profi tability with- out worrying about exchange-rate volatility. These macro aspects had a uniform impact, not one that varied among sectors. The fact that the bureaucracy charged with implementing export promotion was largely insulated from political pressure from fi rms, while there was a simulta- neous close monitoring of fi rms that provided considerable information about their problems. Such policies are not easily emulated in most of the countries that need to expand their industrial base. Other problems emerge as well. For example, Yamamura (1986) (in an exhaustive study of Japanese industrial policy in the early 1950s) identi- ed the criteria used by MITI to identify potential competitors with the fi US. Products encouraged were those with high income elasticities (so that additional supply by Japan would not drive down the initial international price) and a large market so that scale economies could be realized. To implement this policy the Japanese government: (a) provided interest- rate subsidies; (b) protected the domestic market by tariffs; (c) limited or rms would precluded entry of new local competitors so that favoured fi not lose their ability to realize scale economies; (d) forbade foreign direct investment in the promoted sector; and (e) precluded potential local com- petitors from borrowing from local fi nancial institutions to avoid the loss of scale economies by favoured fi rms. Korea and Taiwan, the other model nations that might be cited as having pursued a successful industrial pol- icy, implemented some but not all of these measures. Clearly this is a for- midable set of policies to implement and the complete programme is more

212 Comments and Rejoinder on Growth Identifi cation and Facilitation | 199 complex than simply looking at a nation that is richer. This programme cult to implement in any nation, and especially in those with would be diffi a poor education base, limited government legitimacy and the widespread corruption characteristic of many of the least industrialized nations. When Japan embarked on its policies, it targeted stable products whose characteristics were changing slowly. There are few such sectors nowadays. Even inexpensive clothing and shoes undergo remarkably rapid changes in style which demand that a successful fi rm be part of an international supply chain that can keep the supplier up-to-date on fash- ions and quality standards. Moreover, it is not clear how offi cials trying to foster an individual sector would even choose a product. Looking at nd ‘shoes’ but 50 or more international trade statistics, one does not fi categories, each employing a different technology and requiring different production and marketing skills. How many government employees in a ministry of industry could make such choices and carry out the cal- culation of a social cost/benefi t analysis? Moreover, to choose among products would require extraordinary knowledge of both other sectors and the international prospects of the industry with respect to both likely prices and cost structures. Lin and Monga correctly identify lacunae that the government should address to promote structural transformation, in particular, ‘information, co-ordination and externality issues, which are intrinsic to industrial upgrading and diversifi cation’. Kamal Saggi and I (2006) have provided a partial list of the requisite knowledge to deal with these problems, based on our synthesis of the industrial-policy literature. These include knowledge of: • which fi rms and industries generate knowledge spillovers • rms and industries benefi t from dynamic scale economies—what which fi is the precise path of such learning and the magnitude of the cost disad- vantage at each stage of the learning process • which sectors have a long-term comparative advantage • rms and sectors in order to the size of scale economies of different fi facilitate investment co-ordination • an ability superior to that of individual fi rms to learn about their poten- tial competitiveness • the nature and extent of capital-market failures • the magnitude and direction of inter-industry spillovers

213 200 | New Structural Economics • rms from others and from the relative amount of learning by individual fi their own experience • the extent to which early entrants generate benefi ts for future entrants rms’ learning abilities the extent of heterogeneity of fi • whether fi rms trying to reduce production costs also begin a simultane- • ous effort to improve their product’s quality in order to obtain a better reputation • the potential effects of FDI or international trade in solving some of the coordination problems, including a detailed knowledge of which of tens of thousands of intermediates are tradeable • forecasting which fi rms can create new knowledge and discover better production methods • the spillover effects of FDI as well as the likely intensity of their purchase of domestic intermediates. This is obviously a formidable list. It is unlikely that, even if a govern- ment hired several major international consulting fi rms, they would possess the ability to undertake this programme, despite having many PhDs and MBAs on their staffs. The implication for much more poorly educated and compensated government staff with considerably fewer resources is obvi- ous. If this characterization is valid, alternatives to government direction have to be sought. None of this implies that Lin and Monga are incorrect in their insistence on a positive role of government in building hard infra- structure such as roads and soft infrastructure such as a legal system and an environment conducive to business. But these critical requirements are likely to exhaust the capabilities (and fi nances) of almost all national gov- ernments of the least industrialized nations. Note † Howard Pack is Professor of Business and Public Policy, The Wharton School, University of Pennsylvania. References Noland, Marcus and Pack, Howard (2007) The Arab Economies in a Changing World. Washington, DC: The Peterson Institute for International Economics. Pack, Howard and Saggi, Kamal (2006) ‘Is There a Case for Industrial Policy? A Critical Survey’, World Bank Research Observer, Fall.

214 Comments and Rejoinder on Growth Identifi cation and Facilitation | 201 World Bank (1993) The East Asian Miracle: Economic Growth and Public Policy. New York: Oxford University Press. Yamamura, Kozo (1986) ‘Caveat Emptor: The Industrial Policy of Japan’ in Paul Strategic Trade Policy and the New International Econom- R. Krugman (ed.), ics. Cambridge, MA: MIT Press. Yoo, Jung-Ho (1990) ‘The Industrial Policy of the 1970s and the Evolution of the Manufacturing Sector.’ Working Paper No. 9017. Seoul: Korea Development Institute. † Wonhyuk Lim Development may be conceptualized as the result of synergies between enhanced human capital and new knowledge, involving complementary investments in physical and social capital. The fundamental policy chal- lenge is for the state to work with non-state actors and markets to address innovation and co-ordination externalities while minimizing negative gov- ernment externalities. Since the time of the Industrial Revolution, coun- tries that have effectively responded to the innovation and coordination challenges have become successful. The key is for a country to retain the ownership of its development and progressively build up its capabilities to add value and respond to shocks, even as it actively learns from, and engages with, the outside world. The reinforcement of successful experi- ments through the feedback mechanism of performance-based rewards can lead to dramatic changes over time (Lim, 2011). Developing countries typically export primary commodities or start their industrialisation in the assembly and production segment of the value chain in such labour-intensive industries as garments. Most coun- tries fail to move to higher value-added segments along the value chain (such as product design) or to shift to higher value-added sectors (such as machinery and equipment) for two reasons. They either neglect to address externalities in technical education, R&D, and infrastructure development or rush to promote sophisticated industries without the requisite accumulation of skill and scale economies. International bench- marking based on endowment structures and close consultation between the government and the private sector is key to solving information and incentive problems at this stage, when countries try to upgrade their com- parative advantage.

215 202 | New Structural Economics Drawing on development history and economic theory, Justin Yifu Lin and Célestin Monga offer practical advice to developing countries faced with the challenge of identifying promising sectors and facilitating struc- tural transformation. They note that successful developing countries have typically targeted mature industries in countries with an endowment struc- ture similar to theirs and with a level of development not much more advanced than theirs’ (emphasis added). Specifi cally, they propose that the government in a developing country focus on ‘tradeable goods and services that have been produced for about 20 years in dynamically grow- ing countries with similar endowment structures and a per capita income [measured in purchasing power parity] that is about 100% more than their own, while also paying close attention to successful experiments in other sectors. They also advise the government to encourage the experi- mentation, self-discovery, and scale-up efforts of private enterprises by removing constraints, supporting pilots, and providing direct incentives to pioneer fi rms. Building on the ideas of comparative advantage, self-discovery, and the facilitating state, this set of policy recommendations will help policy-mak- ers in developing countries ‘to reap the advantage of backwardness’ in the early stages of development. However, more is likely to be needed if they are to move beyond ‘the middle-income trap’, when catch-up economies may have to take considerable strategic risks to jump into non-mature industries to compete with advanced economies. This is not an easy task. In fact, countries tend to move through the product space by developing goods close to those they currently produce, and can reach the core ‘only by traversing empirically infrequent distances’, which may explain why poor countries fail to converge with the income levels of rich countries (Hidalgo et al., 2007: 482). Korea’s case is illustrative in this regard. Korea exploited its latent comparative advantage to develop mature, labour-intensive downstream industries in the 1960s, much in line with the advice provided by Lin and Monga. However, it did not just wait for its income and skill levels to rise to move into higher value-added industries. Instead, it systematically studied what had to be done to fi ll the missing links in the domestic value chain and move up the quality ladder, and made conscious and concerted efforts to aim for international competitiveness from the outset. It sought

216 Comments and Rejoinder on Growth Identifi cation and Facilitation | 203 to indigenize intermediate inputs imported from foreign upstream indus- tries, through the acquisition of technology, the development of human resources, and the construction of optimal-scale plants aimed at the global market. For instance, in the chemical-textile value chain, it systematically built the linkages backwards from the export of textiles to the production of synthetic fi bers, and the development of basic petrochemicals. Korea had a strong and increasing revealed comparative advantage in light industries when it made the decision to promote heavy and chemical industries in 1973. After benchmarking advanced industrial nations with natural endowments similar to its own, such as Japan, Korea recognized that it had a potential comparative advantage in machinery and equipment industries and began to remove the obstacles to achieving this objective, such as lack of technicians and engineers with the requisite skills in sophis- ticated industries. The government drafted a plan to increase the supply of technicians from 240,000 in 1969 to 1,700,000 in 1981, and established mechanical technical high schools offering full scholarships to poor but talented young students. National universities were called upon to focus on one specialized engineering fi eld related to a nearby industrial complex. In promoting upstream industries in the 1970s, Korea had to make a strategic choice. It could play safe and develop heavy and chemical indus- ciency resulting from tries for the small domestic market and risk the ineffi sub-optimal scales and entrenched protectionism. Alternatively, it could promote these industries for the global market and risk capacity under- utilization and fi nancial distress. It chose the latter option because, despite considerable risks, this promised a dynamically effi cient growth trajectory if the country managed to develop the requisite skills before the fi nancial burden associated with scale economies and complementary investments became overwhelming. To minimize time and exploit scale economies in establishing capital-intensive industries, the government decided to rely on a select group of state-owned enterprises and family-based business groups (chaebol) with successful track records. It considered that scale economies called for regulated monopoly or oligopoly in these industries until demand became large enough to support effective competition (Lim, 2011). Although Korea’s case is but one example, it shows that industrial upgrading requires much more than international benchmarking based on comparative advantage and self-discovery and scale-up efforts. Innovation

217 204 | New Structural Economics and co-ordination externalities in structural transformation demand stra- tegic risk-taking by the public and private sectors. Note Wonhyuk Lim is Director of Policy Research, Center for International Devel- † opment, Korea Development Institute. References Hidalgo, C.; Klinger, A. B.; Barabasi, A.-L. and Hausmann, R. (2007) ‘The Product Space Conditions the Development of Nations’, Science 317: 482–7. Lim, Wonhyuk (2011) ‘Joint Discovery and Upgrading of Comparative Advantage: Lessons from Korea’s Development Experience’ in Shahrokh Fardoust, Yongbeom Kim and Claudia Sepulveda (eds.), Postcrisis Growth and Development: A Development Agenda for the G-20. Washington, DC: World Bank. Rejoinder by Justin Yifu Lin and Célestin Monga We are grateful to Professors Amsden, Tendulkar and Pack, and Drs Amo- ako and Lim for their insightful comments on our article. We fi rst discuss some of the general themes emerging from their analyses. We then respond to some specifi c comments by each of them. 1 General Comments 1.1 On Scope and Justifi cation. It is useful to start by stressing that every country in the world, intentionally or not, pursues industrial policy. This is true not only of the usual suspects such as China, Singapore, France and Brazil, but also of the United Kingdom, Germany, Chile and the United States. This is surprising only if one forgets that industrial pol- icy broadly refers to any government decision, regulation or law that encourages ongoing activity or investment in a particular industry. After all, economic development and sustained growth are the result of con- tinual industrial and technological upgrading, a process that requires public- private collaboration. The theoretical case for industrial policy is quite strong and has been acknowledged in the literature at least since Adam Smith in the lesser known Book 5 of The Wealth of Nations (in which he discusses factor endowments and infrastructure endowments),

218 Comments and Rejoinder on Growth Identifi cation and Facilitation | 205 and Alfred Marshall, who outlined the analytical framework for under- standing externalities and co-ordination. Nowadays, a new wave of skepticism rests on the idea that industrial (sectoral) policy and competition policy are contradictory or at best sub- stitutes. This argument is implicit in some of the comments by Professors Tendulkar and Pack and Dr Amoako. We believe that industrial policy based on the Growth Identifi cation and Facilitation Framework (GIFF) actually enhances competition. By facilitating co-ordination and address- ing externality issues, industrial policy helps many domestic and foreign rms to enter sectors that are consistent with the country’s latent com- fi parative advantage and turn them into overt comparative advantages, and thereby intensifi es competition within the industries and enhances the economy’s competitiveness internationally (Lin and Chang, 2009). Moreover, as shown by Aghion et al. (2010), competition weeds out bad projects, and thus reduces the danger of picking the wrong winner. Also, fi rms may naturally try to differentiate horizontally in order to increase their competitiveness in the market. In such situations, the more intense product market competition is within sectors, the more innovative and competitiveness-enhancing it will be. 1.2 On Discipline and Implementation. - The political-economy diffi culties of implementing any type of public policy are well known. The comments received highlight some of them: the fact that the body politic may be tempted to ignore economic rationality and pursue more sophis- ticated sectors in its zeal to emulate advanced countries; and the pos- sibility of extending even successful policies well beyond their effective timespan, thus creating opportunities for rent-seeking activities. These general governance issues are increasingly well studied in the economic and political-science literature (Tollison and Congleton, 1995; Robinson and Torvik, 2005). These concerns are legitimate but only for the traditional type of industrial policy which encourages fi rms to enter industries that defy comparative advantage. Firms in these industries are not viable in an open, competitive market. Their entry and continuous operation often depend on large subsidies and protection, which create opportunities for rent-seeking and corruption, and make it diffi cult for the government to

219 206 | New Structural Economics abandon interventions and exit from distortions (Lin and Tan, 1999). The GIFF promotes something quite different: the development of industries that are consistent with the economy’s latent comparative advantage. Firms are viable once the constraints to their entry and operation are removed. The incentives provided by the government to the fi rst movers are to be temporary and small, solely for the purpose of compensating for their information externality. In that context, the issues of pervasive rent-seeking and the persistence of government intervention beyond its initial timetable can be mitigated. c Comments 2 Specifi Professor Tendulkar comments on the distinction between the roles of the state in facilitating growth and in identifying new industries for growth. He accepts the state’s important role in growth facilitation, but is more uncertain about its role in growth identifi cation. Referring specifi cally to the South Asian context, he also asks how the over-enthusiastic govern- ment can be prevented from taking on much more than it can effectively handle. We believe that without identifi cation it is hard to determine the type of facilitation that would be desirable. The appropriate hard and soft infrastructures needed to foster industrial upgrading are often industry- c. For the state to play its role in determining and providing the specifi necessary infrastructure (facilitation), government offi cials must form a judgment and make decisions about which particular industries will need it (identifi cation). The two roles are therefore complementary and some- times diffi cult to disentangle. Moreover, because resources and capacity are limited, governments must prioritise their interventions—and explic- itly or implicitly engage in some form of growth identifi cation. The issue of over-enthusiastic governments is not specifi c to South Asia. Many countries in Latin America, Africa and Asia (even China before 1979) exhibited the zealous state syndrome, with governments doing too much in their attempts to promote development. That risk, which is real, does not invalidate the need to deal with externalities and co-ordination. It simply points to the necessity to set clear, transparent and rigorous criteria that mitigate the propensity of governments to over-intervene or to support uncompetitive industries. We offer the GIFF precisely to advise political

220 Comments and Rejoinder on Growth Identifi cation and Facilitation | 207 leaders and the general public on the right way of carrying out industrial policy, and identify clearly what would be the wrong way, so that the prob- ability for governments being over-enthusiastic is reduced. Professor Amsden argues that industrial policies in countries ranging from the Middle East’s energy belt to Asia’s manufacturing corridor and the so-called BRIC economies (Brazil, Russia, India and China) have been more successful in practice than is portrayed in our article. Admittedly, many OPEC countries avoided the resource curse and managed to reach commendable levels of per capita income. However, most of them failed to use the resource rents to facilitate structural transformation in their countries, as carried out by the other resource-rich countries, such as the Scandinavian countries, the United States, Canada or Australia. We submit that the performance of resource-rich countries could be fur- ther enhanced if they used the GIFF to support structural transformation. This would require them to invest an appropriate share of revenues from their natural resources in human, infrastructural and social capital, and create incentives for domestic or foreign fi rms to facilitate the develop- ment and upgrading of industries in the non-resource sector. Their strategy should not be limited to maintaining good governance, keeping natural- resource revenues in sovereign funds and investing in foreign equity mar- kets to insure against commodity price fl uctuations, as is often the case. Professor Amsden also questions the applicability of the GIFF to a peas- ant export economy where the pace of population growth is rapid and landlessness, unemployment and underemployment are high, but where labour costs are not low enough and manufacturing experience is not deep enough to compete against labour-surplus economies like India. Regarding population growth, the same could have been said about Asian econo- mies before their economic take-off in the 1960s. Children represent old- age insurance for many families in poor countries, and the increase in per capita income generally reduces fertility rates because that insurance is less needed and the opportunity cost of raising children rises with the increase in wages. East Asian economies did not have Mainland China’s restrictive family planning system but they experienced similar reductions in popu- lation growth rates. African governments should have devoted the same focus on promoting economic growth as they did on various interven- tions to reduce child mortality. As for the labour costs, those in the formal

221 208 | New Structural Economics sector may not be low, especially in some African countries, as observed by Amsden. But the informal-sector labour costs are unlikely to be high. Moreover, one way out of that dilemma is for those countries to follow the practice of Mauritius in the 1970s (Subramanian and Roy, 2003), namely, allowing wage fl exibility in the special economic zones so as to promote the development of new competitive, labour-intensive industries. Professor Amsden stresses the importance of experience in manag- ing business organizations, which is important indeed. By facilitating the development of latent comparative advantage industries, the GIFF would allow more entrepreneurs to enter competitive manufacturing sectors, gain experience, and prepare their fi rms to upgrade to higher-level industries. Many successful business giants in Japan (Toyota, Sony, Honda), Korea (Samsung, LG, Daewoo), Taiwan-China (Formosa Plastics), or Hong Kong (Tyco-on Li Kasing) started as small businesses with a few employees and a few thousand dollars of investment. They overcame the odds because their promoters were gifted leaders, but they also acquired experience in business management because they operated in an environment that was conducive to sustained growth. She also notes that our model can be enhanced using industrial policy to invest overseas and attract skills. We agree. In a country that is record- ing dynamic growth, the government can employ outward investments to facilitate: (i) the relocation of fi rms that operate in its sunset sectors to other lower-income countries with a similar endowment structure so as to use those countries as export bases and benefi t from their cheaper labour and/ or to get access to their domestic markets; (ii) the acquisition by domes- tic fi rms of foreign fi rms in related sectors in higher-income countries, in order to get access to their technology, management experiences and mar- ket channels; and (iii) the acquisition of resources by domestic fi rms (in resource-scarce countries) from countries where they are abundant. Dr Amoako argues that successful industrialisation has not always been based on competitive markets, and that African countries have not always succeeded despite following their comparative advantage. The GIFF provides a dual-track strategy for government intervention. Fol- lowing comparative advantage, which is only the fi rst track of the GIFF, is a necessary condition for a successful industrial policy. However, that is not suffi cient. For industrial policy to contribute to a country’s growth

222 Comments and Rejoinder on Growth Identifi cation and Facilitation | 209 and structural transformation, the government also needs to play the rst movers and to help facilitation role by providing incentives to the fi them by removing binding constraints to their growth and by coordinat- ing investments in the soft and hard infrastructures that are needed. The likely reason why some African countries have not succeeded despite fol- lowing their comparative advantage is because their governments failed to play their facilitation role. By arguing that the article pays too much attention to supporting products that follow a comparative advantage and too little attention to the acquisition of technological capabilities and learning, he also seems to assume that the GIFF approach promotes static comparative advan- tage. Actually it does the opposite. Our framework promotes upgrading and diversifi cation to new industries and is therefore dynamic in nature. There is a major difference between the GIFF and the theory of dynamic comparative advantage which Dr Amoako has in mind. The latter typi- cally attempts to help fi rms to enter industries that are a country’s future comparative advantage. Because of endowment constraints, fi rms in those industries would not yet be viable in a competitive market even if the government helped them with the co-ordination and externality compen- rms enter industries with sation. By contrast, the GIFF aims at helping fi latent comparative advantage. Under that scenario, fi rms would be viable and require no subsidies or protection once the government provides co- ordination and externality compensation. It should be noted that if Afri- can countries cannot be successful in industries with latent comparative advantage, their probability of success in industries without comparative advantages will be quite small. With the GIFF approach, developing countries can tap into the potential advantage of backwardness, record higher rates of growth and upgrade their industrial structure, income level, and endowment structure faster than the high-income countries. Once their income levels and endowment structures are close to those of high-income countries, they will have gained comparative advantage in advanced industries, which will enable them to compete directly with and even overtake the high-income countries. There- fore, contrary to Dr Amoako’s prediction that ‘if each country diversifi ed and upgraded its industries only by trying to break into markets that coun- tries ahead of it on the “industrial ladder” were becoming less competitive

223 210 | New Structural Economics in due to rising labour costs, then one would expect the industrial rankings of countries to be rather static over time’, it is actually a faster way for a latecomer to catch up with the more advanced countries. Professor Pack believes that targeting industries in richer comparable countries and then following the country’s comparative advantage accord- ingly is a problematic algorithm. His reasoning is two-fold: fi rst, the eco- nomic structure of the richer country could be the result of distorting policies; and second, a formidable set of policies is required for successful cation of potential products. In policies which go beyond the mere identifi support of his skepticism, he offers the examples of Japan and the USSR (which India tried unsuccessfully to emulate). This is a valid warning. Even in successful cases, industrial policy is never a smooth process. It always involves trial and error from governments that put in place good mecha- nisms and channels to learn from mistakes, adjust economic strategies, and minimize the potential costs of bad decisions. However, our framework recommends not only that target countries be richer but also that they have recorded dynamic growth for a long period and where higher incomes and productivity gains in successful industries eventually raise wages and make them less competitive. If they have suc- ceeded in growing dynamically for several decades, it is unlikely that they have followed strategies that defy their comparative advantage. After the Meiji restoration, Japan took the German kingdom of Prussia as a model. According to estimates by Angus Maddison (2010), Germany’s 1 per capita income in 1890 was US$2,428 and Japan’s $1,012. Japan’s was 42% that of Germany, hence Japan’s strategy was consistent with the approach proposed in the GIFF. While Professor Pack’s summary of MITI’s policies in the 1950s and 1960s is quite instructive, the story behind the numbers is fully consistent with the GIFF analysis as well: Japan’s per capita income in 1950, 1960 and 1965 was $1,921, $3,986, and $5,934 respectively, whereas those of the US were $9,561, $10,961, and $13,419. The ratios were as follows: 20%, 36% and 44%. The numbers for 1960 and 1965 are consistent with the principle of the GIFF. The 1950 fi gure was lower than the normal threshold that the GIFF suggests. This is probably due to the fact that Japan was still recovering from the war and its human capital and soft and hard infrastructure were greater than those indicated by its per capita income; a strong indication is the fact that Japan’s per

224 Comments and Rejoinder on Growth Identifi cation and Facilitation | 211 capita income in the 1930s had already reached about 40% that of the US (for example, $2,120 vs. $5,467 in 1935). In contrast to Japan’s story, the USSR in the 1950s was the wrong model for India for two reasons. First, the two countries did not have a similar endowment structure; the USSR was a resource-rich country while India was a resource-poor country. Secondly, the USSR was too far advanced compared with India. According to Maddison, the USSR’s per capita income in 1955 was $3,313, while India’s was $676 (only 20% that of its reference country). The GIFF recommends that latecomers be realistic (and even modest) in their choice of reference countries and target industries. Professor Pack also observes that world trade has undergone remark- ably rapid changes in style and that there are fewer stable products and industries to be targeted today compared with several decades ago. We believe that, despite changes in style and product customization, the divi- sion of labour among countries at different levels of development is still the same. For example, television evolved from black and white to colour and to fl at panel today. The main producing countries have changed from the US before the 1950s to Japan in the 1960s-80s, to Korea in the 1980s- 2000s, and China today. A latecomer entering the market today could go at-panel TV production fi rst, just as into labour-intensive assembly of the fl forerunners did a few decades ago when they decided to compete success- fully in the black and white and colour TV markets. Globalization provides huge potential for industrialisation through specialization. Several decades ago, many low-income countries faced the constraints of their limited market size, high transportation costs and trade barriers, and could not take advantage of the opportunities offered by large-scale manufacturing. With globalization, virtually any country can identify production activities for which it has overt or latent comparative advantage, scale them up and create its own niche in the world market. Precisely because of globalization, the economic development strategy in every country should follow comparative advantage closely. Multinational fi rms are more likely to exploit any small difference in production costs in the determination of their locations of production or procurement systems. Globalization also makes the government’s role in the facilitation process even more important because only with good hard and soft infrastructures,

225 212 | New Structural Economics which reduce transaction costs, can the cost advantage based on endow- ment structure and specialization be realized. Professor Pack provides an impressive list of knowledge requirements about targeted industries that government offi cials would need to know in order to design a successful industrial policy. He questions the capacity of governments in developing countries to meet those requirements. First, all countries at low-income levels tend to lack high capacity by defi nition. Chang (2008) reminds us that, not so long ago, it was not unusual to refer to ‘Lazy Japanese and Thieving Germans’. With the process of economic development taking place, capacity will be enhanced in any society. More important, some of the requirements he identifi es are likely to be relevant only for more advanced industries in high-income countries. For industries with low technical content, the list should be streamlined considerably. Moreover, instead of analyzing the technical nature of various industries to fi nd out the knowledge underpinning them, the private sector and gov- ernment offi cials can rely on the advantage of backwardness and observe what the dynamically growing countries with similar endowment struc- tures are already doing. These successful countries must have already over- come those knowledge challenges either by trial and error or by analysis. Dr Lim agrees that policy advice based on ideas of comparative advan- tage, self-discovery and the facilitating state will help policy- makers in the early stage of development, but argues that more needs to be done to move beyond the middle-income trap. Korea, he writes, defi ed its comparative advantage by moving into heavy and chemical industries by building specifi c skills, fi lling specifi c gaps in the value chain, and relying on a select set of business groups and strategic choice. We agree with his observation that dynamically growing middle-income countries will have some industries that have already reached the global technol- ogy frontier and will eventually face the challenge of taking risks in technology and product innovation. For such industries, the government should continue to play its facilitating role, and use policy instruments similar to those in high-income countries, such as subsidising the R&D activities of individual fi rms by funding basic research in universities or public institutions, granting patents for new inventions, offering prefer- ential taxes and defense and other government procurements, etc. But for other industries that remain well within the global technological

226 Comments and Rejoinder on Growth Identifi cation and Facilitation | 213 frontier even at that level of development, the GIFF could be used to address externalities and co-ordination issues. The Korean government’s encouragement of the development of more capital/technology-intensive industries in the 1970s, as discussed by Lim, is in fact consistent with the need for industrial upgrading due to the change in comparative advantages. The textile, garment, plywood, wigs and other labour-intensive industries were Korea’s comparative advan- tages and very competitive internationally in the 1960s. The success of these labour-intensive industries allowed the country to accumulate human and fi nancial capital. As a result Korea’s endowment structure was upgraded. That process led to a gradual loss of comparative advan- tage in the original industries and allowed the economy to move into new, more capital- and technology-intensive industries. Lim’s account of Korea’s industrial upgrading process in the 1970s, which targeted mature industries in Japan instead of the most advanced industries in the United States, is in fact a good illustration of how the GIFF approach explains the country’s economic success. Note 1. All the estimates of per capita income here are measured in 1990 International Geary-Khamis dollars, taken from Maddison (2010). References Aghion, P.; Dewatripont, M.; Du, L.; Harrison, A.; and Legros, P. (2010) ‘Indus- trial Policy and Competition: Disproving a Fallacy?’. Unpublished presentation. Washington, DC: World Bank. Chang, H.-J. (2008) Bad Samaritans: The Myth of Free Trade and the Secret History of Capitalism. New York, Bloomsbury Press. Lin, J. Y. and Chang, H.-J. (2009) ‘DPR Debate: Should Industrial Policy in Developing Countries Conform to Comparative Advantage or Defy It?’, Development Policy Review 27 (5): 483–502, (Reprinted in chapter II of this volume.) Lin, J. Y. and Tan, G. (1999) ‘Policy Burdens, Accountability, and the Soft Budget Constraint’, American Economic Review: Papers and Proceedings 89 (2): 426–31. Maddison, A. (2010) ‘Historical Statistics of the World Economy: 1–2008 AD’ (www.ggdc.net/maddison/).

227 214 | New Structural Economics Journal of Public Robinson, J. A. and Torvik, R. (2005) ‘White Elephants’, Economics 89: 197–210. Subramanian, A. and Roy, D. (2003) ‘Who Can Explain the Mauritian Miracle? Mede, Romer, Sachs, or Rodrik?’ in D. Rodrik (ed.), In Search of Prosperity: Analytic Narratives on Economic Growth. Princeton, NJ: Princeton University Press. Tollison, R. D. and Congleton, R. D. (eds.) (1995) The Economic Analysis of Rent-Seeking. Cheltenham: Edward Elgar Publishing.

228 IV Applying the Growth Identification and Facilitation Framework: The Case of Nigeria

229

230 IV Applying the Growth Identification and Facilitation Framework: The Case of Nigeria † with Volker Treichel Introduction Nigeria faces a growing employment crisis. Notwithstanding sustained, high and broad-based growth in the non-oil economy, unemployment has not fallen materially since 1999. More importantly, youth unemployment has markedly risen over the same period. While the number of jobs seems to have grown in line with the labor force, most of these jobs have been created in informal family agriculture. Wage employment, however, has declined. Nigeria needs a strategy aimed at increasing the employment intensity and sustainability of its growth performance. 217

231 218 | New Structural Economics How to promote economic growth has been a topic for economic dis- course and research for a long time. Modern economic growth is a pro- cess of continuous technological innovation, industrial upgrading and cation, and of improvements in the various types of infrastructure diversifi and institutional arrangements that constitute the context for business development and wealth creation. While past theories have long empha- sized that market mechanisms are essential to getting relative prices right and thereby facilitating an effi cient allocation of factors, the growth experi- ence in many countries shows that governments often play a crucial role in facilitating industrial transformation. 1 conceptualizes these aspects of growth by New Structural Economics integrating some of the insights from the old structural economics, namely the need to take into account, on the one hand, structural features of devel- oping economies in analyzing the process of economic development and, on the other hand, the role of the state in facilitating structural change in developing countries. The key innovation of the approach is that it consid- ers structural differences between developed and developing countries to be endogenous to their endowment structure. With the economy’s struc- ture of factor endowment—defi ned as the relative composition of natural resources, labor, human capital and physical capital—being given at each stage of development and different from one stage to another, the optimal industrial structure will be different at different stages of development. To move from one stage to another, the market requires industrial upgrading and corresponding improvements in hard and soft infrastructure. The Growth Identifi cation and Facilitation Framework (GIFF) opera- tionalizes key insights of New Structural Economics by developing a meth- odology for identifying sectors where the country may have a latent com- parative advantage and removing binding constraints to facilitate private fi rms’ entry into those industries. The purpose of this paper is to apply the GIFF to Nigeria. The reason for choosing Nigeria is that, in addition to facing a growing employment crisis, Nigeria is also Africa’s most populous 2 country and a regional growth pole. Following an overview of Nigeria’s recent economic performance and its impact on employment, the paper describes the basic rationale underly- ing the GIFF and its methodology. The third section discusses, based on a range of criteria proposed by the GIFF, which sectors or products would be compatible with Nigeria’s latent comparative advantage and should

232 Applying the Growth Identifi cation and Facilitation Framework | 219 therefore be promoted using industrial policy. The fourth section reviews the binding constraints to growth in each of these sectors and discusses specifi c interventions the government could undertake—in collaboration with the private sector—in order to alleviate these constraints. In view of the fact that shortcomings in governance have in the past often under- mined the effectiveness of policy interventions in Nigeria, this section also discusses how the measures could be implemented to ensure accountability and transparency. I. Recent Economic Developments in Nigeria Since 2001, Nigeria has had the longest period of sustained expansion of the non-oil economy since independence. Growth has occurred across all sectors of the economy and has been accelerating. While non-oil growth averaged about 3–4 percent in 1995-2000, it more than doubled to over 7 percent and rose to 8–9 percent in recent years. Even in spite of the cur- rent global fi nancial crisis, growth of the non-oil economy remained above 8 percent in 2009 and 2010. While the oil economy contracted in recent years owing to unrest in the Niger Delta, since 2009, the contribution of the Niger Delta has improved as a result of positive effects of the amnesty on oil production (table IV.1). Moreover, over the last fi ve years, the growth of Nigeria’s non-oil econ- omy has been superior to that of most oil-exporting and non-oil exporting countries in Sub-Saharan Africa (table IV.2). An analysis of the sources of growth shows that, while total factor pro- ductivity (TFP) seems to have improved signifi cantly since 2000, relative to the US it has been declining and has only recently improved (fi gures IV.1 and IV.2). Table IV.1: Macroeconomic Aggregates, 2003–2009 (percent) Aggregate 2003 2004 2005 2006 2007 2008 2009 Real GDP 10.2 6.5 6.0 6.4 6.00 7.0 10.5 Oil GDP 3.3 0.5 –4.4 –4.5 –6.2 0.5 23.8 Non-Oil GDP 5.8 13.2 8.6 9.4 9.5 9.0 8.3 Infl ation Rate 8.0 14.0 15.0 17.9 (CPI annual average) 5.4 11.6 12.5 Source: World Development Indicators and various IMF reports.

233 220 | New Structural Economics Table IV.2: Real Non-Oil GDP Growth, 2003–2009 (percent per year) Country 2005 2006 2007 2008 2009 2003 2004 8.6 9.5 9.0 8.3 13.2 9.4 Nigeria 5.8 Oil producers 14.1 27.5 20.1 Angola 10.3 8.1 9 14.7 4.9 3.2 2.9 4.1 3.2 3.0 Cameroon 4.9 Gabon 0.8 4.3 4.9 6.2 3.0 2.3 2.3 Chad 6.0 11 4.7 3.1 3.2 –0.5 –0.5 5.4 6.6 5.4 5.9 Congo, Rep. 5.4 3.9 5.0 Equatorial Guinea 15.4 25.8 29.8 47.2 18.1 27.6 3.7 Non-oil producers Ghana 5.2 5.6 5.9 6.4 6.3 7.3 3.5 Kenya 2.9 5.1 6.9 2.1 3.8 5.7 6.1 6.7 6.7 7.1 7.4 6.0 Tanzania 5.7 7.4 3.1 4.8 5.1 5.0 4.8 3.7 –1.8 South Africa WDI/Various IMF reports. Source: Figure IV.1: Evolution of Total Factor Productivity Base Year 1960 = 1 5 4.5 4 3.5 3 2.5 percent 2 1.5 1 0.5 0 1984 1994 1996 1998 1990 1982 1988 1986 1992 1980 2002 2006 2008 2004 2000 China India Indonesia Nigeria United States Source: Bosworth & Collins, 2003.

234 Applying the Growth Identifi cation and Facilitation Framework | 221 Figure IV.2: Total Factor Productivity Relative to the United States 0.45 0.4 0.35 0.3 0.25 0.2 percent 0.15 0.1 0.05 0 2010 1980 1992 1994 1996 1986 1998 1984 1982 1990 1988 2008 2002 2004 2006 2000 India China Nigeria Indonesia Source: UNIDO, World Productivity Database. The growth of the non-oil economy was largely driven by the agriculture sector, which contributed on average more than 50 percent (table IV.3). The contribution of agriculture was followed by that of the wholesale and retail sector (about 20 percent), the manufacturing and fi nancial sectors (4–5 percent), and the telecommunications sector (about 3–4 percent). Since 2001, changes in the services sector have led to a structural change in Nigeria’s economy, manifested in substantial growth of the telecommunications, transportation, hotel and restaurants, construction and real estate, and fi nancial sectors. The fastest-growing sector has been telecommunications (at an aver- age rate of over 30 percent), followed by the wholesale and retail sectors (about 15 percent) and construction (about 13 percent). Solid minerals grew by over 10 percent on average and manufacturing by about 8–9 per- cent. Agriculture grew by 6–7 percent on average, the strongest sustained growth performance in more than a decade. Notwithstanding Nigeria’s strong economic performance over the past 10 years, its export and production structure has shown little

235 222 | New Structural Economics Table IV.3: Contribution to Non-Oil GDP (percent) 2007 2008 Sector 2004 2009 2005 2006 52.3 49.9 Agriculture 55.3 54.5 53.5 51.1 0.4 0.4 0.4 0.4 0.4 Solid Mineral 0.4 5.0 5.0 5.0 5.0 5.0 Manufacturing 5.0 4.4 3.5 2.9 2.4 1.9 Telecommunication 1.6 5.2 Finance & Insurance 4.4 4.6 4.8 5.0 5.5 17.4 Wholesale and Retail Trade 21.7 21.1 20.2 19.2 18.2 2.3 Building and Construction 2.01 2.1 2.1 2.2 2.0 12.1 Others 13.1 13.0 12.7 12.5 12.2 WDI/IMF. Source: diversifi cation. Nigeria’s exports are concentrated in oil and gas (98 per- cent), while the structure of the non-oil economy is dominated by the agriculture, wholesale, and retail sectors that serve the domestic market. How Have Employment and Incomes Responded to This Strong Growth Performance? Table IV.4 shows how the labor force has evolved since 1999. A key feature of Nigeria’s working-age population is the high share (approximately one-fourth) of the population that is not in the labor force. As in other African countries, formal unemployment (measured as nd a job) is extremely low. The vast majority of job seekers who cannot fi people outside the labor force are either discouraged job seekers or have not embarked upon a job search, as they do not consider the prospects to be promising. The share of people outside the labor force is a more suitable indicator of unemployment than the offi cial unemployment rate, which consists of individuals who are looking for, but are unable to fi nd, employ- ment. Nonetheless, actual unemployment will be less than 25 percent given that there are individuals not in the labor force who are genuinely not interested in work. However, given the pervasive poverty in Nigeria, that gure is not likely to be high. fi Table IV.4 shows that despite the high growth performance, the share of the population that is not in the labor force has remained broadly unchanged. That means that the number of jobs has risen broadly in line with the labor force and that unemployment has remained basically unchanged.

236 Applying the Growth Identifi cation and Facilitation Framework | 223 Table IV.4: Labor Force Status (percent, weighted) Labor force status 1999 2004 2006 Not in the labor force 25.3 23.0 25.2 In the labor force 74.7 74.8 77.0 Unemployment status Employed 97.8 97.0 97.4 2.6 Unemployed 2.2 3.0 Francis Teal / Luke Haywood NLSS 2003-2004 and General Household Survey Source: (GHS) 1999–2006. Sample includes population aged 15–65 not in schooling. Table IV.5 shows the evolution of employment broken down by fam- ily agriculture, non-agriculture self-employed (that is, mostly urban) and wage employment. From 1999 to 2006, the most important structural changes that have occurred in Nigeria’s labor force have been a shift into agriculture employ- ment and out of wage employment: the proportion of the sample popula- tion aged 15 to 65 (excluding those in full-time education) with wage jobs declined over this period (from 15 percent in 1999 to 10 percent in 2006). ed as non-agriculture self-employed The same is also true for those classifi (their share of the population fell from 24.1 to 22.9 percent). The cat- egory that saw a major increase in this share of the population was family 3 agriculture, which rose from 30.8 to 37.8 percent. Table IV.6 provides further insights into the development of wage employment since 1999: Wage employment in parastatals, ministries and public companies has declined, while employment in the private sector and others (including NGOs, international organizations and associa- tions) has risen. The decline of wage employment refl ects three developments: (i) the retrenchment of civil servants and the privatization of many parastatals led to a sharp decline in public service employment, which has long dom- inated employment in the formal sector and continues to represent the largest share of wage employment; (ii) many private industries with large wage employment, notably the textile industry, have been in decline for a number of years and have shed a considerable part of their work force; and (iii) sectors of the economy that have grown quickly, such as whole- sale/retail, construction and agriculture, have been, to a signifi cant extent, in the informal sector, while those in the formal sector, for example, the

237 224 | New Structural Economics Table IV.5: Types of Employment as a Percentage of the Sample Population (percent, weighted) Type of employment 1999 2004 2006 30.8 Family agriculture 36.6 37.8 22.9 Non-agriculture self-employed 24.1 25.8 0 0.1 0.1 Non-agriculture unpaid family work Wage employment 15.0 10.4 10.0 1.1 Apprenticeship 2.1 1.9 Unemployed 1.7 2.4 1.9 26.4 25.5 Not in the labor force 23.7 Francis Teal / Luke Haywood NLSS 2003-2004 and GHS 1999–2006. Source: Table IV.6: Types of Wage Employment (percent, weighted) 1999 Type of employment 2004 2006 29.6 25.2 Other 22.8 Parastatals and ministries 45.6 42.2 48.6 20.5 18.0 Private companies 17.0 11.6 Public companies 6.9 12.0 Francis Teal / Luke Haywood NLSS 2003–2004 and GHS 1999–2006. Source: fi nancial services and hospitality industries, either are not very employ- ment intensive or added labor from a very low base, failing to make a signifi cant difference in the growth of wage employment. Two features stand out: Among the young, the share of family agriculture almost doubled from • 1999 to 2006. By 2006, the share of young people outside the labor force in the urban • areas had appreciably increased. A more detailed review of the share of the people outside the labor force suggests that most of them consist of women engaged in household work and men who have never had any employment experience. This picture generally supports the conclusion that youth unemploy- ment has been on the rise since 1999, an alarming trend in view of the strong growth performance in recent years. The pattern of growth in Nigeria and its relation to the evolution of Nigeria’s labor market can be described as follows: • Nigeria’s strong growth in recent years has been dominated by the agri- culture sector. In the labor market this has been refl ected in a shift of

238 Applying the Growth Identifi cation and Facilitation Framework | 225 employment into family agriculture. The considerable growth of employ- ment in the agriculture sector is consistent with the absence of improve- ments in agricultural productivity. Creation of contractual wage jobs in the rapidly growing sectors of the • economy was unable to compensate for the loss of wage jobs in the public sector, parastatals and ministries, leading to a decline in wage employment. With the share of population outside the labor force unchanged for the population as a whole and rising for the lowest age bracket, Nigeria’s growth performance has clearly not responded to the aspirations of its population. Nigeria’s strong growth performance refl ected primarily two factors: (i) sound macroeconomic policies that created a more favorable environ- ment for private investment, and (ii) sectoral policies, such as the banking consolidation exercise that directly boosted growth in specifi c sectors of the economy. Both macroeconomic and structural policies contributed to confi dence in a new era in the Nigerian economy and thus promoted invest- ment, substantially fueled by foreign direct investment and remittances. However, this investment was more focused on capital-intensive than employment-intensive industries. Investment occurred primarily in the oil and gas and the telecommunications industries, where returns were particularly high. Hence, few productivity improvements occurred in sec- tors that are employment-intensive and consistent with the comparative advantage of the economy, such as the labor-intensive manufacturing sec- tor. As a result, the infrastructure constraints became more binding in these sectors of the economy, limiting improvements in their productivity and competitiveness and hence their ability to generate employment. A forward-looking growth strategy needs to focus on improving productiv- ity in the employment-intensive sectors of the economy. The next section identifi es the sectors that Nigeria should target based on the methodology proposed by the Growth Identifi cation and Facilita- tion Framework. II. The Growth Identifi cation and Facilitation Framework New Structural Economics notes that modern economic growth is a pro- cess of continuous technological innovation, industrial upgrading and

239 226 | New Structural Economics diversifi cation, and improvements in the various types of infrastructure and institutional arrangements that constitute the context of business development and wealth creation. At any given point in time, the struc- ture of a country’s endowment, that is, the relative abundance of factors that the country possesses, determines relative factor prices and thus the optimal industrial structure. A low-income country with abundant labor or natural resources and scarce capital will have comparative advantage and be competitive in labor-intensive or resource-intensive industries. Hence, the optimal industrial structure in a country which will make the country most competitive is endogenously determined by its endowment structure. For a developing country to reach the income level of advanced countries, it needs to upgrade its industrial structure to the same relative capital-intensity of the advanced countries. A country’s endowment structure is not static, but will depend on the rate of capital accumulation and technological progress. The change in relative prices associated with these changes will affect the type of industries in which the country has a latent comparative advantage and hence the optimal industrial structure, given that, in order to be competitive, the new industry needs to be consistent with a country’s 4 latent comparative advantage. Of particular importance to the latent comparative advantage is the wage level. By imitating or licensing to obtain technology—a process that is less expensive than inventing the technology on their own—low-income countries will be able to pro- duce the same commodities at a signifi cantly lower cost than developed countries provided the enabling conditions have been created. That way the country can exploit the latecomer advantages by developing matured industries in dynamically growing, more advanced countries with endowment structures similar to theirs. By following carefully selected lead countries, latecomers can emulate the leader-follower, fl ying-geese pattern that has served well all successful economies since the 18th century. The process of upgrading the industrial structure to a higher level consistent with the factor endowment cannot rely solely on the market mechanism. For example, starting a new industry may be diffi cult because of the lack of complementary inputs or adequate infrastructure for the new industry even if the targeted industry is consistent with the economy’s

240 Applying the Growth Identifi cation and Facilitation Framework | 227 comparative advantage. Private fi rms will not be able to internalize those investments in their decision to upgrade or diversify. Therefore, the gov- ernment has an important role in providing or coordinating investments in necessary infrastructure and complementary inputs. In addition, innovation which underlies the industrial upgrading and cation process is a risky process, as it presents a fi rst-mover prob- diversifi lem. Both failure and success of a fi rst mover create externalities. For example, fi rms that are fi rst movers pay the cost of failure and produce valuable information for other fi rst mov- rms. At the same time, when fi ers succeed, their experience also provides valuable information to other table in market participants about the type of industries that can be profi the specifi c country. However, if new fi rms enter on a large scale this may largely eliminate the possible rents that the fi rst mover may enjoy. In a developed country, a successful fi rst mover can in general be rewarded with a patent and enjoy the rent created by a matured industry. However, in a developing country, a new patent may not be available, as the industry may already be located within the global industrial frontier. Therefore, the fi rst mover will not be able to obtain a patent for its entry into a new industry in its economy, and, as a result, some form of direct support by the govern- rms may be justifi able. ment to pioneer fi The GIFF proposes a new approach to help identify industries where the economies may have a latent comparative advantage and remove bind- ing constraints to facilitate private fi rms’ entry into those industries, or facilitate industries that are already active in the country to grow fast. In this context, the GIFF argues that picking winners is inevitable because the binding constraints may be sector specifi c and removing them may not be possible for the private sector alone. Therefore the main issue is to minimize the error margin of picking the wrong industry. The key risk in this regard is that countries target industries that are too advanced and far beyond the latent comparative advantage or target industries in which the country has already lost its comparative advantage. The GIFF proposes a six-step approach to growth identifi cation and facilitation. Three of these steps aim at the selection of sectors. After the sectors are selected, value-chain analyses can be used to identify the binding constraints for private fi rms’ entry and growth in those sectors (box IV.1).

241 228 | New Structural Economics Box IV.1: Applying the GIFF: Comparative Value Chain Analysis A forthcoming report by the World Bank on Light Manufacturing in Africa (World Bank 2011) demonstrates how to implement an innovative form of value chain analysis to both determine the competitiveness of a sector as well as assist governments and the private sector in identifying the constraints which most impact the cost competitiveness of domestically produced products on the global market. In the usual value-chain analysis, the advantages, bottlenecks and policy issues would be analyzed within the country of study with some comparison between sectors within the economy. In the comparative approach, however, China and Vietnam are being chosen as benchmark countries in order to compare the cost competitiveness of African production of particular products, chosen to be as like- for-like as possible. After applying the GIFF to arrive at several sub-sectors which could potentially be successfully produced in SSA (Ethiopia, Tanzania and Zambia were the sample countries), in-depth value chain analyses were conducted for particular products in each of those sub-sectors in order to gain a representative view of the com- petitiveness and constraints of the sub-sector. The analysis included a quantitative breakdown of the proportion and cost of inputs, effi ciency input use, logistic costs, labor productivity, production wastage and effi ciency etc. This data was gathered from a reasonable sample of fi rms in all fi ve countries producing similar products in each of the fi ve identifi ed sub-sectors. Each component which impacts the cost and competitiveness of the fi rms was compared between China, Vietnam and SSA. The results were conclusive in identifying the cost elements which vary signifi cantly between East Asia and SSA, thereby identifying the priority areas for intervention. The results also screened out those sectors where the country does not have com- parative advantage by calculating the domestic resource costs. For Nigeria, value chain analyses have been conducted in recent years for sev- eral key sectors which have been valuable in highlighting the constraints and oppor- tunities in the sectors studied. However, the new approach proposed in this paper is to use the GIFF to identify sectors where Nigeria may have some comparative (Box continued next page)

242 Applying the Growth Identifi cation and Facilitation Framework | 229 Box IV.1: Continued advantage, latent or revealed. A comparative value chain analysis could then be undertaken in those identifi ed sectors which will provide rigorous evidence and support for a prioritized program by government and private sector to overcome key constraints in targeted sectors. For example, the comparative value chain provides some more conclusive evidence on the wage difference in specifi c sectors as well as the difference in labor productivity in those sectors. That way, conclusions can be drawn on both the poverty-reducing employment effect of expansion in a sector, as well as the labor cost advantage (or disadvantage) which can be a crucial aspect in determining competitiveness of a sector. The fi rst step consists of identifying tradable goods and services that • have been growing dynamically for about 20 years in fast-growing countries with similar endowment structures that have a per capita GDP about 100 to 300 percent higher than their own. In many cases, given that wages tend to rise in the growth process, a fast-growing country that has produced goods and services for about 20 years may 5 begin to lose its comparative advantage in this sector. In addition, Nigeria could domestically produce simple manufacturing goods which are labor-intensive, have limited economies of scale, require only small cation investments, and are imported. This step also allows the identifi of industries that are new to the country, but may be good business opportunities for Nigeria. Second, among the industries on the list, the government may give prior- • rms have already entered ity to those in which some domestic private fi spontaneously, and try to identify: (i) the obstacles that are preventing these fi rms from upgrading the quality of their products; or (ii) the bar- riers that limit entry to those industries by other private fi rms. For such industries, the government could also adopt specifi c measures to encour- age foreign direct investment in the higher-income country to invest in these industries. • Third, in addition to the industries identifi ed on the list of opportunities for tradable goods and services in step 1, developing country governments

243 230 | New Structural Economics should pay close attention to successful self-discoveries by private enter- prises and provide support to scale up those industries. The application of this methodology to Nigeria is discussed below. III. Selecting Sectors Selecting a Country with a Per Capita Income 100–300 percent above Nigeria’s Table IV.7 shows a list of countries that have a per-capita GDP of 100 to 300 percent of that of Nigeria. Removing slowly growing countries, i.e., countries growing at less than 6 percent per year, leaves the following countries: Indonesia, China, Vietnam and India. Using the criterion of factor endowment, among these countries Indonesia would stand out as the country with the greatest similarity with Nigeria because it is a natural resource-rich country and a former 6 member of OPEC, but also specializes in labor-intensive production. Indonesia has effectively used both its natural resources as well its abun- dant labor supply to develop industries that correspond to its latent comparative advantage. As discussed in a blog by Justin Lin (March 2011), a resource-rich, labor-abundant country can use both resource- 7 rich and labor-abundant countries as comparators. While not a resource-rich country, Vietnam’s high growth rate makes it an appropriate comparator, especially in view of its labor-intensive Table IV.7: GDP Per Capita PPP in 2009 (constant 2005 international $) Country GDP per capita Percent of Nigeria 100 Nigeria 2,001 Vietnam 2,682 134 148 India 2,970 Philippines 3,216 161 Indonesia 3,813 191 Morocco 4,081 204 Paraguay 4,107 205 Egypt, Arab Rep. 5,151 257 China 6,200 310 Tunisia 7,512 375 Source: World Development Indicators.

244 Applying the Growth Identifi cation and Facilitation Framework | 231 economy. Its strong growth performance and consequent rise in labor costs could also quickly erode Vietnam’s cost advantage in certain labor- intensive industries. Another country that lends itself as a comparator is China. China has a per capita income that is about 300 percent higher than that of Nigeria and is not a natural-resource-rich country. However, given its fast growth, large population size and domestic market, as well as fast ascent on the tech- nological value-added ladder, its production structure may be suited for imitation given that it may be in the process of losing its cost advantage in some of the industries that have in the past driven its growth performance. This is especially true if Nigeria can use the rent from natural resources to improve its infrastructure and education. A further comparator country is India. While India has not consistently followed its comparative advantage of abundant unskilled labor, the avail- ability of skilled labor has been successfully used for several new sectors, such as call centers. Hence in some areas, India’s production structure has been in line with its latent comparative advantage. Which Commodities Do These Countries Export? es industries in these comparator countries where pro- Table IV.8 identifi duction is labor intensive or requires natural resources, and provides brief comments on Nigeria’s potential in these industries. Imports of Labor-Intensive Manufactured Goods Have Limited Economies of Scale, and Require Only Small Investments A review of imports of labor-intensive manufactured goods with limited economies of scale shows the following commodities (at the 4-digit SITC level) (table IV.9). Industries Where the Private Sector Is Already Active and Where Successful Self-Discovery Has Taken Place A third criterion for selection is to choose sectors in which Nigeria’s private sector has become increasingly active and where successful self-discovery has already taken place, such as ICT, light manufacturing, food process- ing, wholesale and retail, construction and car parts, meat and poultry, oil palm, and cocoa. None of these industries currently produces for export.

245 Nigeria potential Tire industry closed several years ago as it could not compete Leather – already private sector momentum, goat/kidskin leather Large domestic production. High potential established in detailed such as the removal of the petroleum subsidy. Also, production refi neries and fertilizer plants, but requires enabling conditions, Indigenous fertilizer plants exist and are growing fast; Nigeria has of petrochemicals needs to match with specifi c type of refi ning IT –Knock-down of computers is successfully taking place. provided land is being made available. in Lagos on CKD basis. Large potential for scaling up exists, Since December 2010, two operators have begun TV assembly capacity. needs better enabling conditions. is the 4th largest export. Industry already in place in Kano that produce at large volumes. well as a small wage differential to comparator countries which with imports is undermined by high costs of power in Nigeria, as Textiles is a failing industry primarily because competitiveness value chain analysis. However, low export value of USD 300,000 in 2009. Large rubber plant exists in Calabar, Cross River. with imports. Natural rubber is Nigeria’s 10th largest export. yarn, fabrics etc. electronically integrated and optical fi ber circuits, insulated wire accessories; textile equipment Printed circuits, sound recording Apparel & clothing Telecommunications & Palm oil equipment yarn, fabrics etc. Apparel & clothing accessories; textile Telecommunications & sound recording Crude rubber Electronic integrated circuits, telecoms. Footwear accessories; textile Apparel & clothing yarn, fabrics etc. fertilizers projectors Offi ce machines & photographic Manufactured automatic data sound recording equipment; manufactures handbags; leather Footwear; travel goods, Telecommunications & Apparel & clothing accessories; textile yarn, fabrics etc.; dyeing China Vietnam India Indonesia & tanning Table IV.8: Identifying Sectors for Growth: Key Exports of China, India, Vietnam, and Indonesia Rubber manufactures 232

246 Already active and growing. Logistical support could help Metal industry in place; but too small and scattered to be cost Organic chemicals industry could benefi t from abundant supply effective. Scaling up could be facilitated through creation of clusters. of raw materials; however, petroleum subsidy is major distortion blocking larger foreign direct investment. assembly already. Industry established; but fragmented. Mergers could help Furniture industry active in Nigeria and rapidly growing. reduce cost. motorcycles and tractors are assembled in a knock-down Onitsha cluster in Anambra state focuses on car parts; accelerate growth. especially power and a cold chain, are required. Both already active in Nigeria; to scale up, enabling conditions, 5th largest export. Food & beverages: booming sector oriented to the domestic market; Cocoa beans are 3rd largest export; frozen crustaceans Paper, paperboard etc. electrical machinery Fixed vegetable oils and spices manufactured fats; coffee, tea, cocoa, Fish, crustaceans prepared Road vehicles; other Organic chemicals power-generating Machinery - electrical, general industrial, vegetables and fruit medicinal and pharmaceutical products Cereals & cereal preparations; transport equipment Furniture and parts thereof Cereals & cereal Machinery - electrical, industrial manufactured Fish, crustaceans tea, cocoa, spices prepared preparations; coffee, Authors’ calculations using declining export shares based on Comtrade data. resins, plastic materials; Organic chemicals; chemical materials & products; artifi cial inorganic chemicals Medicinal and power-generating Machinery - electrical, metalworking or pharmaceutical products manufactures thereof; cork and wood Furniture and parts Road vehicles Paper, paperboard etc. Vegetables & fruit prepared Source: Fish, crustaceans 233

247 234 | New Structural Economics Table IV.9: Nigeria’s Top Imports, 2010 Product (4-digit) 1,000 of US$ percent 863,917 10.7 Cereals and cereal preparations Telecommunications & sound recording 330,136 4.1 Fish, crustaceans, mollusk, preparation 3.4 276,152 3.2 Other transport equipment* 255,846 Medicinal and pharmaceutical products 241,312 3.0 2.7 Manufactures of metal 214,157 Artifi cial resins, plastic materials, cellulose 1.9 151,868 1.3 Essential oils & perfume material 104,932 Professional, scientifi c & controlling equipment 1.3 101,065 1.2 99,125 Dairy products and birds’ eggs 1.2 98,169 Miscellaneous manufactured articles Rubber manufactures, n.e.s. 1.2 96,489 Paper, paperboard, articles of paper 91,269 1.1 Textile yarn, fabrics, made-up articles 1.0 82,120 Beverages 58,480 0.7 Source: COMTRADE database, 4-digit SITC Revision 2. * Since transport equipment may be protected by patents, Nigeria could start by producing generic products. However, all of them have signifi cant employment and growth potential and could be upgraded for exports. Figure IV.3 highlights how the growth and employment potential of a sector may differ on a regional or geographical basis. For example, rice production has a lower employment and growth potential in Kano than in Kaduna. And the wholesale and retail sector has greater growth potential in Lagos, given the very large domestic market, than in Kano, where the market is smaller. Such detailed regional analysis is very important, given the great disparity in Nigeria. In addition to these sectors, there are a number in which successful self- discovery has already taken place. For example, production of suitcases has recently successfully started and is expanding rapidly. At this stage, 60 percent of the required parts are being produced domestically, which has allowed for the unit cost to fall signifi cantly; also about 50 percent of the domestic demand is being met through domestically produced suit- cases. A further area of successful self-discovery is TV assembly, which began as recently as December 2010. Both areas of production could be further expanded rapidly, including for exports, if the government pro- nance. vided assistance toward scaling up, e.g., through better access to fi

248 Key Kano Lagos Kaduna Cross River to State Economy Current Importance State Colour Coding High Leather Tourism Calabar Port ICT Aquaculture Food Processing Oil Palm Construction Solid Minerals Light Manufacturing Cocoa Dairy Rice Rice Feasibility • Capacity of private sector actors • Likelihood of policy reform • Ability to bridge competitiveness gap Meat & Poultry Meat & Poultry Wholesale/Retail Food Processing Wholesale/Retail Light Manufacturing Low High Upside Potential • Employment • Growth • Spillovers Treichel 2010. Source: Figure IV.3: Prioritization of Value Chains for Further Investigation 235

249 236 | New Structural Economics What Are the Sectors in Which Nigeria Has a Potential Comparative Advantage Based on This Analysis? The sections above use three different criteria to identify sectors with high growth and employment potential that could be the subject of targeted interventions. First is the identifi cation of dynamically growing tradable industries in fast-growing countries with similar factor endowment and a per capita income 100 to 300 percent above that of Nigeria. Second is the review of Nigeria’s imports to identify sectors that require only small investments and have limited economies of scale and could therefore be manufactured domestically. Third is the identifi cation of domestic sectors where successful self-discovery has already taken place or that are already growing fast, but have a high employment impact and could grow faster. Upon application of the fi rst criterion, seven sectors emerge quite clearly for further analysis, as they represent industries in countries with a similar endowment structure: footwear, including sports shoes; tex- tiles; TV recorders; aquaculture; motor vehicle parts; vegetable oil; and fertilizers. Additional sectors are motorcycles; meat, meat products, and oil seeds; fertilizers, petroleum products; leather; travel goods; offi ce machines; pharmaceutical products; and organic chemicals. Based on the second criterion, the following four sectors would be pri- oritized: vehicles parts; color TV receivers; tires; and metal manufacture. The third criterion, which focuses on sectors that are already growing fast, yields a list of target sectors that is slightly different from that identi- fi ed through the fi rst two criteria: light manufacturing, food processing, meat and poultry, palm oil and rice, telecommunications, leather, whole- sale and retail, and construction. Nigeria is a country rich in natural resources, in particular oil and gas, but also solid minerals. Industries associated with these natural resources, in particular refi ned petroleum products, petrochemicals, cosmetics and plastics, are currently not particularly active in Nigeria. However, given that they are imported in large quantities, and raw materials are available in abundance, they should be subject to a detailed value chain analysis aimed at assessing whether they can be produced at a comparative advan- tage in Nigeria. How should the targeted sectors be selected from this list? The key criteria will be the upside potential of the sector in terms of growth and

250 Applying the Growth Identifi cation and Facilitation Framework | 237 employment creation, as well as the feasibility of growth in terms of private sector capability and the public sector regulatory framework. These questions can ultimately be answered only through detailed value chain analysis, along the lines of the methodology described in Box 1. However, as a fi rst approximation, the list of potential target industries can be narrowed down further by applying a set of pre-screening crite- ria developed in the context of the forthcoming report on light manu- facturing in Africa. First, sectors with very high capital requirements and only small domes- tic markets should be eliminated, given that Nigeria is not a capital abun- dant country and initially success will be in catering to the large domestic market. Second, goods should ideally be produced by small and medium- size enterprises in the comparator countries, given that large enterprises are currently not prevalent in Nigeria. However, to the extent that Nigeria’s business environment could be made conducive to attracting large-scale foreign direct investment, the goods could also be produced by large com- panies in comparator countries. Third, a supply chain should exist for each product in the domestic market. Fourth, raw materials should be available in the domestic market or be easily imported. And fi fth, labor skills should be easily transferable. Table IV.10 shows whether the pre-selected sectors meet the criteria: Wholesale/retail and construction sectors have not been included, as they are not sectors that would be imitated from other countries, but may still benefi t from targeted interventions to make them more responsive to higher demand and more employment-intensive. Most of these sectors meet the pre-screening criteria, that is, they have some upside potential for growth and meet feasibility criteria. A notable exception is sportswear—unavailability of PVC in the domestic market cre- ates a comparative disadvantage and has already resulted in the closure of domestic production. Its competitiveness may depend on the establishment of a domestic petrochemical industry. Also, in the comparator countries, fertilizer and petrochemical production as well as TV production may not take place in small or medium-size enterprises; however, targeted foreign direct investment may be able to attract investments to establish larger fi rms, including through joint ventures, provided enabling conditions have been put in place.

251 238 | New Structural Economics Table IV.10: Criteria for Screening Potential Subsectors Criteria 3: There is some factor endowment in Nigeria — supply chain Criteria 1: Production Criteria 2: Production exists in the domestic has low capital in comparator market (domestic or countries is by requirements and there imported raw materials); small and medium-size is a signifi cant domestic labor skills should be Product groups market easily transferable. enterprises Leather supply chain Yes Footwear, including Yes exists; however, PVC sports shoes, Travel required for sports bags shoes does not exist. Leather shoes and travel bags are already being manufactured and see strong growth. Yes In some cases Raw materials can be TV electronics easily imported. Not a high-skills type of production. Tires and motor vehicle Yes Yes Rubber and associated parts supply chain exists. International companies were active in Nigeria. Vegetable oil, Yes Yes Yes aquaculture, palm oil and rice; food processing, meat and poultry Yes Yes Yes Motorcycles and tractors Fertilizers, Yes No Nigeria is abundant in oil and gas. Labor skills petrochemicals, organic chemicals are transferable. Light manufacturing Yes Yes Yes. Vibrant domestic industry with relevant skills already in place. Leather Yes Yes Yes Pharmaceuticals Yes Yes Yes Paper board Yes Yes Yes Source: World Bank 2011. In a second step, Nigeria’s basic wage competitiveness in these sectors needs to be reviewed to determine whether that would allow Nigeria to reap the advantage of backwardness. Table IV.11 summarizes wage data for China, Vietnam, and Nigeria on a sectoral basis.

252 China 192–265 206–251 237–296 192–236 78–207 78–130 85–135 Vietnam 117–142 89 26–48 26–52 37–52 Ethiopia Unskilled labor 45 48 35 67 20 alues are mean over sample); Others – Light Manufacturing in Ethiopia (ICA) 82 87 54 67 87 66 79 71 125 127 Nigeria China 265–369 398–442 331–370 393–442 Vietnam 168–233 119–181 181–363 181–259 181 Ethiopia 37–185 89–141 81–119 Skilled labor 82 82 71 154 151 Ethiopia (ICA) 130 107 163 106 102 Nigeria Nigeria – Productivity and Investment Climate Survey, 2009; Ethiopia Investment Climate Survey, Manufacturing 2006 (reported v Ethiopia has been included to provide another African country as a reference for Nigeria. Machinery & equipment Table IV.11: Average Wage, Including Benefi ts, by Industry (US$) Chemicals 212 Textiles 120 Electronics 119 Other manufacturing Sector Non-metallic minerals Food 135 Garments 85 Source: Metal & metal products Wood, wood products Africa (2001), vol. II (values reported are the range reported by sample fi rms). Note: & furn 239

253 240 | New Structural Economics These data confi rm Nigeria’s relative cost advantage in cheap labor in ed above. the industries identifi IV. How Can Growth in the Selected Value Chains Be Promoted? In addition to proposing the above methodology to identify target sectors, the GIFF also identifi es a number of steps to encourage growth in these targeted sectors. As discussed above, the government can try to identify the obstacles that are preventing these fi rms from upgrading the quality of their products or the barriers that limit entry to those industries through value chain analysis or the Growth Diagnostic studies suggested by Hausmann, Rodrik, and Velasco (2005). In addition, the government can adopt spe- cifi rms in the higher-income countries identifi ed c measures to encourage fi in the fi rst step to invest in these industries. Moreover, in developing coun- tries with poor infrastructure and an unfriendly business environment, the government can invest in industrial parks or export processing zones. Such industrial parks or EPZs typically provide conditions that are specifi cally targeted at certain sectors or industries, e.g., IT or light manufacturing, and are often built around already existing industry clusters. Lastly, the government can also provide limited incentives to domestic pioneer fi rms ed in step or foreign investors that work within the list of industries identifi 1 in order to compensate for the non-rival public knowledge created by their investments. These steps may include corporate income tax holidays, directed tax credits, or priority access to foreign reserves to import key equipment. In the literature, the former type of intervention is referred to as soft and the latter as hard industrial policy. The following discusses key constraints in the selected value chains and what specifi cally could be done about it in the Nigerian context. Avail- able value chain studies provide an analysis of the binding constraints to 8 growth in a number of these value chains. The binding constraints can be broadly categorized in 5 categories: (i) physical infrastructure, in par- ticular lack of power and roads; (ii) business environment (cumbersome procedures); (iii) lack of access to fi nance; (iv) lack of a technical and voca- tional education system that corresponds to the needs of the market; and (v) restrictive trade policy. The annex table summarizes binding constraints and the measures that could be undertaken to address them in a number of selected value chains.

254 Applying the Growth Identifi cation and Facilitation Framework | 241 Specifi c measures to be undertaken for each category of constraints to growth could be as follows: Physical Infrastructure. Construction of industrial parks with dedicated power supply and transportation. Construction of independent power plants (IPPs) in geographical areas with high growth potential that already have a high concentration of promising value chains, possibly through the Bank of Industry in close collaboration with state governments. Business Environment. Selective capacity-building in key government agencies, such as the Standards Organization of Nigeria that enforce qual- ity, and reform of business licensing as well as land transactions. TVET (Technical and Vocational Education and Training). Linking growth sectors to skills development through promotion of Innovation Enterprise Institutions (IEI)—an initiative promoted by the Nigeria Board of Technical Education. Developing a National Vocational Qualifi cation framework and calibrating the National Youth Service to assign youth corp participants to jobs matching their qualifi cations. Encouraging the development and adoption of training standards in traditional apprentice- ships through trade associations. Access to Finance. Mobilizing mortgage-based fi nance by enhancing the availability of mortgages through reform of the land allocation system. Introducing directed credit schemes at concessional interest rates. Trade Policy Reform. Import bans and high tariffs adversely affect com- petitiveness of a number of value chains. Replacing import bans that adversely affect certain sectors with the highest growth potential with tariffs could be very benefi cial to the development of industries. Some protection may still be necessary for sectors with high growth potential that still need to develop. Key Constraints for Each Sector One of the most important challenges for government and the private sector is to identify the most important constraint, which, if alleviated, is likely to allow for the sector to grow faster. The annex table highlights various constraints for each value chain and how they could be addressed

255 242 | New Structural Economics in each individual case. The section below discusses some of the fi ndings from meetings with entrepreneurs in Lagos. Lack of electric power is a pervasive constraint in almost every industry and is therefore not specifi - cally mentioned. It is imperative for the Nigerian government to promote construction of IPPs in industrial zones as the main tool for addressing the high cost of power. On a general note, entrepreneurs call for greater protection from imports through tariffs. In the past, Nigeria has consistently protected domes- tic industry through high tariffs or import bans. However, the desired improvement in domestic output has not materialized, as key constraints to greater productivity have remained unaddressed, in particular the lack of power. Against this background, it would be preferable if continued pro- tection was associated with a pre-commitment on the part of government to gradually phase out protection and address some of the key binding constraints in a comprehensive package of measures, e.g., the construction of independent power plants, the establishment of a fast-track window for imports of manufacturers and specifi nancial interventions to facilitate c fi access to fi nance of key value chains. Import bans should be replaced by tariffs, given the fact that most import bans cannot be enforced and only encourage smuggling. Food processing (including fruit juices, meat and poultry, noodles and spaghetti and tomato paste) has experienced strong growth in recent years and producers are confi dent about prospects for further growth. Tomato paste producers indicate that their growth poten- tial would sharply improve if domestic production of tomatoes could c government incentives such as for be scaled up. In addition, specifi research and development, the full operationalization of the Export Expansion Grant (EEG) and assistance in distributing seeds could allow production to further expand. Construction has a very signifi cant potential for job creation. The pri- mary constraint for faster growth is the unavailability of mortgage fi nanc- ing. Specifi c interventions to improve the availability of such fi nancing through reform of the land transaction process and the development of mortgage-related fi nancial instruments would be critical to facilitate faster growth. In addition, the industry suffers from shortages of skilled techni- cal labor. Targeted interventions to substantially upgrade the quality of

256 Applying the Growth Identifi cation and Facilitation Framework | 243 vocational training would help youth unemployment and reduce the cost rms. of construction fi are set for rapid expansion. Key Motorcycle, tractor and TV assembly constraints are the lack of adequate trade facilitation leading to delays in clearance of imports and the need for land to allow expansion of produc- tion and reap benefi ts from economies of scale. is also growing rapidly. Partnership between the Computer assembly public and private sectors to help reduce the skills gap would be crucial to reduce cost. In addition, the government may facilitate the adoption of broadband internet access in universities and schools. Following years of decline owing to lack of competitiveness with imports, the tire industry ceased production in 2008. Key constraints to greater productivity include (i) the need for natural gas to power an IPP; (ii) the need to rehabilitate the Warri refi nery to facilitate availability of black carbon, a key input for tire production; and (iii) the need for a bail- out fund to address the large amount of unserved debt. Gas had not been available owing to the turmoil in the Niger Delta, while the rehabilitation nery had not proceeded on a timely basis. An injection of fresh of the refi capital from the government could be crucial to bailing out the industry, especially if packaged with other measures, in particular the rehabilita- nery and concessional loans (based on a performance tion of the Warri refi agreement with the private contractor). The metal industry has been suffering from power shortages and lower price competition from abroad. Nonetheless, some segments of the produc- tion, like cast iron and manganese steel, have been prospering, while oth- ers, such as aluminum, have been in decline. One of the key obstacles, in addition to the power supply, has been the customs administration which has been delaying the clearance of imported raw materials. However, the most important challenge in facilitating the growth of this industry is the lack of power. V. How Should Governance Issues Be Addressed in Implementing These Measures? One of the most important criticisms against industrial policy is the potential for elite capture of the interventions in a way that could seriously undermine

257 244 | New Structural Economics the effectiveness of any policy intervention. Nigeria has a poor track record in governance, traditionally ranking near the bottom of the global Cor- ruption Perception Index. In this context, it is important to establish some principles that would allow proper management of governance-related issues in the implementation of these targeted policy measures. Based on experience in other countries, the following seem to be components that could help improve the governance aspect of these measures: • Transparency and accountability is best ensured through a public process of agreement and follow-up on the implementation of the agreed mea- sures. For example, as a fi rst step a jobs summit could be held in which private and public sector representatives for key sectors meet, discuss and agree on critical interventions aimed at boosting growth in the individual sector, including selection criteria and appropriate implementation mech- anisms. These Memoranda of Understanding could then be published and their implementation reviewed in public fora on a regular basis. • The agreements should also explicitly specify the results that are to be expected. In addition, they should say that if the results are not achieved, the intervention should be retracted. • A further measure helping to foster transparency and accountability is to limit the scale of the intervention. Smaller interventions stand a greater chance of transparency than large schemes. This is because the poten- tial for elite capture is directly proportional to the magnitude of rents from government subsidies and other forms of protection and therefore smaller interventions stand a better chance. • The selection of industries could be delegated to a consulting fi rm, rather than be handled by the government (as has been the case in Chile). VI. Conclusions This paper has aimed to identify sectors with high growth and employ- ment potential and targeted interventions to remove binding constraints to growth in each of these sectors. The paper concludes that a number of sectors, some of which are already active in Nigeria and some new to Nige- ria, may hold signifi cant potential for growth and employment creation and should be subject to detailed value chain analyses that would identify the type of interventions that would allow Nigeria to effectively compete

258 Applying the Growth Identifi cation and Facilitation Framework | 245 with its competitors. Targeted interventions to promote growth should primarily focus on (i) providing physical infrastructure, especially power, water and sewerage; (ii) improving the business environment; (iii) develop- ing targeted vocational and educational skills programs; (iv) improving the business and regulatory environment; (v) reforming trade policy; and (vi) providing some tax incentives, access to fi nance, and access to foreign exchange for the targeted sectors. Of crucial importance in implementing these targeted interventions is the adoption, in parallel, of a range of measures to support good gover- nance as highlighted above. Provided such policies are undertaken in par- allel, Nigeria should be in a good position to maintain its strong growth performance, as well as to increase the employment intensity of growth.

259 asymmetries Technology and information procedures Onerous administrative X Poor logistics & handling cost infrastructure Unreliable & high XX enforcement Poor environmental practices & on product standards Lack of regulatory enforcement Distortionary trade policy X Finance labor inputs Low quality & unproductive Intervention approach to simplify individual procedures in Initiate administrative reform program through risk-based customs clearances Reform customs procedures, including growth potential that already have a formalizing trade) tariff would maximize incentives for such as rebar steel, with tariffs (15% Replace import bans on key inputs, chains high concentration of promising value relevant agencies to speed up the Construction of independent power plants in geographical areas with high compliance/approval process other key infrastructure areas (tariff and regulatory reform) toral reforms in the power sector and Public-private partnerships and sec- Growth constraints nied by poor border controls, import bans on factor inputs; tariff and duties to protect domestic goods. Distortionary trade policy accompa- Onerous administrative procedures, planning approvals & building per- mits; burdensome business regula- tory compliance, including EEG services system under-developed freight transport port); poor logistics & handling. ture services (power, water, trans- Unreliable and high-cost infrastruc- rising demand class provides signifi cant and emerging middle Potential sectors struction: Real estate & con- upside potential Table IV.A1: Growth-Inhibiting Cross-Cutting Constraints, Interventions and Expected Outcomes Annex 246

260 (continued next page) X X X tional Training Centre of Excellence in by hands-on implementation with stakeholders Redesign or reengineer procedures streamline legal basis as necessary opment of the fi nancial sector reform of land transactions and devel- Explore possibility of using the Innova- tive Enterprise Institutions to deliver training programs Provide strong technical assistance in Develop new and strengthen existing vocational institutions to increase access to unskilled, informal labor sector including programs designed for practical teaching programs (e.g. Voca- through careful process mapping and Lagos for the construction sector) related fi nancial services through Facilitate development of mortgage- the preparation/design stage followed Deliver capacity building program to strengthen inspection activities of relevant government authorities per- taining to quality and safety standards of products in the construction industry Town Planning Authority Inadequate development of the mortgage market Product Standards, SON, Local Lack of regulator Enforcement on labor exhibits low productivity lack of vocational training, unskilled inputs: shortage of skilled labor, Low-quality and unproductive labor 247

261 asymmetries Technology and information X procedures Onerous administrative Poor logistics & handling cost infrastructure Unreliable & high X enforcement Poor environmental practices & on product standards Lack of regulatory enforcement Distortionary trade policy Finance X labor inputs Low quality & unproductive Intervention approach statutory responsibility develop the appropriate level of State Ministry of Environment to Provide technical assistance to the environmental regulations under its the institutional authority to enforce resources and expertise, as well as including through risk-based customs clearances Replace import bans on key inputs with tariffs and reform customs procedures, Industry associations to address information and knowledge transfers to SME processors Growth constraints enforcement; inadequate proce- dures for disposal of construction materials Poor environmental practices and producers Coordination failures, weak Raw material costs sources locally are high because of import bans, leading to high and uncompetitive prices of products, produced with imported raw material linkages between processors and rising (Continued) Potential Sectors and for processed foods non-alcoholic beverages demand for alcoholic and incomes have spurred Food processing: Table IV.A1: 248

262 X (continued next page) XX relevant agencies to speed up compli- ance/approval process by hands-on implementaton with Redesign or reengineer procedures through careful process mapping and streamline legal basis as necessary Provide strong technical assistance in the preparation/design stage followed Initiate administrative reform program stakeholders Deliver an integrated logistics program for the food sector from farm gate to the processing stage covering farm storage, drayage, wholesale markets, rural roads and line haul transport to the processor, complete with tempera- ture control equipment as warranted to simplify individual procedures in temperature control, improper bag- other dedicated infrastructure for industrial areas with high volumes of food processing Onerous administrative procedures, Poor logistics & handling; poor rural roads to dispersed small scale farms; poor handling practices (no land registration ging, storage etc.); under-developed freight transport services system; business environment (power, water, roads) is lacking: IPP and 249

263 asymmetries Technology and information procedures Onerous administrative Poor logistics & handling cost infrastructure Unreliable & high enforcement Poor environmental practices & on product standards Lack of regulatory enforcement Distortionary trade policy Finance labor inputs Low quality & unproductive X Intervention approach and butchers in Dubai, Malaysia, South traditional practices of the domestic Explore possibility of using the Innova- tive Enterprise Institutions to deliver training programs Africa Develop new and strengthen existing vocational institutes to increase access to unskilled, informal labor sector including programs designed for practi- cal teaching programs (e.g. Vocational Training Centre of Excellence in Lagos for the construction sector); MAN and NASSI to provide outreach for private sector interventions Sponsor technical advisory trips to locations that incorporate religious and target groups, e.g. Islamic run abattoirs Growth constraints Low quality and unproductive labor inputs: shortage of skilled labor, lack of vocational training, unskilled labor exhibits low productivity. Lack of skilled labor: support to the technical and vocational training system for food technology training; (Continued) Potential Sectors Table IV.A1: 250

264 (continued next page) X X X X X X Privatization of abattoirs relevant government authorities per- taining to quality and safety standards of products in the agriculture, animal livestock and food industries. In par- ticular, harmonize roles of food inspec- tion agencies for seamless cooperation (State Ministry of Agriculture, State Ministry of Health, and NAFDAC) the institutional authority to enforce Deliver capacity building program to strengthen inspection activities of statutory responsibility environmental regulations under its Provide technical assistance to the State Ministry of Environment to develop the appropriate level of resources and expertise, as well as establishment designs Support with technical training and Removal of import ban on meat government institutions Targeted capacity-building for key Dept of Agriculture; Veterinary Poor environmental practices and NAFDAC Lack of regulator Enforcement on tal control of animal waste at public enforcement, improper environmen- produce inspection and fi sheries; abattoirs vices; Agric Services, Pest control, Product Standards - SON; State Services; Livestock & Poultry ser- and other support services especially for veterinary services Information on technology options and prices competitiveness Import bans are undermining Abattoirs are dysfunctional Public sector institutions are weak, growing created high demand for Aquaculture: meat; fast food industry steadily from a low base refl ecting increased demand very important as source of demand Meat & Poultry: urbanization, emerging middle class, and high income elasticity have 251

265 asymmetries Technology and information procedures Onerous administrative Poor logistics & handling cost infrastructure Unreliable & high XX enforcement Poor environmental practices & on product standards Lack of regulatory enforcement Distortionary trade policy XX X Finance X labor inputs Low quality & unproductive Intervention approach Increase access to term lending for be developed through infrastructure, technology, information systems Commercial stakeholders’ capacity to Construction of Independent Power Plants in geographical areas with high growth potential that already have a high concentration of promising value chains micro and small businesses tariff would maximize incentives for Reform customs procedures, including through risk-based customs clearances formalizing trade) Replace import bans with tariffs (15% Growth constraints ment services Access to fi nance ture services (power, water, trans- port); poor logistics & handling: poor rural roads to dispersed small scale farms Unreliable and high cost infrastruc- Input supply and business develop- trade policy Distortionary trade policy accom- panied by poor border controls: import bans on factor inputs; tariff and dustiest to protect domestic goods; porous borders undermine (Continued) high-quality Potential Sectors upgrading for international products with potential Leather: marketability that needs Table IV.A1: 252

266 X (continued next page) X relevant agencies to speed up compli- ance/approval process Africa Redesign or re-engineer procedures through careful process mapping and streamline legal basis as necessary Provide strong technical assistance in the preparation/design stage followed by hands-on implementation with stakeholders Explore possibility of using the Innova- tive Enterprise Institutions to deliver Initiate administrative reform program to simplify individual procedures in Develop new and strengthen existing vocational institutions to increase access to unskilled, informal labor sector including programs designed for practical teaching programs (e.g. Voca- tional Training Centre of Excellence in Lagos for the construction sector) Sponsor technical advisory trips to locations that incorporate religious and traditional practices of the domestic target groups, e.g. Islamic run abattoirs and butchers in Dubai, Malaysia, South training programs Low-quality and unproductive labor inputs: shortage of skilled labor, lack of vocational training, unskilled dures: burdensome business Onerous administrative proce- regulatory compliance, including EEG labor exhibits low productivity 253

267 asymmetries Technology and information procedures Onerous administrative Poor logistics & handling cost infrastructure Unreliable & high enforcement Poor environmental practices & X on product standards Lack of regulatory enforcement X Distortionary trade policy Finance labor inputs Low quality & unproductive Intervention approach livestock industries Provide technical assistance to the State Ministry of Environment to develop the appropriate level of ing to quality and safety standards of the institutional authority to enforce environmental regulations under its statutory responsibility products in the agriculture and animal Deliver capacity building program to strengthen inspection activities of rel- evant government authorities pertain- resources and expertise, as well as Growth constraints Lack of regulator Enforcement on Product Standards: Veterinary Services; Livestock & Poultry services neries (3 out of 6 have no chemical treatment facilities) Poor environmental practices and enforcement: improper chemical disposal practices by leather tan- (Continued) Potential Sectors Table IV.A1: 254

268 X (continued next page) X X X X X X X Increased access to term lending Targeted training Targeted investment in vocational regulations Specifi c fi nancing windows for the telecommunications sector nications regulator; new streamlined Replace import bans with tariffs Targeted capacity-building for commu- training, especially through IEI’s Industry associations should increase mechanisms by sharing information on industry; improve technology transfer the fl ow of market intelligence to the Power needs of clusters should be met more effectively through IPP’s, dedicated water and roads supply successful innovation capacity of the market. Spectrum to assess proper levels of service charges lacking; confusion over role management plan to be issued. Skills shortages Enforcement capacity and ability Market failures to access debt and equity fi nancing issues new licenses for new High cost of raw materials (uncom- Policy & regulatory reforms: NCC operators to enter exceeding the of telecoms agencies petitive steel producers) Skills shortages Unequal access to information sector investment level results in low level of public businesses especially for micro and small Policy neglect at local government Lack of access to fi nance : rising crucial role in opening of call centers. ICT: increasing productiv- ity for other sectors by reducing communications and transactions costs; strong links with fi nancial sector, large market (metal, wood process- ing, furniture) attracts FDI; recent demand in construction increases demand for structural timber and steel products increase in competition has led to productivity increases & product innovation; opportunities for regional investment; language skills allow for Light manufacturing 255

269 asymmetries Technology and information procedures Onerous administrative Poor logistics & handling cost infrastructure Unreliable & high enforcement Poor environmental practices & on product standards Lack of regulatory enforcement Distortionary trade policy Finance labor inputs Low quality & unproductive Intervention approach Rehabilitate refi nery in Warri Build IPP in industrial zone Reduce training costs Consider establishment of a bail-out Reform fast-track manufacturing line Activate fast-track lane for Facilitate acquiring of land Encourage mergers and acquisitions fund and of a credit line at conces- sional rates; construction of an IPP manufacturers Growth constraints dysfunctional Lack of access to black carbon Skills upgrade Access to land Power and water infrastructure because refi nery in Warri is Customs procedures Access to fi nance Customs procedures Too small and fragmented to oper- ate competitively Power and water infrastructure (Continued) World Bank/DFID (2008). Potential Sectors Tires Pharmaceutical goods assembly Car parts, motor cycle Color TV receivers Source: Table IV.A1: 256

270 Applying the Growth Identifi cation and Facilitation Framework | 257 Notes The authors wish to thank Doerte Doemeland, Hinh Dinh, John Litwack, † Ngozi Okonjo-Iweala, Brian Pinto, David Rosenblatt and Sunil Sinha for com- ments and suggestions. Excellent research assistance was provided by Frances Cossar and Dimitris Mavridis. ce of the Volker Treichel has been a Lead Economist in the World Bank’s Offi Senior Vice President and Chief Economist from December 2010. Prior to that, he served as Lead Economist in Nigeria. Before 2007, he worked at the IMF, including as mission chief for Togo and resident representative in Albania. 1. See chapter I of this volume. 2. See also Global Development Horizons (2011). 3. It is important to note that this fi nding does not necessarily imply that people in wage employment moved into family agriculture. It could also mean that those who previously reported no activity (that is, outside the labor force) but were at least temporarily involved in agriculture, became engaged in agriculture to an extent that they now reported employment in family agriculture. That means they moved from under-employment to employment. Rodrik (2010) fi nds evidence that labor moved from the wholesale/retail sectors (which have a reasonably high productivity) to agriculture. 4. See chapter II of this volume. Countries with a similar endowment structure should have a similar com- 5. parative advantage. The country with a lower wage level than the compara- tor country is hence able to produce a commodity at a lower cost than its competitor. Within the same industry, the complexity of the associated tech- c products a country nology may differ widely; as a result, in some specifi may have comparative advantage and others not. For example, when Korea entered the memory chip industry in the 1980s, Japan’s memory chip indus- tries were still expanding. What made Korea’s entry successful was that it started with simple, technologically matured chips which Japan had pro- duced 10 years ago. Also, an industry can be divided into different segments with different capital-intensity. For example, the IT industry can be divided, according to capital intensity, R&D, chips, spare parts, and assembly. The lower-income countries can enter the industry starting from labor-intensive assembly. 6. The similarities between Nigeria and Indonesia had earlier been recognized in a World Bank publication which reviewed the economic performance of the two countries over the period 1960-85 (see Bevan, Collier, and Gunning 1999). 7. See Lin (2011).

271 258 | New Structural Economics 8. ed binding con- A jobs summit that took place in Abuja in August 2010 identifi straints to growth for each of the key value chains. How to alleviate these con- straints has been agreed upon in a Memorandum of Understanding between the public and the private sector. These measures have subsequently been rati- fi ed by the government and are currently being implemented. References Bevan, David, Paul Collier, and Jan Willem Gunning. 1999. The Political New York: Economy of Poverty, Equity, and Growth: Nigeria and Indonesia. Oxford University Press. Bosworth, Barry, and Susan Collins. 2008. “Accounting for Growth, Compar- ing China and India,” Journal of Economic Perspectives 22, no. 1 (Winter): 45–66. Hausmann, Ricardo, Dani Rodrik and Andres Velasco, 2008. “Growth Disagnos- tic,” The Washington Consensus Reconsidered: Towards a New Global Gov- ernance, eds. Narcis Serra and Joseph E. Stiglitz . Cambridge, Massachusetts . Lin, Justin Yifu. 2011. “Economic Development in Natural Resource-Rich, Labor-Abundant Countries,” Let’s Talk Development Blog, February 28 (http://blogs.worldbank.org/developmenttalk/economic-development-in- resource-rich-labor-abundant-economies). Rodrik, Dani. 2010. “Globalization, Structural Change and Productivity Growth.” Working Paper 17143. National Bureau of Economic Research, Cambridge, MA. Treichel, Volker, ed. 2010. Putting Nigeria to Work: A Strategy for Employment and Growth. Washington, DC: World Bank. World Bank. 2011. Global Development Horizons: Multipolarity: The New Global Economy. Washington, DC: World Bank. World Bank. Forthcoming. “Light Manufacturing-Focused Policies to Enhance Private Investment and Create Productive Jobs.” Washington, DC: World Bank. World Bank/DFID. 2008. “Nigeria Valua Chain Analysis: Sector Choice and Market Analysis Report.” EME consultants, London.

272 V Financial Structure and Economic Development

273

274 V Financial Structure and Economic * Development † with Lixin Colin Xu Introduction nancial Financial structure differs greatly across countries. In bank-based fi systems such as in Germany, Japan, and India, banks offer the main fi nan- cial services in mobilizing savings, allocating capital, monitoring corpo- rate managers, and providing risk management services. In market-based systems such as in the United Kingdom, the United States, and Malaysia, nancial both stock markets and banks play important roles in all fi services. There are vast variations in fi nancial structure. Using compre- nancial structure, Demirgüç-Kunt and hensive cross-country data on fi Levine (2001) classify a large number of countries into four categories— nancially developed or underdevel- bank- or market-based systems in fi oped countries. Bank-based fi nancially-underdeveloped countries include Adapted from a paper presented at the Sixteenth World Congress of the International Economic * Association in Beijing, China, in July 2011. 261

275 262 | New Structural Economics Bangladesh, Nepal, Egypt, Costa Rica, Kenya, Sri Lanka, Indonesia, Colombia, Pakistan, Zimbabwe, Greece, Argentina, Venezuela, India, and Ireland. Market-based fi nancially-underdeveloped countries include Denmark, Peru, Chile, Brazil, Mexico, Philippines, and Turkey. Bank- nancially-developed economies include Tunisia, Portugal, Austria, based fi Belgium, Italy, Finland, Norway, Japan, France, Jordan, Germany, Israel, and Spain. Finally, market-based fi nancially-developed countries include the Netherlands, Thailand, Canada, Australia, South Africa, Korea, Sweden, Great Britain, Singapore, the United States, Switzerland, Hong Kong, and Malaysia. nancial structure? Does the combination of institu- What explains fi nancial system have any impact on tions and markets that constitute the fi economic development? These questions have fascinated economists for decades. One of the earliest attempts to address these questions was Gold- smith (1969), who 40 years ago tried to document the change in fi nancial structure over time and to assess the impacts of fi nancial development on economic development. He states that “one of the most important prob- lems in the fi eld of fi nance, if not the single most important one, almost everyone would agree, is the effect that fi nancial structure and development have on economic growth.” With data from 35 countries for the pre-1964 period, he fi nancial development and nds positive correlation between fi economic growth. But data constraints prevented him from going far on fi nancial structure: he could rely only on careful comparisons of Germany and United Kingdom. Obviously, it is hard to extend the conclusions from case studies to the rest of the world. Since Goldsmith wrote, there has been great progress in the research on fi nancial structure. Having collected comprehensive cross-country data on fi nancial structure themselves (along with their coauthors), Demirgüç- Kunt and Levine (2001) fi nancial systems nd from this new data set that fi become more complex as countries become richer with both banks and markets getting larger, more active, and more effi cient. But in general, the structure becomes more market-based in higher-income countries. They also fi nd strong and consistent evidence that what matters for economic development is the level of fi nancial development, and that the relative mix of banks and stock market does not matter much (Beck et al. 2001).

276 Financial Structure and Economic Development | 263 This conclusion that fi nancial structure is irrelevant for development cant challenge. Several authors argue theoretically that faces signifi nancial structure should matter a great deal. After all, economic devel- fi opment increases the demand for the services provided by securities markets relative to services provided by banks (Allen and Gale 2000; Boyd and Smith 1998). Moreover, banks and stock markets exhibit distinct effectiveness in delivering corporate governance and investor protection (Stulz 2001). In particular, banks are better at reducing the nancing standardized, shorter-run, lower- market frictions related to fi risk, and well-collateralized projects, while security markets are better at fi nancing more innovative, longer-run, and higher-risk projects that rely more on intangible inputs such as human capital (Allen and Gale 2000). Furthermore, the fact that stock markets become signifi cantly more active and important as economies develop also conjures up the notion that stock markets may become more useful as income levels rise. In this paper, we summarize some recent progress, both theoretical and empirical, that suggest that fi nancial structure does indeed matter for eco- nomic development, that banks and stock markets play different roles at countries at different development stages, and that there might be optimal fi nancial structure associated with each development stage. We also offer nancial structure in a country may deviate from evidence that the actual fi its optimum due to politics. In the rest of this paper, we fi rst summarize conventional wisdom and fi ndings about fi nancial structure. We then proceed to discuss some new ideas on this topic, along with empirical support for these new ideas. Traditional View of Financial Structure and Economic Development While the literature on the relative merits of banks versus markets is large, it can be summarized by four views (Beck et al. 2001; Levine 2002; Stulz 2001). The fi rst is the fi nancial-structure-irrelevancy view. In a perfect capital market with risk-neutral agents, the interest rate determines which investment opportunities are worth taking up, and all investment oppor- tunities yielding positive net return (after capital costs) will be taken (Stulz

277 264 | New Structural Economics 2001). If there is imperfect capital mobility, that is, if capital fl ows across c risks, then what borders are hindered by worries about country-specifi cient allocation of resources matters for job creation, fi rm growth, and effi cient fi nancial services and nancial system can provide effi is whether the fi cient access to fi nance; the mix of banks and markets does not matter. suffi According to this view, only fi nancial depth, not fi nancial structure, mat- ters for economic performance. nancial-structure-irrelevance view is A particular version of the fi the law nance view and fi , which argues that the primary determinant of the sound- ness of the fi nancial system is the legal system (La Porta et al. 2000). In nancial particular, this view holds that what is relevant for growth is not fi structure, but rather whether it is bank-based or market-based. The overall fi nancial development is determined by the legal system and the origins of law. The legal system may affect external fi nance because good legal pro- tection increases investors’ confi dence that they would reap at least some return on their investments (managed by fi rms) and as a result, they are rm managers (La Porta et al. more likely to provide investment funds to fi 2000; Stulz 2001). Underlying the fi nancial-structure-irrelevancy view are strong assump- tions that may not hold in reality. When the fi nancial system fails to direct savings to its more effi nancial structure becomes important cient uses, fi (Stulz 2001). Two key market imperfections destroy perfect fi nancial mar- kets (Stulz 2001): managers have an information advantage over investors about the fi rm’s activities (“hidden information”), and managers’ actions cannot be observed by investors (“hidden action”). Hidden information and action allow managers to pursue its own objectives. And managers cannot credibly commit to return investment returns to investors, who in turn may fail to fi nance projects that may have positive returns to them in a perfect-information world. With these two issues in mind, fi nancial structure leads to real consequences when it changes information and transaction costs, affects the cost of capital, and alters the incentives and monitoring of management. The bank-based view emphasizes the positive role of banks in mobi- lizing resources, identifying good projects, monitoring managers, and managing risks, and highlights the shortcomings of the stock market (Beck et al. 2001). One of the pioneers in research on fi nancial structure,

278 Financial Structure and Economic Development | 265 Gerschenkron (1962), suggests that banks are more important than mar- kets in the early stage of economic development when the institutional environment cannot support market activities effectively. The reason is that even in countries with weak legal and accounting systems and frail institutions, powerful banks can force fi rms to reveal information and pay their debts, thereby helping industrial growth. Moreover, banks may be better than markets at providing external fi rms requiring nance to new fi staged fi nancing: banks can more credibly commit to making additional nd it more dif- funding available as the project proceeds, while markets fi cult to make credible, long-term commitments. In contrast, a good stock fi market quickly and fully reveals information in public markets, which decreases the incentives for investors to acquire information. Good mar- ket development may thus impede incentives for identifying innovative projects and thereby hinder effi cient resource allocation. Moreover, liquid markets also lead to a myopic investment sentiment—all investors need to do is to watch stock prices without having to actively monitor fi rm man- agers, which hinders corporate control. In contrast, the market-based view regards stock markets as crucial in promoting economic success (Beck et al. 2001). Markets allow investors to diversify and manage risks more effectively, thereby encouraging more nance. Market-based systems also facilitate compe- supply of external fi tition, which induces stronger incentives for R&D and growth. Thus, market-based systems may be especially effective in promoting innovative and R&D-based industries (Allen and Gale 2000). Liquid stock markets also allow investors to build and seek large stakes, therefore enabling hostile takeovers to discipline shirking or incompetent managers (Stulz 2001). This market-based view also emphasizes the negative roles played by banks. By spending expensive resources on information about fi rms, banks can extract large rents from fi rms, which reduces the incentives for fi rms lose a large rms to undertake high-risk, high-return projects since fi share of the rents to the banks. Moreover, because of the nature of the debt contracts—banks do not benefi t from high returns but are harmed by low returns—banks prefer to fi nance safe and low-return projects, retard- ing innovation and growth. Moreover, powerful banks may collude with fi rm managers to prevent entry by other investors, a practice that reduces competition and effective corporate control and therefore growth.

279 266 | New Structural Economics Conventional Empirical Results Demirgüç-Kunt and Levine (2001) use the new cross-country database on fi nancial structure to document how fi nancial structure evolves with nancial structure by ratios of economic development. They characterize fi banking sector development (measured by size, activity and effi ciency) rela- tive to stock market development (similarly measured), with a higher ratio meaning a more bank-based structure. They then classify countries into bank-based or market-based countries. Austria, France, Germany, Great Britain, Hong Kong, Japan, the Netherlands, and Switzerland have com- paratively large, active banking systems. In contrast, Argentina, Colombia, Costa Rica, Ghana, Nepal, Nigeria, Peru, Turkey, and Zimbabwe have particularly small, inactive banking systems. In terms of stock market development, some countries emerge as particularly well-developed by all measures (Australia, Great Britain, Hong Kong, Malaysia, the Netherlands, Singapore, Sweden, Switzerland, Thailand, and the United States), while other countries, such as Chile and South Africa, are large and illiquid. A few countries, such as Germany and Korea, have active but small stock markets. Demirgüç-Kunt and Levine fi nd that banks, nonbanks, stock markets, and bond markets are larger, more active, and more effi cient in richer countries, confi rming the fi ndings of Goldsmith (1969) with a smaller nancial systems on aver- sample of countries in earlier periods. Thus fi age are more developed in richer countries. In addition, stock markets in higher-income countries tend to be more active and effi cient relative to banks. Furthermore, fi nancial structure is more market-oriented in countries with common law tradition (as distinct from a civil law tradi- tion), strong protection of minority shareholder rights, good accounting systems, low levels of corruption, and no explicit deposit insurance. This is consistent with theories that argue that higher information costs and worse legal protection of property rights tend to favor banks over markets (Allen and Gale 2000; Stulz 2001). Beck et al. (2001) provide comprehensive evidence that fi nancial struc- ture does not matter but that fi nancial depth does. They combine the new cross-country database of fi nancial structure with both fi rm-level and cross-country industry level data. Relying on evidence about fi nancial structure and economic performance at three levels (pure cross-country

280 Financial Structure and Economic Development | 267 comparisons, cross-industry, cross-country methods, and fi rm-level data nd no evi- across many countries), they obtain consistent results. They fi dence that fi nancial structure helps explain country economic perfor- nancially dependent industries mance: “Countries do not grow faster, fi do not expand at higher rates, new fi rms are not created more easily, rms’ access to external fi nance is not easier, and fi rms do not grow faster fi in either market-based or bank-based systems.” In contrast, they write, nancial development does help “distinguishing countries by overall fi explain cross-country differences in economic performance. Measures of bank development and market development are strongly linked to eco- cally, the data indicate that economies grow nomic growth. More specifi nance expand at faster faster, industries depending heavily on external fi new fi rms form more easily, fi rates, nancing rms’ access to external fi is easier, and fi rms grow more rapidly in economies with a higher level of overall fi nancial-sector development.” They also fi nd that the part of fi nancial development explained by the legal system consistently explains rm, industry, and national economic success, consistent with the law and fi nance view of fi nancial structure. fi New Waves of Theoretical Arguments Does fi nancial structure really not matter for development? Recent develop- ments cast doubt that this is the case. First, economists have come to real- ize that there is often no one-size-fi ts-all recipe for development (Kremer 1993). The reform areas with the largest payoffs differ from country to country, and there are often development “bottlenecks,” which can be lik- ened to the famous failure of the space shuttle : with thousands Challenger of components, it “exploded because it was launched at a temperature that caused one of those components, the O-rings, to malfunction” (Kremer 1993). Consistent with this notion of country-specifi c and development- stage-specifi c bottlenecks, some research has found policy complementarity in various contexts. In particular, Xu (2011) summarizes evidence suggest- ing that the effects of the business environment on development tend to be heterogeneous depending on the stage of development, and that in particu- lar, bad infrastructure and labor infl exibility—as in the case of India—tend to be key bottlenecks because of their negative indirect effects.

281 268 | New Structural Economics Second, the theory on fi nancial structure has evolved. Relying on the comparative experience of Germany, Japan, the United Kingdom, and the United States, Allen and Gale (2000) examine whether fi nancial structure matters. They conclude that since banks and markets offer distinct fi nan- cial services, economies at different stages of development require distinct mixtures of these fi nancial services to work effi ciently. They conjecture that a country will require different mixtures of fi nancial services (that is, banks and stock markets) as it grows richer (Boyd and Smith 1998), and that nancial structure differs from the optimal mix- when a country’s actual fi ture of banks and markets, its economy will not obtain the appropriate mix of fi nancial services, hurting economic growth. nancial structure has to Lin, Sun, and Jiang (2011) also argue that fi matter for development but from a different angle. The key reason, they contend, is that an effi nancial structure must refl ect the demand of cient fi the real economy. Fundamentally, factor endowments (labor, capital, and natural resources) determine industrial structure, which in turn needs the support of a certain development-stage-specifi c fi nancial structure. In par- ticular, at each stage of development a country has a specifi c combination of factor endowments. That combination determines factor prices, which in turns determines the optimal industry structure, the nature of its associ- rm size (Lin 2009). Since enterprises that ated risk, and distribution of fi nancing needs, the operate in different industries differ in size, risk, and fi nancial services at some development demands of the real economy for fi stages can be systematically different from those of the same economy at other stages. When the characteristics of fi nancial structure match those of an economy’s industrial structure, the fi nancial system can perform its fun- damental functions most effi ciently and thus contribute to sustainable and inclusive development. Therefore, there is an optimal fi nancial structure for an economy at each stage of development. With respect to developing economies, the key characteristic of their endowment structures is the relative abundance of unskilled labor (and scarcity of capital). Labor-intensive industries and labor-intensive sections of capital-intensive industries have the comparative advantage and should thus dominate in these economies. Since the experiences from developed economies can be mimicked, the industries, products, and technologies that are appropriate in developing economies are relatively mature. With respect

282 Financial Structure and Economic Development | 269 to fi rms in labor-intensive industries are usually smaller, especially rm size, fi rms in capital-intensive industries. The effi in terms of capital, relative to fi - ciency of the fi nancial system in developing countries therefore depends on its ability to serve the fi nancing needs of labor-intensive, small, and mature businesses. Since those fi rms also tend to be more opaque due to their lack nancial information, screening fi rms and monitoring fi rm of standard fi managers become the major concerns for providers of external funds to these fi rms. In such an economic environment, banks, especially small local banks, have more strengths than stock markets, due to their superior abilities to harness local information, assess “soft” information regarding creditworthiness, and engage in long-term relationships with borrowers. In addition, banks are particularly attractive to fi rms in low-income countries because banks represent lower costs of capital to fi rms in such countries: (i) when borrowing from no more than a few banks, these fi rms do not need to have public information such as fi nancial statements and external audit- ing ready for the lenders, thus saving precious capital; (ii) interest rate pay- ments for loans tend to be lower than returns to shares in the stock market due to lower risks associated with bank loans, a fact that again saves pre- cious capital from the perspective of fi rms in developing countries. Thus, if there are no distortions, the fi nancial systems in these economies are likely to be characterized by the dominance of banks. cant role It is likely that small regional banks play an especially signifi in effi ciently serving small fi rms in developing countries. Recent evidence suggests that there is a match between bank size and the size of fi rms that these banks serve. Large banks tend to shy away from small businesses but rather focus on large businesses, while small banks tend to target small businesses. Large banks can save transaction costs if making loans mainly to large businesses—since making a loan, no matter how large or small, involves the same procedures and forms. Making a few large loans to large businesses, rather than many smaller loans to small fi rms, therefore lowers the unit costs of loans for large banks. Serving small fi rms is thus left to small banks in developing countries. In contrast, the key characteristics of the endowment structure in devel- oped countries are the relative abundance of skilled labor and capital. The comparative advantage of these countries is then capital-intensive indus- tries. Firms in such industries tend to be large, demanding more external

283 270 | New Structural Economics fi nancing. Since these countries are already at or near the technological rms there would spend much more on R&D and innovation, and frontier, fi 1 rm bear higher technological and product innovation risks. With larger fi xed costs of providing standard rms can afford the (more or less) fi sizes, fi nancial information to the market, and specialized fi nancial agencies can fi make suffi cient money and become viable in providing specialized fi nancial nancial information avail- and auditing information. Thus with standard fi nance able, stock markets, bond markets, and big banks become the main fi rms. providers to these capital-intensive fi Moreover, there are arguments that stock markets are better suited to richer countries. For fi rms with new technologies or innovative projects, investors do not have much information and often have diverse opinions about the prospects of these new technologies. Decentralized stock markets allow people to agree to disagree about the future prospects of these fi rms, and these fi rms, as a result, are more likely to be funded (Allen and Gale 2000). Furthermore , stock markets can take advantage of the standard fi nancial information—information available only in richer countries—to reduce the information asymmetry between the managers of a fi rm and the external investors, which allows investors to make more informed deci- sions about what fi rms they are more likely rms to invest in and in which fi to have safer returns. Venture capital is often involved in the early stage of high-risk innovative and capital-intensive fi rms, but stock markets remain crucial by providing exit options for venture capital and by fi nancing fur- ther development of these high-tech businesses. Banks can also offer staged investment once venture capital has identifi ed good projects as demon- strated by good initial returns. Thus, for rich countries, the optimal fi nan- cial structure is likely characterized by a large and active stock market, 2 augmented with many large banks. As a result, for a country at a certain stage of its economic develop- ment, some specifi c fi nancial structure will be more effi cient in mobilizing and allocating capital. In other words, there is a certain optimal fi nancial structure at a specifi c stage of development, in which the composition and relative importance of available fi nancial arrangements can most effi ciently allocate fi nancial resources to viable fi rms in the competitive sectors of the optimal industrial structure, which is in turn determined by its endow- ment structure. The optimal fi nancial structure for developing countries

284 Financial Structure and Economic Development | 271 tends to feature a stronger role for banks (especially small banks) than for stock markets, while the opposite is true for developed countries. More- over, the optimal fi nancial structure is dynamic. As the endowment struc- ture changes with physical and human capital accumulation, the optimal nancial structure changes accordingly. There is, therefore, industrial and fi no unique fi nancial structure that fi ts all countries. For future reference, we call this view of optimal fi nancial structure specifi c to each development stage the new structural view. Some New Empirical Results Several fairly recent papers offer evidence supporting the premise that nancial structure matters in various ways for economic development. The fi fi rst and the key piece of evidence is based on a cross-country study by Demirgüç-Kunt, Feyan and Levine (2011). Noting that the past literature has not been successful in identifying the importance of fi nancial structure, they explore whether deviations from an optimal fi nancial structure are associated with the speed of development. They use data from 72 coun- tries from 1980 to 2008 to reassess the role of fi nancial structure in eco- cally, they assess whether the sensitivity nomic development. More specifi of economic development to increases in bank securities market develop- ment change during the process of economic development, and whether each level of economic development is associated with an optimal fi nancial structure. Financial structure here is measured as the ratio of private credit (as a share of GDP) to security market capitalization (as a share of GDP) and some of its variants. The authors use quantile regressions to assess how the sensitivities of economic activity to bank and securities market development evolve as countries grow. The quantile regressions provide information on how the associations between economic development and both bank and securities market development change as countries grow richer. In contrast, the con- ventional cross-country studies tend to focus on the association between economic development and fi nancial structure for the “average” country. The reliance on quantile regression, which implicitly insists that the effects of fi nancial structure have distinct effects for countries at different income levels, proves to be the key for fi nding that fi nancial structure matters.

285 272 | New Structural Economics A measure of optimal fi nancial structure at each level of development nancial structure as a share of is constructed by regressing a measure of fi GDP per capita for the sample of OECD countries, while also controlling for key institutional, geographic, and structural traits of those countries. The maintained hypothesis is that conditional on these traits, the OECD countries provide information on how the optimal fi nancial structure var- cients from ies with economic development. Next, the authors use the coeffi the regression to compute the estimated optimal fi nancial structure for each nancial structure gap,” which country in each year. They then compute a “fi is equal to the natural logarithm of the absolute value of the difference between the actual and the estimated optimal fi nancial structure. nd that as economies develop, both banks and markets become They fi larger relative to the size of the overall economy. More importantly, as countries become richer, the sensitivity of economic development to changes in bank development decreases, while the sensitivity of economic develop- ment to changes in securities market development increases. Thus the rela- tive demand for the services provided by the stock market increases as an economy develops, and these services differ from those the banks provide, as suggested by Allen and Gale (2000). Demirgüç-Kunt, Feyan, and Levine (2011) fi nd some support for the notion that there is appropriate fi nancial structure for countries at distinct stages of development. In particular, deviations in an economy’s actual fi nancial structure from its estimated optimal one (that is, the size of the fi nancial development gap) are associated with reduced economic output. Even when controlling for the level of bank development, securities market xed effects, there is development, a standard set of controls, and country fi a robust and negative relationship between the fi nancial structure gap and economic activity. They also look at whether it matters if the non-optimal fi nancial structure is due to too much bank orientation or too much mar- ket orientation, and fi nd that neither matters. Magnitudes of the effects of deviations from the optimal fi nancial structure are non-trivial: an increase of one standard deviation in the fi nancial structure gap is associated with a drop in log real GDP per capita of 0.06, or a 6 percent reduction in eco- nomic activity. Further controlling for country and period fi xed effects and some standard controls, the magnitude drops by 50 percent but remains signifi cant. The magnitude, at face value, is interesting: it is certainly non- trivial, but the magnitude is not overwhelmingly important.

286 Financial Structure and Economic Development | 273 It is useful to point out that this paper does not deal with the potential nancial structure. If fi nancial structure does respond endogeneity of the fi to income level, as the theory would suggest, then there may be bias for the nancial structure. Moreover, it is not clear that using estimated effects of fi nancial structure is completely convinc- OECD countries to infer optimal fi ing. After all, OECD countries are all rich, and it is unclear that this group of countries can form the base to infer the optimal fi nancial structure in much poorer countries. Still, the paper offers plausible empirical support nancial structure and is a nice fi rst step for the new structural view of fi toward disentangling the mystery of the impact of fi nancial structure. The second clue of the effects of fi nancial structure comes from a large fi rm-level data across 89 coun- rm-level dataset. Cull and Xu (2011) use fi rms vary with their country’s tries, looking at how labor growth rates of fi nancial structure . An important advantage of combining fi rm-level data fi nancial structure is the ability to examine and cross-country indicators of fi how various types of fi rms may be affected differently by fi nancial struc- ture. This allows us to distinguish between the effi ciency-based (that is, the new structural view) and the political-economy based approach of explain- ing the evolution of fi nancial structure. rm-level labor growth rates to country- Cull and Xu (2011) regress fi level measures of bank and stock market development (after controlling for basic fi rm and country characteristics). They are concerned about the nancial structure in the labor growth equation potential endogeneity of fi for two reasons. First, there might be omitted variables that are corre- lated with both fi nancial structure and labor growth rates. Such variables might, for instance, include non-fi nance business environment variables (Xu 2011). Second, causality might go both ways, from fi nance to fi rm growth, or vice versa. They thus resort to instrumental variables to deal with such issues. In particular, they consider potential instrumental vari- ables including natural resource dependence, the level of trust in a society, cereal plantation patterns, settler mortality, and so on, and choose a sub- set of these potential instrumental variables that are related to fi nancial structure yet pass the over-identifying restrictions test. Beside the instru- mental variable approach, they also use the Rajan-Zingales difference- in-difference approach to examine whether fi rms in industries that rely more heavily on external fi nance benefi t more in terms of fi rm growth from fi nancial development at the country level, holding constant both

287 274 | New Structural Economics country and industry fi xed effects, therefore controlling for all country- c factors. This approach signifi and industry-specifi cantly reduces the extent of omitted variable bias. Relating fi rm growth to fi rm and country characteristics and fi nancial structure, and taking into account the potential endogeneity of fi nancial nd that labor growth is swifter in low- structure, Cull and Xu (2011) fi income countries that have a higher ratio of private credit to GDP, and the growth-spurring effects of banking development are especially pronounced nance. In high-income countries, in industries that heavily rely on external fi labor growth rates are increasing in the level of stock market capitaliza- tion. Both patterns are consistent with predictions from the new structural view and some earlier theoretical conjectures (Allen and Gale 2000; Boyd and Smith 1998; Lin, Sun, and Jiang 2011). The third clue about the effects of fi nancial structure emerges from examining the impact of fi nancial structure on poverty. Financial structure might affect poverty because entrepreneurs have trouble obtaining fi nance due to information asymmetry between them and investors—the entrepre- neurs know more about the prospects of the projects than banks and ato- mistic investors in the stock market. A number of researchers argue that banks are better able to reduce this information asymmetry problem than stock markets. One reason is that banks form long-term relationship with borrowers and can benefi t from the value of the information obtained from this long-term relationship. In contrast, well-established stock markets quickly and publicly reveal information, thereby reducing the incentives for individual investors to acquire information. Banks therefore may have better capacity to reduce the information asymmetry issue and make exter- nal fi nancing possible. Moreover, since stock markets rely more strongly on the legal and accounting framework to safeguard necessary returns to investors, the effects of stock markets may depend on institutions to a greater extent, whereas banks can more effectively force fi rms and house- holds to honor their contracts than stock markets (Gerschenkron 1962, Boyd and Smith 1998), and are therefore especially important in poorer countries with weak contract enforcement. Based on the above logic, Kpodar and Singh (2011), using data from 47 developing countries from 1984 to 2008, show that fi nancial deepening through banks is associated with reduced poverty levels, while market-based

288 Financial Structure and Economic Development | 275 measures of fi nancial development are associated with higher incidence of poverty in this sample. In addition, the interaction between institutional quality and the sized-based measures of the importance of stock markets relative to banks is negative and signifi cant in their regressions, indicating nan- that as institutions improve, the positive link between market-based fi cial development and poverty incidence phases out, and even reverses after some threshold of institutional quality is reached. Conversely, the results suggest that in weak institutional environments bank-based fi nancial sys- tems tend to reduce poverty more than market-based ones. The authors nancial structure for poverty by using have dealt with the endogeneity of fi xed effects, the system GMM estimator, which controls for country fi nancial structure and other variables to be endogenous and and allows fi predetermined. The Deviation from the Optimal Financial Structure Besides derived demand based on industrial structure that originates in the endowment structure, there are other determinants of fi nancial struc- ture that cause the actual fi nancial structure to deviate from the optimal one. Earlier research has shown that fi nancial structure is signifi cantly and robustly related to law and legal origins (La Porta et al. 2000; Demirgüç- Kunt and Levine 2001); here we focus on several other factors such as the role of belief and ideas and the role of politics, which have emerged as potentially important in recent studies. 3 In most rst factor stems from the belief of government leaders. The fi developing countries, the government plays a very important role in defi ning the structure of the economy. And the belief of the top gov- ernment leaders will naturally shape the country’s fi nancial structure. A case in point is the fi nancial repression that is widely observed in many developing countries. Countries featuring fi nancial repression tend to adopt policies restricting entry into the banking sector, controlling inter- est rates, and intervening in the allocation of bank loans. As a result, a few big banks tend to dominate the banking landscape, and capital tends to fl ow to large fi rms. Small businesses, which have comparative advan- tages in these economies, have little access to credit and have to make do with internal capital or resort to informal channels for external fi nance.

289 276 | New Structural Economics Why do countries adopt such obviously ineffi cient policies? Inappropri- ate development strategies adopted by the government are likely the main driving force leading to these repressive policies and distorted fi nancial 4 system. If the government’s priority is to promote industries that are inconsistent with the comparative advantages endogenously determined by the economy’s endowment structure, it has to use distortional policies to channel scarce resources into the priority sectors. As a result, govern- nancial system are ment interventions and consequent repression of the fi inevitable. Due to inertia of institutional change, such distorted policies can have prolonged infl uence on the evolution of the fi nancial system. A good example of this practice is China. In the 1950s, the factor endow- ments in the Chinese economy were characterized by extreme scarcity of capital and enormous abundance of labor. The government, however, decided to adopt an ambitious comparative-advantage-defying develop- ment strategy in which establishment and development of heavy industries took the fi rst priority. To push the development of heavy industries, which are very capital-intensive, the government had to deliberately distort prices of various products and production factors including labor, capital, and foreign exchange; replace market mechanisms with a government plan- ning system to control the allocation of production factors; nationalize private businesses; and collectivize agricultural production with the Peo- ple’s Communes. In this centrally planned economic regime, banks were closed or merged into the People’s Bank of China, which became the only fi nancial institution in the whole economy until the end of 1970s. After the reform and opening in the late 1970s, the government adopted a dual- track approach to the transition: on one hand, some transitory protections and subsidies were provided to fi rms in the old priority sectors, and, on the other hand, entry to sectors that were consistent with the economy’s comparative advantages and were repressed in the old strategy were liber- alized. As part of the economic reform, four big state-owned banks were established in the early 1980s. A dozen joint-stock commercial banks were also set up in the late 1980s and early 1990s. But interest rates are still under the control of the state, and domestic entry into the banking sec- tor is rigidly restricted by the government. The market share of the four big state-owned banks has slowly declined, but they still hold a dominant position in the banking system today. Because of this serious mismatch of

290 Financial Structure and Economic Development | 277 fi nancial structure with optimal industrial structure, labor-intensive small nancial credit, a situation businesses have very limited access to formal fi that reduces job creation and contributes to widening inequality of income distribution in China. nancial structure from A second factor behind the deviation of actual fi its optimal one stems from the belief of many policy advisors in the benefi ts of fi nancial liberalization and the possibility of leapfrogging in fi nancial nancial repression, fi nan- development. As a policy prescription to correct fi cial liberalization has been generously prescribed by theorists and exer- cised by many developing countries. While those repressive policies should be removed, some new, less noticeable policy distortions may be intro- duced in the process of fi nancial liberalization. It is not rare that developing countries are advised to establish and develop a fi nancial system similar to those in the advanced economies. The U.K. and U.S. fi nancial systems, where fi nancial markets are highly active, are often taken as the model that developing countries should follow. This model is often justifi ed by the supposed superiority of fi nancial markets. As a result, some small, low- income economies are eager to develop stock markets, consolidate small banks into large banks. and repress the development of local banks. However, as the new structural view has argued—and with support- ing evidence emerging—the optimal fi nancial structure for poor countries is likely to be systemically different from that for advanced economies. Thus imitating the fi nancial model of advanced economies will not lead to improved effi nancial system nor generate better eco- ciency of the fi nomic performance in poor countries. Such imitation may even result in destructive consequences such as fi nancial crises. Such policy advice is also inconsistent with the growth experience of those successful economies in their industrialization periods. For instance, in the British Industrial Revolution, industrial enterprises were typically very small and mainly internally fi nanced at both the start-up and expansion stages. In the case of external fi nance, personal contacts played a crucial role. The role of the banking system in fi nancing long-term investment in industrial sec- tors was insignifi cant. The British banks were typically small and locally based with a limited number of offi ces until at least the mid-19th century. Bank merger movement in England did not develop until the 1860s, with the peak of merger activity occurring in the late 1880s and early 1890s.

291 278 | New Structural Economics When it comes to the role of capital market, history shows that capital nancing industrial sectors markets started to play an important role in fi only at the end of 19th century. In the United States before 1890, indus- trial fi rms were numerous, small, and closely owned. Industrial securi- ties, except in the coal and textile industries, were almost unknown. A capital market for industrial preferred stocks did not develop until the 1887–1904 merger wave. And public markets for common stocks devel- nancial markets are a prominent ele- oped even later. Therefore, while fi ment of the current U.K. and U.S. fi nancial systems, this was not the case at the early stage of economic development in these countries. According to Cull et al. (2006), during the 19th and early 20th centuries, it was a variety of local fi nancial institutions that emerged to meet the needs of rms in the economies of the North Atlantic small- and medium-sized fi Core. These fi nancial intermediaries were able to tap into local informa- tion networks and so extend credit to fi rms that were too young or small to get funds from large, fi nancial center banks. The third factor causing deviation of reality from optimal fi nancial structure is politics, as argued by Calomiris and Haber (2011) in the case of bank crisis—in which case the fi nancial structure is clearly not optimal. Many under-banked economies repeatedly supplied credit imprudently: once a crisis was over, banks appeared to continue misallocating scarce credit to fi rms and households that were prone to default. Why? Calomiris and Haber rely on reasoning rooted in political economy that can explain the prevalence of fragile banking systems that allocate credit narrowly. The key reason is that government actors face inherent confl icts of interest when it comes to the operation of the banking system, and those confl icts can lead to banking instability and undersupply of credit. Specifi cally, govern- ments regulate and supervise banks to limit risk taking but they also rely nance (by borrowing from and tax- on banks as a source of risky public fi ing them). In addition, while governments enforce contracts that discipline bank borrowers, they also depend on bank debtors for votes or political support. Finally, governments distribute losses among creditors when a bank fails, but they also must depend on the largest creditor group—bank depositors—for their political fate. These confl icts of interest imply that regulatory policies toward banks often refl ect the interest of the political coalitions that support the government.

292 Financial Structure and Economic Development | 279 This political economic framework turns out to be very useful for under- standing banking structure in a series of historical case studies (Scotland, England, United States, Canada, Mexico, and Brazil). Indeed, formation of viable political coalitions under different types of government dictated the evolution of the banking structure in each of these countries. Adapting the conceptual framework to the historical case studies leads to a number of conclusions. Foremost, the nature of the coalitions that generate barriers to entry in banking varies across types of political regimes. In an autocracy, it is easier to create a stable coalition in favor of tight entry restrictions, in part because potential borrowers from banks do not have a voice in the political process. Autocracies therefore tend to create banking systems that allocate credit narrowly to the government and to enterprises owned by an elite class of government-selected bankers. The narrow allocation of credit under authoritarian regimes has not resulted in greater banking sector sta- bility, however: in times of economic strife, bank insiders and the govern- ment expropriate fi rms and households that are either loosely or not at all affi liated with the coalition (that is, minority shareholders and depositors). In times of extreme diffi culty, the autocrat can (and has) expropriated bank insiders. Mass suffrage, by giving voices to mass economic actors, makes it harder to sustain a banking system that allocates credit narrowly to an elite group. It does not, however, necessarily guarantee banking stability. Bank borrowers can vote for representatives that expand the supply of credit, improve the terms on which credit is offered, and then forgive those debts when they prove diffi cult to repay. This was largely the story of the U.S. subprime crisis. Under any type of political system, banking systems are fragile. Therefore, only a small share of countries has been able to enjoy stable banking along with broad credit supply, because this outcome requires political institutions that allow for mass suffrage, but also limit the authority and discretion of the parties in control of the government. To shed light on whether real fi nancial structure tends to deviate from the optimal fi nancial structure, Cull and Xu (2011) examine the types of fi rms that benefi t more from private credit market development. In particu- lar, they allow the private credit variable (that is, private credit as a share of GDP) to interact with fi rm characteristics such as fi rm size and capital inten- sity in the labor growth equation, estimated at the fi rm level. The authors

293 280 | New Structural Economics fi rms in low-income countries benefi t most nd no evidence that small-scale fi from private credit market development. Rather, the labor growth rates of large and capital-intensive fi rms increase more with the level of private credit market development. Thus large and more-capital-intensive fi rms seem to benefi t more from banking development in low-income countries. nancial structure likely deviates from the This suggests that the actual fi optimal fi nancial structure. In particular, the likely scenario is that banks in developing countries tend to lend mostly to large and capital-intensive fi rms, allowing a small segment of elite fi rms to grow faster. Such a sce- nario could be due to an over-concentrated banking structure dominated by large banks, which in turn lend largely only to large fi rms (Lin, Sun and Jiang 2011), or political coalition between political and banking insiders restrict entry into the banking sector, resulting in a bank sector dominated by large banks, which lends largely to affi liated inside fi rms that tend to be large and capital-intensive (Calomiris and Haber 2011). Conclusions What explains the vast variations across countries in fi nancial structure? Does fi nancial structure have any impact on economic development? There has been some evolution on these questions. The traditional theoretical views tend to argue that fi nancial structure does not matter. The traditional empirical consensus tends to imply that it is fi nancial depth, not fi nancial structure, that determines aggregate economic performance. nancial services are Several researchers have recently argued that fi endogenous to industrial structure which in turn depends on a country’s relative endowment structure, and optimal fi nancial structure should be specifi c to the particular development stage. And some recent fi ndings seem to support this view. In particular, while both banks and stock markets become larger and more active as a country grows richer, stock markets become relatively more important. Moreover, as economies become richer, the sensitivity of economic development to changes in bank development decreases, while the sensitivity of economic development to changes in stock market development increase, thus the relative demand for the service pro- vided by stock market increases. In addition, deviation of a country from its optimal fi nancial structure is found to be negatively and signifi cantly

294 Financial Structure and Economic Development | 281 related to a lower income level. Firm-level evidence also shows that bank development has particularly strong effects in relatively poor countries, especially in those industries heavily relying on external fi nance, while stock market development has particularly strong effects in relatively rich countries. Banks (relative to stock markets) are also found to be relatively better in reducing poverty in developing countries, especially in institution- ally weak countries. On the other hand, there is no evidence that small fi t more from bank development, due to rms in developing countries benefi the deviation of actual fi nancial structure from the optimal one. ndings have important implications. First, the optimal fi The fi nancial structure changes, becoming more market-oriented, as economies develop. nancial structures Second, new evidence suggests that indeed different fi may be better at promoting economic activity at different stages of a coun- . These fi try’s economic development nancial structure ndings advertise fi as an independent fi nancial policy consideration. And if the optimal mix- ture changes as an economy develops, then this suggests the desirability of appropriately adjusting fi nancial policies and institutions as countries develop. Third, politics, legal origins, and beliefs of government leaders may cause the actual fi nancial structure in a country to deviate from its ciency and welfare losses to the economy. optimal, resulting in some effi Improving the understanding of what the optimal is and the effi ciency and welfare losses due to the deviation from the optimal, therefore, may miti- gate the impact of political and other belief-related factors in the determi- nation of a country’s actual fi nancial structure. Notes † This paper benefi ted from discussions with Robert Cull and Asli Demirgüç- Kunt, and the discussions at the World Bank Conference on Financial Struc- ture held in Washington, D.C., on June 17, 2011. Lixin Colin Xu is a Lead Economist in the Development Research Group of the World Bank. 1. Technological innovation risks are those related to successfully developing new products, while product innovation risks concern those related to successfully getting the new product accepted by the market. 2. There would also be numerous small banks offering services to small labor- intensive fi rms in the non-tradable sectors.

295 282 | New Structural Economics 3. The next four paragraphs draw heavily from Lin, Sun, and Jiang (2011). See also references therein. 4. See Lin (2009) for detailed discussion of development strategy and its impact on the development of fi nancial institutions. References Allen, Franklin, and Douglas Gale. 2000. Comparing Financial Systems . Cambridge, MA: MIT Press. Beck, Thorsten, Asli Demirgüç-Kunt, Ross Levine, and Vojislav Maksimovic. 2001. “Financial Structure and Economic Development: Firms, Industry, and Country Evidence.” In Financial Structure and Economic Growth: A Cross- Country Comparison of Banks, Markets, and Development, eds. Demirgüç- Kunt and Levine. Cambridge, MA: MIT Press. Boyd, John H., and Bruce D. Smith. 1998. “The Evolution of Debt and Equity Markets in Economic Development.” Economic Theory 12: 519–60. Calomiris, Charles, and Stephen Haber. 2011. “Fragile Banks, Durable Bargains: Why Banking Is All about Politics and Always Has Been.” Stanford University, Stanford, CA. Cull, Robert, and Lixin Colin Xu. 2011. “Firm Growth and Finance: Are Some Financial Institutions Better Suited to Early Stages of Development than Others?” World Bank, Washington, DC. Cull, R., L. E. Davis, N. R. Lamoreaux, and J. Rosenthal. 2006. “Historical Financing of Small- and Medium-sized Enterprises.” Journal of Banking and 30: 3017–42. Finance Demirgüç-Kunt, Asli, Erik Feyen, and Ross Levine. 2011. “Optimal Financial Structures and Development: The Evolving Importance of Banks and Markets.” World Bank, Washington, DC . Demirgüç-Kunt, Asli, and Ross Levine. 2001. “Bank-Based and Market-Based Financial Systems: Cross-Country Comparisons.” In Financial Structure and Economic Growth: A Cross-Country Comparison of Banks, Markets, and Development , eds. Demirgüç-Kunt and Levine. Cambridge, MA: MIT Press. Gerschenkron, Alexander. 1962. Economic Backwardness in Historical Perspec- tive, a Book of Essays . Cambridge, MA: Harvard University Press. Goldsmith, Raymond W. 1969. Financial Structure and Development . New Haven, CT: Yale University Press. Kpodar, Kangni, and Raju Singh. 2011. “Does Financial Structure Matter for Poverty? Evidence from Developing Countries.” International Monetary Fund, Washington, DC. Kremer, Michael. 1993. “The O-Ring Theory of Economic Development.” Quarterly Journal of Economics 108: 551–75.

296 Financial Structure and Economic Development | 283 La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert W. Journal of Vishny. 2000. “Investor Protection and Corporate Governance.” Financial Economics 58: 3–27. Levine, Ross. 2002. “Bank-Based or Market-Based Financial Systems: Which Is Journal of Financial Intermediation 11: 1–30. Better?” Lin, Justin Yifu. 2009. Economic Development and Transition: Thought, Strategy, and Viability . New York: Cambridge University Press. Lin, Justin Yifu, Xifang Sun, and Ye Jiang. 2011. “Toward a Theory of Optimal Financial Structure.” World Bank, Washington, DC. Stulz, René. 2001. “Does Financial Structure Matter for Economic Growth? A Corporate Finance Perspective.” In Financial Structure and Economic Growth: A Cross-Country Comparison of Banks, Markets, and Development, eds. Demirgüç-Kunt and Levine. Cambridge, MA: MIT Press. Xu, Lixin Colin. 2011. “The Effects of Business Environments on Development: A Survey of New Firm-Level Evidence.” World Bank Research Observer 26: 310–40.

297

298 VI Development Strategy, Institutions, and Economic Performance

299

300 VI Development Strategy, Institutions, and Economic Performance *† 1 P ART Introduction Since the Industrial Revolution in the eighteenth century, the world’s rst group includes rich, countries have evolved into two groups. The fi industrialized, developed countries (DCs). The second group includes poor, agrarian, less-developed countries (LDCs). The wealth of devel- oped countries results from their industrial and technological advantages. Since the nineteenth century, political leaders and intellectuals alike have After debated how to modernize LDCs (Gerschenkron 1962; Lal 1985). World War II, many LDC governments adopted various policy measures to industrialize their economies. However, only a small number of econo- mies in East Asia have actually succeeded in raising their level of per 1 capita income to the level in DCs. Adapted from “Development Strategy, Viability, and Economic Convergence,” by Justin Yifu * (2003) 51 (2): 277–308. Lin, originally published in Economic Development and Cultural Change Reprinted with the permission of The University of Chicago. Copyright © 2003 The University of Chicago. All rights reserved. 287

301 288 | New Structural Economics I argue here that the failure of most LDCs to converge with DCs in terms of economic performance can be explained largely by their govern- ments’ inappropriate development strategies. After World War II, most LDC governments pursued development plans that placed priority on the development of certain capital-intensive industries. However, an economy’s optimal industrial structure is endogenously determined by that economy’s endowment structure. Often the fi rms in a government’s priority indus- tries are not viable in an open, competitive market because these industries do not match the comparative advantage of their particular economy. As such, the government introduces a series of distortions in its international trade, fi rms. nancial sector, labor market, and so on, to support nonviable fi It is possible with such distortions to establish capital-intensive industries in developing countries, but the economy becomes very ineffi cient because of misallocation of resources, rampant rent seeking, macroeconomic insta- bility, and so forth. Consequently, convergence—that is, convergence of LDC economic indicators to levels akin to those in DCs—fails to occur. I argue that the government of an LDC should focus development efforts on upgrading the country’s endowment structure instead of on upgrad- ing its industry/technology structure. Once the endowment structure is upgraded, profi t motives and competitive pressures will lead fi rms to upgrade their technologies and industries. The upgrading of the endow- ment structure means faster accumulation of capital—both physical and human—than the growth of labor and natural resources in the economy. Capital accumulation depends on the economic surplus (or, alternatively, the profi ts) and the savings propensity in an economy. If an LDC develops its industries in accordance with its comparative advantages, its economy will have the largest possible economic surplus and the highest savings propensities and will therefore achieve the highest possible upgrade in its endowment structure. Following this strategy, an LDC could achieve faster upgrades in endowment, technology, and industrial structures than the DCs and realize convergence. A fi rm’s choice of industry/technology depends on the relative prices of capital, labor, and natural resources in the economy. Therefore, only if the price structure of the economy can refl ect the relative abundances of capital, labor, and natural resources will fi rms choose their industries and technologies according to comparative

302 Development Strategy, Institutions, and Economic Performance | 289 ect the relative abundance of each advantage. The price structure will refl factor only if the prices are determined in competitive markets. Therefore, the government’s primary function for economic development is to main- tain well-functioning markets. I fi rst present a brief overview of recent theoretical developments and debates on economic growth and convergence. I then discuss the deter- rm’s viability and an economy’s comparative advantages minants of a fi and their relations to the economy’s factor endowments. After analyzing a government’s alternative development strategies, I present the statistical measurement of a development strategy and the econometric estimation of the impact of the development strategy on economic growth. The policy implications of the analyses are set forth in a concluding section. Growth Theories: An Overview When the fi eld of development economics started to take shape in the postwar period, development economists encouraged LDC governments to adopt interventional policies to accelerate capital accumulation and to pursue “inward-looking” strategies oriented toward heavy industry or import substitution aimed directly at closing the industry-technology gap with DCs (Chenery 1961; Warr 1994). These economists were strongly infl uenced by the Soviet Union’s initial success in nation-building, by the pessimism surrounding the export of primary products born during the dence in markets, and by neoclassical Great Depression, by the lack of confi growth theory (Rosenstein-Rodan 1943; Prebisch 1959). Since the 1950s, most LDCs in both socialist and capitalist camps have adopted some varia- tion of these strategies (Krueger 1992). According to seminal work by Robert Solow (1956) and others, neo- classical growth theory, with its assumption of the same given technology to DCs and LDCs, has suggested that LDCs would grow faster than DCs and that the gap in per capita income between DCs and LDCs would narrow because of the diminishing returns to capital in DCs. However, empirical evidence shows that, while convergence occurred within the different states in the United States and among the DCs (Barro and Sala-i-Martin 1992; Baumol 1986), most LDCs failed to narrow the gaps

303 290 | New Structural Economics between their per capita incomes and those of the DCs (Pearson 1969; Romer 1994). Unsatisfi ed with neoclassical growth theory’s inability to explain the continuous growth of DCs and the failure of most LDCs to converge with DCs, Paul Romer (1986) and Robert Lucas (1988) pioneered a new growth theory. Their theory treats technological innovation as endogenously deter- mined by the accumulation of human capital, research and development (R&D), learning by doing, and so on. This new growth theory is insightful for explaining the continuous growth of DCs, which use the most advanced technologies. However, the new growth theory cannot satisfactorily explain the extraordinary growth and convergence during the last three decades of the twentieth century of the newly industrialized economies (NIEs) in Asia, including South Korea, Taiwan, Hong Kong, Singapore, and, recently, During the catch-up China (Pack 1994; Grossman and Helpman 1994). process, these NIEs’ investments in R&D, human capital, and learning by doing were much lower than those of DCs. The LDCs generally use technologies that are inside the technology frontier of the DCs (Caselli and Coleman 2000). Technological innovation in a DC that adopts technology on the new frontier can be obtained only through R&D or other knowledge-generating mechanisms. For an LDC, however, technological innovation can be the result of technology transfer or the imitation of existing technology held by DCs. The costs of techno- logical innovation through R&D are obviously much higher than the costs of imitation or other ways of technological borrowing. Therefore, technol- ogy diffusion from DCs to the LDCs will facilitate the growth of LDCs. It is futile, when attempting to understand convergence, to focus primarily on mechanisms that generate new technology. However, the technological gap between DCs and LDCs is fi lled with a whole spectrum of different technologies. An LDC is faced with the ques- tion of which technology is appropriate to imitate or borrow. The idea of appropriate technology was fi rst introduced in neoclassical trade theory by Anthony Atkinson and Joseph Stiglitz (1969), who formal- ized “localized learning by doing.” E. F. Schumacher (1973) made a similar argument in development economics. The study of appropriate technol- ogy has been revived recently by I. Diwan and D. Rodrik (1991), Susanto Basu and David Weil (1998), and Daron Acemoglu and Fabrizio Zilibotti

304 Development Strategy, Institutions, and Economic Performance | 291 2 However, the models based on the idea of appropriate technology (1999). are inconclusive on the issue of convergence. Basu and Weil (1998) con- sider the relatively low capital stock in an LDC as a barrier to adopting the They conclude that an LDC will experience advanced technology of DCs. a period of rapid growth by raising its savings rate to take advantage of the advanced technology. However, their arguments cannot explain why governmental interventions to improve the savings rate in Latin America, Africa, and Asia—excluding the “Four Little Dragons”—failed to acceler- ate the growth rate. In a cross-country study, Francisco Rodríguez and Dani Rodrik show that causality runs from growth to savings, not vice versa. It would be quite diffi cult for a rise in the savings rate to trigger rapid growth. By contrast, Acemoglu and Zilibotti (1999) stress the disad- vantages of importing technology. In their framework, technology in DCs is used by skilled workers. When the technology is transferred to an LDC, the technology is used by unskilled workers. This mismatch between labor skill and technology can lead to sizable differences in output per capita and total factor productivity (TFP). To Acemoglu and Zilibotti, improving the skill base and human capital of workers, the same argument made by Lucas (1993), is critical to income convergence. The assumption adopted by Acemoglu and Zilibotti is, however, too strong: they assume that LDCs always adopt DCs’ frontier technologies rather than some technologies inside the frontier. The appropriate technology argument does not answer the question of the appropriate role of LDC government in the process of economic growth. Although the linkage of knowledge diffusion with an appropri- ate technology suggests an alternative development path that differs from the development practices followed by many LDCs, it is not clear if the government’s intervention matters to economic growth. Moreover, it is not clear if governments should adopt policies to improve the savings rate and human capital stock of the private sector or if they should subsidize the adoption of high technology industries directly. Viability, Comparative Advantage, and Endowment Structure A country’s per capita income is a function of the prevailing technologies and industries found in the country. If two countries have an identical

305 292 | New Structural Economics technology-industry structure, the two countries should have a similar level of per capita GDP. To understand how the income of an LDC con- verges to that of DCs, we need to understand how an LDC can narrow the ne the mean- technology/industry gap between it and DCs. I will fi rst defi rm’s viability and rm’s viability and the relationship between a fi ing of a fi its industry/technology choice. viability with respect to the expected rate of profi t of I defi ne the term rm in an open, free, and competitive market. If, a normally managed fi without any external subsidies or protections, a normally managed fi rm is expected to earn a socially acceptable profi t in a free, open, and competi- rm is viable. Otherwise, the fi tive market, the fi rm is nonviable. It is obvi- ous that no one will invest in a fi rm if it is not expected to earn a socially t. Such a fi acceptable normal profi rm will exist only if the government gives it support. In a competitive market, the management of a fi rm will affect its prof- itability. This statement is a known proposition. However, the expected tability of a fi rm also depends on its industry/technology choice. profi My discussion begins by presenting a simple economy that possesses two given factor endowments, capital and labor, and produces only one good. Each point on the isoquant shown in fi gure VI.1 represents a technology of production or a combination of capital and labor required to produce a given amount of a certain product. The technology represented by A is more labor-intensive than that of B ; C, C 1, D, and D 1 are isocost lines. The slope of an isocost line represents the relative prices of capital and labor. In an economy where capital is relatively expensive and labor is rela- C C 1, the adoption tively inexpensive, as represented by isocost lines and of technology A to produce the given amount of output will cost the least. When the relative price of labor increases, as represented by the isocost D and D 1, production will cost least if technology B is adopted. lines by In a free, open, and competitive market economy that produces only one product as illustrated in fi rm will be viable only if it adopts gure 1, a fi the least-cost technology in its production. In fi gure VI.1, if the relative prices of capital and labor can be presented by C, the adoption of technol- ogy A costs the least. The adoption of any other technology, such as B, will cost more. Market competition will make fi rms that adopt technologies other than A nonviable. Therefore, in a competitive market with given

306 Development Strategy, Institutions, and Economic Performance | 293 Figure VI.1: Relative Price of Production Factors and Technique Choice capital B A C D DC 1 1 labor relative prices of labor and capital, the viability of a fi rm depends on its technology choice. In a competitive market, the relative prices of capital and labor are determined by the relative abundance or scarcity of capital and labor in the economy’s factor endowments. When labor is relatively abundant and capital is relatively scarce, the isocost line will be similar to that of line in C fi gure VI.1. When capital becomes relatively abundant and labor relatively scarce, the isocost line will change to something like line D. Therefore, the rm in a competitive market depends on whether its choice viability of a fi of technology is on the least-cost lines determined by the relative factor endowments of the economy. This discussion can be extended to an economy with one industry that has many different products and an economy that has many differ- I 1, I 2, and I 3 represent the ent industries. As shown in fi gure VI.2, lines isoquants of three different products that have the same output value in I. industry The average relative capital intensity of the three products is gure VI.2, the viability of a fi 3. As shown in fi I rm is I increasing from 1 to determined by whether or not its product and technology choices are on the least-cost line, which is determined by the relative factor endowments of the economy.

307 294 | New Structural Economics Figure VI.2: Product Choice in an Industry capital I C 3 I 3 A 3 C 2 A 2 C I 1 2 A I 1 1 labor An industry can be represented by an isovalue line, which is the enve- lope of the isoquants of all different kinds of products in the industry. On c product in the isovalue line of an industry, each point represents a specifi the industry that is produced by a specifi c technology and has the same value as any other product in the same line. Figure VI.3 shows an econ- omy that has three different industries, which are represented by the three and K, respectively. These three lines have the industrial isovalue lines I, J, same value. If labor is relatively abundant and the isocost line is indicated I C, by J, and and the economy has a comparative advantage in industries I rm will be viable if it enters industry (or a fi ) and adopts a correspond- J I ing technology to produce product 1). Suppose that the relative J 1 (or abundance of capital increases such that the isocost line changes to line D. The comparative advantage of the economy will change accordingly, and J a fi 2 J 1 to rm will be viable if it upgrades its product-technology from 1. The fi rm that in industry J or it migrates to industry K and produces K I produces 1 in industry I will become nonviable. rm’s viabil- From the above discussion, one sees that the concept of a fi ity and the concept of an economy’s comparative advantage are closely tability, while comparative related. Viability refers to a fi rm’s expected profi advantage refers to the competitiveness of an industry in an open economy.

308 Development Strategy, Institutions, and Economic Performance | 295 Figure VI.3: Industry and Product Choices in an Economy capital K K 1 J C 2 J 1 J I 1 C D 1 I labor Both are endogenously determined by the economy’s relative factor endow- ments. In a closed economy, however, the concept of viability is still relevant while the concept of comparative advantage is not. This discussion leads to the conclusion that, if an LDC wants to close the industry-technology gap, it needs to start by focusing on narrowing the factor endowment gap. Alternative Development Strategies The government is the most important institution in any economy. Its eco- nomic policies shape the macro incentive structure that fi rms in the econ- omy face. With the aim of explaining the success or failure of convergence in an LDC, I analyze government economic policies toward industrial development. I group these policies into different development strategies and then broadly divide the development strategies into two mutually exclusive groups: the comparative-advantage-defying (CAD) strategy, rms to ignore the existing comparative which attempts to encourage fi advantages of the economy in their entry/choice of industry/technology, and the comparative-advantage-following (CAF) strategy, which attempts to facilitate the fi rms’ entry/choice of industry/technology according to 3 No one country has ever the economy’s existing comparative advantages.

309 296 | New Structural Economics followed either strategy consistently or without amendment. However, some countries have followed a strategy close enough to be a model of that strategy. A country that follows a particular strategy may also aban- don it. A switch in strategy provides a good opportunity for careful com- parison of the impact of different strategies. The Characteristics of Development Strategies The CAD Strategy. Most LDCs are characterized by relatively abundant labor and scarce capital. Therefore, in a free, open, and competitive mar- ket, fi rms in LDCs enter relatively labor-intensive industries and adopt 4 However, relatively labor-intensive technologies in their production. political leaders and intellectuals in LDCs often equate industrialization, especially heavy industrialization, with modernization and push their countries to develop capital-intensive heavy industries and adopt the most advanced technologies in their production as quickly as possible. They want the economy to develop some industry like and produce product K 1 when the isocost line determined by their endowment structure is C K 5 gure VI.3). With the given endowment structure, a fi rm producing (see fi product K 1 will not be viable in a free, open, and competitive market. If a free, open, and competitive market is maintained, a fi rm following its government’s strategy will incur a loss equivalent to the distance between rm. Because and C isocost lines C 1. I call this loss a policy burden on the fi the government is responsible for the fi rm’s entry-adoption of the industry/ technology, the government is accountable for the fi rm’s loss. Therefore, for implementing the CAD strategy, the government must give the fi rm a policy subsidy to compensate for losses incurred (Lin and Tan 1999; Lin, Cai, and Li 1998, 2001). How large the subsidy needs to be to compensate for the policy burden in the real world depends on how distant the promoted industry-technology is from the economy’s comparative advantages. If the distance is small, the government can rely on tax incentives or direct fi scal transfer to subsidize the fi rm. However, this distance is often very large when the government in an LDC pursues a CAD strategy and special institutional arrangements are required for achieving the strategy’s goal. When an LDC government pursues a CAD strategy, the most frequently used method of subsidy is to suppress interest rates by regulation in order

310 Development Strategy, Institutions, and Economic Performance | 297 to reduce the project’s capital costs. In addition, the equipment for the CAD project, in general, cannot be produced domestically in an LDC and needs to be imported from DCs. Therefore, access to foreign exchange is also required for the CAD project. However, foreign exchange in an LDC is generally scarce and expensive because the LDC’s exports are limited and consist mainly of low-value agricultural products and resources. To lower the costs of equipment imports for the CAD project, governments also tend 6 to overvalue domestic currency and undervalue foreign exchanges. On the one hand, the distortions in the interest rate and the foreign exchange rates will stimulate fi rms in both the priority and nonpriority sectors to demand more capital and foreign exchange. On the other hand, distortions will suppress the incentives to save and export and, thus, reduce the availability of capital and foreign exchange in the economy. Therefore, there will be shortages in capital and foreign exchange, and the government will need to use administrative measures to ration capital and foreign reserves in order to guarantee that the CAD fi rms will have the resources to perform strategic tasks. The resource allocation function of markets is thus constrained, or even replaced by, direct government 7 rationing. Theoretically, the government that adopts a CAD strategy is respon- sible only for giving a subsidy to compensate for the loss arising from the policy burden. Given information asymmetry, however, the government cannot distinguish losses induced by the policy burden from operational rms will use the policy burden as an excuse and use resources losses. The fi to lobby the government for ex ante policy favors, such as access to low- interest loans, tax reductions, tariff protection, legal monopolies, and so on, to compensate for policy burdens. In addition to policy favors, if the rms still incur losses, they will also request that the government offer fi some ex post, ad hoc administrative assistance, such as more preferential loans. The economy will be full of rent-seeking or directly unproductive 8 Because the fi rms can use the policy burdens as an t-seeking activities. profi excuse to bargain for more government support and because it is hard for the government to shun such responsibility, the fi rm’s budget constraints 9 become soft. When a soft budget constraint exists, the manager of the fi rm will have no pressure to improve productivity and will have more on-the- job consumption and other moral hazards. The subsidies could actually

311 298 | New Structural Economics end up much higher than those required to compensate for the original policy burdens. The CAF Strategy. The government in an LDC could adopt the alternative rms to enter the industries for which the coun- CAF strategy to encourage fi try has comparative advantages and to adopt the technology in production rms viable. As discussed above, the industries for that will make these fi which the economy has comparative advantages and the technologies that are appropriate for production are all determined by the country’s relative factor endowments. However, the managers of fi rms, as micro agents, have no knowledge or concern of the actual endowments. Their only concerns are the prices of their outputs and the costs of their production. They will enter the industry and choose the technology of production appropriately only if the relative factor prices correctly refl ect the relative factor abun- dances, which can be achieved only if the markets are competitive. There- fore, when the government in an LDC adopts a CAF strategy, its primary policy is to remove all possible obstacles to the functioning of free, open, and competitive product and factor markets. The above discussions assume that the information about the product markets, industries, and production technologies is freely available to the fi rms in the economy. Therefore, when the factor endowment struc- rms can upgrade their product- ture of the economy is upgraded, the fi technologies or smoothly upgrade from a less capital-intensive industry to a relatively more capital-intensive industry. Such information may not be available, however, so it is necessary to invest resources to search for, collect, and analyze industry, product, and technology information. If a fi rm carries out the activities on its own, it will keep the information private, and other fi rms will be required to make the same investment to obtain the information. There will be repetition in the information investments. The information has a public goods aspect, however. After the information has been gathered and processed, the cost of informa- tion dissemination is close to zero. Therefore, the government can col- lect the information about the new industries, markets, and technology and make it available in the form of an industrial policy to all fi rms. The upgrading of technology and industry in an economy often requires the coordination of different fi rms and sectors in the economy. For example,

312 Development Strategy, Institutions, and Economic Performance | 299 the human capital or skill requirements of new industries and technologies may be different from those used with older industries and technologies. A fi rm may not be able to internalize the supply of the new requirements and will need to rely on outside sources. Therefore, the success of a fi rm’s industry-technology upgrade also depends on the existence of an outside supply of new human capital. In addition to human capital, the fi rms that are upgrading may also require new fi nancial institutions, trading arrangements, marketing, distribution facilities, and so on. Therefore, the rms in government may also use industrial policy to coordinate among fi different industries and sectors for the upgrade of industry and technology in the economy. The upgrading of industry and technology is an innovation, and it is risky by nature. Even with the information and coordination provided by the government’s industry policy, a fi rm’s attempt to upgrade may fail because the upgrade is too ambitious, the new market is too small, the coordination is simply inadequate, and so forth. The failure will indicate to other fi rms that the targets of the industrial policy are not appropriate and, therefore, they can avoid that failure by not following the policy. That is, the fi rst fi rm pays the cost of failure and produces valuable informa- tion for other fi rms. If the fi rm succeeds, the success will also provide rst fi rms, prompting them to engage in similar upgrades. externalities to other fi These subsequent upgrades will also dissipate the possible rents that the rst fi rm may enjoy, so there is an asymmetry between the costs of failure fi and the gains of success that the fi rst fi rm may have. To compensate for the externality and the asymmetry between the possible costs and gains, the government may provide some forms of subsidy, such as tax incen- rms that initially follow the government’s tives or loan guarantees, to the fi industrial policy. It is worth noting that there is a fundamental difference between the industrial policy of the CAF strategy and that of the CAD strategy. The promoted industry/technology in the CAF strategy is consistent with the comparative advantage determined by changes in the economy’s factor endowments, whereas the priority industry/technology that the CAD strategy attempts to promote is not consistent with comparative advantage. Therefore, the fi rms in the CAF strategy should be viable, and a small, limited-time subsidy should be enough to compensate for

313 300 | New Structural Economics the information externality. By contrast, fi rms following a CAD strategy are not viable, and their survival depends on large, continuous policy 10 favors and support from the government. A comparison of the successes and failures of industrial policies on auto- mobile production in Japan, South Korea, India, and China is a good illus- tration of the differences between the CAF and CAD industrial policies. The automobile industry is a typical capital-intensive heavy industry, and its development is the dream of every LDC. Japan adopted an industrial policy to promote its automobile industry in the mid-1960s and achieved great success. Japan’s experience is often cited as a supporting argument by advocates of an industrial policy for heavy industries in developing coun- tries. South Korea instituted an industrial policy for automobile produc- tion in the mid-1970s and has also achieved a limited degree of success in automobile production. The automobile industries in China and India were started in the 1950s, and the industry in both countries has required continuous protection from the government since that time. How does one explain why a similar industrial policy can yield success in one instance and failure in another? This becomes clear once one compares the per capita income of these countries with the per capita income of the United States at the time when they initiated their policies (table VI.1). Per capita income is a good proxy for the relative abundances of capital and labor in an economy. Capital is abundant and wage rates are high in a high-income country. In a low-income country, the opposite holds true. Table VI.1 indicates that, when Japan initiated its automobile pro- duction policy in the mid-1960s, its per capita income was more than 40 percent that of the United States. The automobile industry was not the most advanced, capital-intensive industry at that time, nor was Japan a Table VI.1: Level of Per Capita Income (in 1990 Geary-Khamis $) Year United States South Korea India China Japan 1955 2,695 1,197 665 818 10,970 1965 14,017 5,771 1,578 785 945 1975 10,973 3,475 900 1,250 16,060 Source: Maddison 1995. Note: The Geary-Khamis dollar is a multilateral purchasing power parity measurement of income. The use of the Geary-Khamis technique to convert income in different countries ensures cross-country comparisons, base country invariance, and additivity of the measurements.

314 Development Strategy, Institutions, and Economic Performance | 301 capital-scarce economy. The Ministry of International Trade and Industry (MITI) gave support only to Nissan and Toyota. However, more than 10 fi rms—ignoring MITI’s prompting not to enter the industry—also started automobile production and were successful, even though they did not receive any support from MITI. The above evidence indicates that the Japanese automobile fi rms were viable and that the MITI’s promotion of the automobile industry in the 1960s was a CAF strategy. When Korea initiated its automobile industry development policy in the 1970s, its per capita income was only about 20 percent that of the United States and about 30 percent that of Japan. This may explain why the Korean govern- ment needed to give its automobile fi rms much greater and longer support than the Japanese government did its fi rms. Despite the support, two of the three automobile fi rms in Korea recently fell into bankruptcy. When China and India initiated their automobile industry development policies in the 1950s, their per capita incomes were less than 10 percent that of the United States. The automobile fi rms in China and in India were not viable at all. It is still the case today that their survival depends on heavy 11 government protection. Human Capital and Economic Development The previous discussion focused on the accumulation of physical capital and its determining effect on an economy’s industry-technology upgrad- ing. The role of human capital in the process of development has received much attention in the development literature in recent years. Recent empir- ical work that attempts to explain cross-country income differences has included human capital as an explanatory variable in the production func- tion and has found that human capital has a positive effect on economic growth (Mankiw, Romer, and Weil 1992; Caselli, Esquivel, and Lefort 1996; Klenow and Rodriguez 1997; Barro 1997). What is the role of human capital accumulation in the development strategy of an LDC? If an LDC adopts a CAF strategy, the upgrading of its factor endowments will occur rapidly, and, consequently, the upgrading of its industry/technology will also be very rapid. Upgrading is an innovation by nature, even though the process is an imitation of an existing industry or technology from more advanced countries. The managers/workers will face and will need to handle uncertainty in skills, production, marketing,

315 302 | New Structural Economics and so on in the upgrading process. They will also need to make many t them to local conditions. adaptations in the borrowed technologies to fi Increasing the manager’s or worker’s human capital will increase his or her ability to handle these kinds of uncertainties and to carry out neces- When a developing country narrows its sary adaptations (Schultz 1975). industry/technology gap with DCs, it will move from mature industries and technologies closer to newer, less mature, and more uncertain indus- tries and technologies. The requirement for human capital increases with economic development because human capital becomes increasingly com- plementary to physical capital in the new, frontier industries and technolo- 12 Given the complementary relationship between physical capital and gies. human capital, it is necessary to accumulate human capital along with the accumulation of physical capital in the convergence process. Human capi- tal is not a substitute for physical capital, however. An overaccumulation of human capital will lead to waste. After World War II, many scientists and engineers migrated to the United States from India and Latin America and other developing countries, but they made little direct contribution to the economic growth of their mother countries. These scientists and engi- neers are not to be blamed, however, because the low factor endowment structures in their mother countries made it impossible for many of them to fi nd suitable positions that would utilize their human capital at home. Comparison of CAF Strategy and CAD Strategy The attempt to catch up with DCs is justifi able for any LDC. The CAD strategy is appealing to political leaders and the general public in LDCs, including elite intellectuals, because most people directly observe the dif- ferences between industry-technology structures in DCs and those of their own countries and notice the correlation between industry-technology and per capita income. However, a CAF strategy enables an LDC to catch up with DCs, while a CAD strategy in effect stifl es an LDC’s opportunity to catch up. Many other theories have also attempted to explain an LDC’s success or failure in achieving sustained economic development. The CAF/ CAD strategy framework provides a unifi ed explanation. Capital Accumulation. An economy’s optimal industry-technology struc- ture is endogenously determined by its endowment structure. Therefore, if an LDC wants to attain the industry/technology structure of a DC, it fi rst

316 Development Strategy, Institutions, and Economic Performance | 303 needs to narrow the gap between their respective factor endowment struc- tures. The upgrading of the factor endowment structure means an increase in capital relative to labor. Capital accumulation depends on the size of rms and the rate of savings of economic agents surplus/profi ts accrued by fi rm in an economy in the economy. When, following the CAF strategy, a fi enters an industry in which that economy has a comparative advantage and adopts the least-cost technology in its production, the fi rm will be competitive, occupy the largest market share, and have the largest surplus/ ts. Meanwhile, the capital in the economy employed in the indus- profi tries following comparative advantage will have the highest possible rate of return. Therefore, economic agents’ incentives to save will be highest. Moreover, the government will not distort the prices of factors and prod- ucts, nor will the government use administrative powers to create legal monopolies. Therefore, there will be no scope for wasteful rent-seeking activities. The fi rm will have a hard budget constraint and will need to earn profi ts by improving management and competitiveness. The CAD strategy will result in just the opposite of what the CAF strategy promises regarding competitiveness, rates of return, rent-seeking activities, and the softness of budget constraints in fi rms in the priority industries. Therefore, upgrading the endowment structure will be faster under the CAF strategy than under the CAD strategy. Upgrading the endowment structure in an economy Technology Transfer. will provide the basis for upgrading the industry-technology structure (Basu and Weil 1998). The targeted industry-technology will be new to the fi rms in an LDC and will need to be transferred from DCs. The learning costs will be smaller under the CAF strategy than under the CAD strat- egy because the distance between the new industry/technology and the old industry/technology is smaller under the former strategy than under the Moreover, the patent pro- latter strategy (Barro and Sala-i-Martin 1992). tections for many of the targeted technologies under the CAF strategy may have already expired. Even if a technology is still under patent protection, the license fee will be lower with the CAF strategy than with the CAD strategy because the targeted technology for the CAF strategy is older than the CAD strategy ceteris paribus. In some cases, the fi rm under the CAD strategy will not be able to obtain the technology from DCs and will need to “reinvent the wheel” and invest in costly and risky R&D of technology

317 304 | New Structural Economics by themselves. Therefore, the acquisition costs of the technology will be lower under the CAF strategy than under the CAD strategy. Openness in International Trade. A number of empirical studies show that more open countries exhibit stronger convergence tendencies than do closed countries (Harberger 1985; Dollar 1992; Warr 1994; Ben-David 1993; Sachs and Warner 1995; Harrison 1996; Michaely 1977; Frankel and Romer 1999). International trade is expected to facilitate technol- nds that countries ogy diffusion among countries. Jong-Wha Lee (1995) fi importing more capital goods tend to grow faster, which means that new However, Rodríguez technologies may be embodied in the capital goods. and Rodrik (1999) argue that “methodological problems with the empiri- cal strategies employed in this literature leave the results open to diverse The role of trade policies is unclear. If the importation interpretations.” of equipment facilitates technology transfer, should the government adopt measures to promote it, or is it best to pursue trade liberalization in the sense of lower tariffs and nontariff barriers to trade? In this framework, a country adopting a CAF strategy will rely on importing products for which it does not have a comparative advantage and exporting products for which it has comparative advantage. For this country, openness is endogenously determined by the country’s factor endowment structure instead of by an exogenously determined policy for imports and exports. If the government in an LDC adopts the CAD strat- egy and attempts to substitute the importation of capital-intensive manu- factured goods by domestic production, not only will the country’s import trade be reduced but also its export trade will be suppressed. The latter consequence results from the transfer of resources away from the indus- tries for which the economy has a comparative advantage. Also, exchange rates may be overvalued to facilitate the development of priority industries, effectively hampering export opportunities. Socialist economies, India, and many Latin American countries exemplify this case. The growth perfor- mance of these countries is miserable compared with economies that have followed the CAF strategy more closely. The government in an LDC may adopt the CAD strategy and, at the same time, encourage its fi rms in the priority capital-intensive industries to export. In this case, exports will be unprofi table even though the fi rms may have a high ratio of exports to 13 foreign markets and may achieve fast technology improvements. The

318 Development Strategy, Institutions, and Economic Performance | 305 fi rms’ survival relies on the protection of domestic markets, preferential loans from banks, and other policy support. The country will have poor external accounts, accumulate foreign debt, and be easily affected by exter- nal shocks. It may be better for an LDC to adopt a CAD strategy that encourages exports rather than a CAD strategy that encourages import substitution. However, the overall economic performance of an economy that adopts the export-promotion strategy will be poorer than that of an 14 economy that adopts the CAF strategy. Therefore, it is not true that more 15 exports in a country necessarily lead to higher GDP growth. Since the pioneering works by E. S. Shaw (1969) and Financial Deepening. R. Mckinnon (1973), many researchers have argued that there exists cau- nancial deepening and economic growth in an economy. sality between fi nancial deepening is either the ratio The indicator often used to measure fi of the money supply (M2) to GDP, or the total amount of credits from fi nancial intermediaries to private sector divided by GDP. The empirical fi ndings have supported the above hypothesis (Levine 1997; Rajan and Zingales 1998). However, the degree of fi nancial deepening in an LDC is, to a large extent, endogenous to a government’s development strategy. Under the CAD strategy, the carriers of a government’s development strategy are the large-sized fi rms. To support the fi nancial needs of nonviable large-sized rms, the government often nationalizes the fi scal fi rms and uses direct fi nancial intermediation—to support these fi rms. appropriation—skipping fi Such was the case in the former socialist planned economies and continues to be the case in India and many other LDCs. Even if the government relies on private fi rms to carry out the CAD strategy, the CAD needs of large- sized fi rms will be signifi cant and can only be met by a heavily regulated oli- gopolistic banking system; consequently, interest rates will be suppressed. In either case, the fi nancial system in the country will be underdeveloped. However, the most competitive and dynamic fi rms in LDCs are the labor- intensive small- and medium-size fi rms, which are discriminated against and often denied access to fi nancial services by large banks. The fi nancial system is, thus, very ineffi cient. Moreover, the priority-sector fi rms that receive preferential access to bank loans are not viable and may not be able to repay loans. The banks often accumulate large amounts of bad debts from the large fi rms in the priority sectors, thus contributing to or even

319 306 | New Structural Economics triggering an economic crisis. A precondition for fi nancial deepening in an LDC is, therefore, a change in the government’s development orientation from a CAD strategy to a CAF strategy. The bulk of empirical studies show that vol- Macroeconomic Stability. atility in the macroeconomy could hamper long-run growth (Barro and If the government in an LDC adopts the CAD Sala-i-Martin 1997). rms in priority industries will not be viable and will rely on pref- strategy, fi erential loans, trade barriers, and other policy support for their survival. Because existing comparative advantages are not utilized, the economy as a whole will not be competitive, no dynamic changes in the economy’s com- parative advantage can be sustained, and the economic performance of the economy will be poor. The economy will have a weak fi nancial sector and cits, debt burdens, and fi nancial fragility poor external accounts. Fiscal defi will accumulate, and macroeconomic stability will become unsustainable. A country that follows the CAF strategy will have better external accounts, will have healthier fi nancial and fi scal systems, will be better equipped to resist external shocks, and will have a much better record of macroeco- 16 nomic stability. Income Distribution. The relationship between income distribution and economic development is one of the oldest subjects in development eco- nomics. Simon Kuznets (1955) proposed an inverted-U hypothesis, sug- gesting that inequality tends to widen during the initial stages of economic There is development with a reversal of this tendency in the later stages. mixed evidence for this hypothesis. A number of cross-sectional studies support this hypothesis (Paukert 1973; Cline 1975, Chenery and Syrquin 1975; Ahluwalia 1976). However, the study of 43 episodes in 19 countries by Gary Fields (1991) fi nds that there is no tendency for poorer countries to yield increased rather than decreased inequality or for richer countries to yield decreased rather than increased income inequality, while a case study by John Fei, Gustav Ranis, and Shirley W. Y. Kuo (1979) shows that I propose that the the Taiwanese economy achieved growth with equity. adoption of the CAF strategy in an LDC will alleviate income inequality, whereas the adoption of the CAD strategy will aggravate income inequal- ity. The most important asset that the poor have in an LDC is their own labor. The CAF strategy will result in sustained economic growth through

320 Development Strategy, Institutions, and Economic Performance | 307 the development of more-labor-intensive industries, creating more job opportunities for the poor, increasing wage rates, and allowing the poor to have a share in the benefi ts of growth. In contrast, the CAD strategy, by facilitating the development of more-capital-intensive industries, will reduce job opportunities for the poor and suppress wage rates of the work- ing poor. Growth will not be sustainable, and when the economy breaks down, the poor will suffer the worst hardship, as evidenced by the recent East Asian fi nancial crisis (Stiglitz 1998). The Choices of Development Strategy When development economics started to take shape in the mid-twentieth century, the dominant view among development economists was to advise LDC governments to ignore their own comparative advantages and to adopt an inward-looking variation of the CAD strategy, such as the heavy-indus- try-oriented strategy or the import-substitution strategy. Proponents of the CAD strategy have often confused the causality of the dynamic change of comparative advantage. They have urged LDCs to disregard the constraint of relative capital scarcity in its factor endowments and to establish directly the same capital-intensive industries as those of DCs. They worked with the understanding that economic development can be accelerated if LDCs bypass development of labor- or resource-intensive industries. I argue that the alignment of industry/technology with an economy’s comparative advantage is key to facilitating the international diffusion of appropriate technology, to accelerating the rate of economic growth, and to realizing convergence. The dynamic change in an economy’s compara- tive advantage depends on the dynamic change in the economy’s endow- ment structure, which itself depends on the rapidity of capital accumula- tion in the economy. Capital accumulation, in turn, depends on how well economic agents in the economy exploit existing comparative advantages in their choices of industry/technology. An LDC that responds to com- parative advantages present in its own factor endowments as the guiding principle in its choice of industry/technology will minimize imitation costs, experience faster shifts in its endowment structure, and sustain a continu- ous upgrade in its industrial structures. The development experience of the East Asian Four Little Dragons is a good illustration of the merits of the CAF strategy.

321 308 | New Structural Economics Like many other developing economies, Taiwan, South Korea, Hong Kong, and Singapore were very poor after World War II. In the early 1950s, their levels of industrialization were low, their capital and foreign exchange reserves were extremely limited, and their per capita incomes were very low. Like any other developing economy, they also faced the problem of choosing an appropriate development path. Taiwan, South Korea, and Sin- gapore initially adopted an import-substitution CAD strategy but gave up the attempt to develop heavy industries while still in their initial stages. Instead, based on their factor endowments, they energetically developed labor-intensive industries, promoted exports, and expanded their outward- oriented economies to use their comparative advantages to the full extent. At the time, many European countries, the United States, and Japan were gradually replacing labor-intensive industries with technology- and capital-intensive ones because of an increasing abundance of capital and increases in wage rates. The Four Little Dragons had abundant, inexpen- sive labor. Therefore, when developed countries’ comparative advantages changed to more capital- and technology-intensive industries, the four East Asian countries were able to capitalize on the dynamic opportunities. Through trade linkages and the openness of their economies, labor-inten- sive industries in developed countries were relocated to these Asian econo- mies. Because of the intensive use of their comparative advantages, the Four Little Dragons were very competitive and were thus able to achieve rapid capital accumulation. Along with the accumulation of capital and the change in comparative advantages, they gradually upgraded to more capital- and technology-intensive industries. As a result these four coun- tries were able to sustain more than 30 years of rapid growth, fi rst becom- ing newly industrialized economies and then reaching or nearly reaching the level of developed economies. Their extraordinary achievements in eco- nomic development have attracted worldwide attention. Most developing economies adopted the CAD strategy in the 1950s and maintained that strategy for quite a long time. Why has Hong Kong never tried the CAD strategy, and why did Taiwan, South Korea, and Sin- gapore switch to the CAF strategy shortly after trying the CAD strategy? Are these little dragons just lucky, or is their choice of the CAF strat- egy attributable to the wisdom of their political leaders? Gustav Ranis and Mahmood Syed (1992) attribute the success to their poor natural

322 Development Strategy, Institutions, and Economic Performance | 309 In addition, I propose it is also attributable to their small popu- resources. cient and costly. The length of time lations. The CAD strategy is very ineffi this strategy can be maintained in an LDC depends on how many resources the government can mobilize to support it. The larger the per capita natu- ral resources or the larger the population in an economy, ceteris paribus, the more resources government can mobilize to maintain the ineffi cient CAD strategy. For an economy with poor natural resources and a small population, the adoption of the CAD strategy will precipitate immediate economic crisis. By that time, the government will have no choice but to carry out reforms In effect, infl u- and change its strategy (Edwards 1995). enced by the prevailing economic thoughts in the 1950s and motivated by the dream of nation building, many political leaders and intellectuals in Taiwan and Korea never gave up their desires to accelerate development of capital-intensive heavy industries. However, their per capita natural resources were extremely poor, and their populations were very small. The implementation of the CAD strategy in the early 1950s in Taiwan led to an immediate and enormous fi scal defi cit coupled with high infl ation, and the government was forced to give up the strategy (Tsiang 1984). When South Korea decided to push the heavy machinery and heavy chem- ical industries in the 1970s, similar results occurred and the push was postponed (Stern et al. 1995). Singapore and Hong Kong were both too small in population and too poor in natural resources to implement the CAD strategy. The Transition Strategy If the government adopts a CAD strategy, the development of labor- intensive sectors—in which developing countries have comparative advantage—is repressed. The growth performance during the transition from a socialist economy to a market economy depends, therefore, on the country’s ability to create an enabling environment for the devel- opment of labor-intensive sectors and at the same time fi nd a way to solve the viability issue for fi rms inherited from the previous develop- ment strategy, paving the way for eliminating previous distortions and interventions. However, in many countries that adopted a CAD strategy, many nonviable enterprises exist that were unable to survive in an open and competitive market. If government distortions and interventions are

323 310 | New Structural Economics eliminated abruptly, these nonviable enterprises will become bankrupt. At the same time, the originally suppressed labor-intensive industries will thrive, and newly created employment opportunities in these industries can surpass the losses from the bankruptcy of nonviable fi rms. As a result, the economy can grow dynamically soon after implementing the shock therapy, with at most a small loss of output and employment initially. On the other hand, if the number of nonviable fi rms is too large, the rms will make up too large a share output value and employment of those fi in the national economy and shock therapy may be inapplicable. Its appli- cation will result in economic chaos due to large-scale bankruptcies and dramatic increases in unemployment. To avoid such dramatic increases in unemployment or to sustain these “advanced” nonviable enterprises, some governments, for example in Eastern Europe, continued their protection and subsidies for these fi rms—either explicitly or implicitly—and in the nd itself in an awkward situation of shock without end, the economy can fi 17 therapy (Kolodko 2000). The Chinese government opted for a dual-track approach, which is arguably better than shock therapy (McKinnon 1993). Instead of follow- ing the “macro-institution-fi rst” approach proposed by the Washington Consensus, the Chinese government employed a ‘micro-fi rst’ approach to improve incentives for farmers and state-owned enterprise workers. It adopted the individual household-based farming system to replace the 18 introduced profi t-retention and managerial collective farming system, 19 autonomy to state-owned enterprises, making farmers and workers par- tial residual claimants. This reform greatly improved the incentives and productivity in agriculture and industry (Groves et al. 1994; Jefferson, Rawski, and Zheng 1992; Jefferson and Rawski 1995; Lin 1992; Li 1997; Weitzman and Xu 1995). Then the government allowed collective town- 20 private enterprises, joint ventures, ship-and-village enterprises (TVEs), and state-owned enterprises to use the resources under their control to invest in labor-intensive industries that had been suppressed in the past. Meanwhile, the government required farmers and state-owned enterprises to fulfi ll their obligations to deliver certain quotas of products to the state at preset prices. The former reform improved the effi ciency of resource allocation, and the latter ensured the government’s ability to continue sub- sidizing the nonviable fi rms. Therefore, economic stability and dynamic growth were achieved simultaneously.

324 Development Strategy, Institutions, and Economic Performance | 311 Notes This is the inaugural D. Gale Johnson Lecture, presented at the University of † Chicago on May 14, 2001. I am grateful for the helpful comments by Gary Becker, Kang Chen, James Heckman, Ralph Huenemann, Keijiro Otsuka, George Rosen, Jan Svejnar, Yingyi Qian, Kislev Yoav, Hao Zhou, and partici- pants at the lecture. I am indebted to Mingxing Liu, Qi Zhang, and Peilin Liu, who provided invaluable help in reviewing the literature, assembling the data sets, and running the regressions. 1. Starting from very low levels, Japan’s per capita income, measured in cur- rent U.S. dollars, exceeded that of the Unites States in 1988, and Singapore’s per capita income exceeded that of the United States in 1996. Taiwan, South cantly narrowed the income gap between Korea, and Hong Kong have all signifi themselves and the DCs. 2. Other economists also hold similar ideas about appropriate technology. Drawn from the lessons of the East Asian Miracle, some economists, such as ying geese K. Akamatsu (1962) and Ito Takatoshi (1998), suggested the “fl pattern” metaphor to describe the characteristics of industrial structure and technological diffusion during different development stages. But distinct pol- icy proposals cannot be obtained from this metaphor. 3. Other ways of classifi cation exist. For example, Griffi n (1999) classifi es devel- opment strategies into six alternatives: monetarism, open economy, industrial- ization, green revolution, redistribution, and socialist strategy. For simplicity, I neglect the endowment of natural resources in the discussion. 4. The propositions derived from the discussion remain valid if natural resources are also considered. 5. Heavy industry was the most advanced sector in the past. Nowadays, the pri- ority of the CAD strategy in an LDC is focused on information technology and other high-tech industries, which are the most capital-intensive industries now. 6. The distortions in the interest rate and the foreign exchange rate are univer- sal for LDCs that pursue a CAD strategy. Socialist countries and other LDCs that adopted a development strategy oriented toward heavy industry often distorted the prices of raw materials and living necessities along with wages. The government that adopts a CAD strategy can also ration capital to the fi rms 7. that are not in the priority industries. This is, in fact, the practice in the socialist planned economy. Certainly, the fi rms in the nonpriority industries will receive less capital than if the government does not adopt this strategy. Alternatively, the government can allow the market to allocate capital after the fi rms in the priority industries have been guaranteed rations. The interest rate will conse- quently be higher than it is when all capital is allocated by the market. On the contrary, the wage rate in the market will be lower because of the low labor

325 312 | New Structural Economics absorption of the fi rms in the rms in the priority industries. Therefore, the fi nonpriority industries will adopt a more labor-intensive technology in their production than if there is no government intervention. The above analyses are also applicable to the allocation of foreign exchange to fi rms in nonpriority industries. The distortions of interest rates and foreign exchange rates and the use of administrative allocations will lead to rent-seeking or directly unproduc- tive profi t-seeking activities. The loss from rent-seeking or directly unproductive profi t-seeking activities is 8. estimated to be much larger than the loss from misallocation. Kornai (1986) was the fi rst economist to analyze the phenomenon of the soft 9. budget constraint. He attributed the existence of the soft budget constraint in a state-owned enterprise in a socialist country to the patriarchic nature of the socialist government. However, I argue that soft budget constraints arise from the government’s accountability for the nonviability of an enterprise caused by the government’s development strategy. My hypothesis can explain why the phenomenon of the soft budget constraint also exists in fi rms of nonsocialist countries, such as South Korea’s and why the soft budget constraint chaebols, phenomenon continues to exist in the Eastern European and former Soviet Union countries after their state-owned fi rms were privatized and the socialist system was abandoned. See also Lin and Tan (1999). 10. The dynamic comparative advantage is an often-used argument for the gov- rms (Redding 1999). In our ernment’s industrial policy and support to the fi framework, however, it is clearly evident that the argument is valid only if the government’s support is limited to overcoming information and coordination rm’s externalities to other fi rms. The industry should costs and the pioneering fi rms in the be consistent with the comparative advantage of the economy, and fi new industry should be viable, otherwise they will collapse once the govern- ment’s supports are removed. 11. Most big-push attempts by the LDCs in the 1950s and 1960s failed. However, there has been a renewed interest in the idea after the infl uential articles by Murphy, Shleifer, and Vishny (1989a, 1989b). Their papers show that a gov- ernment’s coordination and support are required for setting up a key indus- try and that the demand spillovers from the key industry to other industries will enhance economic growth. For the “big push” strategy to be successful, however, the industry being promoted must be consistent with the compara- tive advantage, which is determined by the relative factor endowment of the economy, and the fi rms in the promoted industry must be viable after the push. Deviation from comparative advantage in the promoted industries and the consequent lack of viability of the chosen fi rms are the reasons why so many big-push attempts by the LDCs in the 1950s and 1960s failed. 12. In recent years, a variety of papers have argued that different technologies may display different degrees of skilled-labor or unskilled-labor bias. This idea

326 Development Strategy, Institutions, and Economic Performance | 313 of skill complementarity has been employed to explain the increase in wage inequality in the 1980s and 1990s in the United States. 13. I met a senior manager of Hyundai Automobile Company in the United States in the early 1990s. He told me that Hyundai was still losing money after 10 years of successful exportation of cars to the U.S. market. 14. Taiwan and South Korea are good examples for comparison. Taiwan has fol- lowed the CAF strategy consistently, whereas Korea has often attempted to switch from the CAF strategy to the CAD strategy. The GDP growth rate, income distribution, macro stability, and other development indicators in Tai- wan are better than those of South Korea. 15. In the development literature, export promotion and import substitution are often used as a classifi cation of development strategy. There are some similari- ties between this classifi cation. Any country’s cation and the CAF/CAD classifi level of export will be higher under a CAF strategy than that under a CAD strategy. However, the level of trade in any economy is endogenously deter- mined by the economy’s endowment structure. It is inappropriate to use an endogenous variable as a policy target or instrument. 16. In the recent East Asian fi nancial crisis, Taiwan, Hong Kong, Singapore, and Malaysia were affected slightly whereas Korea, Indonesia, and Thailand were hard hit. One reason for the different performances among these two groups of economies is the difference in their development strategies. The fi rst group followed the CAF strategy closely, whereas the latter group leaned toward the CAD strategy; see Lin (2000). The difference in the shares of nonviable fi 17. rms in the economy might explain why the shock therapy recommended by Sachs succeeded in Bolivia but not in the economies of Eastern Europe and the former Soviet Union. Bolivia is a poor, small economy; therefore, the resources that the government could mobilize to subsidize the nonviable fi rms were small, and the share of nonvi- able fi rms in the economy was also relatively small. Stiglitz (1998) questioned the universal applicability of the Washington Consensus. He pointed out that it advocated use of a small set of instruments—including macroeconomic stability, liberalized trade and privatization—to achieve a relatively narrow goal of economic growth. He encouraged governments to use a broader set of nancial regulations and competition policy—to achieve instruments—such as fi a broader set of goals, including sustainable development, equity of income distribution and so on. Stiglitz’s arguments are based on information asym- metry and the need for government to overcome market failures. However, he did not discuss how to deal with the issue of nonviable fi rms in developing and transitional economies and the implications of nonviability for choices of transition path and policies. 18. When reform started at the end of 1978, the government originally proposed to raise the agricultural procurement prices, to liberalize rural market fairs,

327 314 | New Structural Economics and to reduce the size of production teams of 20–30 households to voluntarily- formed production groups of 3–5 households, but explicitly prohibited the replacement of the production team system with an individual household-based farming system. However, a production team in a poor village in Fengyang County, Anhui Province, secretly leased the collective-owned land to individ- ual households in the team in the fall of 1978 and harvested a bumper increase in outputs in 1979. Seeing the effects of an individual-household-based farm- ing system, the government changed its policy and endorsed this approach as a new direction of reform (Lin 1992). Initially, the collectively-owned land was leased to farm households for one to three years, extended to 15 years in 1985, and further extended to 30 years in 1994. The farm household was obliged to deliver certain amounts of agricultural produce at the government-set prices to fulfi ll its quota obligation until the late 1990s. 19. t-retention system The state-owned enterprise reform proceeded from the profi in 1979, the contract-responsibility system in 1986, and the modern corpora- tion system from the 1990s to now. Each system was experimented in a small rst before that system was extended nationwide (Lin, group of enterprises fi Cai, and Li 1994). 20. The TVE was another institutional innovation by the peasants in China dur- ing the transition process. After the Household Responsibility System (HRS) reform, farmers obtained a substantial amount of residuals and saw profi table investment opportunities in consumer-products sector. However, due to the ideological reason at that time, the form of private enterprise was prohibited and the farmers used the collective TVE as an alternative to tap into the profi t- able opportunity. The government initially put many restrictions on the opera- tion of TVEs for fear of TVEs’ competition with state-owned enterprises for credits, resources, and markets. Only after the government was convinced by the evidences that the TVE was good for increasing farmers’ income and for solving the shortages in the urban markets, did the government give green light to the development of TVEs in rural China (Lin, Cai, and Li 1994). References Acemoglu, Daron, and Fabrizio Zilibotti. 1999. “Productivity Differences.” Working Paper 6879 National Bureau of Economic Research, Cambridge, Mass. Ahluwalia, Montek S. 1976. “Inequality, Poverty, and Development.” Journal of Development Economics 3 (December): 307–12. Akamatsu, K. 1962. “A Historical Pattern of Economic Growth in Developing Countries.” Developing Economies, preliminary issue, no. 1 (March–August): 3–25.

328 Development Strategy, Institutions, and Economic Performance | 315 Atkinson, Anthony B., and Joseph E. Stiglitz. 1969. “A New View of Technologi- Economic Journal 79 (September): 573–78. cal Change.” Barro, Robert J. 1997. Determinants of Economic Growth: A Cross-Country Empirical Study. Cambridge, Mass.: MIT Press. Barro, Robert J., and Xavier Sala-i-Martin. 1992. “Convergence.” Journal of Political Economy 100 (April): 223–51. Journal of ———. 1997. “Technological Diffusion, Convergence, and Growth.” Economic Growth 2 (March): 1–26. Basu, Susanto, and David N. Weil. 1998. “Appropriate Technology and Quarterly Journal of Economics Growth.” 113 (November): 1025–54. Baumol, William J. 1986. “Productivity Growth, Convergence, and Welfare: American Economic Review What the Long-Run Data Show.” 76 (December): 1072–85. Ben-David, Dan. 1993. “Equalizing Exchange: Trade Liberalization and Income Quarterly Journal of Economics 108 (August): 653–79. Convergence.” American Economic Caselli, Francesco. 1999. “Technological Revolutions,” Review 89 (March): 78–102. Caselli, Francesco, and Wilbur John Coleman II. 2000. “The World Technol- ogy Frontier.” Working Paper 7904. National Bureau of Economic Research, Cambridge, Mass. Caselli, Francesco, Gerardo Esquivel, and Fernando Lefort. 1996. “Reopening the Convergence Debate: A New Look at Cross-Country Growth Empirics.” 1 (September): 363–89. Journal of Economic Growth Chenery, Hollis B. 1961. “Comparative Advantage and Development Policy.” American Economic Review 51 (March): 18–51. Chenery, Hollis B., and M. Syrquin. 1975. Pattern of Development, 1950–70. New York: Oxford University Press. Cline, William. 1975. “Distribution and Development: A Survey of the Litera- Journal of Development Economics 1 (February): 359–400. ture.” Diwan, I., and D. Rodrik. 1991. “Patents, Appropriate Technology, and North- South Trade.” Journal of International Economics 30: 27–47. Dollar, David. 1992. “Outward-Oriented Developing Economies Really Do Grow More Rapidly: Evidence from 95 LDCs, 1976–1985.” Economic Development and Cultural Change 40, no. 3 (April): 523–44. Edwards, Sebastian. 1995. Crisis and Reform in Latin America: From Despair to Hope. New York: Oxford University Press. Fei, John, Gustav Ranis, and Shirley W. Y. Kuo. 1979. Growth with Equity: The Taiwan Case, New York: Oxford University Press. Fields, Gary. 1991. “Growth and Income Distribution.” In Essays on Poverty, Equity, and Growth, ed. George Psacharopoulus. Oxford: Pergamon, pp. 1–52.

329 316 | New Structural Economics Frankel, Jeffrey, and David Romer. 1999. “Does Trade Cause Growth?” American Economic Review 89 (June): 379–99. Gerschenkron, A. 1962. Economic Backwardness in Historical Perspective. Cambridge, MA: Harvard University Press. Alternative Strategies for Economic Development, 2d ed. Griffi n, Keith. 1999. London: St. Martin’s Press. Grossman, Gene M., and Elhanan Helpman. 1994. “Endogenous Innovation in the Theory of Growth.” Journal of Economic Perspectives 8 (Winter): 23–44. Groves, Theodore, Yongmiao Hong, John McMillan, and Barry Naughton. 1994. Quarterly Journal of “Autonomy and Incentives in Chinese State Enterprises.” 109 (1): 183–209. Economics World Economic Growth. Harberger, Arnold C., ed. 1985. San Francisco: ICS. Harrison, Ann. 1996. “Openness and Growth: A Time-Series, Cross-Country Analysis for Developing Countries.” Journal of Development Economics 48 (March): 419–47. Jefferson, Gary H., and Thomas G. Rawski. 1995. “How Industrial Reform Worked in China: The Role of Innovation, Competition, and Property Rights.” Proceedings of the World Bank Annual Conference of Development Economics. Washington DC: World Bank, pp. 129–56. Jefferson, Gary H., Thomas G. Rawski, and Yuxin Zheng, 1992. “Growth, Effi - ciency, and Convergence in China’s State and Collective Industry.” Economic Development and Cultural Change 40 (2): 239–66. Klenow, Peter, and A. Rodríguez-Clare. 1997. “The Neoclassical Revival in Growth Economics: Has It Gone Too Far?” In NBER Macro Annual 1997, ed. Ben S. Bernanke and Julio Rotemberg. Cambridge, Mass.: MIT Press, pp. 73–114. Kolodko, Grzegorz W. 2000. From Shock to Therapy. Political Economy of Post- socialist Transformation . New York: Oxford University Press. Kornai, Janos. 1986. “The Soft Budget Constraint.” Kyklos 39, no. 1: 3–30. Krueger, A. O. 1974. “The Political Economy of the Rent-Seeking Society.” American Economic Review 64 (June): 291–303. ———. 1992. Economic Policy Reform in Developing Countries. Oxford, U.K.: Blackwell. American Kuznets, Simon. 1955. “Economic Growth and Income Inequality.” Economic Review 45 (March): 1–8. Lal, Deepak. 1985. “Nationalism, Socialism and Planning: Infl uential Ideas in the South.” World Development 13 (June): 749–59. Lee, Jong-Wha. 1995. “Capital Goods Imports and Long Run Growth.” Journal of Development Economics 48 (October 1): 91–110. Levine, Ross. 1997. “Financial Development and Economic Growth: Views and Agenda.” Journal of Economic Literature 35 (June): 688–726.

330 Development Strategy, Institutions, and Economic Performance | 317 Li, Wei. 1997. “The Impact of Economic Reform on the Performance of Chi- Journal of Political Economy 105 (5): nese State Enterprises, 1980–1989.” 1080–106. Lin, Justin Yifu. 1992. “The Needham Puzzle: Why the Industrial Revolution Did Not Originate in China.” UCLA Economics Working Paper 650. UCLA Department of Economics, Los Angeles. ———. 2000. “The Financial and Economic Crisis in Asia: Causes and Long- Term Implications.” In The New Social Policy Agenda in Asia: Proceedings of the Manila Social Forum (Manila: Asian Development Bank), pp. 9–17. Lin, Justin Yifu, and Guofu Tan. 1999. “Policy Burdens, Accountability, and the Soft Budget Constraint.” American Economic Review: Papers and Proceedings 89 (May): 426–31. Lin, Justin Yifu, Fang Cai, and Zhou Li. 1994. “China’s Economic Reforms: Pointers for Oher Economies in Transition.” Policy Research Working Paper 1310, World Bank, Washington, DC. ———. 1998. “Competition, Policy Burdens, and State-Owned Enterprise Reform.” 88 (May): American Economic Review: Papers and Proceedings 422–27. China’s State-Owned Firm Reform, trans. by the authors (Hong ———. 2001. Kong SAR, China: Chinese University of Hong Kong Press, 2001; originally published as Zhongguo Guoyou Qiye Gaige. Taipei: Linking Press, 2000). Lucas, Robert E. 1988. “On the Mechanism of Economic Development.” Journal of Monetary Economics 22 (March): 3–42. ———. 1993. “Making a Miracle.” Econometrica 61 (March): 251–72. Monitoring the World Economy, 1820–1992. Maddison, Angus. 1995. Paris: OECD, pp. 196–205. Mankiw, N. Gregory, David Romer, and David N. Weil. 1992. “A Contribution Quarterly Journal of Economics to the Empirics of Economic Growth.” 107 (May): 407–37. McKinnon, R. 1973. Money and Capital in Economic Development. Washington, D.C.: Brookings Institution. ———. 1993. The Order of Economic Liberalization: Financial Control in the Transition to a Market Economy , 2d ed. Baltimore: Johns Hopkins University Press. Michaely, Michael. 1977. “Exports and Growth: An Empirical Investigation.” Journal of Development Economics 4 (March): 49–53. Murphy, Kevin M., Andrei Shleifer, and Robert W. Vishny. 1989a. “Income Distribution, Market Size, and Industrialization.” Quarterly Journal of Economics 104 (August): 537–64. ———. 1989b. “Industrialization and Big Push.” Journal of Political Economy 97 (October): 1003–26.

331 318 | New Structural Economics Pack, Howard. 1994. “Endogenous Growth Theory: Intellectual Appeal and Journal of Economic Perspectives 8 (Winter): 55–72. Empirical Shortcomings.” Paukert, Felix. 1973. “Income Distribution at Different Levels of Development: A International Labour Review 108 (August–September): Survey of Evidence.” 97–125. Pearson, Lester B. 1969. Partners in Development: Report of the Commission on International Development. New York: Praeger. Prebisch, Raul. 1959. “Commercial Policy in the Underdeveloped Countries.” American Economic Review: Papers and Proceedings 49 (May): 251–73. Rajan. R. G., and L. Zingales. 1998. “Financial Dependence and Growth.” American Economic Review 88 (June): 559–86. Ranis, Gustav, and Mahmood Syed. 1992. The Political Economy of Develop- Cambridge, Mass.: Blackwell. ment Policy Change. Redding, Stephen. 1999. “Dynamic Comparative Advantage and the Welfare Effects of Trade.” Oxford Economic Papers 51 (January): 15–39. Rodríguez, Francisco, and Dani Rodrik. 1999. “Trade Policy and Economic Growth: A Skeptic’s Guide to the Cross-National Evidence.” Working Paper w7081. National Bureau of Economic Research, Cambridge, Mass; subsequently published in National Bureau of Economic Research. 2000. NBER Macroeconomics Annual, 2000, ed. B. Bernanke and K. Rogoff. Cambridge, Mass.: MIT Press, pp. 261–325. Romer, Paul. 1986. “Increasing Returns and Long-Run Growth.” Journal of Political Economy 94 (October): 1002–37. Journal of Economic ———. 1994. “The Origins of Endogenous Growth.” Perspectives 5 (Winter): 3–22. Rosenstein-Rodan, P. 1943. “Problems of Industrialization of Eastern and South- Economic Journal eastern Europe.” 53 (June–September): 202–11. Sachs, Jeffrey D., and Andrew Warner. 1995. “Economic Reform and the Process of Global Integration.” Brookings Papers on Economic Activity, no. 1: 1–95. Schultz, T. W. 1975. “The Value of the Ability to Deal with Disequilibria.” Journal of Economic Literature 13 (September): 827–46. Schumacher, E. F. 1973. Small Is Beautiful: Economics as If People Mattered. New York: Harper & Row. Shaw, E. S. 1969. Financial Deepening in Economic Development. New York: Oxford University Press. Solow, Robert M. 1956. “A Contribution to the Theory of Economic Growth.” Quarterly Journal of Economics 70 (February): 65–94. Stern, Joseph J., Ji-hong Kim, Dwight H. Perkins, and Jung-ho Yoo, eds. 1995. Industrialization and the State: The Korean Heavy and Chemical Industry Drive. Cambridge, Mass.: Harvard University Press. Stiglitz, Joseph E. 1998. “Toward a New Paradigm for Development: Strategies, Polices, and Processes.” 1998 Prebisch Lecture at the United Nations Confer- ence on Trade and Development, Geneva, October 19.

332 Development Strategy, Institutions, and Economic Performance | 319 Takatoshi, Ito. 1998. “What Can Developing Countries Learn from East Asia’s Annual World Bank Conference on Development Economic Growth?” In Economics, 1997, ed. Boris Pleskovic and Joseph E. Stiglitz. Washington, D.C.: World Bank, pp. 183–200. Tsiang, Sho-chieh. 1984. “Taiwan’s Economic Miracle: Lessons in Economic Development.” In World Economic Growth: Case Studies of Developed and Developing Nations, ed. Arnold C. Harberger. San Francisco: ICS. Warr, Peter G. 1994. “Comparative and Competitive Advantage,” c Asian Pacifi Economic Literature 8 (November): 1–14. Weitzman, Martin L., and Xu Chenggang. 1994. “Chinese Township-Village Enterprises as Vaguely Defi ned Cooperatives.” Journal of Comparative Economics 18 (2): 121–45.

333 320 | New Structural Economics * P ART 2 Strategy Choice and Economic Performance: Empirical Testing The previous sections discussed the effects of development strategy on institutional arrangements, economic growth, income distribution, and transition performance in a country. From those discussions, I derive sev- eral testable hypotheses. Hypothesis 1. A country that adopts a CAD strategy will require various government interventions and distortions in its economy. Hypothesis 2. Over an extended period, a country that adopts a CAD strategy will have poor growth performance. Hypothesis 3. Over an extended period, a country that adopts a CAD strategy will have a volatile economy. Hypothesis 4. Over an extended period, a country that adopts a CAD strategy will have less equitable income distribution. Hypothesis 5. In the transition to a market economy, a country’s overall economic growth will be improved if it creates conditions to facilitate the development of formerly repressed labor-intensive industries. Estimating the Choice of Development Strategy Using the Capital Intensity in the Manufacturing Sector To expand the set of countries and testable hypotheses, Lin and Liu ned (2004) propose another technology choice index (TCI), which is defi as follows: LM AVM ,, it it = TCI (VI.1) , it GDP L it ,, it where AV M is the added value of manufacturing industries of country i,t i i is is the total added value of country LM ; t at time GDP at time t ; i,t i,t * Adapted from “Development Strategy, Development and Transition Performances: Empirical Economic Development and Transition: Thought, Strategy, and Viability , by Justin Analysis,” in Yifu Lin. Reprinted with the permission of Cambridge University Press. Copyright © 2009 Cam- bridge University Press. The conclusion was adapted from “Development Strategy, Viability, and Economic Convergence.” See disclaimer on the fi rst page of part 1 of this chapter.

334 Development Strategy, Institutions, and Economic Performance | 321 the labor in the manufacturing industry and L is the total labor force. i,t If a government adopts a CAD strategy to promote its capital-intensive industries, TCI in this country is expected to be larger than otherwise. This is because if a country adopts a CAD strategy, in order to overcome the viability issue of the fi rms in the prioritized sectors of the manufacturing industries, the government might give the fi rms monopoly positions in the product markets—allowing them to charge higher output prices—and pro- vide them with subsidized credits and inputs to lower their investment and AV M operation costs. The above policy measures will result in a larger i,t than otherwise. Meanwhile, investment in the prioritized manufacturing industry will be more capital intensive and absorb less labor— ceteris pari- bus . The numerator in equation VI.1 will therefore be larger for a country that adopts a CAD strategy. As such, given the income level and other con- ditions, the magnitude of the TCI can be used as a proxy for the extent that 1 The data for calculating the TCI a CAD strategy is pursued in a country. are taken from the World Bank’s World Development Indicators (2002) and the United Nations Industrial Development Organisation’s Interna- (2002). The means and variations tional Yearbook of Industrial Statistics of the TCI for each of the 122 countries in the period 1962–99 are reported in annex table VI.1. Development Strategy and Institutions To assess the effects of development strategy on the government’s distor- tions and interventions in the economy—as postulated in Hypothesis 1—I use several proxies for the institutions: 1) the “black-market premium” is used as an index of price distortion; 2) the index of economic freedom (IEF) and the expropriation risk are used as indexes of government’s inter- vention in property rights institution; 3) the number of procedures required for a start-up fi rm to obtain legal status and the “executive de facto inde- pendence” are used as indexes of enterprise autonomy; and 4) the trade dependence ratio is used as an index for openness. The means and varia- tions of each proxy for each country are reported in the annex table. Development Strategy and Price Distortions. The black-market premium of 105 countries is taken from the Global Development Network Growth Database provided by the Development Research Institute of New York

335 322 | New Structural Economics University. The relationship between the TCI and the black-market pre- mium across four decades (1960–69, 1970–79, 1980–89, 1990–99) is shown in fi gure VI.4. and the black-market premium had Figure VI.4 shows that the TCI positive relationships throughout the four decades, which implies—as pre- dicted by Hypothesis 1—that a higher degree of CAD strategy is associated with a larger black-market premium. Development Strategy and Government Intervention in Resource Allocation. To measure government’s intervention in property rights institution, I use the index of economic freedom and the expropriation Economic risk. The observations of IEF from 91 countries are taken from Freedom of the World (Gwartney and Lawson 2007), which are available from 1970 onwards. This index ranges from zero to 10. A higher value means a higher degree of economic freedom. The correlations between the TCI and the IEF averaged across a decade for each country are shown in fi gure VI.5. There is a strong negative relationship between the TCI and the IEF in each of the panels, which is consistent with the prediction that the more aggressively a government pursues a CAD strategy, the more government intervention is required, and the less economic freedom there is. The expropriation risk of 102 countries is adopted from the Interna- tional Country Risk Guide . The expropriation risk is the risk of outright scation and forced nationalization of property. This variable ranges confi from zero to 10. A higher value means that a private enterprise has a lower probability of being expropriated. Figure VI.6 plots the relationship between the TCI and the expropriation risk. Both variables are calculated as the average values from 1982 until 1997. As shown, there is a negative relationship between the TCI and expro- priation risk, which is consistent with the expectation that the more aggressive the government’s CAD strategy, the more likely it is that the government will confi scate or nationalize an enterprise. Development Strategy and Enterprise Autonomy. To analyze the rela- tionship between the government’s development strategy and enterprise autonomy, the study uses two indexes—including the number of pro- cedures and the executive de facto independence used in Djankov and

336 15 15 10 10 TCI TCI 5 5 premium (1990s) premium (1970s) Correlation between TCI and black-market Correlation between TCI and black-market 0 0 0 0 10 60 50 20 30 40 50 60 70 80 90 30 20 90 40 80 70 10 100 100 BMP BMP 15 15 10 10 TCI TCI 5 premium (1980s) 5 premium (1960s) Correlation between TCI and black-market Correlation between TCI and black-market 0 0 0 0 70 30 20 10 50 40 90 80 60 40 70 60 50 80 30 20 10 100 BMP BMP Figure VI.4: The TCI and Black-Market Premium 323

337 TCI Correlation between TCI and IEF (1980s) 0510 8 7 4 9 5 6 3 10 IEF TCI TCI Correlation between TCI and IEF (1970s) Correlation between TCI and IEF (1990s) 0510 0510 9 8 7 6 5 4 3 2 4 5 6 7 8 9 3 10 10 IEF IEF Figure VI.5: The TCI and the IEF 324

338 Development Strategy, Institutions, and Economic Performance | 325 Figure VI.6: The TCI and Expropriation Risk Correlation between TCI and expropriation risk 1 8 E 6 4 2 15 0 5 10 TC Murrell (2002)—to represent the extent of enterprise autonomy. There are 69 countries in the samples. The “number of procedures” is the number of administrative proce- dures that a start-up fi rm has to comply with in order to obtain legal status—that is, to start operating as a legal entity. “‘Executive de facto independence” is an index of “operation ( ) independence of the de facto = chief executive,” descending from 1 to 7 (1 = pure individual; 2 inter- mediate category; 3 = slight to moderate limitations; 4 = intermediate cat- executive egory; 5 = substantial limitations; 6 = intermediate category; 7 = parity or subordination). Both indexes are the average values for the years from 1965 until 1998. The positive relationship between the TCI and the number of proce- dures and the negative relationship between the TCI and the executive gure VI.7 indicate that a high degree of independence shown in fi de facto rms CAD strategy is associated with low enterprise autonomy, which confi the prediction of Hypothesis 1. Development Strategy and Openness. The trade-dependence ratio of 115 countries—taken from Dollar and Kraay (2003)—is used to refl ect the openness of a country. The correlations between the TCI and open- ness averaged across the past four decades in each country are shown in 2 fi gure VI.8. We fi nd that the TCI and openness have a negative relationship, which is consistent with the hypothesis that if a developing-country government adopts a CAD strategy, its economy will become more inward-oriented

339 326 | New Structural Economics Figure VI.7: The TCI and Enterprise Autonomy Correlation between TCI and number of procedures 20 15 10 NP 5 0 0 15 10 5 TCI Correlation between TCI and de facto independence 8 6 4 IND 2 0 0510 TCI than otherwise. This is because the CAD strategy attempts to substitute the import of capital-intensive manufactured goods with domestic produc- tion, causing a reduction in imports. Exports will also be suppressed due to the inevitable transfer of resources away from the industries that have comparative advantage to the prioritized sectors determined by the CAD strategy. The more a country follows a CAD strategy, therefore, the less openness there will be in the country. Development Strategy and Economic Growth Hypothesis 2 predicts that over an extended period, a country adopting a CAD strategy will have a poor growth performance. The following econo- metric model is used to test the hypothesis: GROWTH β X + θ (VI.2) TCI + α C = + i,t i,t

340 15 15 10 10 TCI TCI 5 5 Correlation between TCI and openness (1990s) Correlation between TCI and openness (1970s) 0 0 0 0 20 40 60 80 40 20 80 60 140 100 120 100 120 openness openness 15 15 10 10 TCI TCI 5 5 Correlation between TCI and openness (1960s) Correlation between TCI and openness (1980s) 0 0 0 0 40 60 80 20 20 80 40 60 120 100 100 140 120 openness openness Figure VI.8: The TCI and Openness 327

341 328 | New Structural Economics where GROWTH is the economic growth rate in a certain period in coun- i,t try , X is a vector that includes the initial per capita GDP to control the i effect of the stage of development, the initial population size to control the effect of market size, the indicator of rule of law to refl ect the institutional quality—which was constructed by Kaufmann and Kraay (2002)—the ect openness, the distance from the Equator trade-dependent ratio to refl and whether the country is land-locked. The last two explanatory variables are included to capture the effects of geography. The instrumental vari- able for controlling the endogeneity of institutional quality is the share of population that speaks English and the share that speaks a major European language (Hall and Jones 1999), which are used to capture the long-run impacts of colonial origin on current institutional quality. Similarly, the fi t- ted values of trade predicted by a gravity model are used as the instrument for openness. This approach was proposed by Frankel and Romer (1999) and revised by Dollar and Kraay (2003). In the regressions that use panel data, the instrument for openness is the single-period lagged value of itself. Table VI.2 summarizes the defi nition of each variable and the data source. rst approach, We will use two approaches to test this hypothesis. In the fi the dependent variable is the average annual growth rate of per capita GDP for the period 1962–99, and in the second, the dependent variable is the average annual growth rate of per capita GDP for each decade of the 1960s, 1970s, 1980s and 1990s. rst approach. Regression Table VI.3 reports the estimates from the fi Model 1.1 and Model 1.2 use the OLS approach to obtain the estimates. The explanatory variables in Model 1.1 include only the proxy for the development strategy, LnTCI1, and the initial GDP per capita, LnGDP60, whereas Model 1.2 includes other explanatory variables that capture insti- tutional quality, openness, geographic location and market size. Model 1.3 has the same explanatory variables but the model uses the 2SLS approach in order to control the endogeneity of institutional quality and openness. The results show that the TCI has the expected negative effect and is highly signifi cant in all three regressions. This fi nding supports Hypothesis 2 that the more aggressive is the CAD strategy pursued by a country, the worse the growth performance is in that country in the period 1962–99. The estimated coeffi cients of LnTCI1 have values ranging from –0.66 to –1.25. From the estimates, we can infer that a 10 percent increase from

342 Development Strategy, Institutions, and Economic Performance | 329 Table VI.2: Variable Defi nitions and Data Source Defi nition Mean Sources Variable Std dev. 7.33 0.80 World Development LnGDP60 Log of real GDP per capita World Bank Indicators in 1960 World Bank World Development Log of real GDP per capita LnGDP80 7.91 1.05 Indicators in 1980 Log of real GDP per capita World Bank World Development LnGDP 7.73 1.02 in 1960, 1970, 1980, 1990 Indicators 0.96 0.90 World Bank World Development LnTCI1 Log of the average technology Indicators and UNIDO (2002) choice index from 1963 to 1999 Log of the average TCI per decade LnTCI2 0.85 0.84 World Bank World Development Indicators and UNIDO (2002) in 1960s, 1970s, 1980s, 1990s LnTCI70 Log of the average TCI from 0.91 0.92 World Bank World Development Indicators and UNIDO (2002) 1970 to 1979. If not available, we use the log of the average TCI from 1980 to 1985 TCI Log of the average TCI from Δ 0.07 0.38 World Bank World Development 1999 to 1990 minus LnTCI70 and UNIDO (2002) Indicators RL01 Rule of law in 2000–01 0.003 0.95 Kaufmann and Kraay (2002) LnOPEN1 Log of the average (exports + Dollar and Kraay (2003) –1.11 0.81 imports)/GDP from 1960 to 1999 LnOPEN2 Log of the decadal average Dollar and Kraay (2003) 0.84 –1.30 imports)/GDP in (exports + 1960s, 1970s, 1980s, 1990s Log of the total mid-year World Bank World Development LnPOP1 15.2 2.11 population from 1960 to 1999 Indicators Log of the total initial-year World Development LnPOP2 14.93 2.12 World Bank population in 1960s, 1970s, Indicators 1980s, 1990s LANDLOCK Dummy variable taking value 0.18 0.39 Dollar and Kraay (2003) of 1 if country is land-locked; 0 otherwise LnDIST Log (DISTEQ+1), where DISTEQ Dollar and Kraay (2003) 0.88 2.96 is the distance from Equator, measured as absolute value of latitude of capital city ENGFRAC Fraction of population speaking 0.07 0.24 Hall and Jones (1999), taken English from Dollar and Kraay (2003) EURFRAC Fraction of population speaking Hall and Jones (1999), taken 0.22 0.38 a major European language from Dollar and Kraay (2003) LnFRINST Instrument variable for LnOPEN –2.83 0.64 Dollar and Kraay (2003) Author’s estimation Predicted value of RL01 in the 0.34 0.003 INST cross-section estimation (ENGFRAC and EURFRAC as the instruments)

343 330 | New Structural Economics Table VI.3: Development Strategy and Economic Growth—Model 1 Model 1.3 Model 1.2 Model 1.1 (2SLS) (OLS) (OLS) Constant 3.26 7.32*** 4.66** (1.87) (2.15) (1.60) LnTCI2 –1.25*** –0.92*** –0.66*** (0.18) (0.20) (0.19) –0.99*** –0.54*** LnGDP60 –0.59*** (0.18) (0.20) (0.21) RL02 0.58*** (0.21) 0.22 INST (0.41) LnOPEN2 0.70*** (0.22) 0.93** TRADE2 (0.43) LnDIST 0.20 0.47*** (0.16) (0.16) 0.33*** 0.22** LnPOP2 (0.09) (0.09) 0.07 LANDLOCK 0.46 (0.38) (0.32) 2 Adjusted-R 0.36 0.44 0.56 85 83 83 Observations * indicates signifi cance at the 10 per cent level ** indicates signifi cance at the 5 per cent level *** indicates signifi cance at the 1 per cent level Note: The dependent variable is the yearly average of per-capita GDP growth rate in 1962–99. Standard errors are reported in parentheses. the mean in the TCI can result in approximately 0.1 of a percentage point reduction in the country’s average annual growth rate of per capita GDP for the whole period 1962–99. The regression results also show that the initial per capita income and the population size have the expected signs and signifi cant effects on the growth rate. Rule of law, openness and distance from the Equator also have the expected signs. Rule of law, however, is not signifi cant in the 2SLS regression and distance from the Equator is not signifi cant in the OLS regression. Whether the country is land-locked is insignifi cant in all three regressions. Table VI.4 reports the results from the second approach, in which the dependent variable is the average annual growth rate of per capita GDP in

344 Development Strategy, Institutions, and Economic Performance | 331 Table VI.4: Development Strategy and Economic Growth—Model 2 Model 2.5 Model 2.4 Model 2.3 Model 2.2 Model 2.1 (OLS) (2SLS) (OLS) (2SLS, fi xed effect) (fi xed effect) –0.74 8.36*** 7.15*** Constant –2.70 3.83* (2.37) (2.16) (2.56) (1.61) (2.11) –0.69*** –0.47** LnTCI2 –1.10*** –0.69*** –0.40** (0.22) (0.20) (0.21) (0.19) (0.24) 0.17 –0.86*** –0.17 –1.39*** –0.54*** LnGDP (0.27) (0.18) (0.23) (0.23) (0.25) 1.45*** RL01 1.12*** (0.22) (0.23) –0.67* –0.38 INST (0.38) (0.42) LnOPEN2 0.35 0.24 (0.22) (0.23) TRADE2 0.01 –0.06 (0.29) (0.27) 0.17 LnDIST –0.04 –0.10 0.27 (0.17) (0.18) (0.20) (0.18) 0.27** LnPOP2 0.32*** 0.41*** 0.22* (0.09) (0.12) (0.12) (0.10) –0.23 LANDLOCK 0.08 –0.31 0.02 (0.36) (0.43) (0.46) (0.39) 2 Adjusted-R 0.23 0.36 0.08 0.24 0.08 278 Observations 213 213 278 315 * indicates signifi cance at the 10 per cent level ** indicates signifi cance at the 5 per cent level *** indicates signifi cance at the 1 per cent level Dependent variable is the average growth rate of GDP per capita in the decades 1960s, 1970s, 1980s, 1990s. Notes: Models 2.3 and 2.5 include the time dummy. Standard errors are reported in parentheses. each decade from 1960–99. The regressions to fi t the estimates are OLS for xed effect for Model 2.3, 2SLS for Model Models 2.1 and 2.2, one-way fi 2.4 and 2SLS and one-way fi xed effect for Model 2.5. In the fi xed-effect models, time dummies are added to control the time effects, whereas the 2SLS models are used for controlling the endogeneity of institutional qual- ity and openness. As in the results in the fi rst approach, the estimates for the TCI have the expected negative sign and are highly signifi cant in all regressions. The fi nding is once again consistent with the prediction of Hypothesis 2 that development strategy is a prime determinant of the long-run economic 3 growth performance of a country.

345 332 | New Structural Economics The results for other explanatory variables are similar to those in table VI.3. Development Strategy and Economic Volatility Hypothesis 3 is about the effect of a CAD strategy on the volatility of the economic growth rate. If a country follows a CAD strategy, there could be a period of investment-led growth, but it will not be sustainable and is likely to cause economic crisis. Therefore, a country that follows a CAD strategy is likely to be more volatile than otherwise. In the empirical testing of this hypothesis, the volatility of a country’s per capita GDP growth rate in the period 1962–99 is measured as follows: 2 ⎡ ⎤ g ⎞ ⎛ it = 1999 T 1 − ⎢ ⎥ ⎟ ⎜ T = 1999 ⎛ ⎞ V = ⎢ ⎥ () (VI.3) 138 / ⎟ ⎜ i ∑ ⎜ ⎟ g 38 ⎥ ⎢ it ∑ ⎟ ⎜ 2 2 t 196 = ⎜ ⎟ ⎟ ⎜ ⎢ ⎥ t 1962 = ⎠ ⎝ ⎠ ⎝ ⎦ ⎣ th where g i is the growth rate of GDP per capita of country in year t . it In testing Hypothesis 3, the dependent variable is the log of the above measurement of volatility, V , and the explanatory variables are the same i as those used in testing Hypothesis 2. The approaches to fi tting the regres- sion equation are also similar to those used previously. Table VI.5 reports tting the regression models. As expected, the estimates the results from fi cant in all three regressions. The of the TCI are positive and highly signifi results support Hypothesis 3 and indicate that the deeper a country follows a CAD strategy, the more volatile is the country’s economic growth rate. From the estimates, it can be inferred that a 10 percent increase in the TCI could cause volatility to increase about 4–6 per cent. The estimates for other explanatory variables show that the quality of institutions, the degree of openness, whether the country is land-locked and the population size all have negative effects on economic volatility. The coeffi cients on population size, which is a proxy for the size of the economy, are signifi cant in the OLS and 2SLS models. The estimated coef- fi cients of initial per capita income in 1960 and the distance from the Equa- tor are insignifi cant in all three regressions. The estimated coeffi cients of all other variables are either signifi cant in the OLS model or the 2SLS model.

346 Development Strategy, Institutions, and Economic Performance | 333 Table VI.5: Development Strategy and Economic Volatility Model 3.2 Model 3.1 Model 3.3 (2SLS) (OLS) (OLS) 3.03** 0.49 Constant 3.63** (1.44) (1.56) (1.06) 0.56*** 0.41*** 0.64*** LnTCI1 (0.14) (0.14) (0.13) –0.04 LnGPP60 –0.07 0.17 (0.15) (0.13) (0.14) RL01 –0.33** (0.16) INST –0.20 (0.29) LnOPEN1 –0.46*** (0.17) TRADE1 –0.53 (0.33) –0.15 –0.003 LnDIST (0.11) (0.11) LANDLOCK –0.53* –0.31 (0.24) (0.28) –0.26*** –0.18** LnPOP1 (0.06) (0.07) 2 Adjusted-R 0.37 0.29 0.47 93 Observations 103 93 * indicates signifi cance at the 10 per cent level ** indicates signifi cance at the 5 per cent level *** indicates signifi cance at the 1 per cent level Notes: Dependent variable is the log of the growth rate’s volatility for GDP per capita from 1962–99. Standard errors are reported in parentheses. Development Strategy and Income Distribution In testing the effect of development strategy on income distribution, the following regression equation is used: α GINI = C + TCI + β X + ε (VI.4) i,t 2i,t GINI where TCI is a at time is the index of inequality in country t , i i,t X is a vector of other explanatory proxy for the development strategy and variables. GINI coeffi cients are taken from a revised version of the data set in Deininger and Squire (1996). The data set includes the estimation of GINI

347 334 | New Structural Economics coeffi cients for many countries in the various literature. Some are esti- mated according to the data on income; others are based on expenditure. The coverage differs between the different countries’ GINI data. Deininger cient estimations; only and Squire (1996) assessed the quality of GINI coeffi those ranked as “acceptable” were used in the regression. The original esti- mates of GINI coeffi cients based on income data are left unchanged, but those based on consumption expenditure are adjusted by adding 6.6, which is the average difference between the two estimation methods. Matching this GINI data with the TCI, I end up with a panel of 261 samples from 33 countries. Figure VI.9 shows the relationship between the TCI and the cient. GINI coeffi In order to test alternative hypotheses for the determination of inequality, , I have included the explanatory variables—per capita income, GDPPC i,t and its reciprocal, GDPPC —which test the Kuznets inversed-U hypoth- 1 i,t – cients for these two variables esis. If Kuznets’ hypothesis holds, the coeffi should be signifi cantly negative. Based on the data set of Deininger and Squire (1996), Li et al. (1998) conducted a robust empirical test, and the result showed that the GINI coeffi cient for an individual country was relatively constant across differ- ent periods. Based on this conclusion, the GINI coeffi cient in the initial year in the data set is introduced into the regression, denoted by IGINI. In this way, the historical factors that could affect income distribution and those non-observable factors across countries can be excluded. In the Figure VI.9: Development Strategy and Income Distribution Correlation between TCI and GINI coefficient 65 50 GINI (%) 30 20 012345678 TC

348 Development Strategy, Institutions, and Economic Performance | 335 data set, the year of IGINI differs from country to country. In spite of difference, the higher the IGINI, the higher are the subsequent GINI this cients—regardless of the initial year. As a result, the coeffi cient of coeffi IGINI is expected to be positive. Corruption could also affect income distribution. Two explanatory variables are included in the regression: the index for corruption, CORR , i,t cials, and the quality of offi . The data for these two variables are taken BQ i,t from Sachs and Warner (2000) and they differ from country to country but remain constant throughout the period studied. The larger the value is, the less is the corruption and the higher is the quality of offi cials. The coef- fi cients of these two variables are expected to be negative. Foreign trade could also affect income distribution. It affects the relative prices of factors of production (Samuelson, 1978) and market opportunities for different sectors in the economy. Consequently, trade— through its effect on employment opportunities (Krugman and Obstfeld, 1997)—can affect income distribution. The regression therefore includes , which is the share an index of economic openness, denoted by OPEN i,t of total import and export value in nominal GDP, as an explanatory vari- able. The data are taken from Easterly and Yu (2000). Openness could, however, have different impacts on skilled and unskilled labor, on trad- able and non-tradable sectors and in the short run and in the long run. Its sign is therefore uncertain. Table VI.6 reports the results from fi ve regression models. Model 4.1 includes all explanatory variables: TCI , IGINI , GDPPC , GDPPC 1, – , , and OPEN . As CORR BQ BQ and OPEN are endogenous, CORR other models exclude these variables to control the endogeneity problem. Because IGINI , CORR and BQ are time invariant, the one-way effects model is applied in fi tting the regression of Models 4.1, 4.2 and 4.4. According to Hausmann tests, the one-way random-effect model is used in the regressions of Models 4.1, 4.2 and 4.4, and the two-way fi xed-effect model is used in the regression of Models 4.3 and 4.5. The estimated coeffi cients of TCI are positive and signifi cant at the 1 percent level in all fi ve regression models. These results strongly sup- port the hypothesis that the more a country pursues a CAD strategy, the more severe will be the income disparity in that country. This result holds whether the initial income distribution is equal or unequal.

349 336 | New Structural Economics Table VI.6: The Effect of Development Strategy on Inequality Model 4.5f Model 4.4r Model 4.3f Model 4.2r Model 4.1r CONSTANT 6.46 8.18*** 31.5*** 8.09*** 32.6*** (3.16) (1.75) (2.40) (4.72) (0.97) 1.72*** TCI 1.35*** 1.32*** 1.35*** 1.84*** (0.46) (0.33) (0.31) (0.48) (0.32) 0.71*** 0.71*** 0.73*** IGINI (0.07) (0.08) (0.07) 0.74 GDPPC –0.89 0.43 (11.3) (10.8) (12.6) GDPPC_1 3.21 1.91 0.40 (16.6) (2.11) (1.84) 1.03* CORR (0.58) –0.84 BQ (0.58) OPEN 0.12 (1.68) R2 0.8941 0.8936 0.5780 0.9040 0.5495 1.99 1.19 7.98 23.91 3.32 Hausmann statistics 0.00 Hausmann P-value 0.19 0.28 0.37 0.00 261 observations from 33 countries Sample f: fi xed-effect model r: random-effect model * indicates signifi cance at the 10 per cent level ** indicates signifi cance at the 5 per cent level *** indicates signifi cance at the 1 per cent level Null hypothesis of Hausmann test: there is a random effect in countries and time. Standard errors are reported in Notes: parentheses. cant at The estimated coeffi cients of IGINI are also positive and signifi the 1 percent level in Models 4.1, 4.2 and 4.4. This result is consistent with the fi nding in Li et al. (1998): that is, the initial income distribution will have a carry-over effect in the subsequent period’s income distribution. GDPPC cients of GDPPC and The estimated coeffi _1 in Models 4.1, 4.3 and 4.4 are all insignifi cant and have an unexpected positive sign— except for GDPPC in Model 4.1. Kuznets’ inversed-U hypothesis of income distribution is therefore rejected. has an The results in Model 4.1 show that the coeffi cient for CORR i,t unexpected positive sign. One possible reason for this is that the effect of

350 Development Strategy, Institutions, and Economic Performance | 337 corruption on distribution is not refl ected accurately in the surveys. The cient for bureaucracy quality, , has an expected, but insignifi - coeffi BQ i,t cant, negative sign. The coeffi cient for openness, OPEN , is positive, but cant. not signifi From the results above, it is clear that development strategy and initial income distribution are the two most important determinants of income distribution in a country. As argued above, for a country in which the government follows a CAF strategy, income distribution will become more equal even if its initial income distribution is unequal. In effect, this is the “growth with equity” phenomenon observed in Taiwan and other newly industrialized economies in East Asia (Fei, Ranis, and Kuo 1979). Transition and Economic Performance As mentioned above, the development of labor-intensive sectors—in which developing countries have comparative advantage—is repressed and many institutions are distorted if the government adopts a CAD strategy, resulting in poor resource allocation and ineffi ciency. The growth performance during transition to a market economy depends, therefore, on the country’s ability to create an enabling environment nd for the development of labor-intensive sectors and at the same time fi a way to solve the viability issue for fi rms inherited from the previous development strategy so as to pave the way for eliminating previous dis- tortions and interventions. A CAD strategy is associated with a high TCI. If, after the reform/transition, a country is able to successfully develop labor-intensive sectors, resource allocation and growth performance will improve, and the TCI will decline. A successful transition from a CAD strategy is therefore expected to result in a negative change in the TCI. The larger the negative change is, the higher is the expected growth rate. For the purpose of testing Hypothesis 5, therefore, a variable, TCI, is Δ created to measure the difference between the log of average TCI in the period 1990–99 and the log of average TCI in the period 1970–79—as the transition in socialist countries and the reforms in other developing countries started in the 1980s. The dependent variable in the regressions is the log of the average annual growth rate of GDP per capita in the period 1980–99. In addition to Δ TCI, the explanatory variables include the log of average TCI in the

351 338 | New Structural Economics initial per capita GDP in 1980 and other explanatory variables— 1970s, representing institutional quality, openness and population size—which are similar to those used in testing Hypothesis 1. rst includes obser- Two approaches are used to test the hypothesis. The fi vations from all countries in the data set, while the second includes only ned by Easterly and Sewadeh (2002). Both the developing countries defi approaches try three regressions—two by OLS and one by 2SLS—to con- trol the endogeneity problem of institutional quality and openness. Table VI.7 reports the results from the regressions. Table VI.7: Development Strategy and the Performance of Economic Reform/Transition Model 5.1 Model 5.2 Model 5.3 Model 5.6 Model 5.4 Model 5.5 (OLS) (OLS) (2SLS) (2SLS) (OLS) (OLS) 2.53 –9.03 3.79 Constant –2.94 –4.50 4.28 (4.24) (3.17) (3.63) (6.43) (3.97) (5.01) –1.25** –1.12** –0.91** Δ TCI –1.30** –1.16* –1.02* (0.55) (0.52) (0.51) (0.60) (0.45) (0.66) –0.84** LnTCI70 –0.38 –0.31 –0.52 –0.26 –0.61 (0.48) (0.38) (0.34) (0.45) (0.38) (0.41) –0.12 –1.32*** –0.31 –0.34 LnGDP80 –0.78* –0.04 (0.50) (0.37) (0.35) (0.45) (0.38) (0.57) RL01 1.31*** 1.78*** (0.47) (0.37) INST 0.96 0.44 (1.18) (0.60) LnOPEN1 0.54 0.71* (0.36) (0.49) 1.50** TRADE1 2.23* (1.26) (0.70) 0.16 0.57* LnDIST –0.06 0.34 (0.28) (0.33) (0.36) (0.29) LnPOP1 0.52*** 0.78** 0.44*** 0.79*** (0.16) (0.19) (0.29) (0.17) –0.87 –0.06 LANDLOCK –0.55 0.54 (0.57) (0.68) (0.73) (1.15) 2 Adjusted-R 0.13 0.43 0.27 0.03 0.45 0.24 Observations 72 72 50 49 49 76 * indicates signifi cance at the 10 per cent level ** indicates signifi cance at the 5 per cent level *** indicates signifi cance at the 1 per cent level Notes: Dependent variable is the average growth rate of GDP per capita from 1980–99. The data samples in the regression of Models 6.4–6.6 include only the developing countries defi ned by Easterly and Sewadeh (2002). Standard errors are reported in parentheses.

352 Development Strategy, Institutions, and Economic Performance | 339 As expected, the sign of TCI is negative and the estimates are signifi - Δ cantly different from zero in all six regressions. The results support the hypothesis that a larger reduction in the TCI from the level in the 1970s to the level in the 1990s has a larger positive effect on the average per capita GDP growth rate in the period 1980–99. For a country that adopts a CAD strategy, therefore, growth performance will be improved if the govern- ment manages well the transition from a CAD to a CAF strategy. From the estimates, we can infer that a 10 percent reduction in the TCI level in the 1990s to the level of the 1970s could cause a 0.1–0.13 percentage point increase in the average annual growth rate of per capita GDP in the period 1980–99. The other explanatory variables all have the expected signs; however, except for the population size—which is positive and highly signifi cant in all six regressions—the other variables are either insignifi cant or signifi cant in some regressions but not in others. In a nutshell, as predicted by Hypothesis 5, the entry of small and medium-size fi rms into the repressed sectors under a CAD strategy is essential for the economy to achieve dynamic growth during the transi- tion process. Concluding Remarks This chapter argues that most LDCs follow an inappropriate development strategy and that, as a result, convergence is impeded, economic volatil- ity enhanced, and income distribution more unequal. During economic reform and transition, a country’s economic performance depends on its government’s ability to create an environment that facilitates the growth of labor-intensive industries, which have been suppressed in the past due to the government’s pursuit of a CAD strategy. The temptation to close the industry/technology gap as soon as possible is strong for LDCs. At a low level of factor endowment structure, however, LDCs’ economies do not have the comparative advantages necessary for capital-intensive industries/technologies, and their fi rms will not be viable in an open, free, and competitive market if they enter/adopt these industries/tech- nologies. To give priority to the development of non–comparative advan- tage industries/technologies, the governments in LDCs often adopt a

353 340 | New Structural Economics CAD strategy and give nonviable fi rms policy support through a series of distortions in interest rates, foreign exchange rates, and other prices. They also use administrative measures to directly allocate resources with rms in the priority industries. With the above distorted prices to the fi policy measures, an LDC may be able to establish fi rms that adopt high technologies in advanced industries for which the economy does not have the comparative advantages. However, the development of fi nancial markets will be repressed, foreign trade will be retarded, rent-seeking activities will be widespread, the macroeconomy will be unstable, income distribution will be unequal, the economy will be very uncompetitive, and the country will fail to converge with DCs in terms of income. I argue here that the optimal industry/technology structure of an econ- omy is endogenously determined by the economy’s factor endowment structure and that the CAF strategy is the better one for an LDC’s develop- ment. The CAF strategy will induce the fi rms in an LDC to enter industries for which the country has comparative advantages and facilitate the fi rms adoption of appropriate technology by borrowing at low costs from the more advanced countries. The economy will be competitive. The country will enjoy rapid upgrades in its factor endowment structure and, conse- quently, its industry/technology structure. As such, the CAF strategy will help an LDC achieve and foster a high rate of growth. Convergence will come true. The empirical fi ndings from the cross-country analyses are con- sistent with the above hypothesis. To implement the CAF strategy, a government needs to maintain an open, free, and competitive market. The government can also adopt an industrial policy to facilitate the fi rms’ upgrading of industry/technology. However, the functions of an industrial policy should be limited to infor- mation sharing, investment coordination, and compensation for externali- rst movers. ties produced by fi The government of an LDC plays an especially important role, for bet- ter or for worse, in the country’s economic development. As W. Arthur Lewis (1965) has noted, “No country has made economic progress with- out positive stimulus from intelligent governments . . . On the other hand, there are so many examples of the mischief done to economic life by governments.” Here I would like to propose that, for the government in an LDC to be an intelligent one, its most important task is to get its development strategy right.

354 S.D. Openness (1960–2003) 16.423 6.248 32.905 5.337 69.527 15.824 17.317 4.359 49.479 4.896 90.813 16.009 55.645 14.057 22.414 5.646 48.321 15.940 56.805 15.229 48.559 9.062 40.868 12.273 68.883 10.109 Mean 120.602 25.441 103.668 24.533 118.786 14.759 116.954 9.390 129.182 10.750 (continued next page) 3.140 7.000 3.679 3.692 7.000 3.520 7.000 7.000 Mean de facto Executive (1945–98) independence Mean 6.313 9.686 5.413 7.264 9.036 7.881 9.743 9.379 9.721 5.600 6.763 8.007 6.463 7.793 Expropriation risk (1982–97) S.D. IEF (1970–2005) 5.615 0.142 4.990 0.969 5.483 0.742 4.363 0.481 5.597 0.144 6.235 0.497 6.578 0.681 5.212 0.406 Mean 8.000 7.316 0.179 2.000 7.858 0.282 2.000 7.585 0.461 9.000 7.149 0.545 Mean (1999) Number of procedures 122 countries 4.861 6.492 0.000 0.000 0.000 4.531 0.000 4.533 S.D. premium (1960–99) 7.442 7.423 10.158 10.000 5.536 0.889 7.503 0.000 0.000 0.000 0.000 3.431 3.424 Black-market Mean S.D. (1962–99) capita (%) of GDP per Growth rate Mean S.D. TCI (1963–99) 2.564 0.588 0.915 5.742 40.934 77.874 14.000 5.365 1.172 1.283 0.521 2.449 4.566 4.302 0.902 1.192 4.091 96.876 66.359 1.017 0.122 2.626 1.959 1.372 0.089 1.541 5.288 1.771 0.095 1.713 9.190 1.531 0.199 2.110 2.097 7.018 1.626 0.977 5.993 1.791 0.801 6.421 5.132 13.180 11.245 4.514 4.247 3.278 3.045 3.521 1.929 0.845 1.504 6.985 12.539 12.764 5.373 1.195 2.371 4.076 29.063 36.841 15.000 5.207 0.868 1.083 0.071 2.790 1.831 1.067 0.072 3.256 4.168 26.857 21.769 1.073 0.162 2.150 2.036 2.157 0.979 1.377 8.127 147.937 137.826 7.341 2.905 0.377 3.590 32.334 84.457 20.000 5.915 1.095 Mean 13.694 2.026 0.861 3.185 Economy Data Annex Table VI.A1: TCI Based on Value Added in the Manufacturing Sector Belgium Bulgaria Brazil Botswana Bolivia Bhutan Argentina Belize Canada Cameroon Barbados Bangladesh Bahamas Albania Austria Algeria Australia Benin 341

355 9.927 S.D. 20.106 13.687 13.088 19.420 13.560 Openness (1960–2003) 54.250 10.204 46.041 14.844 26.614 16.924 50.157 9.744 30.873 5.286 47.462 25.698 56.661 70.816 59.607 40.066 10.166 38.959 9.083 64.511 7.827 48.161 93.218 15.593 32.004 10.522 53.552 Mean 104.364 8.351 102.438 19.668 101.288 15.192 104.950 101.192 17.250 4.148 5.074 1.943 2.321 7.000 3.667 3.340 5.283 5.792 7.000 3.519 3.192 Mean de facto Executive (1945–98) independence Mean 7.038 5.206 6.219 6.763 7.350 7.800 8.114 9.721 6.356 9.707 7.481 9.721 8.486 7.556 6.800 5.146 6.047 8.385 Expropriation risk (1982–97) S.D. 1.264 0.755 IEF (1970–2005) 6.468 6.730 6.327 0.680 4.944 0.470 5.963 0.231 Mean 5.000 7.371 0.462 3.000 7.268 0.502 Mean (1999) 21.000 11.000 Number of procedures 122 countries 4.456 0.000 4.064 7.510 18.000 5.282 0.256 4.550 1.939 4.035 5.028 15.000 6.394 0.532 0.000 0.000 15.000 6.645 0.432 S.D. 36.064 48.101 67.249 45.442 premium (1960–99) 3.271 0.000 7.993 2.866 1.740 4.671 1.605 0.000 0.000 6.511 11.907 5.412 Black-market 31.641 72.262 73.517 42.640 39.256 40.799 Mean S.D. 5.232 3.924 5.896 3.925 2.913 3.350 (1962–99) capita (%) of GDP per Growth rate 2.800 3.013 1.190 0.825 1.833 Mean –0.837 2.221 0.368 0.238 1.569 2.614 0.683 S.D. TCI (1963–99) 9.830 2.532 3.878 1.238 1.263 3.381 20.225 24.613 16.000 5.300 0.592 2.012 4.466 0.701 1.780 2.117 6.847 1.237 0.116 2.885 3.009 4.229 2.190 1.337 0.087 3.200 3.878 4.165 1.327 6.003 7.381 71.004 111.533 12.000 5.397 0.525 5.962 2.075 0.071 4.253 248.144 729.713 10.000 5.159 1.390 5.442 3.157 0.595 3.398 2.119 0.759 2.538 10.245 1.106 0.096 2.519 1.664 4.307 1.223 2.595 4.798 38.157 104.680 10.000 6.554 1.345 1.581 0.637 0.884 8.096 37.525 25.826 12.000 5.855 0.680 1.178 0.079 2.100 2.230 1.308 0.310 5.357 4.515 1.564 0.214 1.711 4.700 Mean 17.921 2.621 0.326 7.127 (Continued) Economy Gambia Ethiopia Central African Republic El Salvador Table VI.A1: Ecuador Dominican Republic Greece Denmark Gabon Cyprus Ghana France Croatia Côte d’Ivoire Finland Costa Rica Congo, Rep. Colombia Egypt, Arab Rep. China Fiji Chile 342

356 18.238 52.589 16.870 44.716 16.991 86.700 25.216 87.759 15.792 15.517 6.343 20.925 3.495 40.020 8.718 60.309 7.232 40.325 9.212 79.064 23.283 60.909 8.653 72.982 6.502 39.266 7.322 68.718 16.915 53.775 96.580 11.295 38.814 77.574 18.966 105.765 31.741 122.600 49.604 119.307 14.334 198.318 32.906 116.900 13.110 162.837 27.787 151.372 49.475 209.386 104.600 20.540 112.698 33.884 156.762 28.830 (continued next page) 2.981 3.735 7.000 7.000 5.381 3.250 6.959 7.000 2.208 7.000 1.571 3.684 7.000 3.140 3.333 7.475 9.079 9.721 9.700 7.044 5.156 5.413 8.150 6.556 9.721 9.457 6.406 8.069 6.863 4.686 8.513 7.875 8.569 7.056 8.488 4.694 2.400 5.088 10.000 7.703 0.105 6.180 0.359 6.321 0.542 6.906 1.102 6.669 0.893 6.236 0.663 6.242 0.556 5.956 6.609 0.817 8.000 6.489 1.059 3.000 7.491 0.642 6.000 6.200 1.023 7.000 6.819 0.382 5.000 5.686 1.283 7.000 6.622 0.818 13.000 1.383 0.466 1.423 3.795 7.090 1.634 6.266 8.125 3.350 11.000 7.071 0.316 0.000 16.000 6.422 0.656 2.899 14.000 6.335 0.698 5.448 0.399 24.015 857.111 0.375 1.233 0.600 4.892 1.172 1.750 0.000 3.399 7.233 9.133 2.724 0.001 15.251 –0.416 464.833 4.445 3.615 7.115 5.192 5.797 0.231 0.071 0.493 0.713 3.303 0.279 1.230 2.500 12.346 15.467 0.914 0.101 3.163 3.267 3.183 0.790 0.820 2.946 12.008 26.842 0.802 0.134 2.823 3.809 2.816 3.073 0.408 3.581 3.974 273.451 806.400 11.000 5.863 0.535 1.853 0.507 4.179 2.806 3.248 0.621 0.756 4.339 19.076 17.070 5.373 0.498 –1.041 4.032 15.000 21.331 17.000 5.316 0.599 1.121 0.447 4.355 1.678 1.854 0.191 3.926 3.483 0.384 0.060 2.666 4.375 3.635 0.421 2.573 3.077 26.530 24.692 10.000 5.744 0.729 0.733 0.935 5.216 209.506 270.332 0.836 0.176 3.425 16.053 82.000 127.559 1.936 0.492 1.980 7.193 8.719 2.037 3.935 6.891 1.680 0.083 4.056 3.678 1.287 0.232 2.744 3.677 14.077 17.706 1.090 0.477 –3.916 8.708 1.151 0.183 3.338 4.210 165.435 155.711 1.646 0.577 –2.515 18.460 851.008 2093.052 0.335 0.030 1.241 4.785 15.722 14.031 11.000 5.973 0.786 1.638 0.010 2.893 7.074 1.292 0.134 2.794 2.143 1.143 0.091 5.196 4.244 8.631 2.923 1.309 5.380 36.658 31.917 12.000 5.038 0.397 Hungary Lesotho Italy Honduras Iceland Kuwait India Libya Iran, Islamic Rep. Korea, Rep. Jordan Luxembourg Israel Indonesia Guyana Latvia Japan Ireland Hong Kong SAR, China Macao Jamaica Iraq Kenya Guatemala Mauritius Malta Malaysia Malawi Madagascar 343

357 6.072 S.D. 17.158 Openness (1960–2003) 51.814 5.736 57.134 57.577 11.450 60.521 11.528 49.170 24.309 32.965 4.991 73.425 3.821 51.277 10.001 34.543 5.348 31.384 16.422 56.720 24.885 48.239 20.040 93.117 12.928 82.113 33.297 15.010 80.400 6.756 Mean 100.484 12.498 154.750 27.245 114.971 16.921 120.161 32.221 122.079 25.610 133.300 24.180 3.538 7.000 3.180 3.538 2.784 7.000 4.083 3.241 4.038 1.930 3.611 7.000 3.769 3.333 Mean de facto Executive (1945–98) independence Mean 7.814 9.736 9.006 7.557 9.979 5.300 6.150 6.900 7.469 5.788 6.713 6.063 9.850 6.206 7.321 7.743 7.950 7.857 Expropriation risk (1982–97) S.D. 0.900 IEF (1970–2005) 7.656 6.041 0.405 7.125 0.440 5.448 0.271 6.239 0.351 5.400 Mean 3.000 8.000 7.620 0.305 9.000 4.659 0.915 8.000 5.190 0.632 5.000 4.000 6.890 0.534 7.000 6.811 0.590 8.000 5.648 1.496 Mean (1999) Number of procedures 122 countries 2.417 3.795 0.000 7.944 12.000 6.635 1.028 8.816 15.000 6.159 0.591 0.000 6.987 13.000 5.600 0.526 3.085 0.000 1.061 0.259 S.D. 15.557 premium (1960–99) 0.600 4.263 0.000 9.418 13.474 14.000 6.176 0.725 0.635 0.460 0.000 0.000 7.673 0.000 4.772 0.203 Black-market 15.938 –0.333 Mean S.D. 4.902 1.312 2.906 2.936 (1962–99) capita (%) of GDP per Growth rate 1.177 1.420 3.760 Mean –1.846 1.541 0.110 0.188 0.718 S.D. TCI (1963–99) 7.250 0.767 1.061 1.158 0.204 2.253 1.946 1.704 0.327 3.320 3.604 351.565 270.847 11.000 5.755 1.103 3.814 1.265 0.257 3.684 3.804 3.697 0.860 –0.258 6.501 2.852 0.450 1.598 3.634 25.390 37.524 3.711 –0.226 2.509 1.230 2.130 4.571 1.143 1.304 3.004 6.114 1.221 2.564 2.397 38.871 42.583 1.036 0.151 6.296 16.124 1.595 0.387 4.073 0.611 –1.986 10.241 0.914 0.072 3.090 1.723 3.201 0.383 1.926 4.544 9.338 6.549 0.801 7.314 86.273 109.203 2.738 0.550 2.186 4.133 5.128 1.162 0.783 4.825 36.554 64.825 1.086 0.046 0.400 5.217 169.469 158.714 16.000 5.149 0.711 4.174 0.342 1.359 2.893 33.574 34.464 2.969 0.242 1.982 3.395 Mean (Continued) Economy Panama Oman Poland Namibia Romania Paraguay Moldova New Zealand Portugal Table VI.A1: Norway Netherlands Peru Netherlands Antilles Papua New Guinea Pakistan Puerto Rico Morocco Mexico Philippines Nigeria Nepal Qatar Mongolia 344

358 6.861 5.272 5.868 10.682 19.001 18.725 11.093 17.137 13.155 70.775 17.046 49.873 64.777 8.952 35.186 9.543 62.380 14.527 60.693 28.782 53.964 16.555 50.766 56.640 77.325 13.388 35.180 12.544 30.608 16.899 48.043 9.123 69.147 22.894 35.923 7.589 58.955 13.210 83.123 18.104 89.984 48.373 51.880 94.879 29.815 28.406 5.289 118.000 12.035 110.931 146.657 30.193 6.176 7.000 7.000 3.000 2.735 4.643 7.000 3.421 3.039 7.000 2.796 3.471 3.000 4.712 5.943 1.625 7.000 2.257 3.808 5.093 6.538 9.979 9.764 5.925 4.800 9.986 6.025 7.644 9.394 7.350 8.500 6.888 7.288 6.938 9.550 6.506 6.669 9.500 6.500 7.294 5.708 7.106 6.944 5.413 5.169 4.019 0.766 0.273 0.407 0.537 0.900 0.708 7.626 8.135 5.720 6.364 4.979 0.326 5.254 4.994 8.000 5.000 4.000 9.000 5.811 0.453 5.000 3.912 0.855 9.000 6.514 0.386 7.000 8.364 0.365 7.000 8.179 0.168 9.000 6.000 6.856 0.681 9.000 5.613 0.557 6.000 5.653 1.336 20.000 14.000 3.172 0.000 0.000 0.000 6.880 0.988 4.531 16.000 5.506 0.494 7.128 2.889 2.235 11.000 6.750 0.508 4.531 8.356 0.000 20.051 50.224 11.191 62.964 independence are not available. 576.479 211.522 308.869 de facto 0.000 0.000 0.000 0.418 3.431 4.239 0.800 2.344 0.000 3.431 30.029 50.615 26.885 –1.255 520.000 128.798 129.831 4.713 8.110 7.645 1.979 1.795 8.022 3.562 1.983 5.760 2.043 2.193 2.149 2.559 0.924 2.831 –3.028 –1.259 –0.780 0.446 0.013 0.108 0.108 0.154 0.755 0.162 1.101 0.004 0.341 0.843 0.365 1.475 0.999 1.588 1.406 0.203 5.576 4.289 1.358 8.914 2.469 0.003 4.200 7.201 2.613 4.641 3.640 3.817 0.733 2.008 4.193 11.283 2.058 1.675 1.071 0.112 2.123 4.236 10.000 0.992 0.086 1.393 2.265 5.118 1.358 0.450 5.847 52.239 56.792 1.853 2.826 1.176 2.728 1.267 0.199 3.332 2.698 5.909 1.694 –0.776 4.695 85.435 119.817 6.236 0.376 2.259 3.224 198.418 301.088 11.000 5.332 1.640 1.206 0.124 2.198 1.993 3.233 0.370 1.297 2.384 86.952 92.424 13.000 5.213 1.183 6.761 1.119 5.531 87.922 155.904 2.409 0.532 0.217 6.114 14.683 2.036 0.430 0.887 4.408 11.699 26.516 10.000 6.304 0.426 9.660 2.364 1.270 6.390 4.586 0.968 1.937 4.124 18.921 20.025 13.000 5.181 0.812 2.891 1.243 3.117 3.613 27.354 41.695 Standard deviations for number of procedures, expropriation risk, and executive Tunisia Uruguay Togo Singapore United Arab Emirates Spain Venezuela, R. B. Russian Federation Sweden Syrian Arab Rep. Swaziland Turkey Sierra Leone United States Trinidad and Tobago Senegal Zambia Thailand Suriname Sri Lanka Saudi Arabia United Kingdom Zimbabwe South Africa Switzerland Slovenia Uganda Tanzania Note: Slovak Republic Sudan 345

359 346 | New Structural Economics Notes Lin (2003) constructs another index—based on the ratio of capital inten- 1. sity in the manufacturing industry and the capital intensity in the whole economy—as a proxy for measuring the degree with which a CAD strategy is pursued. That proxy is correlated highly with the current proxy and the results of empirical analyses based on that proxy are similar to the results reported in this section. The data for capital used in a country’s manufactur- ing industry are, however, available for only a small number of countries. To enlarge the number of countries in the studies, I therefore use the proxy based on the added value of manufacturing industries. 2. The samples are 86 for the 1960s, 97 for the 1970s, 107 for the 1980s and 114 for the 1990s. These fi 3. ndings are similar to the result presented in Lin (2003) using the TCI based on the ratio of capital intensity in the manufacturing industry and the capital intensity in the whole economy. Lin (2003) uses a two stage estimation to estimate the effect of the choice of the development strategy on growth. In rst stage, TCI is regressed on several variables that capture an economy’s the fi factor endowment. The residuals from this regression are used as a proxy for an economy’s deviations from a CAF strategy. They are expected to be zero if the government adopted a CAF strategy and nonzero if the strategy is CAD. The second stage consists of a cross-country growth equation where the depen- dent variable is the annual growth rate of per capita real GDP. The results show that the proxy for the development strategy has the expected negative sign cant in all regressions. The magnitudes of the impact and is statistically signifi of the development strategy on the per capita GDP growth rate in the period 1970–80 are twice those of the period 1980–92. The results suggest that if a developing country, for example, India, as shown in the appendix, adopted a CAD strategy causing TCI to be 8.47 and the residual to be 3.60, the annual per capita real GDP growth rate would be reduced 0.47 percent per year over the period 1970–92. References Deininger, K., and L. Squire. 1996. “A New Data Set Measuring Income Inequal- ity.” World Bank Economic Review 10 (3): 565–91. Djankov, Simeon, and Peter Murrell. 2002. “Enterprise Restructuring in Transi- tion: A Quantitative Survey.” Journal of Economic Literature 40 (3): 739–92. Dollar, David, and Aart Kraay. 2003. “Institutions, Trade, and Growth.” Journal of Monetary Economics 50 (1): 133–62. Easterly, W., and M. Sewadeh. 2002. Global Development Network Database, http://www.worldbank.org/research/growth/GDNdata.htm.

360 Development Strategy, Institutions, and Economic Performance | 347 Easterly, William, and H. Yu. 2000. “Global Development Growth Network Database.” Technical Report, World Bank, Washington, DC. Fei, John, Gustav Ranis, and Shirley W. Y. Kuo. 1979. Growth with Equity: The Taiwan Case, New York: Oxford University Press. Frankel, Jeffrey, and David Romer. 1999. “Does Trade Cause Growth?” American Economic Review 89 (June): 379–99. Gwartney, James, and Robert Lawson, eds. 2007. Economic Freedom of the . Vancouver: Fraser Institute. World 2007 Annual Report Hall, R. E., and C. Jones. 1999. “Why Do Some Countries Produce So Much More Output per Worker than Others?” Quarterly Journal of Economics 114 (1): 83–116. Kaufmann, Daniel, and Aart Kraay. 2002. “Growth without Governance.” Policy Research Working Paper No. 2928, World Bank, Washington, DC. Krugman, Paul, and Maurice Obstfeld. 1997. International Economics: Theory and Policy , Fourth Edition. Reading, MA: Addison-Wesley. Lewis, W. Arthur. 1965. Theory of Economic Growth (New York: Harper & Row), p. 376. Li, H. Y., L. Squire, and H. Zou. 1998. “Explaining International and Intertem- poral Variations in Income Inequality.” Economic Journal 108: 26–43. Lin, Justin Yifu, 2003. “Development Strategy, Viability, and Economic Con- vergence.” 51(2): 276–308. Economic Development and Cultural Change (Reprinted in part in this chapter.) Lin, Justin Yifu, and Mingxing Liu. 2004. “Development Strategy, Transition and Challenges of Development in Lagging Regions.” China Center for Economic Research Working Paper Series 2004–2. Peking University, Beijing. Sachs, Jeffrey D., and Andrew M. Warner. 2001. “The Curse of Natural Resources.” European Economic Review 45 (4–6) (May): 827–38. Samuelson, Paul A. 1978. c Papers of Paul A. Samuelson , The Collected Scientifi edited by H. Nagasani and K. Crowley. Cambridge, MA: MIT Press. United Nations Industrial Development Organisation. 2002. International Year- book of Industrial Statistics. Edward Elgar Publishing. World Bank. Various years. World Development Indicators. Washington, DC: World Bank.

361

362 VII Epilogue: The Path to a Golden Age of Industrialization in the Developing World nance has now ended,” as Barry Eichengreen com- “The golden age of fi 1 In my view, however, mented recently in reference to the Great Recession. the golden age of industrialization in the developing world has just begun. nancial crisis is still looming large over Europe. Newspa- The global fi pers are carrying daily reports on the anemic recovery, stubbornly high unemployment rates, downgraded sovereign credit ratings, and recurrent debt crises occurring in the wake of the recession in advanced countries on both sides of the Atlantic. Political leaders the world over are just waking up to the fact that over reliance on making fi nancial deals to maintain a high standard of living, without building and rebuilding a strong industrial base, is just a mirage. For a sustainable global recovery and robust growth in the coming years, the world needs to look beyond the Euro Area and sovereign debt wor- ries to the promise inherent in structural transformation, which, as defi ned 349

363 350 | New Structural Economics in this volume, is the process by which countries climb the industrial ladder and change the sector employment and production compositions of their economies. Except for a few oil-exporting countries, no countries have ever rst. During my travels in gotten rich without achieving industrialization fi the past three and a half years as the Chief Economist of the World Bank Group, I have been struck by the potential for less developed countries to take a page from the playbook of more successful industrializing East Asian countries, such as China, Indonesia, Japan, Korea, Malaysia, Singa- pore, and Vietnam, and to dramatically improve their development perfor- mance. My belief in the coming of a golden age of industrialization in the devel- oping world is based on the potential to rapidly expand industrial sectors in developing countries, including those in Sub-Saharan African countries, and on the dynamic relocation of industries in a multipolar growth world. The fi rst force can be envisioned through an improved understanding of the mechanics of economic transformation in modern times ushered in by the Industrial Revolution in the 18th century. In advanced countries technological innovation and industrial upgrading require costly and risky investments in research and development, because their vanguard tech- nologies and industries are located on the global frontier. By contrast, a latecomer country can borrow technology from the advanced countries at low risk and cost. Hence, if a developing country knows how to tap the advantage of backwardness, its industrial upgrading and economic growth can proceed at an annual rate several times that of high-income countries for decades as the country closes its industrial and income gap with advanced countries. The second force is the rapid wage increase in the dynamically growing emerging market economies and the unavoidable relocation of their labor-intensive manufacturing industries to other lower- income countries. Take China, for example: its monthly wage for unskilled worker is about $350. China is likely to maintain high growth in the com- ing decades (Lin 2011a). Its monthly wage for unskilled worker will reach at least $1,000 in 10 years. Such wage dynamics means China will need to upgrade to higher value added, more capital-intensive sectors, opening up a huge opportunity for other countries with income levels lower than China’s to enter the labor-intensive manufacturing industries.

364 Epilogue: The Path to a Golden Age of Industrialization in the Developing World | 351 In the UNU-WIDER annual lecture I delivered in Maputo, Mozambique, in May 2011, I explained how developing countries can capture these opportunities to achieve rapid industrialization and economic growth. The winning formula is for them to develop tradable industries that are expand- ing rapidly in countries that have been growing dynamically for decades and that have higher income and similar endowment structures to theirs. The pattern of fl ying geese is a useful metaphor to explain my vision. Since the 18th century, the successfully catching-up countries in Western Europe, North America, and East Asia all followed carefully selected lead countries that had per capita income about twice as high as theirs, and emulated the ying-geese pattern in their industrial upgrading and diver- leader-follower fl sifi cation before becoming advanced countries themselves (Lin 2011c). The emergence of large market economies such as Brazil, Russia, India and China (BRIC) as new growth poles in the multipolar world and their likely continuous dynamic growth in the postcrisis world offers an unprec- edented opportunity to all developing economies with income levels cur- rently below theirs—including those in Sub-Saharan Africa—to develop manufacturing and jump-start industrialization. China, for example, hav- ing been a “follower goose” in East Asia, is on the verge of graduating from low-skilled manufacturing jobs. Because of its size, however, China may become a “leading dragon” for other developing countries instead of a “lead goose” in the traditional fl ying geese pattern of the international diffusion of industrial development. China will free up 85 million labor- intensive manufacturing jobs, compared with Japan’s 9.7 million in the 1960’s and Korea’s 2.3 million in the 1980s (Lin 2011c). The benefi ts of reallocating labor-intensive manufacturing jobs from China and other dynamically growing emerging market economies, such as India and Brazil, to low-income countries, most of which are located in Sub-Saharan Africa, could be enormous. In 2009 alone, China exported $107 billion worth of apparel to the world, compared with Sub-Saharan Africa’s total apparel exports of $2 billion (2 percent of Chinese apparel exports). If only 1 percent of China’s production of apparel is shifted to lower-wage African countries, African production and exports of apparel would increase by 47 percent. Similarly, employment gains could be sig- nifi cant. Africa’s population (north and south of the Sahara) is 1 billion,

365 352 | New Structural Economics slightly less than India’s 1.15 billion. In 2009 manufacturing value added was 16 percent of GDP in India, 13 percent in Sub-Saharan African coun- tries, and 16 percent in North African countries such as Egypt, Morocco, and Tunisia. India’s employment in manufacturing was 8.7 million in 2009. Hence, based on a back-of-the-envelope calculation, it is reason- able to assume that total manufacturing employment in Africa is at about 10 million (Lin 2011c). This suggests that relocation of even a small share of China’s 85 million labor-intensive manufacturing jobs to Africa would provide unprecedented opportunities for Africa. But why are Chinese fi rms and lower-income country governments that t substantially from a reallocation of fi rms from China and would benefi other emerging market economies not yet organizing themselves to seize these opportunities? From my frequent interactions in the past three years with policy makers in low-income countries in Africa and Asia, as well as with business people and government offi cials in emerging market econ- omies, I know that policy makers and business communities would be rms from emerg- interested in pursuing this opportunity. Some individual fi ing markets have linked up with entrepreneurs in low-income countries to develop various labor-intensive manufacturing industries. Still, many industrialists in emerging markets are hesitant to relocate abroad, espe- cially to Africa. They cite the following concerns: (i) social and politi- cal instability; (ii) differences in labor laws and qualifi cation; (iii) poor logistics; and (iv) the lack of adequate infrastructure and business con- ditions. These soft and hard infrastructure concerns add to the risks of their investments, increase the transaction costs of their operations, and outweigh the potential benefi ts of low labor costs in Africa and other low-income countries. How to deal with these infrastructure problems? The fi rst two issues can be mitigated through the commitment and broad-based support of recipi- ent governments; the latter two could be addressed effectively through the development of sector-specifi c cluster-based industrial zones. Why is the latter sector-specifi c approach—sometimes called “picking winners”— desirable? First, the required infrastructure improvements are often industry spe- cifi c. The cut fl ower and textile industries, for example, require different infrastructure for their exports. Because the government’s fi scal resources and implementation capacity in a developing country are limited, the

366 Epilogue: The Path to a Golden Age of Industrialization in the Developing World | 353 government has to prioritize the infrastructure improvement according to the targeted industries. Second, to compete in the globalized world, a new industry not only must align with the country’s comparative advantage so that its factor costs of production can be at the lowest possible level, but the industry also needs to have the lowest possible transaction-related costs. Suppose a country’s infrastructure and business environment are good and industrial upgrad- ing and diversifi cation happen spontaneously. Without the government’s rms may enter into too many different industries that are coordination, fi all consistent with the country’s comparative advantage. As a result, most industries may not form large enough clusters in the country and may not be competitive in the domestic and international market. Only in the wake of many failures might a few clusters eventually emerge. Such “trial and error” is likely to be a long and costly process, reducing the individual domestic and foreign fi rms’ expected returns and incentives to enter new industries or relocate to other countries. This in turn can slow down or even stall a country’s economic development. But there exists a long list of failed attempts to pick winners. These failures, as discussed in the previous chapters, were often the result of the inability of government to come up with good criteria for identifying industries that are appropriate for a given country’s endowment structure and level of development. In fact, governments’ propensity to target indus- tries that are too ambitious and not aligned with a country’s comparative advantage largely explains why their attempts to “pick winners” resulted in “picking losers” (Lin, 2011d). The recipe to economic success therefore is the one that helps policy makers in developing countries identify the industries in which their economies may have a latent comparative advantage and remove binding constraints to facilitate private domestic and foreign fi rms’ entry to and operation in those industries. Chapter III of this book provides the govern- ments in developing countries with a pragmatic and easy-to-follow growth identifi cation and facilitation framework to do so. Many low-income countries have an abundance of natural resources. They may also benefi t from the industrialization opportunity provided by the industrial upgrading in dynamically growing emerging market economies by following the “fl ying geese” pattern. Resource-intensive industries, such as extraction, provide very limited job opportunities.

367 354 | New Structural Economics In a visit to Papua New Guinea in 2009, I found that its famous OK Tedi copper mine generated 40 percent of the country’s public revenues and 80 percent of its exports but provided only 2,000 jobs in 2009. Most Guinea’s 6.6 million people still live on subsistence agri- of Papua New culture. Their wage rate is low, and wages constitute the major cost of production for labor-intensive industries. Low wage, natural-resource- rich countries could therefore develop labor-intensive industries, creat- ing much needed jobs. Indonesia is a good example showing that this is possible. Labor-intensive manufacturing industries not only offer the potential to absorb surplus labor from the rural subsistence sector, but the development of such industries can also pave the way through con- tinuous upgrading to higher value added industries. Finland’s Nokia, for an example, started as a logging company and diversifi ed its operation to the labor-intensive business of producing rubber boots; it then became the original equipment manufacturer of household electronics for Phil- lips before venturing into mobile phones. Still, resource-rich countries often suffer from the Dutch Disease, as export receipts from natural resources push up the value of the cur- rency, thus adversely affecting the competitiveness of their other exports. Sometimes also the wealth from natural resources is captured by powerful groups, turning resource richness into a curse. At the same time natural- resource rents can provide a great opportunity for development if man- aged in a transparent way and prudently invested in human and physical capital, such as infrastructure, and used to diversify to nonresource sectors as suggested in the growth identifi cation and facilitation framework. These investments, if well chosen, can increase labor productivity, reduce pro- duction and transaction costs and ultimately cure the Dutch Disease, and turn the abundance in natural resources from a curse to a blessing. This is because such countries have opportunities to accumulate capital, upgrade endowments, improve infrastructure, transform industrial structure, and subsequently raise incomes faster than labor-abundant, resource-poor countries (Lin, 2011b). The discussion so far has been on the opportunity of and ways to achieve rapid industrialization in low-income countries. The new struc- tural economics also offers new insights to middle-income countries about how to upgrade their industries and achieve dynamic growth. A unique

368 Epilogue: The Path to a Golden Age of Industrialization in the Developing World | 355 feature of middle-income countries is that some of their industries will still locate within the global frontier and some of their industries will locate on the frontier because of the graduation of higher-income countries from those industries. For the former industries, the government can follow the growth identifi cation and facilitation framework to assist the private fi rms to tap into the potential of the latecomer’s advantage, and for the latter industries, the government should adopt the same measures as those in the advanced countries for supporting innovation in technology and industries. Commonly used measures include support for basic research, providing patent protection, mandated use of new technology/products, and direct government procurement of new products. If a middle-income country can implement these measures to facilitate private fi rms’ industrial upgrading and diversifi cation, the country can not only avoid the middle- income trap but also achieve dynamic growth and catch up to advanced countries in a generation. The discussion so far has not discussed the technological innovation and productivity improvement in agriculture. In low-income countries, where most people work in agriculture, improving agriculture will be important not only for reducing poverty but also for generating economic surplus to support industrialization. Governments need to facilitate the innovation and extension of agricultural technology and improvement of infrastruc- ture for agricultural production and commercialization. Finally, as stated in the introduction, I am convinced that, every develop- ing country, including those in sub-Saharan Africa, has the potential to grow at 8 percent or more continuously for several decades, to signifi cantly reduce poverty, and to become a middle-income or even a high-income country in the span of one or two generations, if its government has the right policy framework to facilitate the private sector’s development along the line of its comparative advantages and tap into the latecomer’s advantages. I hope this book will help developing countries to realize their growth potential. A world without poverty will then become a reality instead of just a dream. Note 1. http://www.project-syndicate.org/series/the_next_financial_order/long_ description.

369 356 | New Structural Economics References Lin, Justin Yifu. 2011a. “Demystifying the Chinese Economy,” Cambridge, UK: Cambridge University Press. ———. 2011b. “Economic Development in Resource-Rich, Labor-Abundant Economies,” Feb. 28, 2011, Let’s Talk Development Blog, http://blogs. worldbank.org/developmenttalk/economic-development-in-resource-rich- labor-abundant-economies. ———. 2011c. “From Flying Geese to Leading Dragons—New Opportunities and Strategies for Structural Transformation in Developing Countries,” WIDER Annual Lecture 15, Helsinki: UNU-WIDER. (A shorter version of this paper is forthcoming in Global Policy. ) ———. 2011d. “Picking Winners,” Let’s Talk Development Blog, Oct. 18, 2011, http://blogs.worldbank.org/developmenttalk/node/670.

370 Index gures, notes, and tables are indicated with b, f, n, and t following the page number. Boxes, fi alternative development strategies, 295–310. A See also comparative advantage acceleration of economic growth, 1 characteristics of, 296–301 access to credit. See fi nancial structure choices of, 307–9 Acemoglu, Daron, 89, 290, 291 capital intensity in manufacturing Advanced Technology Program (U.S.), 153 sector as basis, 320–21 See also c regions and countries Africa. specifi economic growth and, 326–32, comparative advantage, t 329–31 208–9 t economic volatility and, 332, 333 economic growth, 84 empirical testing, 320–39 failures of state to facilitate industrializa- f enterprise autonomy and, 322–25, 326 tion, 18, 159, 208–9 government intervention in resource industry-specifi c nature of investment in, allocation and, 322 68–69 income distribution and, 333–37, international trade costs, 31 f, 334 336 f light manufacturing report (World Bank), f price distortions and, 321–22, 323–25 228–29 b proxies used for institutions, 321, manufacturing industries, 352 341–45 t African Center for Economic trade openness and, 325–26, 327 f Transformation (ACET), 195 transition and economic performance, Aghion, P., 205 337–39, 338 t agriculture transition strategy, 309–10 rural productivity growth, 49 Amoako, K. Y., 183–84, 192, 205, 208–10 technology and, 22–23, 169 n 7, 355 Amsden, Alice, 183, 184, 188, 207–8 aid effectiveness, disappointment with, 20 apparel industry. See textiles and garments 2 Akamatsu, K., 311 n industry Allen, Franklin, 268 357

371 358 | Index Arrow, Kenneth J., 60 C c regions and countries Asia. See specifi CAD (comparative-advantage-defying) Asian Tigers, 2, 3, 35, 90 compara- strategy, 296–98. See also asparagus market, 175 37 n tive advantage Atkinson, Anthony, 290 CAF (comparative-advantage-following) automobile industry, 127, 133, 156, 157, com- See also strategy, 298–301. 13 300–301, 313 n parative advantage 44 Ayyagari, M., 176 n Calomiris, Charles, 278 capital accumulation, 24, 28, 29, 35, 40 n 3, B 123, 302–3, 307 balanced budgets, 29 capital-intensive industries, 8, 18 Balassa-Samuelson theorem, 75 comparative-advantage-defying approach Baldwin, Robert E., 50 and, 28, 55, 118 Bangladesh’s garment industry, 174 n 36 comparative-advantage-following fi bank-based structure. nancial structure See approach and, 135 n 7, 83 Barro, R.J., 41 fi nancial structure changes in response to Basu, Susanto, 290, 291 growth of, 33, 269–70, 280 Beck, Thorsten, 266 3, LDC governments focusing on, 40 n Becker, G.S., 89 288, 309 best practices, 92 market failures in developing countries, 11 “big push,” 120 n 1, 312 n 27, 117 binding constraints, identifi cation of, progression from labor-intensive indus- 163–67 tries to, 157 Growth Diagnostics framework and, 12 capital investment, 27, 42 n 19–20, 92–93 capital/labor ratio, 21–22 methodology for, 107–8 12 n 13 n Carneiro, P., 42 t Nigeria, 240–43, 246–56 n Cass, D., 107 6 physical infrastructure, 30 catch-up strategies, history of, 151–60, “black glass ceiling,” 192 194–95 “black-market premium,” 321–22, 323 f, Chang, Ha-Joon, 119, 120, 130, 137, 151, 341–45 t 153, 211–12 n Bluestone, B., 170 17 Chang, Morris, 191 Bolivia, shock therapy’s success in, 313 n 17 children, costs and additions to economy bond markets. fi nancial structure See from, 207–8 Botswana Chile economic growth, 84, 91 comparative advantage and government natural resource wealth, 42 10 n support, 157–58 Bourguignon, F., 176 n 45 economic growth, 13, 84 brain drain, reversal of, 191–92 rms’ advantage, 71 large fi Brazil’s economic growth, 2, 13, 84 wine industry, 174 35 n BRIC economies (Brazil, Russia, India and China China), 183, 207, 351 automobile industry, 300–301 Britain and industrialization, 151, 154, 156, average wage by sector, 239 t 277 compared to Nigeria, 220–21 f, c nature of Burkina Faso, industry-specifi 230, 230 t investment in, 68 economic growth, 2, 13, 84, 90–91

372 Index | 359 economic reform favoring privatization, as part of catch-up strategy, 195 72 transition strategy to CAF approach, 309–10 351 t, exports of, 55, 68, 77–78, 232–33 comparative-advantage-following (CAF) nancial structure, 276–77 fi 36 foreign direct investment in, 174 n approach, 117, 131–32, 133, 135–37 industrialization and comparative advan- tage, 77–78, 119 characteristics of, 298–301 c investment as economic industry-specifi compared to CAD strategy, 302–7 growth strategy, 69 ned, 295 defi information industry, 127 competitive advantage vs., 24, 27, 30, 19, 118–19 n 108–9 labor-intensive manufacturing industries, n East Asian Four Little Dragons, 308–9. 24, 351 172 Hong Kong SAR, China; labor surplus, 67–68 See also rst” approach to economic “micro-fi Korea, Republic of; Singapore; Taiwan, China development, 310 facilitating with vision and realism, new growth theory and, 290 Pareto improvement contributing to 136–37 factor endowments and, 24, 28, 50, 75, economic development, 127 136, 353 poverty rate, 2 as foundation for industrial policy, 25, steel industry, 136 transition from “follower goose” to 50–51, 101–2, 105–6, 129, 288 growth and industrial upgrading, 6, 14, “lead goose” status, 351 23–24, 99–100, 113–14 18–19 nn transition strategy, 310, 313–14 higher-income country exploiting lower- TVEs (township-and-village enterprises), n income country’s comparative 20 310, 314 advantage, 40 wage dynamics, 250 classical economists, 86 historical development of theory, 121 cation of countries in “old” structural latent, 52, 75–76, 78, 145, 155, 163 classifi market failures and, 55, 71, 114–16 economics, 28 “old” structural economics’ failure to collective action, 25, 50, 69, 148 college wage premium, 89 properly account for, 48, 66 risks of deviating from, 131–32, 138 Colombia’s investment in Ecuador, 174 n state assistance to private sector to 36 comparative advantage, 8, 113–39 exploit, 116–17 adjustment costs and technological success of entrepreneurs linked to, 75 differences undermining, 126–27, viability’s relationship to, 291–95, 292–93 f, 295 f 130–31 choice of strategy. See alternative devel- b comparative value chain analysis, 228–29 opment strategies compensation mechanisms lacking for comparative-advantage-defying (CAD) trade-adjustment process, 122 competitive advantage, 24, 27, 30, 41 n 5, approach, 54, 117–18, 127, 16, 175–76 n 41, 185 n 76, 108–9 138–39, 170 n 19, 118–19 characteristics of, 296–98 competitive markets compared to CAF strategy, 302–7 price system in, 24, 288–89 defi ned, 295 technology adoption and, 292–93, drawbacks of, 205 293 f

373 360 | Index convergence vs. divergence, 16, 20, 89–91, neoclassical growth theory and, 146 289–90 collective co-ordination by government. See historical background, 287 action human capital and, 301–2 co-ordination problems for developing industrial policy, 23, 113–39 countries, 115 labor growth rates, 274, 280 corruption, 335 macroeconomic crises, 35 Corruption Perception Index, 244 nance, 262 market-based systems of fi cost-benefi t analysis, 49, 70 migration and unemployment, 67 Cull, Robert, 273–74, 278, 279 path to industrialization, 24, 349–56 returns on investment projects, D 41–42 n 8 Daewoo (Korean manufacturer), 174 36 n successes not attributed to development Deaton, A., 89 policies, 4 technology and. technological capa- See Decision Tree approach, 19–20, 92 bilities; technology transfer defense-related R&D funding in U.S., 153 development thinking, 6, 66–79. Deininger, K., 333–34 See also New Structural Economics Demirgüç-Kunt, Asli, 261, 262, 266, choices of strategies, 307–9 271, 272 Deng Xiaoping, 69 developing countries, premises about, dependency theories, rejection of, 28 29 n 172–73 free market approach, 19 developed countries (DCs). See high-income goals of new development thinking, 4 countries developing countries, 287–347 historical overview of, 17 import substitution, 18 backwardness as advantage, 6, 23, 28, structuralist approach to, 17 36, 73, 102, 109 n 21, 114, 150, n 173–74 as subdiscipline of economics, 3 33, 350 bank structure advantageous to, 33, Washington Consensus, 19, 38, 57, 66 distortions, role of, 36, 38, 51–52, 55, 288, 261–62, 269, 277, 281 compensation mechanisms lacking for 6 297, 311 n trade-adjustment process, 122 “black-market premium,” 321–22, 323 f, 341–45 coordination problems for, 115 t economic growth, 2 Diwan, I., 290 Dollar, David, 328 fi nancial repression and, 33 new structural economics approach Domar, E., 16 to, 30 Dutch Disease, 354 trajectory, unique vs. general, 26–27 E endowment structure, 22 East Asia failures of state to facilitate industrialization, 144, 159, fi nancial crisis (late 1990s), 74, 307, n 16 173 313 n 30 government-driven industrialization, 155 gap with industrialized countries, 3, 14, newly industrialized economies (NIEs), 28, 53 157, 290 endowment gaps, 147 role models in industrial policy, 191, GDP per capita, 90 knowledge gaps, 58–59 192–93

374 Index | 361 Easterly, William, 335, 338 F Eastern Europe factor endowments, 21, 22, 24, 28, 29, protection of nonviable fi rms, 310 39 n 2, 50, 75, 136, 146–47, 194, 9 n soft budget constraint and, 312 299, 340, 353 n 8, owers, 169 Ecuador’s exporting of cut fl Far East role model, 189 36 174 n FDI. See foreign direct investment education policy, 37, 61, 197 Fei, John, 306 effi cient markets Feyen, Erik, 271, 272 facilitating state and, 103, 145–50 Fields, Gary, 306 ciency, 60 static vs. dynamic effi fi nancial deepening, 305–6 Egypt’s industrialization program, 160 fi nancial development and new structural Eichengreen, Barry, 349 economics, 32–33 n endogenous growth theory, 16, 39–40 3, nancial structure, 7, 261–81 fi 59, 62 1, 73–74, 88, 98–99, 116, n bank-based systems, 261–62, 266 290 advantageous to developing countries, endowments and economic development, 33, 269, 277, 281 21–24, 38, 58. See also factor match in size of business to size of endowments; natural resource bank, 269 wealth bank-based view, 264–65 change with capital accumulation or belief of government leaders and, 275–76 population growth, 24, 208 conventional empirical results, 266–67 comparative advantage and, 28, 50, deviation from optimal fi nancial struc- 146–47 ture, 275–80 gaps between developing and developed differences from country to country countries, 147 by stage of development, 267, international trade and, 35–36 270–71, 281 LDCs’ need to focus on upgrading endogeneity of, 273, 275 endowment structure, 288 fi nancial liberalization policies, 277 as limitation, 59, 288 nancial-structure-irrelevancy view, fi total endowments, 39 n 2 263–64 energy costs, 42 n 8 importance to development, 263 Enlightenment, 86 labor-intensive vs. capital-intensive indus- entrepreneurship, 59, 60, 70, 74–75, 208 tries and, 33, 269–70, 280 Ethiopia law and fi nance view, 264, 275 average wage by sector, 239 t market-based systems, 261–62, 266 industry-specifi c nature of investment, advantageous to developed countries, n 8, 175 n 37 68–69, 169 270, 272, 277 owers, 175 Ethiopia’s exporting of cut fl n 37 fi rms benefi ting from, 279–80 19 n European Commission, 153, 171 market-based view, 265 Europe’s industrial policies, 153–54, nancial services as country mixtures of fi 171 n 19 grows richer, 268 exchange rate policies, 60, 75, 78 new empirical results, 271–75 Export Processing Zones Development OECD countries and, 273 Authority (Mauritius), 158 politics, effect of, 278–79 international trade; See exports. c specifi poverty and, 274–75 countries theoretical arguments, 267–71

375 362 | Index fi nancial structure ( continued ) nancial crisis, 14, 57, 58, 84, global fi traditional view of, 263–65 n 1, 143–44, 349 106 types of, 261–62 globalization, 28, 35–36, 59, 69, 96–97, Finland 211–12 government assistance to Nokia, 128–29, Goldsmith, Raymond W., 262 131, 132–33, 354 Greenwald, B., 58, 59 state-led industrialization, 154 n Griffi 3 n, Keith, 311 rst movers’ risk, 100 fi gross domestic product (GDP) per capita scal policy and new structural economics, fi n 6 aws in, 63 fl 29–31 gap between developing and developed owers, fl ower industry, exporting of cut fl countries, 90 n 37 n 8, 174 n 36, 175 169 history of growth of, 1, 84, 145–46, 187 fl ying-geese pattern of economic develop- regional comparisons, 86–87 2, n ment, 126, 131, 135, 136, 311 3, Growth Commission, 27, 95, 106 n 351, 353 n 108 The Growth See also 16. foreign direct investment (FDI), 34–35, 74, Report 78–79, 101, 126, 174 39, n 36, 175 n Growth Diagnostics, 19–20, 92–93, 161, 191, 193 165–67, 240 former Soviet Union, 72 cation and Facilitation Growth Identifi 9 n soft budget constraint and, 312 Framework (GIFF), 6–7 framework for growth identifi cation and Amoako comments on, 192–96 ca- Growth Identifi See facilitation. Amsden comments on, 188–92 tion and Facilitation Framework comparative value chain analysis, (GIFF) b 228–29 framework for rethinking development, Lim comments on, 201–204 New Structural 13–42. See also Nigeria, 218, 225–30 Economics Pack comments on, 196–201 France and industrialization, 151, 154, 156, rejoinder to comments, 204–14 n 20 171 state’s role, 69–70, 78, 160–67 Frankel, Jeffrey, 328 steps in, 161–62, 181–82 Frederick the Great, 152 Tendulkar comments on, 186–88 free market approach, 19 te Velde comments on, 181–86 Fundacion Chile, 158, 174 n 36 The Growth Report: Strategies for Sustained Growth and Inclusive (Commission on Development G Growth and Development), 85, Gale, Douglas, 268 See garments industry. textiles and garments 94–98 industry ndings, 101–4 key fi Germany recommendations of, 96–97 value of, 105 industrialization, 152, 156 warnings against “bad ideas,” 98, as model for post-Meiji Japan, 210 n 8 103–4 vocational education and training, 169 11, 265 n Gerschenkron, A., 26, 42 growth theories Growth Identifi See GIFF. cation and Facili- challenges for, 105 convergence vs. divergence, 16, 20, tation Framework Glaeser, E., 89 89–91

376 Index | 363 n 3, endogenous growth theory, 16, 39–40 investment in textiles and garments 59, 73–74, 116 industry in Mauritius, 158 in historical perspective, 15–16, new growth theory and, 290 85–88 household consumption behavior, 30 modern economic growth, 144, Hsingchu Science-based Industrial Park 146, 218 36 n (Taiwan, China), 174 new directions in applied growth human capital, 24, 36–37, 301–2 research, 89, 91–93 Hume, David, 86 overview, 289–91 Hungary’s attempted catch-up strategy, 158 4 n structural change and, 169 Hyundai Automobile Company, 313 n 13 systematic growth analysis, 16 technological change, theory of, 16 I IEF. See index of economic freedom imperfect factor mobility, 126 H import substitution, 18, 36, 51–52, 55, 197, Haber, Stephen, 278 n infrastructure 15 hard infrastructure. See also 313 2 n income distribution, 306–7 examples of, 168 n incubation programs, 162 Harrison, B., 170 17 index of economic freedom (IEF) and Harrod, R.F., 16 25, expropriation risk, 321, 322, Hausmann, R., 19–20, 161, 165, 172 n f, 324–25 190, 240 341–45 t Heavy and Chemical Industrialisation pro- India automobile industry, 300–301 gramme (Korea), 133 business development’s effects on heavy industries, 197, 212, 309, 311 n 5 13, 89 n Heckman, J.J., 42 economic development, 267 compared to Nigeria, 220–21 230, 230 f, Heckscher-Ohlin-Samuelson theory, t economic growth, 2, 13, 84 121–23, 126, 130, 193, 194 Heckscher-Ohlin theory, 59 emulating Soviet industrialization, 197, 210 high-income countries nance, 262 bank-based systems of fi exports of, 232–33 t capital-intensive industries as characteris- fi nancial regulations, 58 tic of, 269–70 labor-surplus economy, 190 economic development, 2–3 manufacturing industries, 352 endowment structure, 22 specialization areas in information 37 24, 175 n industry, 172 n historical background, 287 labor growth rates, 274 trade policies, shortcomings in, 304 nance, 262 market-based systems of fi Indonesia compared to Nigeria, 220–21 state’s role in facilitating industrialization 230, f, cation, 151–55 6 n and diversifi t, 257 230 Hirschman, A.O., 61 nancial crisis (late 1990s), East Asian fi 16 313 Hong Kong SAR, China n comparative-advantage-following (CAF) economic growth, 13, 84 approach, 308–9 exports of, 232–33 t East Asian fi nancial crisis (late 1990s), labor-intensive industries, 354 16 n 313 industrial clusters/parks, formation of, 79, n 27 economic growth, 2, 172 162, 174 n 36, 240, 353

377 364 | Index An Industrial Competitiveness Policy for as binding constraint on growth, 30, 352 (European ciencies, effect of, 31 ineffi the European Union c nature of, investment in, industry-specifi Commission), 153 high-income 25, 50–51, 68, 70–71, 101, 352–53 industrialized countries. See countries as part of economy’s endowments, 22, 198 industrial policy in developing countries, 113–39 11. n innovation, 55, 114–16, 152–53, 170 See also research and development comparative-advantage-defying (R&D) approach, 18, 28, 117–18 International Commission on the comparative vs. competitive advantage, 118–19 Measurement of Economic Performance and Social Progress, dynamic process of industrial upgrading, 6 14, 134–36 63 n international trade growth and industrial upgrading, 6, 18, 49, 113–14 comparative advantage and, 121, 197 15 n export promotion, 313 fi nancial structure changes, 33 market failures that block innovation, infrastructure ineffi ciencies, effect of, 31 new structural economics and, 35–36 114–16 3 n openness in, 36, 56, 304–5 mobility of capital and, 40 recovery from global economic crisis and, trade dependence ratio as index f, for, 321, 325–26, 327 349–50 state assistance to private sector to 341–45 t exploit comparative advantage, post-World War II views of, 17–18 116–17 pros and cons of trade liberalization, Industrial Revolution, 151 122, 131 protection and trade liberalization, 28, industrial upgrading comparative advantage and, 6, 14, 36, 52, 78, 124, 126 intertemporal trade, 34 23–24, 113–14, 125 investment climate as obstacle to business continuous process of, 135–36, 144, 209, 44 n development, 163–65, 176 290 rms, 51, 71 favoring of larger fi human capital and, 301–2 J knowledge needs intrinsic to, 199–200 Jäntti, M., 154 public support for, 14, 134–35 Japan risk in, 23, 137, 148, 149 automobile industry, 156, 300–301 state’s role See state’s role in. n economic growth, 2, 170 17 timing in economic development process, Germany as post-Meiji model for, 210 51, 67 government support for industrialization, ciencies, effect of, 31 ineffi 198–99, 209–210 ciency, problem of, 24 x-ineffi large fi rms’ advantage, 71 information fl ow for industrial upgrading, Meiji Period and industrialization, 25, 29, 50, 54, 70, 100–101, 115, 76–77, 152, 170 n 15, 197 148, 298 MITI policies in 1950s and 1960s, 198, n information industry, 126–27, 175 37, 210, 301 n 311 5 trade protection, 124, 133 infrastructure J-curve vs. L-curve growth path, 72

378 Index | 365 Jiang, Ye, 268 L Jones, C.I., 27, 94 labor-intensive industries 3 Ju, J., 40 n comparative advantage of, 268 n 20 Juppé-Roccard report (France), 171 moving to capital-intensive industries, 157, 269–70, 280 K in natural-resource-rich countries, 354 Kaldor’s characteristics of 20th century relocation to lower-income countries, growth, 94 350, 351 Kaufmann, Daniel, 328 rm sizes and, 269 smaller fi Keyder, C., 167 unskilled, 67, 70, 76, 148–49 Keynesianism, 3, 4, 18, 29 labor migration of skilled workers as Klinger, B., 172 25 n means of technology transfer, knowledge 151 gaps between developing and developed labor regulation as binding constraint, countries, 58–59 175 n 38 needs intrinsic to industrial upgrading laissez-faire economics, 15, 48–49, and diversifi cation, 199, 291 115, 121 6 n Koopmans, T.C., 107 Lange, Oskar, 138 Korea, Republic of Laos’s economic reform, 72 automobile industry, 301 late-comer status as advantage, 6, 8, 23, comparative advantage and, 77, 119, n 21, 114, 150, 28, 36, 73, 102, 109 131–33, 139, 202–4 33, 350 173–74 n comparative-advantage-following (CAF) Latin America. See also c countries specifi 14 approach, 308–9, 313 n debt crisis, 4, 18, 19, 57 nancial crisis (late 1990s), East Asian fi economic growth, 2, 84 16 313 n failures of state to facilitate industrializa- economic growth, 2, 13, 60 tion, 18, 159 government support for industrialization, international trade and, 17 157, 198, 212–13 saving and investment rates, 97 heavy machinery and heavy chemical trade policies, shortcomings in, 304 industries, 197, 212, 309 leadership and governance, 97 industrialization and trade policy, “learning society,” 59, 60, 73, 124 127–28 Lee, Jong-Wha, 304 industrial upgrading as continuous Leibenstein, H., 24 process, 135 Leipziger, Danny, 95 labor surplus and migration, 68 Lesotho and industry-specifi c nature of new growth theory and, 290 investment, 68–69 steel industry, 135–36, 139 n less developed countries (LDCs), 40 3, targeting industries in Japan, 156, See also 287–347. developing 197, 203 countries Kornai, Janos, 312 9 n Levine, Ross, 261, 262, 266, 271, 272 Kpodar, Kangni, 274 Lewis, W. Arthur, 115, 189–90, 340 Kraay, Aart, 328 Lim, Wonhyuk, 184, 201, 212 Krueger, Anne, 48–53, 66–72 Lin, Justin Yifu, 40 n 3, 105, 113, 120, 134, Kuo, Shirley W.Y., 306 204, 268 Kuznets, Simon, 306 Lisbon Strategy, 153

379 366 | Index 12 n List, F., 151, 170 Mckinnon, R., 305 Lucas, Robert, 35, 37, 89, 113–14, 15, n Meiji Period in Japan, 76–77, 152, 170 290, 291 197 Memoranda of Understanding for sector, 244, 258 M n 8 Mexico’s debt crisis of 1980s, 19 macroeconomic crises, 35 macroeconomicists, 86 microeconomic analyses, 31, 166 Middle East macroeconomic policy of industrialized countries, 197–98 economic growth, 196 macroeconomic stability, 97, 102, 306 role models in industrial policy, 190–91 5, 172 n 27, n Maddison, A., 84, 158, 169 210, 211 middle-income countries, 2, 126–27 “middle-income trap,” 2 Malaysia migration East Asian fi nancial crisis (late 1990s), of rural labor to unskilled labor-intensive 313 16 n economic growth, 13 industries, 67 government reform of industrial of skilled labor with knowledge, 151 policy, 191 MITI policies in 1950s and 1960s, 198, 210, 301 information industry, 127 saving and investment rates, 97 MIT Poverty Lab, 21, 93 modern economic growth, 144, c nature of invest- Mali and industry-specifi 146, 218 ment, 68–69 Mokyr, J., 90 Malthusian conditions, 85 17 n Mondale, Walter, 170 market allocation, 4, 97, 103 market failures monetary policy and new structural economics, 31–32 comparative advantage not utilized in, 71 effect of, 58, 69 Monga, Célestin, 78, 83, 143, 181, 204 innovation blocked by, 55, 114–16 multilateral institutions and development agencies, 4, 19 as lessons to others, 100, 149, 160 Murphy, Kevin M., 312 11 reasons for, 201 n n Myint, Hla, 189, 190 reforms need to address, 107 11 state’s role and, 155 N Marshall, Alfred, 205 10, 59, 74, n natural resource wealth, 31, 42 Marxism, 15 207, 353–54 mature industries, in state-led structural- CAD strategy and, 309 change strategies, 144, 156, 202 Mauritius neoclassical growth theory, 14, 16, 21–25, comparative advantage and government 30, 88 entry into new industry and, 124 support, 158 foreign direct investment and, 34 27 n economic growth, 13, 84, 91, 172 gaps between LDCs and DCs and, industry-specifi c nature of investment, 289–90 68–69 limits of subscribing to, 133–34, 137, 38, 208 n labor regulation, 175 rms’ advantage, 71 138 large fi Mauritius Industrial Development Author- new industry, entry into, 101, 124, 148. See ity (MIDA), 158 also industrial upgrading

380 Index | 367 procedures required to obtain legal prioritization of value chains, 235 f status, as index of enterprise promoting growth in selected value f, autonomy, 321, 322–25, 326 chains, 240–43 t screening potential subsectors, 238 341–45 t n 10 viability of, 312 sectors in which Nigeria has potential New Structural Economics (NSE), 5, 6, comparative advantage, 236–40 13–42 technical and vocational education and benefi ts of, 38 training, 241 differences from “old” structural eco- t top imports, 234 nomics, 27–29, 54 trade policy reform, 241 transparency and accountability, 244 difference with earlier literature, 26–27 fi nancial development and, 32–33 Nokia, 124–25, 128–29, 132–33, 354 North, Douglass, 89 scal policy and, 29–31 fi foreign capital and, 34–35 O goals of, 38 O’Brien, P., 167 human development and, 36–37 OECD countries and fi nancial structure, international trade and, 35–36 273 Krueger comments, 48–53 Ohno, Kenichi, 77 monetary policy and, 31–32 OPEC development role model, 190, 207 policy insights and, 29–37, 60 openness principles of, 98–101 to global economy, 96–97 rejoinder to comments, 67–79 in international trade, 36, 56, 304–5 Rodrik comments, 53–56 similarities to “old” structural econom- P ics, 27, 218 Pack, Howard, 184, 196, 205, 210–12 state’s role See state’s role in. n 8 Pakistan’s power generation costs, 42 Stiglitz comments, 56–65 Papua New Guinea and natural resource Nigeria, 7, 217–56 10, 354 n wealth, 42 access to fi nance, 241 patents, 149–50, 151, 152, 153 average wage by sector, 239 t Penn World tables, 89, 105 binding constraints, identifi cation of, per capita income 240–43, 246–56 t Japan, Singapore, and other Asian coun- business environment, 241 1 n tries, 311 comparison of countries with per capita as proxy for abundant capital and labor, income 100-300 percent above 300–301, 300 t 5 t, 257 n Nigeria, 230–35, 230 perfect factor mobility, 122–23 t, economic growth, 219–20 219–25, 37 n Peru’s asparagus exports, 175 f, 222 t 220–21 pioneer countries as models and targets, employment and incomes, 217, 222–25, 156, 210 t 223–24 pioneer fi rms, 25, 29, 41 n 6, 50, 79, governance issues, 243–44 147–50, 156, 162, 229, 240 Growth Identifi cation and Facilitation Pohang Iron and Steel (Korean company), Framework, 218, 225–30 135 Memoranda of Understanding for sector, politics, effect on fi nancial structure, n 244, 258 8 278–79 physical infrastructure, 241

381 368 | Index n 5, 76, 118, 190 Porter, M.E., 41 S POSCO (state-owned steel mill in the SABIC (Saudi Arabia’s state-owned petro- Republic of Korea), 133 chemical company), 191 poverty and poverty reduction 17, 335 n Sachs, Jeffrey D., 313 countries moving out of vs. countries Saggi, Kamal, 199 remaining in poverty, 13–14 Sala-i-Martin, X., 83 fi nancial structure and, 274–75 Samsung, 133, 136 2 n global fi nancial crisis’s effect on, 106 saving and investment rates, 97, 102–3, growth’s relationship with poverty 197, 291 reduction, 61 Schultze, Charles, 160–61 Prebisch, R., 17–18 Schumacher, E.F., 290 price system in competitive markets, 24, Schumpeter, Joseph, 62 n 1 288–89, 292–93, 293 f semiconductor industry, 127, 157, 191 Pritchett, L., 107 n 11 Sewadeh, M., 338 profi ts and surplus, 116, 147 shadow prices, 20, 166 public sector quality, 60 Shaw, E. S., 305 11 n Shleifer, Andrei, 89, 312 R shock therapy, 310, 313 n 17 Rajan-Zingales difference-in-difference Singapore approach, 273 comparative-advantage-following (CAF) randomized control trials (RCTs), 21, 93, approach, 308–9 14 n 108 nancial crisis (late 1990s), East Asian fi Ranis, Gustav, 306, 308 16 n 313 rational expectations revolution, 19 economic growth, 2, 13 rational expectations theory, 4 human capital development, 37 research and development (R&D), 153, new growth theory and, 290 212–13, 270, 290, 303, 350 saving and investment rates, 97 research support from state, 41 n 6 Singer, H., 17–18 Ricardian equivalence trap, 30 Singh, Raju, 274 Ricardo, David, 121, 126 Slovenia’s economic reform, 72 risk in industrial upgrading and Smith, Adam, 204 diversifi cation, 23, 100, 137, socialist systems 148, 149 7 n CAD strategy and rationing, 311 Rodríguez, Francisco, 291, 304 collapse, 4, 18–19 Rodrik, Dani, 19–20, 53–56, 66, 67, soft budget constraint and, 312 n 9 75–78, 109 n 20, 193, 240, 290, trade policies, shortcomings of, 304 291, 304 societal transformation, 61 Romer, David, 328 9 soft budget constraint, 312 n Romer, P.M., 27, 40 n 3, 84, 89, 94, 155, Solow, Robert M., 58, 86, 89, 289 290 Solow-Swan model, 16, 87, 107 n 5 Rosenstein-Rodan, P., 17 South Asian attempts to facilitate Rostow, W.W., 26, 86, 186 industrialization, 18, 159, 188, 206 rural development, 49 Southeast Asian saving and investment Russia’s attempted catch-up strategy, 158 rates, 97 Rwanda and industry-specifi c nature of South Korea. See Korea, Republic of investment, 68–69 Spence, Michael, 95

382 Index | 369 Squire, L., 333–34 developing countries’ differences from ation, 4, 18 developed ones, 53 stagfl 4 n in growth literature, 169 state’s role, 3, 143–68 importance of, 3, 5, 14, 23 in advanced economies, 150–55 state’s role in dynamics of, 143–68 Amoako comments on, 192–96 Sub-Saharan Africa Amsden comments on, 188–92 agricultural sector, 3 economic importance of, 295 comparative value chain analysis, facilitating state as best enabler for indus- b 228–29 trial upgrading and technological economic growth, 2, 3, 196 advance, 103, 114, 116–17, 121, nancial policies, 33 fi 145–50, 353–55 investment climate as obstacle to business cation and Facilitation Growth Identifi Framework, 69–70, 78, 160–67 development, 163 identifi cation of binding constraints, 20, manufacturing industries, 352 state’s role See subsidies. 163–67 index of economic freedom (IEF) and Sun, Xifang, 268 sustainable growth, 1, 3 expropriation risk as indexes for, ts and, 61 t citizens’ long-term benefi f, 341–45 321, 322, 324–25 commitment needed for, 96 industrial upgrading supported by poverty reduction and, 14 state subsidies, 134–35, 137, Sweden’s compensation mechanisms for 149–50, 162 trade-adjustment process, 122 in industrialized nations, 152–53 Switzerland and watch-making industry, 59, Lim comments on, 201–204 73–74 market failures and, 55, 114–16, 155, 159–60, 192, 288, 353 Syed, Mahmood, 308 in new structural economics, 15, 29, 48, 54, 100–101 T in “old” structural economics, 28 Taiwan, China Pack comments on, 196–201 comparative advantage and, 190 public sector quality, 60 comparative-advantage-following (CAF) recipe for success, 29, 57, 155–60 14 n approach, 308–9, 313 rejoinder to comments, 204–14 nancial crisis (late 1990s), East Asian fi n research support from state, 41 6 n 313 16 in structuralist approach to economic economic growth, 2 development, 17, 67 government support for industrialization, Tendulkar comments on, 186–88 198–99 te Velde comments on, 181–86 income distribution, 306 types of government interventions, large fi rms’ advantage, 71 145, 168 new growth theory and, 290 steel industry, 135–36, 139 targeting industries in Japan, 156, 197 Stiglitz, Joseph E., 56, 58, 59, 66, 67, Takatoshi, Ito, 311 n 2 72–75, 78, 290, 313 17 n technological capabilities. See also See also stock markets, 263, 270. nancial fi technology transfer structure catch up with more advanced countries, structural change 123 categories of level of growth for, 26 industry-specifi c nature of, 125

383 370 | Index technological capabilities. See also history of economic growth, 186, continued ) technology transfer ( 278, 308 misplaced focus on, 290 incentives for innovation and research, not equal in all countries, 194 152–53 technological change in growth analysis, 16, research and development (R&D) 22–24, 87–88 funding, 153 3 n in less developed countries (LDCs), 40 subprime crisis, 279 nontradable goods and services and, unskilled labor-intensive industries, 67, 4 41 n 70, 76 n 3 ned, 169 technology, defi UNU-WIDER Lecture (May 4, 2011), 8 technology choice index (TCI), 320–37, industrial upgrading See upgrading. 334 323–27 f f, USSR’s industrialization, 197, 210, 211 See also late-comer technology transfer. Uzbekistan and economic reform favoring status as advantage privatization, 72 CAF vs. CAD strategy, 303–4 from DCs to LDCs, 290 V economic policies to encourage, 60, 198 Vartiainen, J., 154 industrial upgrading and, 24–25 Velasco, A., 19–20, 240 mismatch in LDC when technology used venture capital, 270 by unskilled workers, 291 viability, 292–95 skilled labor migration as means of, 151 costs and, 292–93 f Tendulkar, Suresh D., 183, 186, industry and product choices, 295 f 204, 205–6 Vietnam te Velde, Dirk Willem, 181 t average wage by sector, 239 textiles and garments industry, 69, 71, 77, compared to Nigeria, 230, 230 t 36, 189, n 135, 152, 157, 158, 174 economic growth, 13, 91 201, 213, 351 economic reform favoring privatization, Thailand 72 East Asian fi nancial crisis (late 1990s), t exports of, 232–33 n 313 16 n Vishny, Robert W., 312 11 economic growth, 13 power generation costs, 42 n 8 W 17 n Thurow, L., 170 3 n Wang, Y., 40 tradable goods, identifi cation of, 41, 162, Warner, Andrew M., 335 173 n 32, 192, 202, 229, 351 Washington Consensus, 19, 38, 57, 66, trade. See international trade 17 n 313 transition strategy, 72, 309–10 The Wealth of Nations (Smith), 204 transparency, 244 Weil, David, 290, 291 Treichel, Volker, 181, 217 Western industrial success, history of, 151–55, 186–87 n wine industry, 174 U 35 World Bank Unilever, 192 United Kingdom. See also “Doing Business” indicators, 164 Britain and industrialization study (2005) on complexity of economic fi nancial structure as model, 277 growth, 91–92 United States World Development Report (1998–99), 58 nancial structure as model, 277, 278 fi World Trade Organization, 189

384 Index | 371 X Z ciency, problem of, 24 11 n Zagha, R., 107 x-ineffi n Xu, Lixin Colin, 261, 267, 273–74, 279 Zambia and pioneer fi rms, 169 9 Zellner, A., 98 3 Y n Zhang, P., 40 Zilibotti, Fabrizio, 290, 291 Yamamura, Kozo, 198 “yellow glass ceiling,” 191–92 Yu, H., 335

385 ECO-AUDIT Environmental Benefits Statement The World Bank is committed to preserving Saved: endangered forests and natural resources. The • 12 trees Office of the Publisher has chosen to print • 4 million BTU of total New Structural Economics: A Framework energy on for Rethinking Development and Policy • 1,149 pounds of net recycled paper with 50 percent post-consumer greenhouse gases (CO 2 waste, in accordance with the recommended equivalent) standards for paper usage set by the Green • 5,184 gallons of waste Press Initiative, a nonprofit program support- water ing publishers in using fiber that is not sourced • 329 pounds of solid waste from endangered forests. For more informa- tion, visit www.greenpressinitiative.org.

386 New New Structural Economics “This splendid collection of essays, by one of the world’s outstanding experts on economic development, puts to work a newly emerging view, which he has helped to shape, of why in recent decades some countries have prospered while Structural others have languished. Lin’s focus is on countries that were all economically underdeveloped six decades ago, but his analysis offers strong hints about future prospects of the rich world as well. His style is dispassionate and unadorned by drama, which makes the essays all the more moving and illuminating.” Economics — Sir Partha Dasgupta Frank Ramsey Professor Emeritus of Economics, University of Cambridge “ New Structural Economics is a truly important and ambitious book. Justin Lin, A Framework for Rethinking with some help from other distinguished scholars, has succeeded in laying out the complex structural microeconomic dynamics of economic growth, diversification Development and Policy and development, and in capturing the crucial complementary roles of government as investor, regulator, coordinator of activity and expectations, and guide. All of this is set in a global economy that is itself in the midst of massive structural change. This book will become an essential reference for scholars and for policy makers not only in developing countries, but also, increasingly, in developed countries.” — Michael Spence 2001 Nobel Prize in Economics William R. Berkley Professor in Economics and Business, New York University Leonard N. Stern School of Business “The World Bank has long been committed to the goal of achieving a world In this brilliant volume, its Chief Economist, Justin Yifu Lin, lays without poverty. out an economic agenda for how to make this dream a reality. He argues that the successes of China can be achieved elsewhere around the world, and explains clearly and forcefully the structural transformations that will be required and The book the role that government can and must play in that transformation. Justin Yifu Lin will be a landmark in rethinking development. It provides an alternative to the now discredited Washington Consensus policies that guided the Bretton Woods Institutions for years. Justin Lin’s ideas have already stirred discussion and debate. This book will ensure that they will continue to be central in the reexamination of developmental policy.” — Joseph Stiglitz 2001 Nobel Prize in Economics Justin Yifu Lin University Professor, Columbia University ISBN 978-0-8213-8955-3 90000 3 38955 9780821 SKU 18955

Related documents

CityNT2019TentRoll 1

CityNT2019TentRoll 1

STATE OF NEW YORK 2 0 1 9 T E N T A T I V E A S S E S S M E N T R O L L PAGE 1 VALUATION DATE-JUL 01, 2018 COUNTY - Niagara T A X A B L E SECTION OF THE ROLL - 1 CITY - North Tonawanda TAX MAP NUMBER ...

More info »
Numerical Recipes

Numerical Recipes

Sample page from NUMERICAL RECIPES IN C: THE ART OF SCIENTIFIC COMPUTING (ISBN 0-521-43108-5) Permission is granted for internet users to make one paper copy for their own personal use. Further reprod...

More info »
book

book

Copyright Cambridge University Press 2003. On-screen viewing permitted. Printing not permitted. http://www.cambridge.org/0521642981 You can buy this book for 30 pounds or $50. See http://www.inference...

More info »
UL White Book

UL White Book

GUIDE INFORMATION FOR ELECTRICAL EQUIPMENT THE WHITE BOOK 2015-16 UL PRODUCT CATEGORIES CORRELATED TO THE 2011 AND 2014 NATIONAL ELECTRICAL CODE® UL’s General Guide Information is updated daily. To co...

More info »
625137

625137

2018-19 Nebraska All-Sports Record Book - Nebraska Communications Office -

More info »
Price Book: Canvas Office Landscape Wall and Private Office

Price Book: Canvas Office Landscape Wall and Private Office

Price Book Y Canvas Office Landscape ® Prices effective January 7, 2019 Published May 2019 Wall and Private Office page 2 Introduction Canvas Office 3 Landscape Office Wall and Private 5 Walls Work Su...

More info »
LaneEtAlPrivacyBigDataAndThePublicGood

LaneEtAlPrivacyBigDataAndThePublicGood

This is a prel version of the book Privacy, Big Data , and the Public Good : iminary er, and for Lane, Victoria Stodden, Stefan Bend nt, ed. Julia Helen Niss enbaum Frameworks Engageme Unive rsity Pre...

More info »
~edelman.indb

~edelman.indb

Deuteronomy–Kings emerging as Authorit Ative BooK s A Conversation Edited by Diana V. Edelman riente Ancient n ast m onographs – m onografías sobre el Antiguo Cercano o ear e society of Biblical Liter...

More info »
Not as small

Not as small

A Faithful Reply to the from a Former CES Employee NEW AND IMPROVED! JIM BENNETT WITH 37% MORE Wri$ng for KEOKUK, IOWA REFERENCES! CANONIZER.COM A September, 2018 line-by-line response to the October ...

More info »
20181003 AuthorsAllianceGuidePublicationContracts

20181003 AuthorsAllianceGuidePublicationContracts

Brianna L. Schofield & Robert Kirk Walker, Eds. Bridge · Diaz · Hagen · Kuksenkova · Nikogosyan Samuelson Law, Technology, and Public Policy Clinic UNDERSTANDING AND NEGOTIATING BOOK PUBLICATION CONTR...

More info »
academic books and their futures jubb1

academic books and their futures jubb1

THE ACADEMIC BOOK OF THE FUTURE Michael Jubb s Academic Book and the ir Future June 2017 Library A Report to the AHRC & the British

More info »
History and Authenticity of the Bible

History and Authenticity of the Bible

History and Authenticity of the Bible By Dr. David Hocking Brought to you by The Blue Letter Bible Institute http://www.blbi.org A ministry of The Blue Letter Bible http://www.blueletterbible.org

More info »
AndersBehringBreivikManifesto

AndersBehringBreivikManifesto

2011 , London – By Andrew Berwick

More info »
CHURCH HISTORY IN THE FULNESS OF TIMES Student Manual

CHURCH HISTORY IN THE FULNESS OF TIMES Student Manual

HURCH C ISTORY H HURCH C H ISTORY IN THE ULNESS F IN THE ULNESS F OF T IMES OF IMES T S tudent M anual S anual M tudent RELIGION 341 THROUGH 343

More info »
CESReply

CESReply

A REPLY FROM A FORMER CES EMPLOYEE FOREWORD As a freshman at the University of Southern California, I was first exposed to what is commonly referred to as “anti-Mormon literature.” I read “The Godmake...

More info »
2017 2018 Bluebook Complete

2017 2018 Bluebook Complete

IDAHO BLUE BOOK 2017-2018 Published by SECRETARY OF STATE LAWERENCE DENNEY for the STATE OF IDAHO

More info »
508Chap4book.book

508Chap4book.book

U.S. Geological Survey Techniques of Water-Resources Investigations Book 9 Handbooks for Water-Resources Investigations National Field Manual for the Collection of Water-Quality Data Chapter A4. COLLE...

More info »