32692 Economic Growthin the1990s Learning from a Decade of Reform THE WORLD BANK Economic Growth in the 1990s Learning from a Decade of Reform Economic Growth in the 1990s Learning from a Decade of Reform Washington, D.C. © 2005The International Bank for Reconstruction and Development /TheWorld Bank 1818 H Street, NW Washington, DC 20433 Telephone 202-473-1000 Internet www.worldbank.org E-mail feedback@worldbank.org All rights reserved. 1 2 3 4 08 07 06 05 This volume is a product of the staff of the International Bank for Reconstruction and Development / TheWorld Bank.The findings, interpretations, and conclusions expressed in this paper do not necessar- ily reflect the views of the Executive Directors ofTheWorld Bank 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 judge- ment 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 publication is copyrighted.Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law.The International Bank for Reconstruction and Development /TheWorld Bank encourages dissemination of its work and will normally grant permis- sion to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, 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,TheWorld Bank,1818 H Street NW,Washington,DC 20433,USA;fax:202-522-2422; e-mail: pubrights@worldbank.org. ISBN-13: 978-8-213-60434-5 ISBN-10: 0-8213-6043-4 e-ISBN: 0-8213-6044-2 Library of Congress Cataloging-in-Publication Data Economic growth in the 1990s : learning from a decade of reform. p. cm. This report was prepared by a team led by Roberto Zagha, under the general direction of Gobind Nankani. Includes bibliographical references and index. ISBN 0-8213-6043-4 (pbk.) 1. Developing countries--Economic policy. 2. Economic stabilization--Developing countries. 3. Privatization--Developing countries. 4. Developing countries--Politics and government. I. Zagha, Roberto. II. Nankani, GobindT., 1949­ III.World Bank. HC59.7.E295 2005 338.9'009172'4--dc22 2005043405 Contents Foreword xi Acknowledgments xv 1 Overview 1 Part 1: Facts of the 1990s 2 Grist and the Mill for the Lessons of the 1990s 29 Country Note A: Economic Growth from the Very Long-Term Perspective of History 56 3 Something Special about the 1990s? 59 Country Note B: Lessons from Countries That Have Sustained Their Growth 78 Part 2: Development Controversies of the 1990s 4 Macroeconomic Stability: The More the Better? 93 Country Note C: Poverty and Inequality:What Have We Learned from the 1990s? 120 5 Trade Liberalization: Why So Much Controversy? 131 Country Note D: The Middle East and North Africa: Performing below Potential 156 6 Privatization and Deregulation:A Push Too Far? 163 Country Note E: Eastern Europe's Transition: Building Institutions 196 7 Financial Liberalization: What Went Right,What Went Wrong? 203 Country Note F: Lessons and Controversies from Financial Crises in the 1990s 238 v vi C O N T E N T S 8 Policy Reforms and Growth Performance: What Have We Learned? 249 Country Note G: Africa's Growth Tragedy: An Institutional Perspective 270 9 Improving Public Sector Governance:The Grand Challenge? 275 Country Note H: Natural Resources:When Blessings Become Curses 304 10 Does Democracy Help? 309 Bibliography 331 Index 357 Tables 2.1 Growth Regressions and "Policy" Impacts, with Two Country Examples 36 2.2 Despite Their Rapid Growth, China,Vietnam, and India Rank Low on Many Measures of Institutional Quality 39 2.3 Growth Rates Differ Enormously across Countries over Periods from One Decade to FortyYears 42 2.4 A Growth Rate of 5.7 Percent perYear Higher for 20Years Would Roughly Triple a Country's per Capita Income 43 2.5 Episodes of Rapid Growth Set in Context 45 2.6 "Syndromes and Symptoms" Summary of the Empirical Growth-Regression Literature 48 3.1 Growth in Developed and Developing Countries, 1960s­1990s 60 3.2 Developing Countries' Growth, 1990s: Regional Perspectives 61 3.3 Global Integration, 1980­2000 62 3.4 Unprecedented Growth of World Trade, 1990s 63 3.5 Exports and Imports of Goods and Services as Shares of GDP, 1980­2000 65 3.6 Diversification Took Place before the 1990s 66 3.7 Reduction in Nontariff Barriers in Developing Countries, 1990s 72 3.8 Tariff Dispersion Decline in the 1990s 73 3.9 Capital Account Restrictions Were Progressively Dismantled, 1970­97 74 B.1 Growth Successes in the 1990s 80 B.2 Economic Successes: Steady Growth, 1960­2002 83 B.3 InflationVolatility 84 B.4 Real Exchange RateVolatility 84 B.5 Total Factor Productivity Growth of Successful Countries, 1960­2000 86 B.6 Progress on Social Indicators, 1980­2000 89 C.1 Surveys Tracking Individuals over Time Show Only a Small Portion of Poverty Is Accounted for by People Who Are Always Poor 124 5.1 Tariff Reductions and Changes in Goods Trade Integration, 1990­2000 136 5.2 Rich Countries Levy Higher Tariffs on Poor Countries' Exports 150 D.1 Progress on Social Indicators, Middle East and North Africa, 1980­2000 157 D.2 Economic Growth in the Middle East and North Africa, 1960­2003 158 D.3 Three Major Middle Eastern and North African Country Groups 158 D.4 Economic Growth in the Middle East and North Africa:Impact of Natural Resources 159 TA B L E S vii 6.1 Winners and Losers from Divestiture in 12 Case Studies 171 6.2 Performance Changes in Privatized Firms in Mexico 172 6.3 Investment in Infrastructure Projects with Private Participation in Developing Countries, 1990­2002 175 6.4 Market Shares of the Three Largest Generation,Transmission, and Distribution Companies inVarious Countries, 2000 180 F.1 Financial Inflows and Major Financial Shocks 243 F.2 Debt Dynamics in Crisis Economies, Cumulative Change,ThreeYears Before . . . and ThreeYears After 246 8.1 Sectoral Example of Direct and Background Institutions of Policymaking 255 8.2 Examples of Misuse of Discretion 256 8.3 Efforts to Limit Government Discretion 256 8.4 Example of the Dependence of Appropriate Policy on Institutional Conditions 259 8.5 Policy Reform and Growth: Sources of Differential Impacts 262 G.1 African Growth in Context:Average Annual Growth Rates of Real per Capita GDP, 1960­2001 271 G.2 Annual Growth in 17 African Countries, 1975­2003 274 9.1 Examples of Efficiency Gains from ICT 289 9.2 Types of States and Entry Points for Strategic Interventions:A Governance Typology 299 Figures 1.1 Worldwide Growth in Real GDP per Capita, 1000­Present 1 1.2 Economic Growth in Perspective, 1960­2002 3 1.3 Regional Perspectives on Growth in the 1990s 4 1.4 Fraction of World Inequality Accounted for by Differences across Countries 4 2.1 Depth and Duration of the Transformational Recession: Eastern European and Former Soviet Union Countries 31 2.2 Depth of the Recession, Ratio of Current to Pretransition Output, and Relationship with Distance from Brussels 32 2.3 Interest Rate Differentials Did Not Predict the Magnitude of the Impending Devaluation of Three East Asian Currencies 33 2.4 Growth Was Much Slower in the 1980s and 1990s than Predicted by Empirical Models That Linked Growth to Policy Reform 34 2.5 Although Nearly Every Country in Latin America and the Caribbean Has Pursued Economic Reform, Growth Has Been Slow 35 2.6 Accelerating Growth in China, India, andVietnam 38 2.7 Poverty Reduction Was Rapid in India, China, andVietnam in the 1990s 39 2.8 Enrollment Rates of Children Aged 7­14 in Brazil Rose Substantially for All Income Groups--But Most Dramatically for the Poorest 40 2.9 Stable Growth in Industrialized Countries 41 2.10 Fraction of World Income Inequality Explained by Differences across Countries 42 2.11 There Is Some, but Weak, Correlation of Growth Rates across Decades 44 2.12 Growth Episodes in Mauritius, 1950­2000 46 3.1 Growth in Developed and Developing Countries, 1963­99 59 3.2 Growth Slowdowns in Developed and Developing Countries 63 3.3 Export Shares of GDP, 1980­2002 64 3.4 Faster Integration into World Trade during the 1990s 64 viii F I G U R E S 3.5 More Competitive Real Exchange Rates 64 3.6 Developing Countries Diversified into Manufactures, 1960s­1990s 65 3.7 Developing Countries' Exports of Manufactures, 1981­2001 66 3.8 No Large Terms-of-Trade Shocks for Developing Countries, 1990s 66 3.9 Decline in Nonenergy Commodity Prices 67 3.10 Oil Prices Were Lower in the 1990s than in the 1970s and 1980s 67 3.11 Capital Flows to Developing Countries Expanded in the 1990s 67 3.12 Capital Flows Were Driven by a Surge in FDI and Portfolio Equity Flows 68 3.13 Nominal and Real Interest Rate, 1980­2002 68 3.14 Developing Countries' Parts and Components Exports Grew Faster in the 1990s 69 3.15 Capital Flows to a "Median" Developing Country as a Percentage of GDP, 1970­2002 70 3.16 Developing Countries Paid Less Interest On External Debt in the 1990s 70 3.17 Reduction in Tariffs in Developing Countries, 1980­2000 72 3.18 Financial Sector Liberalization, 1973­96 73 3.19 Effects of Liberalizing the Financial Sector, Developing Countries 75 B.1 Investment as a Share of GDP 81 B.2 External Debt as a Share of Gross National Income 84 B.3 Integration with the World Economy, 1970­2000 85 B.4 Average Annual Growth of per Capita Income of Different Income Groups, 1980 to Mid/Late 1990s 88 B.5 Ratio of Real per Capita Income of Bottom 40 Percent to That of Top 20 Percent, 1980 to Mid/Late 1990s 88 4.1 GDP GrowthVolatility, 1966­2000 94 4.2 Structure of GDP GrowthVolatility, 1961­2000 94 4.3 Inflation Rates, 1991­99 95 4.4 High Inflation in Developing Countries, 1961­99 95 4.5 Current Account, 1966­2000 96 4.6 Developing Countries' Overall Fiscal Balance 96 4.7 Primary Fiscal Balance, 1990­2002 97 4.8 Developing Countries: Seigniorage Revenues, 1966­2000 97 4.9 Real Exchange RateVolatility, 1961­2000 98 4.10 Developing Countries: Exchange Rate Crises, 1963­2001 98 4.11 Volatility of Net Capital Flows, 1977­2000 99 4.12 Developing Countries: Sudden Stops in Net Capital Inflows, 1978­2000 99 4.13 Government Debt, 1990­2002 101 4.14 Total Fiscal Costs of Systemic Banking Crises as a Percentage of GDP 102 4.15 Developing Countries' Foreign Currency Debt, 1997 and 2001 103 4.16 Emerging Markets Bond Index Spreads for Latin and Non-Latin Borrowers 103 4.17 Pro-Cyclicality of Public Consumption, 1980­2000 105 4.18 Central Bank Independence in Developing Countries, 1975­98 107 4.19 Dollarization of Deposits, 1996 and 2001 107 4.20 Ex Post Real Interest Rates, 1990­2001 108 4.21 Incidence of Systemic Banking Crises, Developing Countries, 1981­2000 109 C.1 Infant Mortality Rates 123 5.1 Temporary Labor Mobility, Underused Mode of Trade in Services 134 5.2 Changes in Export Shares of GDP and Changes in Tariffs, 1990­2000 137 D.1 Median and GDP-Weighted Economic Growth in the Middle East and North Africa, 1961­2000 157 F I G U R E S ix D.2 Investment in the Middle East and North Africa, 1960­2002 159 D.3 Tariffs in the Middle East and North Africa, 1980­2000 160 D.4 Trade Outcomes in Egypt, Morocco, and Tunisia, 1960­2000 161 D.5 Diversification in Egypt, Morocco, and Tunisia, 1960-2002 161 6.1 Market Shares of Revenues for U.S. Long-Distance Carriers, 1984­98 165 6.2 Access to Basic Services, by Income Group--Ghana, Mexico, and Peru 174 6.3 Infrastructure Projects with Private Participation, 1990­2001 176 6.4 Growth in Latin American Telecom Lines 176 6.5 Optimal Size of U.S. Generating Plants 177 6.6 Postprivatization Labor Productivity in Electricity Distribution in Chile, Argentina, and the United Kingdom 178 6.7 Railroad Cargo in the Transformation Period, 1985­97 181 6.8 Labor Productivity in the Water Sector 184 6.9 Local Exchange Carriers in Latin America 185 6.10 Cost-Coverage Ratios of Electricity Prices in the CIS Countries 186 6.11 Eastern Europe's Population Is Significantly Older than Latin America's 189 6.12 Coverage is Greater in Europe and Central Asia than Latin America and the Caribbean, but IntraregionalVariation is Wide 190 6.13 Eastern Europe Has Greater Fiscal Need for Reform than Latin America 190 6.14 Reforms in Latin America Reduced Debt: Projected Pension Debt (Explicitly Accumulated after 2001) 191 6.15 Bolivia's Pension Reform Was Unexpectedly Costly 191 6.16 No Marked Rise in Participation in Latin America 191 E.1 Key Indicators for Transition Countries 198 7.1 Increase in Average Deposits/GDP in Major Countries, by Regions, 1960s­90s 204 7.2 Changes in the Ratios of Bank Assets and Liabilities Plus Capital to GDP, 1990s 208 7.3 State Ownership in Banking, 1998­2000 212 7.4 Selected Financial Crises, 1980­99 216 F.1 Interest Rate Spreads and Real Exchange Rates in Crisis Countries 240 8.1 Simulated Impacts of Policy Reform on the Level and Growth Rate of Output 251 8.2 The Elements of Policy Action 254 8.3 Diagnosing the Problem of Low Levels of Investment and Entrepreneurship 264 G.1 Africa: Getting Poorer over Decades 270 G.2 Africa's Seven Biggest Economies:Volatile and Unstable 273 G.3 Africa: Rebounding in the Late 1990s 273 9.1 Citizens and Politicians 276 9.2 Politicians and Policy Makers and the Bureaucracy 278 9.3 Bureaucrats and the Citizenry 278 9.4 Classification of States by Governance Profile 297 H.1 Natural Resources and Growth, 1970­89 305 10.1 Electoral Rules in Richer and Poorer Democracies 312 10.2 Political Systems in Richer and Poorer Democracies 313 10.3 Indicators of Political Market Imperfections in Countries Holding Competitive Elections, 1995 314 10.4 Newspaper Circulation and Corruption 314 10.5 ContinuousYears of Competitive Elections and Corruption 316 10.6 Gross Secondary School Enrollment and ContinuousYears of Competitive Elections 317 x B OX E S Boxes 2.1 Per Capita Growth in the 1990s: Forecast and Actual 30 2.2 How Money and Power Can Influence Patterns of Institutional Development 51 4.1 Devaluation of the CFA franc 106 C.1 Perceptions of Fairness in Allocating Opportunity Are Central (Case Study: Sri Lanka) 127 5.1 Trade Policy over the Centuries 133 5.2 The Trade and Growth Debate 135 5.3 The Impact of Foreign Direct Investment on Growth 138 5.4 Jamaica and Mauritius: Institutions and Macroeconomic Stability Make the Difference 140 5.5 Behind Chile's Success:A Less than Orthodox Approach 145 6.1 China: Stealth Privatization 168 6.2 Cellular Phone Operators in Rural Bangladesh 177 6.3 Brazil: No Rain or Privatization Gains 179 6.4 Controversial Power Purchases in Indonesia 180 6.5 Problems with Unbundling in Railroads 182 6.6 India's Regulatory Capacity versus Effectiveness 185 6.7 Definitions 187 6.8 Argentina: Private Accounts Do Not Protect Workers from Government Risk 192 7.1 India--A Successful Liberalizer with Strong Capital Markets 207 7.2 Nonbank Financial Intermediaries (NBFIs) in the 1990s 213 7.3 Problems with the Process of Financial Liberalization 217 7.4 Indonesia: Early Liberalization andWeaknesses Related to Political Connections 218 7.5 Bank Privatization in Mexico 220 7.6 Bank Restructuring and Privatization in Sub-Saharan Africa 221 7.7 Extending Credit for Small Borrowers 230 9.1 Bright Line Rules versus Standards 280 9.2 Integrated Justice Sector Reforms: The Jamaican Case 281 9.3 Fiscal Transparency and Developing Countries 284 9.4 SUNAT in Peru:A Modest Success 286 9.5 The Enclave Conundrum in Uganda 286 9.6 Building a Sense of Calling and Commitment in Public Service Delivery: Ceará, Brazil 287 9.7 Mexico's e-SAT Program for More Efficient Tax Administration 288 9.8 What It Takes to Create a Successful Anticorruption Agency 290 9.9 The Report Card Survey in Bangalore, India: Stimulating Administrative Reforms 292 9.10 Service Delivery and Civil Society in the West Bank and Gaza 293 9.11 Procurement Watch:Working with the "Enemy" 293 9.12 Hard Budget Constraints:The Challenge of Fiscal Decentralization in Argentina 295 9.13 Investigative Journalism: Lifestyle Checks of Public Officials 296 10.1 Clientelism, Credibility, and Politics 315 10.2 Political Parties and Reform 325 Foreword When you get right down to business, there aren't too many poli- cies that we can say with certainty deeply and positively affect growth. --Arnold Harberger, July 2003 (IMF Survey, 216) Therefore, the real lesson for the architects of growth strategies is to take economics more seriously. --Dani Rodrik, September 2003 (Growth Strategies, 30) A T THE START OF THE 1990S, smaller role for governments more generally. Pri- economists thought the road vatization and deregulation were taking hold in ahead was clear.What for many the United Kingdom,in the United States,and in countries had been the"lost decade"of the 1980s Eastern Europe and the former Soviet Union. made it evident that government interference in Williamson had emphasized that the Consensus the economy--through price controls, foreign was to be applied judiciously, not mechanistically, exchange rationing, distorted trade regimes, but it quickly took on a life of its own, becoming repressed financial markets, and state ownership the expression of what economists both inside and of commercial enterprises--wasted resources and outsideWashington thought developing countries impeded growth. Hence, the logic went, rolling needed for growth and development.This think- back the state would lead developing countries to ing guided much of the advice by theWorld Bank sustained growth. and was reflected in the conditionality associated Much of this vision was reflected in the with adjustment loans. Some of its key aspects "Washington Consensus." Articulated by John were reflected in the World Bank's 1991 World Williamson in 1990,the Consensus was meant to Development Report, although that report stressed synthesize the reforms that most economists in the importance of achieving the right balance the World Bank, the International Monetary between government and market, rather than Fund, the U.S.Treasury, and some of Washing- choosing between them, and was generally more ton's think tanks believed were needed to rescue nuanced on the impact of specific reforms.The Latin American countries from cycles of high Washington Consensus was not the only point of inflation and low growth. view among economists. But it was the dominant When the Consensus was formulated,the cur- view,making it difficult for others to be heard,and rent of opinion was already shifting toward a it provided the framework for many of the reforms xi xii F O R E WO R D implemented during the 1990s by a wide spec- the art of economic policy making.The range of trum of countries around the world. options puts the onus on economic analysis to The results of these reforms were unexpected. guide policy making effectively. In dealing with They exceeded the most optimistic forecasts in growth processes, economists have no formula. some cases and fell well short of expectations in They have broad principles and tools--in the others.Although implemented in a manner that same way that principles and tools can be used to departed from conventional wisdom--in terms build an airplane. If those are not appropriately of speed and design of reform, large presence of put to use, the airplane may not fly, or may not the state and, until very recently, high levels of weather storms well.The manner and sequence import protection--domestic liberalization and in which economic principles and tools are used outward orientation were associated with spec- will determine whether specific growth country tacular growth, poverty reduction, and social strategies will succeed or not. progress in East and South Asia.At the same time, This volume is part of a three-pronged exer- booms and busts continued in Latin America and cise theWorld Bank undertook to learn from the extended to East Asia and other regions as well. experience of the 1990s from three perspectives: For most countries emerging from the former (1) analytical (this book); (2) policy (13 policy Soviet Union,the 1990s will be remembered as a makers who were at the forefront of policy imple- costly and traumatic decade. Sharp declines were mentation in the 1990s drew lessons from their followed by a prolonged and as yet incomplete experience during a one-year cycle of lectures at recovery,with results varying from relative success the Bank); and (3) operational (13 former Bank in the Czech Republic, Hungary, and Poland to country directors drew lessons from their work at costly transitions in most other countries.Africa the Bank in a series of papers, to be published did not see the take-off that was expected at the separately). From all three perspectives, growth beginning of the decade, although many coun- was at the center of the discussion.An institution tries showed signs of recovery in the late 1990s. whose primary business is finance and advice for Costly financial crises rocked Mexico (1994),East poverty reduction needs to understand what Asia (1997),Brazil (1998),the Russian Federation causes growth and what sustains it. Poverty (1998), Turkey (2000), and Argentina (2002). declines rapidly where growth is rapid and sus- Some countries managed to sustain rapid growth tained.Poverty stagnates where growth is tepid.A with just modest reforms, and others could not few exceptions notwithstanding, the unambigu- grow even after implementing a wide range of ous impact of rapid growth on poverty reduction reforms. was confirmed again in the 1990s. However, Interpreting the reasons for this wide varia- growth is difficult to predict because it implies tion is the central task of this report.A common social transformation: a break with past trends, interpretation has been that countries that grow behaviors, and institutions that reflect deep forces have reformed enough, and countries that have in societies and how they organize themselves. not achieved sustained growth have not The findings of the analysis confirm and build reformed enough.But for many economists and, on the conclusions of an earlierWorld Bank report, perhaps more important, the policy makers they The East Asian Miracle (1993), which reviewed advise, this interpretation is not entirely satisfac- experiences of highly successful East Asian tory. Unquestionably, macroeconomic stability, economies. They confirm the importance for domestic liberalization, and openness lie at the growth of macro-stability,of market forces govern- heart of any sustained growth process. But the ing the allocation of resources, and openness. But options for achieving these goals vary widely. they also emphasize that these general principles Which options should be chosen depends on ini- translate into diverse policy and institutional paths, tial conditions,the quality of existing institutions, implying that economic policies and policy advice the history of policies, political economy factors, must be country-specific and institution-sensitive the external environment, and last but not least, if they are to be effective. F O R E WO R D xiii The central message of this volume is then that an appropriate response to this challenge, and that there is no unique universal set of rules. Sustained while economic policy advice should be cog- growth depends on key functions that need to be nizant of the strengths and weaknesses of institu- fulfilled overtime: accumulation of physical and tions and downside risks, it should not be human capital, efficiency in the allocation of influenced by mistrust. In September 2004, 16 resources, adoption of technology, and the sharing well-known economists gathered in Barcelona of the benefits of growth.Which of these functions and issued a new consensus that reflects their views is the most critical at any given point in time, and on growth and development.1 The Barcelona hence which policies will need to introduced, Consensus echoes much of the findings of the which institutions will need to be created for these World Bank's work, which in turn reflects recent functions to be fulfilled, and in which sequence, academic research by several of the signatories. varies depending on initial conditions and the We expect this change in thinking to influ- legacy of history.Thus we need to get away from ence operational decision making in the World formulae and the search for elusive"best practices," Bank and aid agencies in general. In the Bank in and rely on deeper economic analysis to identify the last few years, these perspectives have been the binding constraints on growth.The choice of translated into new analytical and operational specific policy and institutional reforms should instruments such as poverty and social impact flow from these growth diagnostics.This much analysis and country-driven poverty reduction more targeted approach requires recognizing strategies, which seek to bring analytical rigor country specificities, and calls for more economic, and empirical accuracy to the evaluation of pol- institutional, and social analysis and rigor rather icy reforms, and country specificity into growth than a formulaic approach to policy making. strategies. To mainstream this approach to the The messages in this book were well received formulation of growth strategies needs persistent during the extensive consultations that we held efforts and willingness to experiment.The new during its preparation.While there is a sense of perspectives also have implications for behav- discomfort associated with the ending of a con- ior--in particular the need for more humility. viction, there was a strong sense that the findings And,last but not least,they highlight the need for of the report spoke to the experience of the 1990s a better understanding of noneconomic fac- and helped its understanding. There was also tors--history, culture, and politics--in economic appreciation and recognition that the complexity growth processes.The operational implications and diversity of growth experiences are not of these perspectives will be explored separately. amenable to simplistic policy prescriptions.They require more refined and rigorous economic Gobind Nankani analysis.There was general acceptance for the real- Former Vice President and Head of Network ization of the multiple ways in which policies and Poverty Reduction and Economic Management institutions can fulfill the functions of growth.At now Vice President for Africa the same time there was concern that these World Bank degrees of freedom could be misused by policy makers and interpreted as "anything goes." It was Washington, D.C. recognized, however, that rigid formulas were not March 2005 1. Olivier Blanchard, Guillermo Calvo, Daniel Cohen, Stanley Fischer, Jeffrey Frankel, Jordi Galí, Ricardo Hausmann, Paul Krug- man,Deepak Nayyar,José Antonio Ocampo,Dani Rodrik,Jeffrey D.Sachs,Joseph E.Stiglitz,AndrésVelasco,JaimeVentura,and John Williamson.The Barcelona Consensus is online at http://www.barcelona2004.org/eng/eventos/dialogos/docs/agenda_eng.pdf Acknowledgments T HIS REPORT WAS PREPARED BY Bernard Hoekman, Richard Newfarmer, and a team led by Roberto Zagha, Alan Winters commented on trade and other under the general direction of selected aspects of the report. Isher Ahluwalia, Gobind Nankani when he was vice president of Danny Leipziger, Edwin Lim, Carlos Antonio the World Bank Poverty Reduction and Eco- Luque, Samir Radwan, Arvind Virmani, John nomic Management Network (PREM).Current Williamson, and Adrian Wood provided valuable PREM vice president Danny Leipziger's support suggestions and advice at different stages of the and encouragement made it possible to complete report. Joelle Chassard commented extensively this volume.The team consisted of J. Edgardo on key parts of the report.Patricia ClarkeAnnez's Campos (chapter 9), James Hanson (chapter 7), and Indermit Gill's ideas and suggestions Ann Harrison (chapter 5), Philip Keefer (chapter throughout the different phases of preparation 10), Ioannis Kessides (chapter 6), Sarwar Lateef were invaluable. Comments from Theodore (chapter 9), Peter Montiel (chapter 4), Lant Ahlers, Emmanuel Akpa, Mahmood Ayub, Milan Pritchett (chapters 2 and 8 and country notes),S. Brahmbhatt, Jean-Jacques Dethier, Shahrokh Ramachandran (chapters 6 and 7 and country Fardoust,Farrukh Iqbal,Kathie Krumm,Pradeep notes), Luis Servén (chapter 4), Oleksiy Shvets Mitra, and Sudhir Shetly are gratefully acknowl- (chapter 3 and country notes), Helena Tang edged. François Bourguignon's interest in this (chapter 5), and Roberto Zagha (chapter 1 and work, as well as his ideas, comments, and sugges- country notes). Major contributions were also tions were particularly insightful and useful, and made by Ihsan Ajwad,Takako Ikezuki, Richard helped articulate the main messages of the report. Messick, and Shilpa Pradhan. Montek Ahluwalia, Masood Ahmed, Eliana Peer reviewers were Rui Coutinho, Ricardo Cardoso, Sudhir Chittale, Michele de Nevers, Hausmann,Ravi Kanbur,and Devesh Kapur.The Ian Goldin, Frannie Leautier, and Nicholas report was discussed during the World Bank's Stern provided valuable suggestions at the start Poverty Reduction and Economic Management of the exercise.The report also benefited from Week and benefited from comments by Robert discussions on earlier drafts in Washington at Buckley,Jeffrey Hammer,Carlos Felipe Jaramillo, USAID, in Delhi at the Indian Council for Deepak Mishra, Peter Moll, and Dina Umali- Research on International Economic Rela- Deininger.It also benefited from comments from tions, in Geneva at the International Labour several World Bank chief economists, in particu- Organization, at the DAC senior economist lar Shanta Devarajan, Alan Gelb, Homi Kharas, meeting in Stockholm, at the Economic Michael Klein, Mustapha Nabli, Guillermo Research Forum in Cairo, at a seminar in the Perry, and Guy Pfeffermann. Uri Dadush, Department of Economics of the Universidade xv xvi AC K N OW L E D G M E N T S de São Paulo, and workshops in Dar-Es-Salaam nated by the World Bank's Office of the Pub- and Kampala. As part of the consultation and lisher under the supervision of Stephen discussion process, comments from the public McGroarty with crucial support from Mark on the draft volume were gathered through the Ingebretsen, Nancy Lammers, and Santiago World Bank's external Web site. Muriel Dar- Pombo-Bejarano. Todd Pugatch dealt with a lington ably handled the management and wide range of substantive and logistical issues. logistical aspects of the report. The report Alfred Friendly and Rachel Weaving were design, editing, and production were coordi- responsible for editing. Chapter 1 Overview E CONOMIC GROWTH IS A RECENT reforms. Section 5 sketches operational implica- event in the history of human- tions.Subsequent chapters set out the facts about ity. During most of 4 million growth in more detail, and then examine the years of evolution, people made limited eco- main areas in which economic and institutional nomic progress and their material well-being reforms concentrated during the 1990s--macro- changed very little. In the last few centuries, economic stabilization,trade,financial sector,pri- however, goods and services started to be pro- vatization and deregulation, modernization of duced at increasingly lower cost in hours of the public sector,and political reforms.The chap- effort. The hours of work needed to produce ters aim to draw lessons from gaps between basic goods such as water or heat at the dawn of expectations and outcomes. Most chapters are civilization were several hundred times those also followed by a Country Note that expands needed today (DeLong 2000). Similar increases on issues insufficiently dealt with in that chapter, in productivity have been achieved for an or that considers country-level perspectives. expanding range of goods and services. Most of this progress has taken place in the last two cen- turies, during which technological progress has been exceptionally rapid, and economic growth FIGURE 1.1 unprecedented (figure 1.1). Worldwide Growth in Real GDP per Capita, 1000­Present It is only in the last 50 years that mainstream economics has focused on the determinants of 100 Adam Smith's"natural progress of opulence"and GDP 90 on how growth could be accelerated. Many 80 questions about growth still lack satisfactory orld w 70 answers.Yet few issues are more important for cent) ealr 60 the world's future than the ability of developing in (per 50 countries to raise both productivity and the rate 40 at which they accumulate capital. wtho capita 30 gr This overview chapter first briefly reviews per 20 our understanding of growth before turning, in 10 section 2,to the facts and controversies of growth 0 Century's and policy reforms in the 1990s. Section 3 draws 11th 12th 13th 14th 15th 16th 17th 18th 19th 20th ­100 the broad lessons coming out of the growth Century experience of the 1990s, and section 4 offers les- Source: DeLong 2000. sons specific to key policy and institutional 1 2 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s 1. Understanding Economic Up to then, thinking about growth had Growth been dominated by the Solow model, the basic model with which we still think about eco- Absent definitive theories, views on growth have nomic growth, in which growth is a function been shaped by facts and changed by experience. of the accumulation of capital, accumulation of Until the 1970s, the growth strategies of devel- labor, and productivity growth. This model oping countries focused on accelerating the rate leaves out much of what needs explaining. In of capital accumulation and technological adop- particular, it views long-run growth as entirely tion. Import substitution, state-owned enter- determined by exogenous factors, independent prises, controls over the financial sector, central from structural characteristics of the economy planning,and a variety of price controls and state such as openness, scale, and saving rate, and, interventions in the economy were some of the most important, from the policies influencing policies that governments used to take the"com- such variables. Also, while left unexplained in manding heights" of the economy and guide the model, productivity growth drives the resource allocation to areas thought to be most empirical story. Solow himself estimated that conducive to long-term growth. Confidence in technological change explained more than half governments was born from their (partial) suc- of per capita output growth in the first half of cess in addressing the Great Depression, in the 1900s in the United States. Calculations by expanding production during World War II, and the World Bank indicate that it explained one- reconstructing Europe and Japan. Economists third of the increase in per capita income in and policy makers saw that market forces dis- East Asia up to the early 1990s (World Bank rupted growth and that governments were able 1993). Other exercises reach similar conclu- to restore it, and to expand capacity efficiently. sions on the large role of productivity gains in The generation of economists that followed, growth experiences. however,familiar with experiences of developing At first the New Growth Theory seemed to countries in the 1970s and 1980s, saw the waste hold the promise of linking policies to growth of enormous resources in ill-conceived govern- performance.It appeared at a time when evidence ment initiatives, the costs of poor macroeco- was accumulating--from the growth experience nomic management, and the ease with which of the 1970s and 1980s--suggesting that the accu- well-intentioned public policies could be mulation of capital was not a panacea, and that diverted to serve narrow political or economic misguided policies were costly for growth.The interests. Understandably, this later generation of new evidence provided the conceptual founda- economists and policy makers came to believe tion for aggregate cross-country regressions, that the cost of government failures was consid- which throughout the 1990s sought to capture erably larger than the cost of market failures, that the effect of policies on long-term growth (Barro government interventions interfered with devel- 1991;Temple 1999) and provided the strongest opment,and that containing the role of the pub- intellectual foundation for the view that better lic sector in the economy, reducing its use of policies would deliver faster growth. resources, and limiting its discretion were essen- A number of empirical problems became evi- tial for economic growth. dent, however, related to the crude manner in which policy variables enter the cross-country regressions; the fact that differences in the insti- New GrowthTheory tutions underlying policy design and policy This shift in views was supported by a new strand implementation are not captured; the lack of of academic research that started in the second robustness to changes in time periods and speci- half of the 1980s and gathered impetus during fications; the crudeness of the assumption that the 1990s, when there was a resurgence of aca- the same model explaining growth in the demic and empirical work on growth. Republic of Korea or Brazil could be used for OV E RV I E W 3 Bolivia or Rwanda; and the poor predictive greater attention to the role of institutions, and power of policies as indicators of performance. studies brought the issue of inequality--both If,as suggested by the growth regressions,poli- within and between countries--increasingly to cies matter for growth, policy improvements the fore. should lead to higher growth. Both in the 1980s and 1990s, policies improved relative to other Growth in Developing Countries: decades, but growth performance remained well Divergence,Variability, and Unpredictability below that of the 1960s and 1970s (Easterly 2001). More recently, empirical research has Research during the 1990s was able to extend argued that when a measure of"institutional qual- the availability of data over long periods.This ity" is included in cross-country regressions, the made it clear that growth was not a linear explanatory power of other variables, including process, and that it did not conform to the the- all measures of "policies," becomes negligible oretical prediction that per capita income in (Acemoglu, Johnson, and Robinson 2001; developing countries would eventually converge Rodrik, Subramanian, and Trebbi 2002; Easterly with that of industrialized countries. In fact, and Levine 2003; and IMF 2003e).This suggests there has been"divergence big time"in the evo- that "good" institutions matter more for growth lution of per capita incomes (Pritchett 1997), than "good" policies--that "institutions rule." both between industrialized and developing In hindsight,the breakthroughs expected from countries and among developing countries the New Growth Theory have not materialized. themselves.This is the case whether the period Nonetheless, in the process, greater clarity was being considered is the last 40 years (figure 1.2) reached on the facts about growth, analysts paid or the last 10 (figure 1.3). FIGURE 1.2 Economic Growth in Perspective, 1960­2002 (GDP per capita) 20 capita 15 per GDP 10 OECD fo 5 cent erP 0 1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 Latin America and the Caribbean Sub-Saharan Africa South Asia Middle East and North Africa Europe and Central Asia East Asia and Pacific Source: WDI 2003. Note: Regions' GDP per capita is shown as a percentage of the OECD GDP per capita (total regional GDP over total population). 4 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s As a result, worldwide inequality has changed from being the result almost exclusively of differ- FIGURE 1.4 ences among people within countries to being the Fraction of World Inequality Accounted for by Differences across Countries result primarily of differences across countries (fig- ure 1.4). 70.0 The consideration of growth over longer peri- 60.0 ods also highlights the variability of growth in developing countries.The experience of Latin 50.0 America since the 1980s, the collapse of growth inequality 40.0 in Africa in the last two decades, and the eco- cent) total 30.0 nomic collapse of Eastern Europe after several fo (per decades of sustained growth stand in sharp con- 20.0 trast to the stability of growth among industrial- 10.0 action ized countries, which have grown at roughly a Fr 0.0 constant rate (except for the interruption of World War II and recovery years) for more than 1800 1820 1840 1860 1880 1900 1920 1940 1960 1980 2000 100 years. It also contrasts with the experience of Years East Asian countries.What is remarkable about Theil coefficient Mean ln deviation EastAsia is not that it experienced a crisis in 1997, Source: Bourguignon and Morrison 2002. but that it experienced so few crises over the pre- ceding decades. By and large, developing coun- tries have one year of negative per capita growth The variability of growth helps to explain roughly once every three years. In East Asia, the why growth in the developing world is so diffi- average is half that rate. Korea has only had only cult to predict. Instances of economists (includ- three years of negative per capita growth since ing, for example, 1977 Nobel Laureate James 1961.1 Meade on Mauritius) making highly inaccurate predictions have become part of the economic folklore. Many of the economic successes of today--Bangladesh, Indonesia, Korea, or Mauri- FIGURE 1.3 tius--were considered "basket cases" in the Regional Perspectives on Growth in the 1990s 1960s, when Africa's growth prospects were seen as superior to those of overpopulated Asia--a 200 view captured in Asian Drama (Myrdal 1972). In 180 the later 1990s, just before the second most dra- 160 matic economic crisis in its history, Argentina was seen as a model for developing countries and (1990=100) 140 believed to have found the path to sustained GDP 120 growth. At a more technical level,World Bank 100 growth projections, as well as growth projections capita 80 by other forecasters, tend to be systematically erP overoptimistic (a point that was highlighted in 60 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 the World Bank's World Development Report East Asia and Pacific South Asia 1991). Eastern Europe and Central Asia Sub-Saharan Africa While a rare occurrence thus far, sustained Latin America and the Caribbean Organisation for Middle East and North Africa growth has improved the lives of millions. Coun- Economic Co-operation and Development (OECD) tries where sustained growth has taken place Source: WDI 2003. (mostly in South and East Asia, including Bangladesh, China, Indonesia, India, and Viet- OV E RV I E W 5 nam) account for a large proportion of world inherent in new industrial activities; price con- population. Out of 117 countries with popula- trols did not have serious economic conse- tions of more than half a million people, only 18 quences because the concentration of wealth have been able to sustain growth rates exceeding precluded the redeployment of resources in industrialized countries' growth and hence nar- response to changes in demand (Seers 1962). row their per capita income gap with those While there are some functions that institu- countries.2 tions need to perform in any society, the form through which institutions can perform these functions can vary considerably (Virmani 2004). Institutions Most of the empirical work on the importance Defined as the rules and norms constraining of institutions leaves open the question of how to human behavior (North 1990), institutions improve institutional performance. Merely include the informal rules and norms that gov- adopting some other country's laws and formal ern personal and social behavior and the formal regulations is no guarantee of achieving the same rules and norms governing economic, social, and institutional performance. Recently, accession to political life. Institutions enable societies to the World Trade Organization and integration organize themselves and function in an orderly into regional supranational entities such as the manner by solving problems central to life in European Union and the New Economic Part- society, particularly agency problems, contain- nership for Africa have strengthened incentives ment of predation by individuals or the state,and for institutional improvements. East Asian coun- collective decision making. Societies' perform- tries have long realized the importance of insti- ance depends on how effectively their institu- tutional change and innovation, and the 1997 tions resolve these problems. crisis made this realization all the more acute, The importance of institutions for economic creating renewed impetus to modernize institu- prosperity is not a novelty learned from the tions, including political institutions. But for 1990s. From different perspectives,Adam Smith, most developing countries, improving the qual- Karl Marx, and Max Weber highlighted the role ity of their institutions remains a challenge. of institutions in the development of a market economy and formation of a capitalist society. Fairness, Growth, and Institutions Economists dealing with development in the 1950s and 1960s were aware that the develop- Another important strain of ideas in the 1990s ment challenges faced by a plantation economy came from the resurgence of interest in inequal- differed from those faced by a society where eco- ity as an apparent influence on growth and insti- nomic and political power were not concen- tutional performance.A recent body of literature trated (Rostow 1952,1960;Adelman and Morris suggests several channels through which inequal- 1965). Latin American economists of the Struc- ity affects economic growth. Fairer societies offer turalist school saw in the legacy of colonialism, their citizens more public goods,more social sup- embedded in institutions serving the interests of port,and more social capital.Hence they are more a small,landed elite,the source for economic per- capable of sharing the costs and benefits of formance inferior to that of the United States or improving economic policies, and in turn facili- Canada (Furtado 1963).This imbalance formed a tating consensus building and decision making part of the justification for an activist state: infla- (Deaton 2003a). Fairness also facilitates agree- tion helped to mobilize resources from the ment on the provision of public goods that have wealthy elite who resisted more efficient forms strong beneficial side effects on society, such as of taxation; states sponsored investments in man- health services, water supply, or waste disposal. ufacturing, particularly in capital-intensive Other channels through which inequality affects industries, because old economic interests resis- growth are market structures and microeconomic ted change and were unwilling to take on risks incentives.A better distribution of wealth reduces 6 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s credit constraints,and broader availability of credit economic interests,were unable to render fair and is found to have a significant and positive effect equitable judgments. World Development Report on growth rates. If individuals are limited in their 2001 provides other examples of how economic borrowing capacity, reallocating capital toward incentives affect the emergence of institutions the poorest will increase aggregate productivity. that sustain the functioning of markets, and the Even if one concludes that greater equality different coordination or risk-reducing problems influences growth positively, there is still consid- that they are meant to resolve. erable ignorance about the means through which greater equality can be achieved. Governments have long sought, with varied degrees of success, 2. Facts and Controversies of the to redistribute income through land redistribu- 1990s tion, employment programs, subsidies, and pro- motion of broad access to credit, infrastructure, At the beginning of the 1990s, most economists health,and education.The large,underresearched working on development and many policy mak- area for further study includes questions related ers shared the conviction that more efficient use to the impact of public spending on equity, both of resources would lead to growth. This was in a static sense (incidence of public spending) believed to require, first, macroeconomic pru- and a dynamic sense (changes in individuals' dence, domestic liberalization, and outward ori- earnings potential). entation, which in turn required freeing market Recent literature has emphasized the impor- incentives and opening the economy. Hence fis- tant links between the distribution of assets in a cal deficit reduction, realignment of exchange society and the institutions that emerge. Knowl- rates to eliminate black market premia, lifting edge is still rudimentary about how institutions controls on prices, deregulation of interest rates emerge and are established in a society, but eco- and liberalization of the financial sector, and nomic research in the 1990s has provided some reduction of tariffs and other restrictions on insights. First, economic incentives influence imports all became central to the policy reform what type of institutions emerge and when.The programs implemented in the 1990s. enforcement of property rights to land,for exam- Second, conventional wisdom held that to ple, will depend on the benefits of enforcement achieve greater efficiency required a reduction in relative to its costs,which for each owner depends the role of the state.There was evidence that the on the extent to which other owners enforce state discretion that was inherent in growth their property rights. In an extractive economy, strategies based on infant industry,import substi- for example, if landowners in general do not tution policies, and the growth of public enter- enforce their property rights, it is uneconomical prises had been misused more often than for one landowner to enforce his: workers will anticipated, had often been captured by narrow find it attractive to exploit land and appropriate interest groups, and served as the source of the rents for themselves. Only when this coordi- endemic corruption. Addressing this problem nation problem is resolved will economic incen- required reducing state discretion, downsizing tives be sufficient for enforcement of property governments, and encouraging a much greater rights (Hoff and Stiglitz 2001). Second, concen- role for the private sector. Hence privatization, trated economic and political interests influence deregulation, elimination of quantitative restric- institutions.This can be seen from experiences tions and of licensing requirements, and disman- with land distribution in Latin America, and also tling agricultural marketing boards and other from the United States in the early 1900s, when forms of state monopoly all became central to the government decided to regulate matters hith- reform programs. Seeing the need to strengthen erto left to private parties and the courts; the rea- the organizational effectiveness of the state, and son for the shift was a perception that judges and the efficiency with which the state used public the courts, having been corrupted by powerful resources, reformers rationalized government OV E RV I E W 7 functions and undertook civil service, legal, and of the economy.In Bolivia,reforms by Paz Esten- budget reforms. Democratic processes were soro that had brought hyperinflation to a halt in expected to provide checks and balances and fur- the mid-1980s were continued in the 1990s, ther incentives to this process. regardless of the parties in government.In Africa, Third, it was believed,reforms had to be rapid. the devaluation of the African Financial Com- Earlier, some of the first authors to argue in favor munity (CFA) franc increased competitiveness of abandoning the dirigiste framework of early and many other reforms were implemented development economics (notably Little, Sci- throughout the region. In Tanzania, President tovsky, and Scott 1970; McKinnon 1973) had Mkapa started an ambitious program of reforms. argued explicitly in favor of a gradualist reform In South Africa, the transition to a multiracial strategy (in respect to trade and the financial sec- democracy was followed by steps toward liberal- tor, respectively).3 But in the course of the 1980s izing the economy. the economics profession began to be influenced Leaders such as Rawlings of Ghana and by the enthusiasm of leading politicians for "the Museveni of Uganda strengthened fiscal funda- magic of the market."Arguments in favor of "big mentals,achieved macroeconomic stability,liber- bang" and "shock treatment" became prominent. alized the economy, and reduced the role of the By the time that the transition to a market econ- state. Privatization, retrenchment of the public omy got under way in the former socialist sector, and liberalization of trade were the focus economies, "a belief in gradualism had almost of economic policy changes in countries as become tantamount to a confession of a lack of diverse as the Central African Republic, Ghana, reforming virility"(Williamson and Zagha 2002). and Tanzania. Reforms in the Middle East and North Africa were less ambitious but were nonetheless significant in the Arab Republic of A Decade of Significant Change Egypt, Jordan, Morocco, and Tunisia. On the The 1990s provided ample opportunity for these political front,democracy spread in former com- views to be implemented.The Russian Federa- munist countries and Africa, and was consoli- tion, Eastern Europe, and Central Asia embraced dated in Latin America.These and other changes capitalism and a new generation of leaders made gave rise to expectations that the 1990s would it a priority to rebuild their economies on the accelerate growth and social progress in the basis of capitalist principles, markets, and priva- developing world. tized firms.Regarding the speed of reform,while there were divergences, the balance of opinions Rapid Growth, Take-offs, and Social Progress supported rapid rather than gradual reform. India and China,together accounting for 40 per- China, the largest developing economy contin- cent of the developing world's population, grew ued the reforms it had begun in 1978 with fur- fast in the 1990s for a second decade in a row, as ther liberalization of the domestic economy, and did many other countries in South and East Asia, increased openness.After its crisis in 1991, India, including Bangladesh, Sri Lanka, and Vietnam. the second-largest developing economy, speeded Chile continued to grow in Latin America, up liberalization started in the 1980s. President Tunisia in North Africa,and Botswana and Mau- Collor of Brazil announced a radical program of ritius in Africa. New high performers appeared economic reform aimed at reducing hyperinfla- and annual gross domestic product (GDP) tion and reversing several decades of state-led growth for the decade was rapid in an array of import-substituting industrialization. In countries: Mozambique (7.8), Uganda (6.8), Argentina, President Menem set the country on Dominican Republic (6.0), Tunisia (5.0), and a course of eliminating hyperinflation through a Poland (4.5). Countries affected by the crisis in currency board as well as ambitious market East Asia made an unexpectedly rapid recovery. reforms, which saw the privatization of state- Notwithstanding unevenness across regions, owned businesses and liberalization and opening the incidence of poverty continued to decline 8 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s throughout the 1990s--more rapidly in EastAsia bank restructuring experts, and emerging-mar- than in South Asia and more rapidly in South ket private traders rolled from crisis to crisis:from Asia than in Latin America. In Africa, however, Mexico during 1994­95, to Korea, Malaysia, the incidence of poverty increased slightly. Thailand, Indonesia during 1997­98, Russia in Growth was the main force behind virtually all 1998, Brazil in 1998, and Turkey in 2001 to the cases of significant reductions of poverty, includ- latest and perhaps most worrisome of all, in ing in China and India. But particularly in Latin Argentina during 2001­02. The evolution of America, there were instances such as Brazil and spreads in the months preceding the financial Bolivia where social indicators improved with- crises suggests that few were anticipated. out significant growth. Alongside these positive developments, how- Delay in Recovering Growth, Particularly in ever, there were several negative surprises. Latin America It was hoped that the "lost decade" of the 1980s "Transition Recession" in the Former Soviet would be reclaimed in the 1990s. Macroeco- Union and Eastern Europe nomic stabilization, fiscal austerity, trade liberal- The transition from a communist, centrally ization, and privatization were expected to lead planned economy to a capitalist one was to rapid growth.Although growth was the fastest expected to be difficult.But the depth of the out- in two decades until 1998, its collapse thereafter put collapse was not widely predicted.The length following the reversal in capital flows created the of the transition--in which many countries in general perception that the growth payoffs have 2003, more than a decade later, remain far below been smaller than expected. their previous levels of output--was not widely forecast.Nor was the variability among countries Argentina: The Collapse of the Hard Peg in the depth and duration of the output collapse. Argentina was the most successful example of a Though recoveries have started to emerge--in trend in the 1990s to create macroeconomic sta- the Czech Republic, Hungary, and Poland, for bility by legal and institutional changes intended example--it will take years, and in some cases, to reduce the scope and latitude of government's decades, for most former Soviet countries to discretion.Exchange-rate arrangements that set a regain their per capita income levels prevailing at fixed rate for peso convertibility were not only the beginning of the transition. incorporated into law but also made especially difficult to alter, and changes were made in the Continued Stagnation in Sub-Saharan Africa operation of the central bank to make these lim- The failure of growth in Africa--either of pow- itations a reality.As part of a package of reforms, erful and rapid growth in a single large country or this convertibility plan eliminated Argentina's in a substantial number of smaller ones--was a hyperinflation and, for a period, it restored eco- surprise.Despite good policy reforms,debt relief, nomic growth. continued high levels of official assistance, prom- Once Argentina achieved stability with ising developments in governance,and a relatively growth, there was considerable discussion--par- supportive external climate, no take-off has ticularly after the devaluation of Brazil's real in ensued. While some positive developments January 1999--as to whether the country should occurred in the later 1990s, such as in Mozam- abandon its rigid exchange rate system.Views bique,Tanzania, and Uganda, and still persist at diverged among economists. Looking back, the the time of writing,it is too early to conclude that former Governor of Argentina's Central Bank Africa has turned the corner. described the abandonment as a marriage to be broken when it was going well (Mario Blejer, Financial Crises World Bank 2005b). For fear that markets could Financial crises in the 1990s were less predictable overreact,Argentina's authorities maintained the than in the 1970s and 1980s. Macroeconomists, system.When the plan collapsed, the result was OV E RV I E W 9 politically and economically costly by design. resources, and foreign direct investment and Thus the damage was not a surprise. But the financial flows to developing countries were demise of the convertibility plan itself was a sur- much larger. If commodity prices affected devel- prise, for two reasons. First, its initial successes oping countries adversely, the damage was not had suggested longevity was possible; it had dramatic and should have been offset by the reduced rapid inflation and initiated a boom in increasing share of manufacturing exports-- the early 1990s, and it had weathered the except in a small group of least developed and "Tequila" after-shocks of the Mexican crisis rea- Sub-Saharan African countries that remained sonably well. Second, while the end of convert- highly dependent on agricultural exports. ibility was costly by design, its actual cost All in all, while external factors played a role, exceeded the most pessimistic forecast. explanations of performance must be sought pri- marily in developing countries' domestic poli- cies. Interpreting the Results Good performance has been associated with From a growth perspective, the net result of the domestic and external liberalization; Chile, India, contrasting experiences of the 1990s is that China, and other countries in East Asia are all developing countries as a group grew faster than more open than in previous decades and have in the 1980s. In East and South Asia this reduced moved toward greater reliance on market forces. the income gap with industrialized countries,but But many aspects of these countries' policies are in other regions, the gap increased. In Latin still far from compliant with conventional wis- America, there were clear gains up to 1998, dom. For example, India has registered fiscal reversed in the late 1990s and early 2000s (figures deficits several times higher than Brazil's or 1.2 and 1.3). Argentina's, with lower inflation and lower inter- Analyzing policy reforms of the 1990s, several est rates.While this fiscal trend is clearly unsus- studies (Loayza, Fajnzylber, and Calderón 2002; tainable in the long run, and measures have been Lora 2001a;Easterly 2001) find that countries that taken to correct it, it is clear that there is more to improved their policies--strengthening macro- macroeconomic stability than a superficial read- economic management, opening up their ing of the size of the fiscal deficit. China has built economies, liberalizing their financial sectors-- extremely large contingent liabilities related to grew faster in the 1990s.However,they also find a unfunded pensions and nonperforming loans in large unexplained negative effect associated with the banking system.While,again,this is not a sus- both the 1990s and the preceding decade.Together tainable situation, it suggests that economies do with analysis of individual country experiences not operate in mechanical ways, and that and overoptimistic forecasts by international dynamism in one sector can offset the cost of financial organizations and private entities, these inefficiency in others. Similarly, India's and studies give an empirical base to perceptions that China's industries, though increasingly competi- the economic policy reforms of the 1990s yielded tive in export markets,remain protected and state results below expectations. enterprises still play a large (though declining) It has been suggested that lower growth in role in these economies. OECD (Organisation for Economic Co-opera- The mismatch between predictions and tion and Development) countries might have results, and the successes of China, India, and depressed developing countries' growth in the Vietnam where there were substantial deviations 1990s (Easterly 2002). In reality, the 1990s was from the full package of reforms, suggest several favorable for developing countries, even if not possible explanations. First, sufficient time may every country found ways to benefit. Exports not have yet elapsed for results to emerge in all from developing countries as a group grew much countries. Over time, market-oriented reforms faster than in previous decades.Real interest rates may ultimately yield the results expected. were lower. Debt obligations claimed fewer Growth rates in African and other developing 10 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s countries have rebounded since 1997;Argentina ural resources while enabling corruption to is experiencing its second year of rapid growth flourish. after the collapse of 2001­02;and growth rates in These explanations are not mutually exclu- Eastern Europe have increased. Second, perhaps sive; one or more may apply to specific country the reforms implemented in the 1990s were not circumstances.The experience also holds some sufficiently ambitious. Insufficient fiscal adjust- deeper lessons. For example, while at one level ment in Latin America, very partial privatization Argentina's experience teaches that fixed in Africa, and insufficient openness to interna- exchange regimes require a very demanding set tional trade in the Middle East and Northern of conditions, a deeper lesson is that rigid rules Africa may explain performance below expecta- are no substitute for credibility, and that govern- tions in these regions. A third possible explana- ment's discretion needs to be checked, not tion is that there were incoherencies in the replaced with rules.Another deeper lesson is that implementation of policies. Argentina intro- the reforms of the 1990s did not focus on the duced a rigid exchange rate without the fiscal binding constraints. For example, they reduced and financial conditions needed to sustain it. Fis- fiscal deficits when perhaps the binding con- cal adjustments in some African countries were straints were lack of public capital and aggregate achieved at the cost of reducing productive pub- demand.Or they reduced tariffs on imports when lic spending. Open capital accounts encouraged perhaps the binding constraint was the workings pro-cyclical flows. Correction of these inco- of the financial sector. Or they focused on cor- herencies may enable growth to resume. recting government failures, when the binding Perhaps most important,while reforms in the constraints were market failures. 1990s focused on increasing the role of markets and decreasing the role of the state, they tended to neglect the role of institutions. Francisco Gil 3. Lessons from the 1990s Diaz, Mexico's Minister of Finance (as quoted in Krueger 2004), recently suggested that Promote Growth, Not Just Efficiency The policies that have been undertaken Reforms need to go beyond the generation of are not even a pale imitation of what mar- efficiency gains to promote growth.The policy ket economics ought to be, if we under- focus of reforms in the 1990s enabled better use stand market economics as the necessary of existing capacity but did not provide sufficient institutional framework for a sound econ- incentives for expanding that capacity. While this omy to operate and flourish. What has emphasis on efficiency was warranted at a time been implemented throughout our conti- of extremely large distortions and waste, it also nent is a grotesque caricature of market explains the frequent instances of stabilization economics. without growth or liberalization without growth.The experience highlights the impor- State enterprises were privatized without tance of the investment climate,and of providing much attention to the operation of the markets predictable conditions for investors and other in which they would function. Financial liberal- economic agents. ization swelled the resources, foreign and It also highlights that growth entails more than domestic, that ineffective intermediaries chan- the efficient use of resources.Growth entails struc- neled to state enterprises and related borrowers, tural transformation,diversification of production, contributing to the massive crises.In some cases, change, risk taking by producers, correction of lack of competitive political forces and such both government and market failures,and changes institutions as a free press allowed those who in policies and institutions. It is also a process of were politically well connected to take advan- social transformation:people will change activities tage of privatizations and to take control of nat- and live in different places. Social relations will OV E RV I E W 11 change, and the informal networks of rural life exaggerated the gains from improved resource will be lost as other more formal networks and allocation and their dynamic repercussions, and organizations are established. Entrepreneurs will proved to be both theoretically incomplete and invest in new machinery to produce new products contradicted by the evidence. Expectations that and adopt new organizational forms.Farmers will gains in growth would be won entirely through adopt new farming methods and change their policy improvements were unrealistic. Means product mix. The economy will produce and were often mistaken for goals--that is, improve- demand different goods and services. These ments in policies were mistaken for growth strate- changes take place over time,alongside changes in gies,as if improvements in policies were an end in institutions that render them possible.Any growth themselves. Going forward, the pursuit of policy strategy needs to include actions,on both the pol- reforms for reform's sake should be replaced by a icy and the institutional front,that address and sup- more comprehensive understanding of the forces port this process of change. underlying growth. Removing obstacles that Better policies can bring efficiency gains, and make growth impossible may not be enough: may increase incentives for investment,but with- growth-oriented action, for example, on techno- out amounting to a growth strategy.They will logical catch-up,or encouragement of risk taking not necessarily induce the behavior by private for faster accumulation, may be needed. investors and the public sector that is needed to put an economy on a sustained growth path. For Common Principles and DiverseWays to this, faster accumulation of physical and human ImplementThem capital by both the private and the public sector is essential, as are gains in productivity. Another mistake often made in the 1990s has been This may explain why the growth impact of the translation of general policy principles into a the reforms of the 1990s was smaller than unique set of actions.The principles of the 1991 expected.The incentives needed to expand pro- World Development Report,"macroeconomic stabil- ductive capacity ("expanding the frontier" in ity; domestic liberalization, and openness," have economists' parlance) differ from those that are been interpreted narrowly to mean "minimize fis- needed to use existing capacity better ("move- cal deficits, minimize inflation, minimize tariffs, ments toward the frontier").What matters for maximize privatization, maximize liberalization of growth is less the degree to which policies finance,"with the assumption that the more of these approximate the ideal than "the extent to which changes the better, at all times and in all places-- a given development strategy is able to mobilize overlooking the fact that these expedients are just the creative forces of society and achieve ever- some of the ways in which these principles can be higher levels of productivity" (Alejandro Foxley, implemented. in World Bank 2005b). And, in Albert There are many ways of achieving macroeco- Hirschman's words (1958): nomic stability,openness,and domestic liberaliza- tion. As seen above, for example, the goal of Development depends not so much on achieving macroeconomic stability does not finding optimal combinations for given imply a need to minimize fiscal deficits at all resources and factors of production as on times. A lower fiscal deficit achieved today calling forth and enlisting for develop- through off-budget contingent liabilities, or ment purposes resources and abilities that through cutting back public investments and thus are hidden, scattered, or badly utilized. reducing long-run growth and the future tax base,may mean a higher fiscal deficit in the future. In retrospect, it is clear that in the 1990s we A lower fiscal deficit does not even guarantee often mistook efficiency gains for growth.The greater macro stability if it is based on external "one size fits all" policy reform approach to eco- borrowing in which interest rates are reduced at nomic growth and the belief in "best practices" the cost of greater vulnerability to exchange rate 12 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s fluctuations, or if it is based on building off- underlying growth strategies.This nonformulaic budget liabilities through the banking system, result holds not only for the eight East Asian which eventually translate into an increase in economies featured in the 1993 study,but also for a public debt--as Latin American countries and larger set of countries 10 years later.Countries with Turkey found to their cost in the 1980s and remarkably different policy and institutional frame- 1990s.Similarly,trade integration can be achieved works--Bangladesh, Botswana, Chile, China, through various means that offset the effect of Egypt, India, Lao PDR, Mauritius, Sri Lanka, tariffs and reduce the implicit tax on exports. Tunisia,andVietnam--have all sustained growth in Duty rebate schemes, subsidized credit to per capita income at rates above the U.S.long-term exporters, and other forms of export promotion, growth rate of close to 2 percent a year. export processing zones,infrastructure,and trans- Second, common to all successes is that four port corridors have all helped China, India, functions have been fulfilled: rapid accumulation Korea, and Mauritius to integrate into the world of capital, efficient resource allocation, technolog- economy while keeping their tariffs relatively ical progress,and sharing of the benefits of growth. high in the initial phases of integration and reduc- Rates of progress in these four functions have not ing them gradually over time. Thailand's and always been uniform,but successful countries have Indonesia's foreign domestic investment regimes achieved a balance among them over time, and had few restrictions, whereas those of Korea and disruptions have ensued when the balance was not India had many until very recently--but both achieved.While there can be substitution tem- Korea and India found alternative instruments to porarily, the balance will need to be reestablished access and adopt modern technologies. Financial at some point. intermediation can be increased by relaxing entry For example,Korea's policies in the 1960s and restrictions in the banking system, or by improv- 1970s sought to encourage risk taking by the pri- ing the workings of the legal system, particularly vate sector. Import protection and priority lend- those parts that deal with the repossession of col- ing contributed to higher levels of capital lateral. accumulation, at the cost of efficient allocation, To sum up,"getting the policies right" mis- which became a more important priority in the takes means for ends. Clearly not everything can 1980s.The Soviet Union, well into the 1960s, be right at once, and not everything needs to be grew rapidly on the basis of sacrificing consump- "right"for growth to take place--as witnessed in tion, accumulating capital, and maintaining a rel- examples from Bangladesh, China, India, atively equitable income distribution. But its Indonesia, and many other countries. considerable progress in science and technology was not effectively deployed in production and, more important, resource allocation was enor- Common Functions and DiverseWays to mously wasteful.Eventually,the costs of this inef- AchieveThem ficiency and the political reforms of the late To sustain growth requires key functions to be 1980s combined to bring growth to collapse. In fulfilled, but there is no unique combination of India, a "big push" in capital formation in the policies and institutions for fulfilling them.The decades following independence was comple- successful growth experiences in eight East Asian mented in the 1980s--when evidence of misal- economies, reported in the World Bank's East location and low productivity growth began to Asian Miracle (World Bank 1993), resulted from emerge--by policies that gradually freed market diverse policy and institutional paths, but com- forces and increased efficiency in resource alloca- mon functions were fulfilled along these paths.4 tion (Virmani 2004), thus ensuring not only the This perspective has several implications. sustainability of growth but also its acceleration. First,different policies can yield the same result, Factoring these four functions into analyses of and the same policy can yield different results, growth makes it easier to understand why both depending on country institutional contexts and policies and institutions play a role. For example, OV E RV I E W 13 capital accumulation by the public sector requires Lanka in the 1960s and 1970s,the market was too sound tax policies and administration, sustainable small;the benefits of inward-looking industrializa- macro policies, and a bureaucracy that is capable tion were negligible and did not justify its costs. of formulating and managing public expenditure Sri Lanka became successful only after it began to programs effectively and of choosing programs liberalize imports in late 1977 and follow export- with high returns.Accumulation by the private oriented policies.Thus,the same inward-industri- sector requires at least reasonably secure private alization policy produced different outcomes property rights, stable expectations about the because country characteristics differed. future, a stable macroeconomy, and access to Conversely, a given policy can yield different finance. One country might strengthen private results because of institutional variation. In Japan investment by, say, improving expectations, during the Meiji industrialization and, more whereas another country could achieve the same recently,in Korea,public institutions were able to result by,say,reforming the financial sector.5 Sim- resist pressures from narrow interest groups.Pub- ilarly, efficiency in allocation requires not only lic enterprises were run efficiently, and state reasonably sound policies--such as competitive ownership built capacity in sectors that the pri- exchange rates and an open trade regime--but vate sector had not entered because of perceived also institutions that can enforce contracts and high risks.The same policy in Bolivia, however, enable markets to function (World Bank, World where public enterprises were run for the bene- Development Report 2001).Technological catch- fit of narrow interest groups,did not play a strate- up requires not only investment and trade policies gic role in the industrialization process,and most that enable a country to attract foreign direct of the enterprises were liquidated when Bolivia investment (FDI) and import equipment,but also had to stabilize its economy in the 1980s. In the institutions that, depending on the country's case of India, it has been shown that in the pres- development stage, promote adaptive research or ence of poor institutions, liberalization can lead a patent regime. Indeed, in some instances, it is to less growth than expected (Virmani 2004). institutions and political realities that define the Like that of policies, the effect of institutions set of feasible policies, as testified by Russia's for- depends on the context. Security of ownership mer Minister of Finance Yegor Gaidar (World rights has been achieved in different ways and to Bank 2005b): different extents in different country contexts.In Soeharto's Indonesia, securing returns depended If I were the tsar of Russia, I would have on connections with the ruling elite. In contem- done everything differently . . . But if I porary China, the definition and enforcement of were deputy prime minister and finance property rights depend on party and local gov- minister, in a government without a par- ernment support--and only recently have initia- liamentary majority and under many pres- tives been taken in this direction. And in India, sures,I would have done more or less what success depends on the functioning of a judiciary we did. modeled after western legal systems. Different policies can have the same effect,and Sharing the benefits of growth has been the same policy can have different effects,depend- important in all sustained growth experiences, ing on the context.In large economies,with access and particularly in countries with authoritarian to foreign technology and equipment, competi- forms of government, where it has helped to tion and economies of scale lessen the efficiency legitimize regimes that often were neither fully cost of trade restrictions and markedly widen the representative nor democratic.Various policies scope for successful inward-oriented industrializa- have been used to promote the sharing of the tion.Brazil,China,and India were able to develop benefits of growth.They include land reform and manufacturing, many segments of which became redistribution of other assets;public expenditures internationally competitive, whereas in small on infrastructure (the 8-7 program in China); countries such as Jamaica and Uruguay, or Sri social spending (Tunisia); policies to increase 14 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s opportunities to economically underprivileged banks to providing infrastructure and social ser- groups (affirmative actions for bumiputra in vices. Improving institutions that support the Malaysia); and poverty-targeted programs (food implementation of policies, and strengthening stamps in Sri Lanka or employment programs in checks on the use of discretion, are more prom- India and Bangladesh). ising guiding principles than seeking to eliminate To sum up, diversity in the form of successful government discretion. growth experiences should be no surprise. Each Much of the complexity encountered in the successful country was successful in its own way. realm of economic institutions is also found in the institutions governing political life.The for- Government Discretion Needs to Be mal institutions of democracy, for example, do not necessarily ensure appropriate checks on dis- Managed and Checked, Not Replaced cretion, nor are those checks always absent in by Rules authoritarian regimes. Mechanisms and levels of Because developing countries' societies resolve accountability can take very different forms, agency,predation,and collective decision-making rarely amenable to the simplicity of formal polit- problems less effectively than do those of indus- ical institutions. Much of the growth success of trialized countries, much of the reform effort in East Asian countries can be attributed to these the 1990s sought to introduce policies that would countries' ability to allow discretion by different limit the discretion of national authorities in government agencies, alongside checks on this growth strategies and minimize demands on insti- discretion that made them accountable. The tutions. Privatization, financial liberalization, and forms of these checks varied: an authoritarian removal of quantitative restrictions on imports are development-oriented political leader in some examples of policy reforms meant not only to cases (Soeharto's Indonesia, Korea in the first improve incentives for more efficient allocation decades of its take-off), the checks and balances but also to reduce the need for government dis- inherent in complex one-party systems (China), cretion. Dollarization, fiscal rules, and integration or the normal checks and balances of a demo- in larger economic unions are examples of insti- cratic regime (India, Sri Lanka). tutional reforms meant to replace government discretion by rigid rules; they are consistent with Prudent Macroeconomic Management Is at the sense that,on balance,the costs of failures out- the Heart of Successful Growth Strategies weigh the benefits of discretion in the workings of an activist, developmental state. Avoidance of busts usually requires avoidance of However, government discretion cannot be booms.The costs of the crises of the 1990s in dispensed with altogether, so it is important to terms of forgone growth, social distress, and pub- find ways in which it can be exerted effectively. lic debt highlight once again the importance of What was learned in the 1990s is not only that prudent macroeconomic management.They also sound policies do not necessarily engender the stress the importance of avoiding macroeconomic institutions of a modern economy--that institu- vulnerabilities,and the risks associated with indis- tions are not entirely endogenous--but also that criminate opening of the capital account.Last but institutions can prevent the adoption of growth- not least,they stress the importance of responding oriented policies or offset their impact. Experi- quickly to downturns. One difference between ence showed how much institutions matter, and successful and less successful growth experiences how hard it is to work around their absence or to is the frequency of downturns:virtually nonexist- improve their quality. Above all, the experience ent for China, Korea, or Malaysia, but numerous showed that government discretion cannot be for Argentina, Brazil, andTurkey. bypassed. It is needed for a wide range of activi- In addition to dealing with crises effectively, it ties that are essential for sustaining growth, rang- is also important to reduce financial fragilities and ing from regulating utilities and supervising hence vulnerability to shocks.The financial crises OV E RV I E W 15 of the 1990s differed from the many that preceded The booms and busts of the 1990s are remi- them because of their cost and their suddenness, niscent of some of the crises of the 1980s.They and they were much harder to predict.The risks of teach several important lessons.First,as with most financial integration had been underestimated and liquidity surges, busts inevitably follow booms: its gains overestimated. In the 1990s in emerging avoiding the bust requires avoiding the boom and market economies, the opening of the capital strengthening the fundamentals. Countries such account to financial inflows triggered large surges as Chile, India, or Malaysia that managed inflows, that lowered the costs of sovereign and private including through the imposition of restrictions, borrowing and helped reduce inflation. For those were able to weather the crises much better than reasons, governments (with exceptions such as countries that took no such precautions. Can a those of Chile, India, Korea in the early and mid- boom be distinguished from a favorable lasting 1990s,and Malaysia) encouraged these inflows.Of trend, ex ante? In most cases the distinguishing the 10 economies that received the largest inflows, factors are the volume of the surge,the pressure it however, 7 suffered severe crises that took the puts on the exchange rate,and its impact on bank form of large output declines,higher incidence of credit. Second, the crises of the 1990s highlight poverty,and large exchange rate devaluations.The the extent to which banks can amplify the conse- three exceptions were China, India, and Hong quences of a crisis, and the risks associated with Kong (China). currency mismatches, including mismatches on Each crisis was preceded by a large surge in the borrowers' balance sheets. Third, sovereign inflows that either led to appreciation of the real borrowing in foreign currencies is risky.While exchange rate and increased current account sovereign borrowing should, in theory, help a deficits or,as in some East Asian countries,created country to access external resources on better an external debt maturity profile excessively terms, in practice it has encouraged governments biased toward the short term, and exposed and private firms to take excessive risks. unhedged commercial banks to currency and maturity risks.In current account crises--many of Move Away from Formulaic Policy which arose in the context of stabilization pro- Making and Focus on the Binding grams anchored on a nominal exchange rate--the Constraint(s) sequence of events followed a remarkably similar pattern: a surge in capital inflows put pressure on A vital lesson for policy formulation and policy the exchange rate to appreciate; the current advice is the need to be cognizant of the shadow account deficit of the balance of payments prices of constraints, and to address whatever is increased; private-sector and government debt the binding constraint on growth, in the right exposures fed resistance to letting the exchange manner and in the right sequence.This requires rate adjust;governments sought to sustain the rate recognizing country specificities, and more eco- by drawing down reserves, but the policy lacked nomic analysis and rigor, than does a formulaic credibility, or reserves were insufficient to sustain approach to policy making. Policy makers face it;a large devaluation followed;and the tightening the practical problem that no scientific method effect of the devaluation was amplified by the con- permits ex ante identification of the most sequences of currency mismatches for the balance important constraint(s) binding growth in spe- sheets of banks or those of their borrowers and of cific country circumstances, and hence the spe- firms.In Indonesia and Korea,for example,where cific measures that are needed to address it current account deficits were relatively small, the (them). During the 1980s and 1990s, China's trigger was the need for a large debt rollover at a approach was to"cross the stream by groping for time when investors were retreating from emerg- the stones."Constraints were identified and dealt ing markets and when risk perceptions were on with as the growth process unfolded, through the rise.This was accentuated,in Indonesia,by the experimentation and trial and error (chapters by uncertainty of the political transition. Lim and Huang, in World Bank 2005a). 16 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s Which policy should be introduced, and of resources through macroeconomic stabiliza- when, varies considerably from case to case tion, liberalization of trade and the financial sec- depending on initial conditions and institu- tor, privatization, and deregulation. Deregulation tional endowments. For example, one can gen- and reduction in the role of government were erally assume that where hyperinflation is expected to improve the governance of the pub- raging, or public debt demands high real inter- lic sector through improvements in incentives for est rates--as it does in Argentina, Brazil, performance, more transparency, and fewer Jamaica, and Turkey, for example-- macroeco- opportunities for rent seeking. Institutional nomic stabilization is the first priority.Where reforms focused on improving collective deci- trade restrictions are extreme and hinder uti- sion making and solving agency problems lization of existing capacity, as in many coun- through democratization, decentralization, and tries of the Middle East and North Africa, public sector reforms aimed at enhancing the reducing them will be essential.Where there is efficiency, transparency, and accountability of uncertainty regarding the future course of eco- government activities. From 60 countries choos- nomic policies, as in Bolivia, Democratic ing their leaders through competitive elections Republic of Congo, and Nigeria, financial sec- in 1989, the number rose to 100 by 2000. Dele- tor liberalization will do little to channel gation to subnational levels of government of resources to private investment.Where property political, administrative, and financial powers has rights are poorly defined and enforced, and reg- taken place not only in federated states such as ulation prevents the movement of domestic India, Brazil, and Russia but also in smaller states resources across sectors, as is still the case in and centralized states such as Bolivia and the some Central Asian and some African countries, Czech Republic. Deregulation and privatization trade liberalization will be of little effect. have been trends virtually everywhere, even Experimentation and learning is hence an though the intensity of the reforms has varied important part of the growth process.The East significantly from region to region. Asia Miracle study highlighted that behind the What have we learned from a decade of miracle was the East Asian countries' willingness reforms in these areas? to experiment, and ability to learn from, not to persist in, their mistakes.This approach helped Macro Stability Needs to Be Achieved in them identify, at any point in time, the constraint a MannerThat Is Sustainable and that most severely limited growth, and the right Pro-Growth sequence of policies needed in each situation. There may be situations in which a country The rise in real interest rates in the late 1970s and needs to address many constraints at once,as dur- early 1980s, combined with a variety of com- ing the transition of Eastern European countries. modity price shocks, had rendered unsustainable These situations are rare, however. In most cases, the fiscal stances, debt levels, and exchange rate countries can deal with constraints sequentially,a regimes of most countries in Latin America, East few at a time. Success in addressing one or a set and South Asia,Africa, and the Middle East. Per- of constraints makes it easier to deal with the formance differed sharply between countries that others, and may help establish virtuous circles. rapidly adjusted to these shocks (Korea and East Asian countries in general) and those that did not (Brazil, Nigeria, and many other countries in 4. Lessons from Policy and Latin America and Africa). Institutional Reform As a result,the Structuralist view that inflation Experiences in the 1990s and macro instability were inevitable companions of structural transformation and growth was The economic policy reforms of the 1990s replaced in the 1990s by the strong belief that focused on improving efficiency in the allocation macroeconomic stability was needed for growth. OV E RV I E W 17 The 1990s indeed saw considerable progress in financial crises (as in Indonesia, Turkey), or this area:fiscal deficits declined in most countries, because of the cost of contingent liabilities being exchange rates were adjusted to reflect market shifted to the public sector (pensions in realities, black markets for foreign exchange dis- Argentina), or because of high real interest rates appeared, and inflation declined virtually every- on the public debt (as in Brazil and Jamaica). where. However, while macroeconomic policies Other design flaws help explain why the search as conventionally measured improved in a major- for macro-stability may in some cases have actu- ity of countries, the growth benefits failed to ally been inimical to growth. A preoccupation materialize. In addition, financial crises were with reducing inflation led some countries to numerous, with severe adverse effects on eco- adopt exchange rate regimes that ultimately nomic growth and poverty. proved destabilizing--price stabilization was The openness of the capital account was a key achieved at the cost of appreciating exchange source of fragility that, combined with unsound rates.Fiscal adjustment was often based on highly policies in the financial sector (such as currency distortionary taxes (for example, on external mismatches on banks' or final borrowers' balance trade or on domestic financial transactions); or sheets in the absence of hedging instruments) and on cuts in spending on productive infrastructure appreciation of the real exchange rate, helps to or human capital that proved detrimental to sus- explain many of the crises of the 1990s. Coun- tained growth; or on borrowing abroad where tries such as India have avoided appreciation of interest rates were lower but currency exposure the real exchange rate, and made the opening of increased risks.Hence a single-minded pursuit of the capital account a medium-term goal, to be macro-stability sometimes came at the cost of realized contingent on strengthened economic public spending that might have both increased performance (including fiscal adjustment),export growth and made stability more durable. diversification,and achievement of a sound bank- There are two lessons to draw from this expe- ing system.Chile and Malaysia,among others,did rience. First, even with macroeconomic stability, not hesitate to tax capital inflows when excessive macroeconomic vulnerabilities induced by pol- liquidity threatened to destabilize the economy, icy flaws can be serious, and these can have and they maintained the competitiveness of the tremendous costs. However, indicators of sustain- real exchange rate.These countries fared much able macro-stability are less self-evident than better than those that opened themselves to common indicators of fiscal and external stance external liquidity surges. Notwithstanding the suggest. The inability of financial markets (as theoretical arguments in favor of capital account measured by country risk premia) to predict openness,the evidence on growth is inconclusive most of the financial crises of the 1990s provides and volatility clearly increased.A major lesson of further evidence that unambiguous indicators of the decade is that restrictions should be placed risk are difficult to find. not so much on outflows as on inflows. Obvi- Second,the institutions underlying macroeco- ously, differentiating an unsustainable boom from nomic outcomes and stability matter as much as a positive sustainable trend can be difficult, but stability itself.There is ample evidence that budg- standard indicators of vulnerability such as etary processes influence fiscal outcomes and that indebtedness, evolution of the real exchange rate, countercyclical fiscal policy rules strengthen and current account deficits have proven to be macroeconomic stability. Centralized budget reliable, if imprecise, tools. processes lead to better balanced fiscal outcomes Macroeconomic stabilization programs often over time, and countercyclical fiscal policies suffered from other design flaws, which created shorten cycles and narrow their amplitude. Few serious macroeconomic fragilities. While pri- governments find it politically appealing to run mary deficits did decline over the 1990s, public fiscal surpluses during good times, however. debt increased in most countries, whether Transparent fiscal rules, with stipulated penalties because of the bank recapitalization costs of for noncompliance,may be effective in some con- 18 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s texts.In others,the creation of institutions such as developing countries significantly reduced tariffs oil stabilization funds may be needed to save on imports and dismantled other forms of trade windfalls. One promising example is Chile's restrictions. As in the case of macroeconomic Structural Surplus rule, which establishes fiscal reforms, however, the results varied and, in gen- policy targets adjusted for the variation in growth eral, fell short of expectations.Whereas openness over the cycle.Other proposals,yet to be adopted, helped efficiency and growth in many cases (East have focused on creating an independent fiscal and South Asian countries, Botswana, Chile, policy council, modeled along lines similar to an Mauritius,Tunisia),it failed to do so in many oth- independent central bank, that would set annual ers. Several lessons emerge. deficit limits.Another institutional dimension of First,openness to trade has been a central ele- fiscal policy is transparency. Uncertainty about ment of successful growth strategies. Although the state of the fiscal accounts probably played a the paths taken toward greater integration with large role in generating the volatility of the risk the world economy were far from uniform dur- premiums that developing-country borrowers ing the 1990s, the most successful developing faced during the 1990s.There is also evidence that countries reduced barriers to international trade more transparent budgetary processes brought and foreign investment during the decade. down deficits and debt. Second,trade is an opportunity,not a guaran- For monetary policy, institutional arrange- tee.Trade reforms in some countries yielded few ments are equally important to ensure that low gains in terms of export expansion or increased and stable rates of inflation are achieved and economic growth,while creating social and eco- maintained,and that they last. However,there are nomic adjustment costs.Liberalization of trade in no magic institutional shortcuts to monetary Argentina in the 1980s and 1990s, and in Chile credibility, which has to be earned through anti- in the early 1980s,for example,was accompanied inflationary performance.The institution of an by an appreciation of the real exchange rate that independent central bank--with a commitment reduced the competitiveness of domestic indus- to price stability that takes the form of a publicly tries, and incentives to exports--with adverse announced inflation target--has succeeded consequences for the balance of payments and among emerging market economies during the the real economy. In some countries of Eastern past decade (Brazil, Chile, Colombia, Korea, Europe in the 1990s, trade was liberalized while Mexico, Peru, South Africa, and Thailand).This property rights were not well defined, and the institutional arrangement has the important institutional base for a market economy was not advantages of flexibility (since the central bank is well developed. These, and other institutional not constrained in how it attains its inflation tar- issues preventing the free movement of resources, get) and commitment (since the central bank's often meant that trade reforms did not expand prestige is publicly put on the line). economic opportunities but restricted them instead (Bolaky and Freund 2004). Such experi- Trade Openness, a Key Element of ences do not imply that less trade reform would Successful Growth Strategies, have been desirable, but that trade reform must be done sensibly, as part of an effective growth Can Be Achieved in ManyWays strategy. During the 1980s, the performance of countries Third,countries that have successfully opened that responded to shocks by increasing their out- their economies have done so following a strik- ward orientation (East Asian countries) con- ing variety of policy approaches. They have trasted sharply with that of countries that did not opened up different sectors at different speeds (Latin America, Africa, most countries of the (for example, Bangladesh and India). Some, such Middle East and North Africa). Most policy as China and Mauritius,have achieved partial lib- makers concluded that openness mattered for eralization through the establishment of export growth and, as a result, during the 1990s, most processing zones, and some have combined uni- OV E RV I E W 19 lateral trade reforms with participation in port biases and also to higher barriers to trade in regional trade agreements (Mexico and coun- developing-country markets), and they can be tries in Central and Eastern Europe that have addressed through collective actions. Those now joined the European Union).These differ- actions are best achieved through the Doha ences, and differences in the range of comple- Round and the World Trade Organization mentary policies adopted,make it difficult to pin (WTO).Although there is a role for nonrecipro- down the statistical relationship between trade cal preferences and for reciprocal regional integration and growth.The academic debates approaches, such preferential arrangements are on whether openness to trade causes higher economically arbitrary; they come at a cost to growth are riddled with problems of mea- excluded countries and are not the best way to surement, reverse causation (faster-growing generate the right incentives for investment. countries tend to open their markets more quickly),and omitted variable bias (countries that Design Privatization and Deregulation successfully lower tariffs and increase growth also with Regard to Institutional Strengths and adopt other complementary policies). Weaknesses Fourth, the distributive effects of trade liberal- ization are diverse,and not always pro-poor. Trade Privatization and deregulation have a potentially reforms were expected to be pro-poor because in large efficiency impact and can benefit the pop- most societies the relatively wealthy and urban ulation at large,including the poor.But there is a classes have been more successful at using protec- need to keep expectations realistic as to what tion for their own benefit.The expectation was they can achieve, to establish the institutions that that trade reform would increase the incomes of are key to success, and to design privatization the unskilled.Yet evidence from the 1990s on the strategies taking into account institutional relationship between trade reforms and poverty is strengths and weaknesses. to date mostly indirect. Even where trade policy Privatization and deregulation were key areas has reduced poverty, there are still distributive of reform in the 1990s. Commercial public issues.An important policy lesson is that countries enterprises, development banks, and other forms need to help the affected workers move out of of public interventions in the economy, even shrinking (import-competing) sectors into when meant to address market failures, had expanding (exporting) sectors. become discredited because in many instances Fifth, the preservation and expansion of the they had failed to work well in practice. State world trade system hinges on its ability to strike activist policies using discretion, combined with a better balance between the interests of indus- weak accountability in public sector organiza- trialized and developing nations.Though more tions and weak political accountability of states supportive of development than at the beginning to citizens, were producing costs that were just of the 1990s,the world trade system is still biased too high.The end of communism,and the dereg- against the poor. Notwithstanding a decade of ulation revolution in the United States and the significant expansion of international trade, United Kingdom, added further impetus for the global markets are most hostile to the products wave of privatization that swept across the indus- the world's poor produce--agricultural products, trialized and developing world in the 1990s. textiles, and labor-intensive manufactures--and The results varied. In most countries, privati- problems of escalating tariffs, tariff peaks, and zation brought unambiguous gains in terms of quota arrangements systematically deny the poor more efficient use of resources, more investment, market access and skew incentives against adding and enhanced welfare for consumers.At the same value in poor countries. These problems are time, however, privatization itself failed to bring embedded in the remaining structure of protec- about all the gains for investment and growth that tion in both industrial countries and developing were expected of it. It is also clear that too much countries (the latter owing to their own antiex- was expected from privatization, particularly in 20 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s some areas of infrastructure, but also in terms of address, (2) the risks of infrastructure investment the governance improvements it would bring. In were not appreciated,and (3) governments could cases where the overall package of reform failed not credibly commit to a policy and regulatory to bring about the expected growth,even the effi- regime. At the beginning of the 1990s, the pri- ciency gains of privatization were put in ques- vate sector was expected to enter virtually all tion--a problem that is particularly serious in areas of infrastructure, including roads, but expe- Latin America and Africa, where it has in some rience has since shown that the risks involved in cases derailed the privatization process. infrastructure investment are often too large to In addition, the process of privatization has be taken up by the private sector. often been less than fully transparent and com- Second, if privatization is overstated as a petitive,and this has left sequels that in some cases means of severing the link between economics can be costly to repair--particularly where pri- and politics, regulation as a means of restoring vatization has led to concentration of economic the link is underappreciated.The clearer the sep- power, as it has in many parts of Africa and East- aration between economics and politics, the bet- ern Europe. ter it is for each:commerce will be more efficient Privatization is not just about finding "better and politics less corrupt. But the more complex owners" than the government but about chang- the regulatory issue,the more likely are mistakes, ing governance to separate the commercial from and the less likely that bad regulation (and cap- the political.As is now widely accepted, govern- ture of the regulators by vested interests) will be ment ownership of a commercial firm makes this detected.Even if detected,the poorer the institu- separation difficult. But privatization does not tions, the less likely it is that bad regulation will automatically ensure this separation.The well- be corrected. publicized difficulties of doing business in coun- Third, the reform experiences of network tries such as Russia show that a government can utilities clearly show that there is no universally use a wide range of laws to influence a firm's appropriate reform model, and that privatization decisions without ownership. Separating the is not necessary or indispensable for every coun- commercial from the political requires institu- try. Every restructuring and privatization pro- tions that define and limit government powers. gram needs to explicitly consider the specific Notwithstanding the claims of some privatiza- features of each sector (its economic attributes tion advocates, institutions to support a well- and technology) as well as the country's institu- functioning market economy will not spring up tional, social, and political characteristics. Impor- quickly in response to demand.The lack of effec- tant lessons in this respect are as follows: tive institutions permits predation through sev- · Regulatory reform should promote competi- eral avenues, not just the government: in tion, not control; competition is the most extractive industries, for example, the mining or effective regulator. petroleum firm and the government are beholden to each other, and either could act or · Getting the economics right is key. Under- collude at public expense.Vested interests could standing the source of benefits helps in struc- act through either the public or private sector, turing the reform.A pricing policy that does and poor shareholder oversight over a firm, as not allow adequate revenue cannot improve well as poor public oversight over governments, the situation even if a utility is privatized or an permits misappropriation. independent regulator is established. For Turning to utilities, the second major area of example, as of 2000, in almost all Common- privatization during the 1990s, there are three wealth of Independent States (CIS) countries, main lessons.First,expectations of private invest- household electricity prices covered less than ment in infrastructure have been overly opti- 50 percent and industrial prices were less than mistic--because (1) underpricing continued to 70 percent of the long-run marginal costs of be a problem that governments did not fully supply. OV E RV I E W 21 · Institutions differ, and hence regulatory agen- down interest rates, limiting competition, and cies cannot be easily transplanted. Countries relying on governments to allocate credit and differ greatly in their economic structures and toward more market-based, internationally open in their institutions--the whole chain that systems. Financial liberalization reflected the includes courts (where appeals are made),leg- reaction to the costs, corruption, and inefficien- islatures (where laws are passed), the press cies of financial repression; the demands of gov- (which informs the electorate), an engaged ernment and the public for more financial public (which demands more from govern- resources and services; and the pressures from ments), and academia (which trains regulators greater trade, travel, and migration, and better and encourages studies of problems).These telecommunications. institutional differences across countries Contrary to expectations,financial liberaliza- determine why what is sound regulation in tion did not add much to growth, and it appears one country is ineffectual in another.They are to have augmented the number of crises. As analogous to the differences in performance expected, deposits and capital inflows rose of state-owned firms: they are disappointing sharply as a result of liberalization. But, other in some countries (India, Mexico) but not in than in a few East Asian and South Asian coun- others (Sweden or France). tries, capital markets did not provide resources for new firms. Numbers of stock market listings Pensions are an area in which the private sec- declined, even in the newly created markets in tor's contribution has most clearly fallen short of the transition countries that were sometimes expectations. Eastern European countries with used for privatizations. Also, although relevant almost universal pension coverage trimmed the time-series data on access are weak,and contrary benefits of their defined benefit schemes, out of to expectations,it appears that access to financial fiscal necessity. Many Latin American countries services did not improve substantially after liber- sought to phase out their defined benefit schemes alization. and replace them with mandatory coverage by The explanation for these disappointing out- private providers through defined contribution comes lies largely in weak institutions, concen- schemes. Few of these schemes have lived up to trated economic and political power, and their billing: despite favorable demography their macroeconomic shocks. The implicit and coverage remains low because of the small size of explicit guarantees that were extended to depos- the private formal labor market; their administra- itors and investors weakened the market disci- tive costs have been high, partly because insur- pline that might have limited the activities of ance costs are included and partly because of weak lenders. By the end of the 1990s, much of start-up costs, and they remain dependent on the deposit growth had been absorbed by central government finances because there are few secu- bank debt and government deficits. The state rities besides government paper to invest in.There banks, which remained important during the was really no way to isolate these countries'social 1990s, and financial industrial conglomerates security and pension schemes from their govern- used their increased deposits to expand lending ments without allowing them a greater range of to state enterprises, well-connected borrowers, investment in external markets. and other parts of financial-industrial conglom- erates. Regulation and supervision were weak, The Impact of Financial Liberalization on reflecting not just technical problems but also political pressures for leniency. Eventually the Growth Depends on Underlying Institu- poor quality of lending was exposed in crises, as tions and on Macroeconomic Management were the weaknesses of the bank privatizations in Over the 1980s and 1990s, as part of the general the context of the weak institutional environ- shift to a more market-oriented economy, the ment and the exclusion, in many cases, of inter- approach to finance shifted away from holding national banks. 22 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s The lack of improved credit access reflected incentives to banks for increasing their offshore not only the preemptive borrowing by the public borrowings. All attempts at limiting excessive sector and central banks but also weak informa- inflows depend on political will to restrict them tional and legal frameworks. Lack of information during a boom. In practice, countries often have on borrowers hindered lenders and gave borrow- eased restrictions on capital inflows to prolong a ers no incentive to maintain a good credit record. boom with negative consequences when the Weak legal and judicial frameworks (designed to flows necessarily slowed. protect borrowers and often responsive to eco- Improvements are being made in regulation nomic and political elites) reduced the incentives and supervision in an attempt to limit financial to service debts; they made it difficult for new crises, but experience in the industrial countries, borrowers to gain access to finance by pledging where political and economic power is more dif- collateral effectively and made it difficult for fuse than in developing countries, suggests that lenders to execute collateral. this will not be easy. In the United States, for The 1990s reinforced the old lesson that suc- example, financial economists have raised con- cessful financial liberalization depends on macro- cerns about some U.S.banks being too big to fail. economic management. No banking system, Also in the United States, political forces and reg- however sound in principle,can withstand a seri- ulatory forbearance are often cited as contributory ous macroeconomic crisis. Dealing with a bank- factors in the savings and loan crisis. In many ing crisis is quite complex, involving highly developing countries a few large banks, often political issues of liquidity support to banks state-owned, dominate the system. Bankers and (which can easily contribute to capital flight and major borrowers are often one and the same.Lim- devaluation), bank closure, and handling of its on connected lending are a problem because explicit and implicit guarantees to depositors; the industrial-financial groups are also the main experience in the 1990s suggests that it is difficult entrepreneurs in many countries, even large ones. to avoid socializing the losses and a fall in output. If problems of loan quality develop, the political Further, open capital accounts and volatile inter- strength of the economic and political elite will national capital flows place a large premium on likely lead to regulatory forbearance in loan classi- sound macroeconomic management. Interna- fication and provisioning standards.These kinds of tionally, few attempts have been made to reduce problems suggest that attempts to improve the reg- the volatility of capital inflows (reducing volatility ulatory and supervisory framework need to depends on limiting the upside, not just trying to include a substantial effort to improve market dis- stop outflows when a crisis develops). Chile's cipline, through better information on the banks implicit taxes on short-term inflows appear to and credible limits on deposit guarantees. have had some success in extending maturities, Increased entry of well-known foreign banks, reducing inflows, and limiting volatility against which have a reputation to protect, can also small shocks, albeit at the cost of reducing credit improve the functioning of the system. availability to the private sector (Edwards 1999; Thus, in finance, the 1990s may best be Forbes 2003). Part of India's success in avoiding a regarded as a transition period.The high expec- 1997 crisis stemmed from its limits on banks'(and tations for liberalization were met only in firms') offshore borrowing, even as it allowed resource mobilization. Resource allocation, inflows into the stock market and eased direct which makes a key contribution to development, foreign investment. Indonesia's limits on state did not generally improve.However,much of the banks' external borrowing did reduce their debris of the old financial system was removed growth,but excessive inflows to private banks and by the crises, albeit by government recapitaliza- corporations were a major factor in the 1997 cri- tion bonds that now represent much of the sys- sis. Except for Chile's taxes on short-term flows tem's assets. and some attempts to hold down interest rates, As the connecting link between savers and countries have made few attempts to remove the investors, the contribution of finance to growth OV E RV I E W 23 depends not only on macroeconomic stability and many countries and donors in the 1990s focused reasonable interest rates, but also on the quality of largely on reforming legal and judicial systems-- financial intermediaries and information and of a channel of political accountability that seemed the legal and regulatory framework. Improving more amenable to technocratic solutions, often the contribution of finance to development will using models directly transplanted from industri- depend not only on market-based finance but also alized countries. on sound institutions, appropriate incentives for Most of the reforms had little effect on lenders, further improvements in informational behavior.The ills that they sought to treat--non- and legal frameworks, and, ultimately, on a more meritocratic civil services, weak financial con- competitive political system that is able to reduce trols, opaque or incoherent budget the power of political-economic elites and their processes--are deeply rooted in local political ability to tap the financial system. and institutional arrangements that favor the sta- tus quo. The decade was not all discouraging, how- Pragmatic, Incremental Approaches to Public ever. Homegrown initiatives gave hope for Sector Governance Are More Effective improving government performance. In some Economic performance depends partly on gov- instances,civil society engagement and participa- ernance, which in turn is shaped by underlying tion and innovative applications of information institutions, defined broadly as the "rules of the technology led to improvements in transparency game" that shape the behavior of organizations and accountability in public decision making and and individuals in a society (North 1990, 3).6 A consequently to some increase in government crisis of governance of varying intensity pervades responsiveness, efficiency, and effectiveness.The much of the developing world, with the poor challenge is scaling up these initiatives, given paying the heaviest price for it. political constraints and historical inertia. Public sector reforms in the 1990s sought to The designs of governance reform strategies change the structure of organs of the state, and in the 1990s typically fell into two broad types: incentives within them, in the hope of improv- "big bang" or ad hoc incrementalism. Big bang ing government efficiency and responsiveness. approaches proved to be largely inconsistent with From mega-reforms such as decentralization to capacity constraints and political realities.Their less sweeping reforms in budget or personnel main results were major changes in formal rules: management, the aim was to find a balance new or amended constitutions, new legislation, between the discretion of politicians and bureau- ostensibly independent courts and audit institu- crats over policy making and policy implementa- tions,and so forth.Meanwhile,the informal rules tion and their accountability for decisions and shaping the incentives that face politicians, actions.The fall of authoritarian regimes and the bureaucrats, and citizens remained in place. consequent spread of democratic processes con- Ad hoc incrementalism has also been problem- strained the previously wide discretion of many atic. Many of the ad hoc reforms were symbolic, governments. Decentralization sought to further intended to preserve the old informal rules while limit central government discretion while grant- pretending to reform. Some represented well- ing local governments more managerial auton- motivated attempts of individual or small groups omy. Legal, judicial, and legislative reforms were of reformers who,for lack of support,were under- initiated to establish institutional checks on exec- mined by jealousy, intrigue, or fatigue. More utive power. Public management reforms sought important, they tended to be unrelated to a more to give public managers more flexibility in deci- coherent reform strategy and thus over time many sion making while demanding greater accounta- lost their steam. bility from them for their decisions. Perhaps An important general lesson is that techno- partly because of the immense difficulty of cratic responses to the governance crisis work only addressing problems in political institutions, in very auspicious settings--where there is com- 24 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s mitted leadership, a broad-based coalition in sup- norms. Elected governments are most likely to port of reform,and sufficient capacity to carry the make policies at the expense of the majority and reform process forward. Clearly, these conditions in favor of narrow segments of the population exist in only a few developing countries,and rarely when citizens are badly informed about what in those that most need governance reform. government does, when political competitors State building is a complex process that cannot make credible promises to voters, and requires time, leadership, and social capital. Gov- when society is polarized. Evidence shows that ernance reforms have to find a delicate balance uninformed or polarized citizens and noncredi- consistent with the country's politics, history, and ble politicians undermine the connection culture.What may be needed are highly focused, between voters and politicians in democracies. pragmatic interventions that may be termed Long-run economic growth and the provision of "strategic incrementalism."These interventions public goods are significantly higher in democra- are opportunistic because they exploit the will- cies with more credible politicians, better ingness to reform,but they are grounded in polit- informed citizens, and less social polarization. ical realities and consistent with the capacity Nondemocracies vary substantially as well: those constraints of the country concerned. that have internal checks on the exercise of dis- cretion by the executive seem to perform better, Politics: Checks and Balances Are Central in terms of both growth and public policy per- to Accountability and Results, butThere Is formance, than others. The lesson here is that governments of all kinds,elected or not,are most No SingleWay to AchieveThem credible and most likely to respect property Institutions resolve a number of problems in rights when they face checks and balances on society, of which two are particularly important: their decision making. collective decision-making processes and princi- Another lesson of the 1990s is that policies pal-agent problems.7 Not all preferences can be fail when citizens cannot hold politicians represented in collective decision making, and accountable for poor performance and when principal-agent problems can be reduced but governments cannot make credible commit- never resolved. ments. Credible, sustainable reform depends on Both theory and evidence suggest that the the checks and balances provided through polit- formal rules of democracy do not ensure effi- ical institutions. In democracies, checks and bal- cient, accountable, and credible government, and ances and elections prevent arbitrary policy conversely that nonelected governments are not reversals by governments. But they are not the incapable of responding to citizens or of acting only means to hold governments accountable: accountably.Though the number of elected gov- broad-based political parties can in some circum- ernments grew significantly in the 1990s, the stances substitute for democratic checks and bal- decade produced no clear evidence that elected ances in one-party states. governments perform better in delivering poli- cies benefiting average citizens than do non- elected ones. The experience did confirm, 5. Operational Implications however, that relative to the situation in richer democracies, private investors in most develop- The complete operational implications of this ing-country democracies receive less enforce- study still need to be fully developed. Some pre- ment of their contractual and property rights, liminary ideas are outlined below. and average citizens are not as well treated by the state as special interests. For Analysis By the close of the 1990s, we had begun to understand the complicated interaction of for- On the analytical front,the first implication is the mal political institutions with informal rules and need to redress the balance between analysis of OV E RV I E W 25 policy instruments and analysis of strategies--under- acknowledgment that these two assumptions do standing strategies as coherent sets of actions that not always hold. Convergence is much less a are intended to initiate and sustain growth. Over force now than anticipated a decade or more the years, in institutions such as the World Bank, ago.Within countries such as Brazil, China, and the focus of research gradually has shifted away India, income differences across regions are as from country-specific growth experiences to large as income differences across countries, and focus increasingly on policies--trade, finance, even in relatively small Bolivia, income differ- macro, privatization to name a few--with sec- ences between the lowlands and the highlands ondary importance given to country contexts.8 are large.This recognition implies a need to pay At the same time, outside the World Bank there much greater attention to the forces driving has been increasing emphasis on individual agglomeration and migration, both within and country experiences (for example, Rodrik across countries. 2003b, and the research programs sponsored by the Global Development Network). For Research The second implication is the need to recog- nize country specificities in country economic On the research front, two issues in particular analysis, acknowledging that policies are con- warrant further examination.The first relates to ceived and implemented within a specific insti- development agencies' role in aid-dependent tutional, social, and historic context. Recent countries.The agencies' large role in financing economic and sector work at the World Bank the budget has forced them to be involved in already seeks to achieve a better balance between budget processes, weakening national decision country specificities and the lessons from coun- making and rendering the concept of "owner- try experiences, but more is needed fully to rec- ship elusive in practice"(Kwesi Botchwey,World ognize that country-specific market structures Bank 2005a), particularly in aid-dependent and institutions have a strong influence on policy Africa. Clearly, forms of engagement developed outcomes. In particular, this recognition calls for for project finance do not apply to budget harder and more rigorous economic, institu- finance.There may be a need to explore new tional, and social analysis. approaches to the transfer of resources to these Third,analytical work needs to change its ori- countries, rooted in public finance, such as those entation,away from seeking to assess how far poli- typically used in federated nations that have cho- cies diverge from optimality, to seeking to assess sen rule-based, arms-length systems of transfers. what policy and institutional conditions--for Second,the unit of analysis for economic and capital accumulation, shared growth, productivity social development has traditionally been the growth,and risk taking in a country-specific con- nation state, reflecting the assumptions (outlined text--are needed to set the growth process in above) that nations are homogeneous and that all motion. nations would be able to catch up to the income levels of industrialized countries.There is a rich research agenda on these assumptions that needs For Strategy to be articulated.All nations may not succeed in There is a need to rethink the focus of growth reaching industrialized countries' income levels strategies and of development assistance. Up to within a reasonable time frame--partly because now, that focus has been on the nation state institutions can take such a long time to develop, with the implicit assumptions that (1) develop- but also because the economics of agglomeration ment outcomes within the boundaries of a and poles of development do not necessarily fol- nation state are homogeneous, and (2) all devel- low national boundaries. Research in this area oping countries' per capita incomes could and may yield important implications for the role of should converge with those of industrialized nations and migration, and also for the optimal countries.There is now greater evidence and degree of discretion regarding national policies. 26 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s For Operations 3. This point of view was reinforced by the fiasco of the collapse of the Southern Cone stabilization programs On the operational front, the recognition that of the late 1970s, in which strong currency appreci- not everything needs to be right for growth to ations combined with rapid reductions in tariffs to succeed, and that partial success may sometimes create an adverse shock to industry,ultimately derail- ing the stabilization programs. Most analysts soon be a more pragmatic goal than optimal policies, blamed the collapse on excessive speed, leading to has obvious consequences for the type and extent faulty sequencing of the reform program: they of conditionality associated with development argued that capital accounts had been liberalized too lending. Again in this case, more rigorous eco- soon, without waiting until fiscal probity had been nomic analysis should help to distinguish what established and both trade and the domestic financial are binding constraints, and thus to inform deci- system had been successfully liberalized. 4. This study reviewed the growth experience of eight sions.The record suggests that forecasts need to economies: China, Hong Kong (China), Indonesia, be realistic and mindful of the forces driving Japan, Korea, Malaysia,Taiwan,Thailand, and Singa- growth. pore.The report highlighted the variations in policy and institutional environments under which these economies reached unprecedented rates of growth. For Behavior It emphasized that, with few exceptions, the state in the economies studied had taken an activist role to On the behavioral front, if solutions must be stimulate risk taking in both the private and the pub- found in specific-country contexts, rather than lic sector.It concluded that while highly successful in applied from blueprints, those who advise or East Asia, the institutions needed for replicating this finance developing countries will need more activist role may not be present in other contexts. humility in their approaches, implying more 5. It has been argued, for example, that the increase in India's growth rate in the early 1980s was less the result openness on the range of solutions possible,more of the reforms introduced at that time than of the pri- empathy with the country's perspectives, and vate sector's changing expectations regarding the more inquisitiveness in assessing the costs and future--where the government was credible in ensur- benefits of different possible solutions. ing reduced expropriation risks and a more welcom- ing environment (Rodrik and Subramanian 2004). 6. Public sector governance refers to how the state Notes acquires and exercises the authority to provide and manage public goods and services. Corruption, 1. See Country Note C,"Poverty and Inequality:What which refers to the use of public office for private HaveWe Learned from the 1990s?" gain, is the mirror image of governance: bad gover- 2. Countries successful at "converging" include most nance invariably leads to corruption; but corruption South Asian countries (Bangladesh, Bhutan, India, can likewise perpetrate bad governance. Nepal,Sri Lanka);many EastAsian countries (China, 7. The principal delegates the implementation of a task Indonesia, the Republic of Korea, Lao PDR, to an agent but must monitor the agent efficiently to Malaysia,Thailand,Vietnam); and Botswana, Chile, ensure that the task is accomplished. the Arab Republic of Egypt, Lesotho, Mauritius, and 8. A recent World Bank research project focusing on Tunisia. See Country Note B,"Lessons from Coun- individual country experiences is Aid and Reform in triesThat Have SustainedTheir Growth." Africa (World Bank 2001c). Part 1 Facts of the 1990s Chapter 2 Grist and the Mill for the Lessons of the 1990s T HE ECONOMIC CHANGES OF THE in response to events,2 and this continuous 1990s conformed to no theo- learning makes the empirical sources of lessons ries. In this chapter we review very difficult to isolate in retrospect. Hence this the lessons that can be drawn from the economic chapter attempts to measure the events of the events of the decade,and also the new ideas, the- 1990s against the conventional wisdom of the ories, and issues that were born of those events mainstream of development economists.To pin and of efforts to order and understand them.To down that elusive concept we choose the spe- be sure, facts and ideas so clearly affect one cific expression of the zeitgeist as the World another that it is difficult to separate them Bank's 1991 World Development Report (WDR cleanly: what constitutes the relevant "facts" is 1991). So, had someone known in 1990 the determined by ideas, while new ideas are often direction and magnitude of the changes in poli- the result of attempts to grapple with the facts. tics, policy, and institutional reform, and known Nevertheless, distinctions are helpful to organize how the global economic environment the discussion, so section 1 reviews the facts unfolded in the 1990s, and had they used about developing countries' economic perfor- roughly the same model of market-friendly mance that form grist for the mill of lessons, and development as the WDR 1991, which of the section 2 discusses the ideas that came on to the economic outcomes of the 1990s would they development agenda in the 1990s. not have predicted? On this basis, the 1990s produced five disap- pointments and three pleasant surprises.The five 1. Events of the 1990s: disappointments are: Disappointments and · The length, depth, and variance across coun- Pleasant Surprises tries of the output loss in the transition from planned to market economies in the former Perhaps the most important experiences of the Soviet Union (FSU) and Eastern European 1990s are those that defied not just forecasts but countries. conditional forecasts. Lessons, pleasant and unpleasant alike, emerge from unexpected · The severity and intensity of the interna- occurrences.1 tional and domestic financial crises that rolled Assessing whether outcomes are surprising through East Asia. requires a model that implicitly or explicitly · Argentina's financial and economic implo- links causes with outcomes.Thoughtful people sion after the collapse of its currency convert- continually update their working mental models ibility regime. 29 30 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s · The weakness of the response of growth to · The strong progress in noneconomic indica- reform, especially in Latin America, and the tors of well-being in spite of low growth in unpopularity of many of the reforms. some cases. · The continued stagnation in Sub-Saharan · The resilience of the world economy to stresses. Africa, the paucity of success cases there, and the apparent wilting of optimism around the Five Disappointments "African Renaissance." The three pleasant surprises are: 1. Output Losses during the Transition in the · Bright spots of sustained rapid growth, espe- FSU and Eastern Europe cially in China, India, and Vietnam, through- Everyone knew that the transition from a com- out the decade (box 2.1). munist, centrally planned economy to a capitalist BOX 2.1 Per Capita Growth in the 1990s: Forecast and Actual T he figure below compares actual per capita Africa, Latin America and the Caribbean, Eastern gross domestic product (GDP) growth in the Europe (excluding the former Soviet Union), and 1990s with the forecasts either offered in Eastern Europe and Central Asia (including the for- the WDR 1991 or made in the early 1990s. The fore- mer Soviet Union), and underestimated for India casts correctly predicted the rough direction--that (and South Asia) and China (and East Asia). The Mid- Africa would grow slowly and East Asia fast--but dle East and North Africa, a region about which WDR made mistakes in exactly the regions one would 1991 said little, grew at almost exactly the pace expect. Growth was overestimated for Sub-Saharan forecast. Forecasts for the 1990s--and Reality 10.0 8.0 . %/yr 6.0 4.0 capita, per 2.0 GDP 0.0 ­2.0 ­4.0 -WDR91 EE-GEP92 C-WDR91SAS-WDR91dia-WDR90 EAP-WDR91Chin a-WDR90 SS A-WDR91ECA -GEP93 LA MENA In Forecast Actual Over/underestimate Note: SSA: Sub-Saharan Africa; ECA: Eastern Europe and Central Asia; EE: Eastern Europe (excluding Russian Federation); MENA: Middle East and North Africa; LAC: Latin America and Caribbean; SAS: South Asia; EAP: East Asia and Pacific. G R I S T A N D T H E M I L L F O R T H E L E S S O N S O F T H E 1 9 9 0S 31 economy of one type or another would be neither smooth nor easy. Anticipation that adjustment FIGURE 2.1 costs would cause output to fall and then to rise Depth and Duration of the Transformational Recession: Eastern European and Former Soviet Union Countries led to the expectation of some "transformational recession" (Kornai 2000c), but the depth and duration of the recession were hard to forecast. Output In fact,the depth of the contraction in transi- tion countries is striking. At the trough, their GDP per capita (unweighted) was a mere 42 per- cent of its pretransition peak (figure 2.1).The contractions in individual countries ranged from Depth 20 percent in some countries to about the aver- age in the Russian Federation and to more than Duration 60 percent in Ukraine.3 Data through 2002 show that for most of the FSU/Eastern European countries, the transition Beginning Time of transition has lasted more than a decade, and that for many it will last much longer.4 While some countries Source: Author's own elaboration (for illustration purposes only). (for example,Poland,Hungary) now have output greater than their pretransition levels, on average the Eastern European/FSU countries are only at 1938 level of output by 1953--eight years 84 percent of their pretransition output. For after the end of the conflict. example, even if Ukraine managed to grow steadily at 5 percent a year, starting in 2002, it Not even the most pessimistic observers in would take until 2017 to regain its previous 1990 foresaw that the typical transition recession peak--implying a transformational recession of would be substantially larger than the Great more than a quarter of a century at best. Depression in the United States and that the time A few historical and contemporary reference taken to recover would be more than twice as long points provide useful perspective to the fall in as for the defeated countries afterWorldWar II. output and the length of the transition: A further surprise is the enormous variation in the depth and length of the transition across · In OECD-country recessions, the typical countries.A substantial part of this variation can peak-to-trough fall in GDP since 1950 has be attributed to the speed and depth of policy been only 2.3 percent. reform (see, for example,World Bank 2002c) or · In Indonesia, the worst-hit of the countries suitability for capitalism. Almost no one is sur- that were affected by the 1997 Asian crisis, prised that the transitional recession was shallow GDP per capita fell by 17 percent, and and short in the Czech Republic, Hungary, or regained its previous level four years after the Poland,all of which had the advantages of a more onset of the crisis. European heritage--and hence were eligible for early discussion of accession to the European · In the United States during the Great Depres- Union--and being "good reformers." More sur- sion, output per capita fell by 31 percent, and prising is an apparent U-shaped relationship recovered to its precrisis level in 10 years. between countries' proximity to Europe and the · While the data are obviously somewhat depth and duration of the transition (Mukand and uncertain, the output fall from pre­World Rodrik 2002). Conditions were much worse in War II peak (1938) to postwar trough was 51 Georgia and Ukraine than in more distant parts percent in (West) Germany and 45 percent in of the former Soviet Union such as Uzbekistan, Japan; both of these countries regained their Kyrgyz Republic, andTurkmenistan (figure 2.2). E C O N O M I C G ROW T H I N T H E 1 9 9 0 s FIGURE 2.2 Depth of the Recession, Ratio of Current to Pretransition Output, and Relationship with Distance from Brussels Peak to trough fall in output per capita Current relative to pretransition output o d 90 140 egr a a gi ten olanP a a 80 on Rep. Geor M Rep. 120 enistan gary Sloveni ovin o ova d e FYR an unH Albani 70 egr old ajikistanT urkmT Uzbekistan Slovak a ain M Average 84.5 erzegH a, a Czech a a ten a eni Ukr Azerbaijan 100 d a stan a e oni . . a ag ani Belarus Serbi 60 on an M a a oati Fed Rep. a aced Estoni allf aniu Cr d Fed Rep. enistan Arm eni a azakhK Bulgari verA M Rom an stan e an aniu ansition gyz Latvi an gyz Arm Lith 50 Average 42.3 a a a a ag yrK etr 80 Bosni Lith urkmT Latvi yrK oati a Cr Serbi azakhK verA Russi Russi centage 40 FYR Albani Estoni Azerbaijan Belarus ent/pr e erP a, a Rep. 60 ain oni ani Bulgari curr a ova 30 a gi old ajikistanT Ukr Rep. d aced Rom Slovak M gary M 40 Geor Ratio 20 Uzbekistan olanP Sloveni unH Czech 20 10 0 0 Depth of transition recession and distance from Brussels Current GDP per capita and distance from Brussels 90 1.8 Slovenia Bosnia and Herzegovina 4 1.6 80 Georgia 10 x Czech Rep. 1.4 70 Tajikistan Hungary Moldova 000s, Slovak Rep. Ukraine Azerbaijan 1.2 Estonia 60 allf Armenia PPP Poland 1.0 CroatiaLithuania Latvia 50 Turkmenistan Latvia Moldova Lithuania Kyrgyz Rep. Russian Fed. capita, 0.8 Russian Fed. centage Serbia and Montenegro Kazakhstan Bulgaria erP40 Croatia Estonia Albania per Romania Belarus 0.6 Macedonia, FYR Bosnia and Herzegovina Kazakhstan Bulgaria Belarus 30 GDP Ukraine Turkmenistan Albania Slovak Rep.Romania 0.4 Macedonia, FYR Azerbaijan Armenia 20 Slovenia Georgia Uzbekistan 0.2 Uzbekistan Hungary Moldova Kyrgyz Rep. Czech Rep. Tajikistan 10 0.0 Serbia and Montenegro 400 800 1200 1600 2000 2400 2800 3200 3600 400 800 1200 1600 2000 2400 2800 3200 3600 Distance from Brussels Distance from Brussels Source: Mukand and Rodrik 2002; EBRD 2003. 32 G R I S T A N D T H E M I L L F O R T H E L E S S O N S O F T H E 1 9 9 0S 33 2. East Asian Financial Crisis predicted it. One way of illustrating its wholly The 1990s saw a string of financial crises in unexpected magnitude, and the speed with which the exchange rate, banking system, and which it came on, is to compare the nominal internal and external debt interacted in ways that interest-rate differentials, between borrowing in sharply depressed output--with adverse effects local currency and in U.S. dollars, with the real- on wages, poverty, jobs, and living standards-- ized depreciations (figure 2.3). and caused large losses in the banking system. Even as late as June 1997, the interest rate dif- Macroeconomists, bank restructuring experts, ferential was less than 10 percentage points.Yet and the emerging-market private traders rolled between June and December 1997 the currencies from crisis to crisis--notably in Mexico during of all three countries depreciated by more than 80 1994­95; the Republic of Korea,Thailand, and percent. To be sure, uncovered interest parity Indonesia during 1997­98; Russia and Brazil in often fails as a predictor of exchange rates.But the 1998; and Turkey in 2000--to the most recent magnitude of the difference and the fact that pri- and perhaps most worrisome of all, Argentina vate sector actors were making huge, unhedged during 2001­02. transactions at these interest rate differentials It is worthwhile to discard any presumption emphasize that the world's financial markets, and that all of these crises teach the same lesson, or not just complacent government bureaucracies or that they necessarily teach new ones.There are hidebound multilateral institutions or academics, two reasons why. were caught unawares. First, that there were financial crises in the The crisis in East Asian countries was surpris- 1990s cannot count as a surprise.Every decade of ing because it did not share the characteristics of the 20th century has seen a financial crisis in at many previous exchange rate crises: slow growth least some major countries. Crises have been or declining output, large and growing public more common in the period of floating exchange sector fiscal imbalances, large public sector rates (since the early 1970s) than previously (Eichengreen 2002). But the boom-and-bust FIGURE 2.3 cycle of exuberant capital inflows followed by Interest Rate Differentials Did Not sharp curtailments of lending was a continuing, Predict the Magnitude of the Impending not a new, phenomenon in the 1990s. Devaluation of Three East Asian Currencies Second, some of the crises of the 1990s rein- force old lessons. The links between financial 100.0% crises and banking sector crises reinforced lessons 90.0% from the 1980s, in which a number of financial 80.0% crises in LatinAmerica led to large banking losses 70.0% (Caprio and Honohan 2001);the 1990s'financial 60.0% crises required large shares of GDP to reestablish 50.0% sound banks.Turkey's crisis, as does that in the Southern Cone in the 1980s,teaches the dangers 40.0% of exchange-rate-based stabilization programs 30.0% with inflation inertia and open capital accounts. 20.0% Arguably, the Russian crisis teaches the old les- 10.0% son that if one loses control of the fiscal situation, 0.0% sooner or later the economy will spiral out of Thailand Indonesia Korea, Rep. of control. And, except for its speed and intensity, Interest rate Nominal devaluation, the Mexican crisis of 1994 was not fundamen- differential, June 1997 June­December 1997 tally surprising. Source: Staff calculation from World Development Indicators 2003 By contrast, however, the crisis in East Asia and International Financial Statistics 2003. was a surprise. Even by June 1997 no one had 34 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s indebtedness, or obvious substantial and persist- by the "found decade" of the 1990s. Surely the ent overvaluation of the currency. Even with the substantial and painful first-generation economic benefit of hindsight, economists had a hard time reforms--macroeconomic stabilization, fiscal creating empirical models that predicted it austerity, trade liberalization, privatization-- (Radelet and Sachs 1998) and even observers would pay off with rapid growth and poverty who argue that the crisis was driven by "funda- reduction.Today, the general perception is that mentals" concede that its timing and intensity the growth payoff has been smaller than expected were not anticipated. (figure 2.4). An index of economic reform (Lora 2001a) 3. Collapse of the Convertibility Regime in suggests that during the 1990s the economic cli- Argentina mate improved substantially for nearly every Economically, the decade known as the 1990s country in the region. Not only did the region- could be said to end with the Argentina crisis of wide mean improve, but the variance among 2001.This crisis deserves special mention as a sur- countries declined as well (figure 2.5).This index prise because Argentina had provided the clearest suggests that policies were better in nearly every and, for the better part of the 1990s, most suc- country in LatinAmerica in 1999 than they were cessful example of a trend to reinforce macroeco- in Chile in 1985. nomic stability by reducing the discretion of the Growth in GDP per capita did not reflect government through legal and institutional these improvements in policy. In the early 1990s changes. The exchange rate arrangements that it appeared that the policy changes were finally made the peso convertible at a fixed rate were paying off, but by 1995 the Mexican crisis had a made part of the legal environment (and a part dampening effect on the region. Then when that was especially difficult to alter) and changes another recovery seemed to be in the making, were made in the operation of the central bank to the international financial crises and their reper- make the convertibility immutable.As part of a cussions pushed per capita growth rates to about package of reforms, the convertibility plan was zero, where they have fluctuated since 1998. enormously successful at eliminating Argentina's Loayza, Fajnzylber, and Calderón (2002) hyperinflation and, for a period, in restoring eco- assess with depth and care the extent to which nomic growth. It is no surprise that the demise of the con- vertibility plan was messy politically (the presi- FIGURE 2.4 dent resigned before the end of his term), or Growth Was Much Slower in the 1980s and 1990s than Predicted by Empirical Models economically, since the demise had been made That Linked Growth to Policy Reform very costly by design.What is surprising is the demise itself. First, the plan's initial successes had 4.0% suggested that longevity was possible.The plan 3.5% Predicted from succeeded in reducing rapid inflation and initiat- panel growth 3.0% regression ing a boom in the early 1990s, and it weathered the "Tequila" aftershocks of the Mexican crisis 2.5% reasonably well. Second, the plan was popular 2.0% domestically and praised internationally during 1.5% nearly all of the 1990s, and everyone knew that Actual growth 1.0% per capita average ending it would be costly.5 0.5% 0.0% 4. Lack of Rapid Growth, Particularly in Latin ­0.5% America 1960s 1970s 1980s 1990s Hopes were high that the so-called lost decade of Source: Easterly 2002. the 1980s in Latin America would be followed G R I S T A N D T H E M I L L F O R T H E L E S S O N S O F T H E 1 9 9 0S 35 FIGURE 2.5 Although Nearly Every Country in Latin America and the Caribbean Has Pursued Economic Reform, Growth Has Been Slow Distribution of reform index for 16 Latin America and Caribbean countries (box plots showing mean, 10th, 25th, 75th, and 90th percentiles) 0.75 0.65 0.55 reform of Chile = .49 0.45 Index 0.35 0.25 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Latin America and Caribbean regional growth in GDP per capita (percent) 8 6 4 2 0 ­2 ­4 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Source: Lora (2001a) for data on reform; WDI 2003 for growth. the growth outcomes in Latin America are a sur- All other variables are similarly calculated. prise. The authors do regressions that relate The results thus raise two striking points. growth to transitional convergence and cyclical First, they do not measure up to expectations reversion, structural policies and institutions, sta- about the effectiveness of policy reform. For bilization policies, and external conditions.They instance, for Brazil they suggest that the impact find that the growth rate changes between any of all structural and stabilization policies (except two decades can be attributed to changes in pol- for education) was to slow the country's growth icy outcomes across the two periods,but that the rate during the 1990s by 0.34 percent per year. effect is very small. Most Brazilian policy makers, if not most Brazil- As shown in column 2 of table 2.1, the ians, would probably be surprised to learn that authors find that the coefficients on all of the the policy environment in the 1990s was (net of classes of variables (excepting the institutional education) less conducive to economic growth indicators) have the expected signs and statistical than in the 1980s.7 This unexpected result may significance.Their analysis suggests, for instance, partly reflect the fact that actual growth coeffi- that because of the increase in secondary enroll- cients are in some sense smaller than popularly ment rates between the 1980s and 1990s,growth conceived,or than were reported in"selling"pol- should have increased by 0.7 percent per year.6 icy reform; after all, the link between policy TABLE 2.1 Growth Regressions and "Policy" Impacts, with Two Country Examples Contributions to growth (% per year) of the various growth correlates--calculated as the difference across the two Estimates decades in the variable times the regression coefficient (coeff., Brazil Bolivia Category Variable t-stat.) 1990s vs.1980s 1990s vs. 1970s 1990s vs. 1980s 1990s vs. 1970s Cyclical and convergence Initial GDP per capita ­.018 0.03 ­0.68 0.11 0.13 (3.80) Cyclical recovery ­.227 0.89 ­0.31 ­0.02 ­0.58 (8.52) Growth rate of TOT .072 0.27 0.24 ­0.12 0.04 (4.98) Structural Log "policies" .017 0.7 1.21 0.11 0.47 and "institutions" (secondary enrollment) (6.7) Log (private domestic .0066 0.13 0.07 0.81 0.87 credit/GDP) (4.28) Log (SATI/GDP) .0096 0.41 0.37 0.33 0.28 (3.14) Log (government ­.015 ­0.72 ­0.91 ­0.26 ­0.28 consumption/GDP) (3.18) Log (main telephone .0071 0.36 0.87 0.36 0.39 lines per capita) (2.71) PC ICRG indicators ­.0012 -- (.68) Stabilization "policies" Log (100+inflation rate) ­.0048 0.14 ­0.51 0.88 0.04 (1.89) Std. dev. output gap ­.277 0.14 0.24 0.08 ­0.06 (3.76) RER overvaluation ­.0061 ­0.13 ­0.02 0.17 0.19 (3.90) Systemic banking crisis ­.029 ­0.67 ­0.96 0.58 0 (7.42) Unexplained period effects ­0.48 ­1.72 ­0.48 ­1.72 Contribution to shifts in growth Structural policies .88 1.61 1.35 1.73 Stabilization policies ­0.52 ­1.25 1.71 0.17 Total policies 0.36 0.36 3.06 1.9 Total policies less education ­0.34 ­0.85 2.95 1.43 Projected change in growth 1 ­2.12 2.54 ­0.23 Actual change in growth rate 1.49 ­4.68 3.48 ­0.14 Actual growth 1990s 1.07 1.07 1.53 1.53 Actual growth 1980s (col. 3, 5)/1970s (col. 4, 6) ­0.42 5.75 ­1.95 1.67 Source: Loayza, Fajnzylber, and Calderón 2002, tables II.2, D3, D4. Note: TOT stands for terms of trade; SATI stands for structurally adjusted trade intensity, and measures openness to trade; PC stands for principal component, which extracts the most salient features of the various governance indicators measured by the ICRG, the International Country Risk Guide (www.icrgonline.co); RER stands for real exchange rate. 36 G R I S T A N D T H E M I L L F O R T H E L E S S O N S O F T H E 1 9 9 0S 37 actions and policy outcomes and growth was tinued unprecedented levels of official assistance, often not explicitly quantified.The regression pressure for policy reform, promising develop- implies that reducing inflation from one standard ments in governance,and a not terribly unfavor- deviation above the mean to the mean--that is,a able external climate, no widespread and reduction in inflation of 60 percentage points, definitive take-off has occurred.Living standards from 80 percent per year to 20 percent per and real incomes have declined precipitously in year--would lead growth to increase by 0.2 per- many countries. No country has achieved sus- cent per year (barely a tenth of a cross-national tained growth sufficient to transform its econ- standard deviation in growth rates).8 Certainly, omy and pull its neighbors along. A particular no one has ever advocated a stabilization package disappointment has been the failure of South on the basis of a 0.2 percent per year gain in Africa and Nigeria--the two largest economies long-run growth. and potential growth engines for their respective Second, this careful econometric analysis of regions--to develop into economic power- growth emphasizes that slower growth in the houses. 1990s remains a mystery.The growth regressions include"unexplained"period variables that allow Four Pleasant Surprises growth to be lower, all else being equal.The esti- mated impact of the period variable for the 1990s The more positive developments of the 1990s versus that for the 1970s is 1.72 percent per year; also hold lessons. thus a country with exactly the same policies in the 1990s as in the 1970s would grow 1.72 per- 1. Sustained Rapid Growth in China, India, cent per year more slowly in the 1990s than in Vietnam, and Several Other Countries the 1970s.The implications can be seen from col- The adoption of market-oriented and globaliz- umn 6 of table 2.1,for Bolivia:while policies pre- ing reforms paid off in extraordinarily rapid dict Bolivia's growth to be 1.9 percent per year growth and rapid poverty reduction in the 1990s faster in the 1990s than in the 1970s, the net pre- in formerly socialist and planned economies of dicted growth in the 1990s is actually slower by Asia, including India and China, which together 0.23 percent per year,because the positive impacts account for 40 percent of the developing world's of policy are offset by the period effect of 1.72 population (figures 2.6 and 2.7). percent per year (and negative cyclical reversion The methodological details of the measure- impacts). Bolivians may well ask,"Wait a second. ment of poverty generate substantial disagree- We did all these stabilization and structural policy ment,10 but there is no question that China, changes and grew at 1.53 percent per year in the India, andVietnam have drastically reduced des- 1990s, whereas in the bad old 1970s we grew at titution (consumption-expenditure poverty 1.67 percent per year--¿qué pasa?"The answer based on the dollar-a-day standard) and poverty this empirical analysis gives is that without policy (measured using national standards). Headcount reform, Bolivia's economy would have con- poverty at the international standard of roughly tracted--because of a large, unexplained reduc- US$1 per day has been halved in a single decade. tion in growth in the 1990s that is common to all In Vietnam, 30 percent of the population has countries.This hardly provides a satisfactory reso- moved out of absolute poverty (defined using a lution to the question of slower growth. national standard) since 1993--a historic accom- plishment. 5. Continued Stagnation in Sub-Saharan The successes of these three countries during Africa the 1990s are particularly important because they The failure to create real engines of growth in preclude any facile reaction to the experiences of Sub-Saharan Africa must count as a disappoint- the former Soviet and Eastern European coun- ment, if not a surprise.9 Despite declared good tries and Latin America.If one believes that mar- intentions, a historic process of debt relief, con- ket-friendly and globalizing policies will increase FIGURE 2.6 Accelerating Growth in China, India, and Vietnam Evolution of GDP per capita, India, 1950­2000 8.0 7.8 1990­2000: 4.4% PPP 7.6 1980­1990: 3.8% 7.4 capita, 7.2 per 1950­1980: 1.7% 7.0 GDP 6.8 (ln) 6.6 6.4 1951 1953 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 Evolution of GDP per capita, China, 1952­2000 8.4 PPP 8.0 1990­2000: 7.0% 7.6 1978­1990: 5.7% capita, 7.2 1952­1978: 1.8% per 6.8 GDP (ln) 6.4 6.0 1951 1953 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 Evolution of GDP per capita, Vietnam, 1976­1999 5.8 1989­1999: 5.6% PPP 5.6 5.4 1981­1988: 4.1% capita, 5.2 per 5.0 1976­1981: 1.0% GDP (ln) 4.8 4.6 1951 1953 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 GDP per capita Trend GDPPC, by period Note: PPP stands for purchasing power parity; GDPPC stands for GDP per capita. Source: Author's calculations from Aten, Heston, and Summers (2001). 38 G R I S T A N D T H E M I L L F O R T H E L E S S O N S O F T H E 1 9 9 0S 39 FIGURE 2.7 Poverty Reduction Was Rapid in India, China, and Vietnam in the 1990s India China Vietnam 40 35 Rural 70 35 Rural 30 Urban 60 30 25 Urban 50 25 20 40 headcount 20 headcount $1/day) headcount 15 30 (national) 15 10 (national) 10 (int'l, 20 Poverty Poverty Poverty 5 5 10 0 0 0 1993 1999 1990 2000 1993 1998 2002 Source: World Development Indicators 2003. Note: The 1993 data are from the 50th round of the National Sample Survey; the 1999 data are from the 55th round. growth, then perhaps the three Asian countries Third, after growth in all three countries represent the expected rule and the others repre- accelerated in the 1980s, it slowed down in the sent the exception.There is much to be said for late 1980s and many observers thought that the this view, but there are three senses in which the growth spurt had had its day.But it then took off Asian countries may not match the conventional again, even more rapidly, in the 1990s (see figure wisdom. 2.6 above). First, the reforms these countries undertook Despite the vagaries of the world economy,sev- in the 1990s were pursued in a gradual, piece- eral other countries experienced take-offs and real- meal,and,many would argue,heterodox fashion. ized substantial and sustained economic growth in China dramatically reduced the fraction of pro- the 1990s. New performers included Chile, with duction supplied by state-owned enterprises, but annual GDP growth of 6.4 percent; the Domini- much less by privatizing existing assets than by can Republic (6.0);Poland (4.5);Bangladesh (4.9); allowing the entry of new firms.Especially in the Sri Lanka (5.1);and Uganda (6.8). early stages, the new firms were not private enterprises in the usual sense but township and 2. Improvements in Social Indicators despite village enterprises.And though India undertook Economic Stagnation and/or Crisis trade reform, it did so in a very gradual way: Social indicators--particularly basic education though its average tariffs fell dramatically, it and child health--have continued to improve, retained some of the highest tariffs in the world. Second, while they were near the top of the charts on growth performance, these countries TABLE 2.2 were far from perfect in their policies and insti- Despite Their Rapid Growth, China, Vietnam, and India tutions during the 1990s. Table 2.2 ranks the Rank Low on Many Measures of Institutional Quality three countries on four indicators of the quality Government Rule Control of Regulatory of governance that are often thought to be Country effectiveness of law corruption quality Growth important for growth. On all four indicators these countries ranked either near the middle of China 58 103 63 94 3 the range of countries or in the bottom half. For Vietnam 80 107 105 135 4 India 79 73 86 101 14 example,while China ranked 3rd in the world in Out of: 180 164 151 180 136 growth, it was only 63rd in the world in control of corruption (by these measures). Source: Kaufmann, Kraay, and Mastruzzi 2003. 40 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s often in spite of a lack of substantial progress in to mitigate the impacts with social safety net pro- economic output and in spite of stagnant or grams in education, health, nutrition, and falling wages. employment (Suryahadi, Sumarto, and Pritchett Particularly in a number of Latin American 2003).The relatively small impact on key social countries, enrollment and grade attainment rates indicators of even a large economic crisis is a improved significantly in the 1990s. Brazil took pleasant surprise--as many observers had just 10 years to raise the enrollment rate of the doubted that such mitigating responses were poorest 20 percent of children from 75 to 94 per- politically or administratively feasible or could be cent (figure 2.8).This progress was the result of a of any economic consequence. thoroughgoing educational reform that changed the flow of fiscal funds and responsibilities among 3. Resilience of the World Economic the center,states,and municipalities.The surprise Environment is that the reform was implemented successfully in a difficult economic environment. The biggest misjudgment that I can In many instances, negative social impacts of remember making . . . was the sense of crises were avoided.During Indonesia's deep and profound pessimism about Russian eco- dramatic economic crisis, enrollment in both nomic reform that I had in the fall of primary and secondary school fell only modestly 1998, and . . . if you had said that by 2003, in the first year and then quickly regained or they would be issuing Eurobonds at 300 exceeded precrisis peaks.A recent study tracking basis points spreads, I would have thought the same households over time found that enroll- that it was absolute madness. ment rates for children aged 7­15 were higher in --Lawrence Summers,"Speaking from 2000 than in 1997, before the crisis--and sub- Experience," lecture at the stantially higher for the poor (Strauss et al.2004). World Bank, February 2, 2004 The crisis was accompanied by aggressive efforts While the volatility of capital flows in interna- tional capital markets made policy management FIGURE 2.8 difficult and imposed large costs,the international Enrollment Rates of Children Aged 7­14 in Brazil Rose economy in the 1990s proved robust to a number Substantially for All Income Groups--But Most of negative shocks (see chapter 3). Dramatically for the Poorest First, the overall global economy allowed for (percentage) reasonably stable growth in exports from devel- oping countries.This was despite the large risks of 1999 a major recession in the OECD (had the cycles of 1998 1997 1997 the major economic powers coincided), enor- 1995 1993 1994 mous swings in exchange rates, and large prob- lems in Japan. In the 1990s the annual income 1987 growth of the high-income countries was 6.8 1983 percent--faster than in either the 1970s or 1980s. Second,capital flows were resilient.While the 1975 volatility of financial flows is a major risk and 1992 2001 source of vulnerability,quick recovery of flows in 5th quintile 2nd quintile the aftermath of a crisis can smooth the transi- (richest 20%) 1st quintile tion path. 4th quintile (poorest 20%) Third, in many instances, recoveries from cri- 3rd quintile sis were quite rapid. One of the most frequently Source: "Education in Brazil, 1995­2002." Presentation by Paulo Renato de Souza, for- mentioned features of globalization is the speed mer Minister of Education of Brazil. with which money and information can rocket G R I S T A N D T H E M I L L F O R T H E L E S S O N S O F T H E 1 9 9 0S 41 around the globe.11 An examination of the speed in comparable purchasing power units both over of output recovery shows that the cost of finan- time (created by Angus Maddison [2002]) and cial crisis to the trend growth of output ranged across countries (from the World Bank and the from a minor hiccup (as in Korea) to a long-term PennWorldTables project on price comparisons) deceleration (as in Indonesia).As the impressions augmented the attention paid to the basic facts of of policy makers such as Lawrence Summers the growth process. In the 1990s the research illustrate, the quick recovery of economic activ- emphasized four characteristics of that process. ity (and lowering of spreads) in Russia counts as Growth fact 1: Among the economically most a pleasant surprise indeed. advanced countries, growth has been steady and nearly equal across countries for more than 100 years (except during World War II and subse- 2. A Mill for the Lessons of the quent recovery) (figure 2.9).The average annual 1990s growth of GDP per capita in these 16 countries was almost exactly the same during 1890­1910 During the 1990s three interrelated strands of (at 1.5 percent) as it was during 1970­90 (at 1.8 research provided lessons about economic policy. percent). Except for a boom, with growth aver- They focused on: aging more than 3 percent during 1950­70, the · The theory and empirics of economic growth rate has been very stable. And, except growth; during and just after World War II, growth rates have varied little among the leading countries, · The role of institutions; and with the fastest-growing countries (90th per- · The issue of inequality within and across centile) usually growing only 1­1.5 percent a countries. year faster than the slowest (10th percentile). Growth fact 2: Over the long historical sweep, All three contributed to, deepened, and in the steady growth of the industrialized countries some instances changed the ideas emerging from has led to widening gaps between them and the the 1991 World Development Report. GrowthTheory, Resurgent, Meets Facts FIGURE 2.9 about Development Stable Growth in Industrialized Countries The 1990s saw the resurgence of economic Distribution of growth rates across OECD by period growth theory.To take stock in a few pages of a PC 0.08 theoretical and empirical literature that spans thousands of individual papers,the following dis- GDP 0.06 cussion groups the lessons into four categories: wth,o · New, stylized facts about the growth process 0.04 gr in developing countries; eary 0.02 · The new growth theory itself; per 0.00 · Findings that emerge from the growth- cent regression literature; and erP­0.02 · Problems with the empirical growth-regres- 1860 1880 1900 1920 1940 1960 1980 sion literature. Twenty-year periods Note: The figure is a box-plot diagram of the growth rates of GDP per capita of 16 New, Stylized Facts of the Growth Process industrialized countries for 20-year periods. The resurgence of interest in economic growth, Source: Author's calculations based on data from Maddison (1995). combined with increasingly reliable data on GDP 42 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s poorer countries (Pritchett 1997). Looking at centage points for decades, more than 5.5 per- income inequality among all individuals in the centage points for 20-year periods, and 4.5 per- world,figure 2.10 from Bourguignon and Morri- centage points for the 40-year period. son (2002) shows the fraction of the world distri- Simultaneously, some countries are booming, bution of income that is due to differences across some are growing slowly, some are caught in countries versus the fraction that is due to differ- stagnation or a poverty trap, and some are expe- ences within countries.At the onset of modern riencing sharp declines. economic growth, in the 1820s, only about 10 Large and sustained differences in growth percent of the inequality was due to differences in rates lead to large differences in material well- average incomes across countries. But between being. If a country with a per capita income of then and roughly 1950, this proportion grew US$1,000 (at purchasing power parity) were to steadily,so that today more than 60 percent of the accelerate its growth by 5.7 percent a year--rais- income inequality in the world is attributable to ing its position from the 10th to the 90th per- differences in incomes across countries.Thus in centile in the country growth ranking--then, 1820 one's position within the income distribu- after a 20-year period, its per capita income tion of one's own country was much the most would be triple what it would have been other- important factor, but by 1960 the country one wise. According to every indicator of material lived in was the most important. well-being--from child mortality to consump- Growth fact 3: Growth rates differ enormously tion of electricity--countries at triple the level among the developing countries.Table 2.3 shows of income are qualitatively different places to live the differences in the growth rate of GDP per (table 2.4). capita between the rapid and slow-growing Growth fact 4: Enormous changes in growth countries during periods of 10 years, 20 years, rates occur in nearly every developing country. and for 1960­2000--a period for which data Three facts emerging from research suggest that exist for nearly all countries. In any given period countries sustain episodes of growth and make the difference between the countries in the 10th transitions from one growth episode to another. percentile and in the 90th percentile of the dis- tribution of growth rate is enormous: 6.5 per- TABLE 2.3 Growth Rates Differ Enormously across Countries over Periods from One Decade FIGURE 2.10 to Forty Years Fraction of World Income Inequality Explained by Differences across Countries Difference in growth rates in percent per year Range from to 10th to 90th Two standard 70.0% Period percentile deviations due 60.0% 50.0% 1960s 6.03 4.61 countries 40.0% 1970s 6.96 5.55 inequality 30.0% 1980s 6.81 5.06 across 20.0% 1990s 6.07 5.76 total 10.0% Average for decades 6.47 5.25 of 0.0% 1960­80 5.41 4.07 1970­90 6.23 4.64 differences 1800 1820 1840 1860 1880 1900 1920 1940 1960 1980 2000 1980­2000 5.59 4.34 Fraction Years Average for two decades 5.74 4.35 Theil coefficient Mean ln deviation 1960­2000 4.52 3.83 Source: Bourguignon and Morrison 2002. Source: Author's calculations from Aten, Heston, and Sum- mers (2001). G R I S T A N D T H E M I L L F O R T H E L E S S O N S O F T H E 1 9 9 0S 43 TABLE 2.4 A Growth Rate of 5.7 Percent per Year Higher for 20 Years Would Roughly Triple a Country's per Capita Income GDP per capita, $ purchasing Under-5 child Primary school Poverty Access to Electricity usage Country power parity mortality rate completion ($1/day) improved water (kWh/capita) Countries about $1,000 Benin 1,020 158 39 50 43 Eritrea 950 111 35 7 Nepal 1,350 91 65 37.7 44 39 Countries about $3,000 Indonesia 2,990 45 91 15.2 62 329 Ecuador 3,130 30 96 20.2 70 611 Sri Lanka 3,390 19 100 6.6 46 227 Countries about $9,000 Chile 9,180 12 99 4.2 85 2,011 Malaysia 8,280 8 -- 89 2,352 Source: WDI 2003. The three facts are a lack of persistence of growth Roy 2001),but growth in Mauritius has been far rates over time (Easterly et al.1993);a large decel- from steady.Using the method outlined in Haus- eration of growth in the 1980s (Ben-David and mann, Pritchett, and Rodrik (2004) for dating Papell 1994); and large changes in countries' growth episodes, it is shown that Mauritius has growth rates, often around specific episodes of had two episodes in which growth accelerated, acceleration or deceleration (Hausmann, Pritch- beginning in 1971 and again in 1983, with ett, and Rodrik 2004).While it had long been growth petering out after the first but continuing emphasized that growth was volatile over the after the second (figure 2.12). business cycle of three to five years, growth rates The existence of growth episodes, often have now been found highly volatile over the around identifiable periods of reform or deliber- medium run (10 to 20 years).Unlike most indus- ate policy action, pointedly raises the question of trial countries,which grow at a remarkably steady whether something beyond laissez-faire is feasi- pace, growth in most developing countries ble and desirable to kick-start growth. involves booms, busts, and periods of stagnation alongside periods of rapid growth (figure 2.11) New Growth Theory Very few developing countries have been able to Because it postulated a relationship between sustain growth for longer than two decades.12The policies and growth, the new growth theory ini- accelerations and decelerations in growth rates tially seemed very promising for development from one period to another are often as large as economists. In hindsight, however, its contribu- the differences across countries. Therefore tions to development economics have been few. research has focused not only on average growth Romer and many others succeeded in creat- rates over arbitrary periods (5, 10, 20 years) but ing models in which incentives for purposive also on the initiation of periods of decline and of behavior in innovation were compatible with acceleration.Among the many episodes of rapid equilibrium steady states--that is, in which tech- growth, some end in busts, some revert to slow nological progress was endogenous to growth.A growth, and some continue (table 2.5). recent excellent review by Jones (2004) points For example, Mauritius is an African country out, however, that the "first generation" new- that has achieved rapid growth (Subramanian and growth models had two serious empirical defects. FIGURE 2.11 There Is Some, but Weak, Correlation of Growth Rates across Decades Growth rates in the 1960s versus the 1970s 0.08 TWN SYC SGP JOR ECU KOR HKG 0.06 GNB IDN BRA MUS MYS SYR BRB 1970s PRY LSO TUN ISL TTO COG THA 0.04 CMR NOR ROM GAB PHL DZA the TCD DOM IRL MWI AUT MAR CANMEX ITA GRC JPN MLI URY IJICOLCRIGTM KEN USA FIN BEL CYPBRT FRA RWA BOL CHN TUR in CPV PAN ESP 0.02 BFA HNDLKA GUY GBR NLD PAK GMBNAM ISRCIV GIN NPL LUX EGY ZAF ZWE IND SLV SWE TZA ARG AUSDNK PNG ETH PER 0.00 CHL NZL CHE BDI Growth NGA BGD SEN GIN TGO ZMB BEN GIN NIC IRN MDG JAM ­0.02 OAF VEN GHA NER SLE COM UGA ­0.04 ZAR ­0.04 ­0.03 ­0.02 ­0.01 0.00 MOZ0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 Growth in the 1960s GNQ Growth rates in the 1970s versus the 1980s 0.08 ROM KOR TWN 0.06 CHN THA CYP HKG 1980s LUX BRBMUS CPV PAK SGP 0.04 IND JPN COG IDN the EGY LKA GBR TUR PRT FIN IRL USAITA in ESP AUT NOR DNKNPL SWE GER CAN MYS 0.02 BGD AUS FRABEL MAR UGA HUN JAM CHE CHL ISE ISL COL BDI TUN GNB NLL BFA KEN COM DZA DOM CMR LSO BRA PRY BEN SEN GRC ZWE 0.00 SLE GHA GAB PNG SYC Growth NGA FJI IRN ZAF GMBHND SLVGIN RWA URY CRIMWI JOR VEN TGO NAM SYR MDG PAN GTMMEX ECU ETH TZA PHL TTO CIV ­0.02 NER MLI CAF ZAR ZMB PER HTI GUY BOL ARG PHL GIN TCD NIC HTI ­0.04 MRT ­0.04 ­0.03 ­0.02 ­0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 Growth in the 1970s Growth rates in the 1980s versus the 1990s GNQ 0.08 CHN IRL GUY KNA 0.06 DOM TWN CHL LUX 1990s IRN KOR ARG POL MYS IND MUS 0.04 UGA URY TUN the NPINOR CPVBWA THA PER SYR MWI MLI AUS BGD SLV USA GRD FINPRI EGYLKA in IDN ISR NLDDNK CAN ESP GBR BRB ATG 0.02 NZL VCT ETH BEN CRC ISL TUR HKG PHLPAN JOR GIN LSO GNB CRI BELSWE AUT BRA BLZ HUN BOL MEX FRA GER ITA SYC COL PAK MOZ TTO GTM FJI BFA GHA SEN JPN LCA MAR NAM CHE 0.00 ROM MRT ZAF PRY DZA Growth TZA HNDPNGGAB KEN TCD NERCIV MDG ECU GMB CMR JAM ZWE VEN ZMB ­0.02 COG NIC RWANGA COM TGO ­0.04 BDI ­0.04 ­0.03 ­0.02 ­0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 CAF Growth in the 1980s 44 G R I S T A N D T H E M I L L F O R T H E L E S S O N S O F T H E 1 9 9 0S 45 TABLE 2.5 Episodes of Rapid Growth Set in Context Countries with an episode of rapid growth and year of the initiation of the episode, listed according to growth rates 7 years before the initiation of rapid growth and 10 years after. Growth Growth rate before episodeb rate after (percent) episodea (percent) Negative before (< 0) Slow before ( 0­2) Above average before ( 2) Negative < 0 (after) Ghana (1965) Ecuador (1970) Congo, Rep. of (1978) Gambia (1969) Mali (1972) Algeria (1975) Jordan (1973) Malawi (1970) Indonesia (1987) Nigeria (1967) Rwanda (1975) Panama (1975) Chad (1973) Trinidad and Tobago Romania (1979) (1975) Syria (1974) Slow Dominican Republic Argentina (1963) Zimbabwe (1964) Brazil (1967) 0­2 (after) (1969) Australia (1961) Colombia (1967) Israel (1967) Pakistan (1962) Great Britain (1982) Lesetho (1971) Paraguay (1974) Uganda (1977) Nicaragua (1960) New Zealand (1957) Thailand (1986) Uruguay (1974) Above average Chile (1986) Canada (1962) Spain (1984) Belgium (59) Tunisia (68) 2 (after) Cameroon (1972) Peru (1959) India (1982) Botswana (69) Taiwan (61) Egypt (1976) Portugal (1985) Ireland (1958) Spain (59) Finland (58) Indonesia (1967) Syria (1969) Ireland (1985) Finland (67) Israel (57) Morocco (1958) United States (1961) Korea (1962) Japan (58) Korea (84) Mauritius (1971) Sri Lanka (1979) Mauritius (1983) Malaysia (70) Singapore (69) Thailand (1957) China (1978) Nigeria (1957) Congo, Pakistan (1979) Rep. of (1969) Panama (1959) Denmark (1957) Source: Hausmann, Pritchett, and Rodrik 2004. Note: An episode of rapid growth is a seven-year period in which growth accelerates by at least 2 percent per year over the previous trend, to a rate that is 3.5 percent per year or faster. a. Growth rate in the 10 years from 7 years after the initiation of the growth episode (t+7 to t+17) (with at least 7 years of data--no episodes after 1986) b. Growth rate in the seven years before the initiation of the episode of rapid growth (t, ­?7) First, nearly all these models have scale effects long run and on incentives for expanding the that predict that larger economies will grow technological frontier. It is not particularly useful faster, but (as is clear from figure 2.9 above) the for most developing countries, whose primary long-run growth of the industrial countries has interest is in short-to-medium-term growth and been very steady, and it is difficult to make this technological catch-up. In particular, only a tiny prediction match the data.13 If there are scale fraction of the observed variation in growth rates effects, either they are very small or they are off- over medium to long periods can possibly be set by many other factors working to reduce explained by differences in the steady-state growth growth. rates of the technological frontier (Bernard and Second, since the new growth literature was Jensen 1999). Essentially, the steady-state growth primarily about the steady-state growth of the of the technological frontier cannot be less than richer industrial countries, it focused on the very zero for theoretical reasons (the economy would 46 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s the policy, institutional, and structural correlates FIGURE 2.12 of growth,sometimes examining causal channels. Growth Episodes in Mauritius, 1950­2000 Decompositions into proximate determinants of growth. A substantial amount of empirical 9.6 research examined the extent to which growth 9.4 was explained by the measured accumulation of 1990­2000: 4.2% observable factors of production (principally 9.2 physical capital,labor,and human capital/school- 1983­1990: 5.5% ing) versus a residual (Senhadji 2000; Bosworth 9.0 and Collins 1996, 2003;World Bank 1993; and many others).This literature found that: 8.8 1971­1978: 6.7% · While measured factors, particularly physical 8.6 capital, are strongly correlated with growth, 1950­1971: 1.3% 1978­1983: 0.9% they explain at most half of the cross-country 8.4 variance in growth (Easterly and Levine 8.2 2003). · While for many reasons one would have 8.0 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 expected faster growth in the developing countries, the growth rate of the residual is Source: Hausmann, Pritchett, and Rodrik 2004. puzzlingly low in most of them: negative in many and less than the OECD rate in nearly disappear), and it cannot be more than about 1 all (Bosworth and Collins 1996, 2003). percent a year (since empirically this is about as · A large debate about the residual in East Asia high as any long-run estimate of total factor pro- concluded that there was no particularly East ductivity growth in leading countries).This limi- Asian pattern.15 tation implies that if a country were, by some means,to accomplish a shift from the lowest to the The main point to be learned from this litera- highest steady-state growth of the technological ture is that the empirical findings of growth frontier,its growth would accelerate by only about accounting do not have any particular policy 1 percentage point a year. Since even at the 40- implications. The findings did not resolve the year horizon, the 10th/90th percentile range of question of causality or of the determinants of growth rates is 4.5 percentage points a year,14 dif- accumulation.First,the proportion of growth that ferences in the steady-state growth of productivity can be attributed to increases in capital, rather cannot account for much of the observed variabil- than in productivity, depends on the way one ity of growth rates across countries even over a counts the correlated components (Klenow and period as long as 40 years. Rodriguez-Clare 1997).If one attributes to capi- tal all of the increase of growth accounting, then Empirical Findings from the capital accounts for much of growth.By contrast, Growth-Decomposition and Growth-Regres- if one attributes to capital only the component of sion Literature growth that is due to changes in the capital/out- One of the principal, if unintended, benefits of put ratio (capital deepening), and attributes the the new growth theory for development is that it remainder of capital stock growth to increases in legitimated empirical work into the determi- productivity, then productivity shifts appear to nants of economic growth.Indeed,it unleashed a drive much more of growth. veritable flood of such studies.One branch of the Second, naming the residual from growth literature decomposed growth into its proximate accounting something such as total factor produc- determinants, and a different branch examined tivity (TFP) has its dangers. Equating the residual G R I S T A N D T H E M I L L F O R T H E L E S S O N S O F T H E 1 9 9 0S 47 withTFP (and particularly then equatingTFP with a current account imbalance might emerge, some notion of technological progress) implies that reserves might be low, the country might restrict the measurement is correct in every other respect. imports in order to cope with the shortage of A good deal of research has emphasized how vul- foreign exchange, a black-market premium nerable the TFP calculation is to a variety of might develop,and/or the country might pursue methodological problems. The functional form ambitious import substitution behind protective and the share assigned to capital affect the results a barriers to save foreign exchange.The same syn- great deal.And the use of cumulated investments as drome and set of symptoms could be set in a proxy for capital, particularly public capital, has motion if relative prices fail to respond to a fall in no firm theoretical foundation and can create large the terms of trade. In the 1990s examples of this difficulties in deciding whether to attribute a lack syndrome often ended with a large recession of growth to "low productivity with a large (after a period of slow growth) and/or a crisis amount of factors" or "low efficacy of investment followed by a substantial devaluation and a stabi- in creating factors"(Pritchett 2000). lization program. If all of these symptoms (slow "Growth" regressions and the correlates of growth. export growth, import barriers, black-market An enormous literature16 relies on linear regres- premium, exchange rate instability, and so forth) sions of growth on explanatory factors X and the were caused by the same underlying syndrome, lagged level of income.Here the explanatory fac- the data and regressions would not be able to dis- tors included in "X"17 can be characterized as: tinguish which particular symptom "caused" the slow growth.19 · "Policy outcome" or "policy" variables such as inflation, trade shares, or exchange rate Problems with the Empirical Growth-Regres- overvaluation; sion Literature · "Institutional" variables such as the rule of law, This is not the place to review the myriad governance indicators,or corruption;and methodological problems of the cross-national growth-regression literature.20 But from a policy · "Structural" variables such as geographic loca- point of view, it is useful to point out three main tion. problems. To summarize the lessons from this literature First, growth regressions cannot predict turn- without getting bogged down in detail, one needs ing points.The basic problem is that most indica- to take a "syndrome" rather than a "symptom" tors of policies, institutions, and structure are approach to understanding the correlates of much more stable than indicators of growth per- growth.18 The growth regression literature has formance (Easterly 2003a).This leads to two fur- identified five syndromes that lead to low growth: ther problems. First, it is very difficult to that is, five phenomena for which the overall distinguish causality since, unlike characteristics weight of the evidence suggests an important rela- such as the rule of law or effectiveness of the tionship, even if it cannot be identified precisely bureaucracy, growth episodes often have discrete (table 2.6). starting dates.Second,a finding that over a period In a sense,growth regression results have been of,say,30 years the rule of law is on average asso- unfairly criticized for a lack of robustness when ciated with higher growth does not give much they are able to indicate "syndromes" but not guidance as to how to initiate and sustain an "symptoms."An example is persistent exchange episode of growth. rate overvaluation, a common and particularly Second, in spite of the name, growth regres- well- documented syndrome of the 1970s. A sions are really not about growth but about the country that pegged its exchange rate but had level of output. One of the puzzles of the growth domestic inflation in excess of international lev- literature is that even though in a mechanical els saw its real exchange rate become overvalued. sense a growth regression explains growth,nearly In such a situation its export growth might slow, all of the functional forms used are simply 48 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s TABLE 2.6 "Syndromes and Symptoms" Summary of the Empirical Growth-Regression Literature Generalizations that cannot be made Low growth syndrome Description of the syndrome Symptoms based on robust evidence Governance and Governments that are not High corruption, ineffective Democracy is good (or bad) for institutions developmental (for example, bureaucracy, low rule of law, growth predatory states, weak states, high risk of expropriation, Authoritarian governments/ "captured" states, elite- high transaction costs, dictatorships are good (or bad) dominated states) political instability for growth Uncertain property rights Insufficient private investment Need for formal western-style definition and enforcement of property rights Macroeconomic Inability to maintain a reliable High and variable inflation/money Reducing inflation will increase and stable means of payment supply growth/exchange rate growth internally and externally depreciation, high fiscal deficits, Reducing a fiscal deficit will persistent episodes of exchange- increase growth rate overvaluation, periodic financial crisis, debt-service problems External policies Policies that inhibit the ability Low growth of imports/exports, Free trade will raise growth of goods, ideas, and finance disincentives to existing and from abroad to contribute to new export products, persistent increasing productivity exchange rate overvaluation, "irrationally" distorting trade measures Financial sector Financial sectors that cannot Low monetary depth, high Immediate financial provide credit to private penetration of central/state- liberalization is necessary sector investors owned banks, legal systems for growth that do not facilitate contract enforcement Bad luck Geographic location or natural Landlocked, continent indicators, endowment that creates susceptibility to disease pressures inimical to conditions, point-source development resource dependence Source: Author's own elaboration. dynamic variants of a model in which levels of policies can be sustained and implemented in policy or institutional variables affect levels of the absence of adequate public sector organiza- economic output.21 tions and institutions. Third, their specification of policies is incor- rect.22 Recent empirical research has found that, Institutions when a measure of institutional quality is included in cross-country regressions, the Well before the 1990s, Adam Smith and Max explanatory power of other variables, including Weber from their different perspectives high- all measures of policies, becomes negligible lighted the role of institutions in the develop- (Acemoglu, Johnson, and Robinson 2001; ment of a market economy and the formation of Rodrik, Subramanian, and Trebbi 2002; Easterly a capitalist society. In the 1950s and 1960s, econ- and Levine 2003; IMF 2003f).This reasoning omists writing about development were aware suggests that good institutions matter more for that the challenge faced by a plantation econ- growth than do good policies. From a syndrome omy, or a dual economy, differed from that faced viewpoint, it is easy to see that this is not an by a society with no concentration of economic assertion that "policies don't matter"--of course and political power (Rostow 1952, 1960; Adel- they do. Rather the question is whether good man and Morris 1965). And Latin American G R I S T A N D T H E M I L L F O R T H E L E S S O N S O F T H E 1 9 9 0S 49 economists of the Structuralist school saw in the did not become discredited because of some ide- legacy of colonialism, embedded in institutions ological shift that made development banking serving the interests of a small landed elite, the intrinsically taboo, or some theorist's discovery source of economic performance inferior to that that in principle activist policies could not of the United States or Canada (Prado 1972;Fur- improve on laissez-faire. Development banks tado 1963).In turn,their perception formed part became discredited because in many instances of the justification for an activist state: inflation they did not work in practice: activist policies helped to mobilize resources from the wealthy using discretion, combined with public sector elite who resisted more efficient forms of taxa- organizations and institutions with weak tion;the state sponsored investments in manufac- accountability (including that of states to citi- turing,particularly in capital-intensive industries, zens), produced costs that were just too high. because old economic interests resisted change Thus the lesson of the 1990s is not that insti- and the risks inherent in new industrial activities; tutions matter, but rather: and price controls did not have serious economic · How much they matter; consequences because the concentration of wealth precluded the redeployment of resources · How difficult it is to work around their in response to changes in demand (Seers 1962). absence or to make transitions in institutions; In Rosenstein-Rodan's words, the challenge and of development has long been how to make sure that "nature makes a jump" (Rosenstein-Rodan · How difficult it is to improve institutional 1984;see also Meier and Seers 1984).Some coun- quality. tries have radically transformed and modernized In the 1990s it was hoped that the strength of institutions through revolutionary and authori- policies could overcome the weaknesses of institu- tarian means (as in Russia in the 1920s,Turkey in tions, and that policies capable of generating eco- the 1930s, and China in the 1950s) or through nomic prosperity would ultimately generate large-scale nationalization (as in Bolivia and incentives for establishing effective institutions. In Madagascar in the 1960s, and former Zaire and response to the costs and perceived inefficacy of Sri Lanka in the 1970s). In others, the state has interventions where institutions were weak,much taken on a developmental role--as in Korea, of the reform effort of the decade sought to limit Brazil,Turkey, and India in the 1950s, 1960s, and governmental discretion in decision making. On 1970s--acting as entrepreneur on a large scale balance, the risks of failure were deemed larger and also introducing the incentives needed for than the benefits of allowing discretion to an import-substituting industrialization. activist developmental state, and this led to an Import substitution policies, command and emphasis on rules that reduced discretion: for control,central planning,"big push,"a coordinat- example, dollarization, fiscal rules, or integration ing role for the state, balanced growth, linkages, in larger economic unions. However, as discussed all have a strong economic rationale, which was below in chapters 8 and 9,it is virtually impossible persuasively put forward in the early develop- to eliminate the discretion exercised by the nation ment literature (Rosenstein-Rodan 1943; state.A better way forward is to look for institu- Hirschman 1958; Gerschenkron 1962; Rostow tions to control the exercise of discretion rather 1962).23 These big ideas found a particularly than for policies or rules to eliminate discretion, receptive environment in the 1950s and 1960s. which have proved to have a risky downside. But though the interventions generally suc- ceeded in igniting growth, they failed to sustain Improving Institutional Quality it--a failure that has discredited strategies based In any society, institutions need to perform cer- on active inducements to industrialization. tain core functions:ensuring the security of peo- This is where "institutions" come into play. ple and property, establishing mechanisms for For example, the notion of development banks collective decision making,and organizing a state 50 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s capable of carrying out key government func- Inequality can affect economic growth tions.An important realization of the 1990s was through several channels."Equal societies have that the design of institutions for these core func- more social cohesion, more solidarity, and less tions can take a broad range of forms. Most of stress;they offer their citizens more public goods, the empirical work on the importance of institu- more social support, and more social capital" tions has focused on the link between institu- (Deaton 2003a), and hence are more capable of tional performance and economic performance, sharing the costs and benefits of improving eco- and almost none examines the link between nomic policies--which facilitates forming con- institutional design and performance.Yet it is now sensus and decision making. More equality also broadly acknowledged that merely adopting facilitates agreement on the provision of public some other country's laws and formal regulations goods,such as health,water supply,and waste dis- is no guarantee of producing the same institu- posal, that have strong externalities.25 tional performance, and that different arrange- Aghion, Caroli, and Garcia Penalosa (1999) ments can lead to equally successful outcomes. explain the positive impact of equality on growth For example, China's arrangements for secur- by reference to market structures and microeco- ing property rights differ from India's, yet both nomic incentives.They find that a better distri- countries offer relative security to investors. In bution of wealth reduces credit constraints, and Soeharto's Indonesia, by contrast, the enforce- that greater availability of credit has a significant ment of one's property rights depended on one's positive effect on growth.If individuals have lim- closeness to the ruling elite. Similarly, financial ited borrowing capacity, reallocating capital systems in the United States and the European toward the poorest will increase aggregate pro- Union have different institutional foundations, ductivity.They also find that better distribution but both perform at comparable levels of effi- of wealth will reduce instability at the individual ciency. As another example, different democra- level and hence at the aggregate level,and conse- cies perform very differently, showing that the quently will mitigate the impact of instability on formal institutions of democracy are insufficient aggregate growth. to ensure a government's accountability and While there is clear evidence that greater credibility. While in some countries these insti- equality augments growth, there is much igno- tutions have delivered satisfactory outcomes, in rance on how greater equality can be achieved.A others they have not (see chapter 10 below). large agenda for deeper research exists on how to Within countries, institutions do not function achieve greater equality, including investigating homogeneously: De Soto (forthcoming) has the impact of public spending on equity, in both shown that within a country the enforcement of a static (incidence of public spending) and a property rights varies across income and social dynamic sense (changes in individuals' earnings groups, with the least security for the least privi- potential). leged, and he has documented the ensuing adverse consequences for investment incentives Inequality and Institutions--A Two-Way and for incomes. Street Recent literature has emphasized the important links between the distribution of assets in a soci- Fairness and Growth ety and the institutions that emerge. Knowledge Another important strain of ideas in the 1990s about how institutions emerge and are estab- was a resurgence of interest in inequality and lished is still rudimentary, but economic research equity.This important concern has many dimen- in the 1990s has provided some insights. sions, but we focus here on the impact of First, economic incentives influence what inequality on economic growth and on the type of institutions emerge and when.For exam- interrelationship between inequality and institu- ple, the enforcement of property rights to land tions.24 will depend on the benefits of enforcement rela- G R I S T A N D T H E M I L L F O R T H E L E S S O N S O F T H E 1 9 9 0S 51 tive to the costs--a ratio that depends on the to economic and social opportunities. In 1800, extent to which other landowners enforce their Argentina's per capita income was equivalent to property rights. In an extractive economy, if all that of the United States, whereas Brazil's, landowners enforce their property rights, the Chile's, Mexico's, and Peru's were only 40­50 alternatives for laborers decline, and so do their percent of that of the United States.Two cen- wages, and as a result, rents on land increase. If turies later,Argentina's per capita income is one- landowners in general do not enforce their prop- fifth that of the United States, and Brazil's, erty rights,it is uneconomical for one of them to Mexico's,and Peru's are one-fifth or less,whereas enforce his or hers: the alternatives for laborers, Chile's has remained about the same.The reason and hence their wages, will be greater because for this divergence in economic performance is they can exploit land where property rights are that the United States,where access to economic, not enforced. Only when this coordination social, and political opportunities was much problem is resolved do economic incentives broader,was able to create a much greater flow of become sufficient for enforcement of property economic opportunities.27 Because population rights (Hoff and Stiglitz 2001).26 densities were much lower in the United States, Second, the concentration of economic and there were fewer incentives to establish preda- political power influences the breadth of access tory institutions oriented toward extracting rents BOX 2.2 How Money and Power Can Influence Patterns of Institutional Development S ocieties' choice of institutions depends on a exploded; trains hurtled off tracks; bridges collapsed; variety of contextual variables, including his- locomotives collided in a grinding scream of steel. tory as embedded in existing institutions, the Railway law and tort law grew up, then, together. In distribution of economic and political power, and the a sense, the two were the same" (Friedman 1985, type of problems these institutions seek to solve. quoted in Glaeser and Shleifer 2003). Glaeser and Shleifer (2003) show how money and Traditional theories of regulation--justifying reg- power subverted the workings of justice in the United ulation on the grounds of market failures--fail to States in the late 1800s and early 1990s, leading to explain this evolution. Glaeser and Shleifer show that the creation of regulatory agencies to handle matters a fundamental change made it more efficient for previously resolved by courts. American society to increase its reliance on regula- Before 1900 numerous commercial and other dis- tions: "Commercialization and industrialization of putes in the United States were resolved through pri- the economy in the second half of the 19th century vate litigation: "Courts ruled on such matters as created firms with vast resources. As the scale of corporate liability in industrial accidents, on anti- enterprise increased, the damage from industrial competitive practices such as railroads' rebates, on accidents rose proportionately, as did the incentives safety of foods and medicines, and even on the con- to avoid paying damages. The cost of influencing jus- stitutionality of income tax." Private litigation was tice, however, did not rise as fast. As a consequence, the principal way to deal with the socially harmful individuals and small companies were unlikely to pre- acts that had been accelerated by the industrial rev- vail against "robber barons . . . . Woodrow Wilson olution: "Trains were also wild beasts; they roared repeatedly complained about the failure of the courts through the countryside, killing livestock, setting to stand up to large corporations because, he said, fire to houses and crops, smashing wagons at grade `The laws of this country do not prevent the strong crossings, mangling passengers and freight. Boilers from crushing the weak.'" Source: Glaeser and Shleifer 2003. 52 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s for the benefit of a small elite. Except in the Particularly when taking into account the substantial United States and Canada, growth in former increases in inequality during the transition, it seems that the median household is potentially even worse European colonies has been influenced by the off than the evolution of the mean incomes suggests. concentration of economic and political power, 4. See Country Note E,"Eastern Europe's Transition: which has restricted access to economic and Building Institutions." social opportunities, created less secure property 5. The "precommitments" in the Argentine case were rights,and influenced the course of development as credible and were fought for as creditably as one for several centuries. could wish. No one could argue that Argentines should have been asked to suffer more to defend the Some recent illustrations of how inequality convertibility plan--and fail. influences institutions and economic growth 6. However, the regression estimated impact (0.017) come from India and the United States.In India, times the change on ln (secondary enrollment) in the state ofWest Bengal,tenancy reform in the (0.41--this is in natural logs so it is roughly a per- late 1970s increased the share of output that ten- centage increase) is that 0.7=(0.017) × (0.41) × 100. 7. The general impression (Birdsall 2002) and most ants could retain, and strengthened tenancy indicators of policy change (Lora 2001a) suggest rights;a sharp increase in yields ensued (Banerjee widespread and substantial policy reform in Latin et al. 2001; Banerjee, Gertler, and Ghatak 2002; America in general, and in Brazil in particular. Hoff 2003). Another instance of concentrated 8. That is, mean of Log(100+inflation rate)=4.79, stan- economic and political interests influencing dard deviation is 0.4047.The growth impact of a institutions comes from the United States in the reduction of one standard deviation is ­0.0048 × 0.4047=0.0019, which corresponds to a reduction early 1900s, when the government decided to from 80 percent to 20 percent inflation. regulate matters hitherto left to private parties 9. There was hope that with the passing of the first gen- and the courts.The reason for this shift was a per- eration of political leaders, their successors could ception that judges and the courts had been so effect a transformation. For instance, President Clin- corrupted by powerful economic interests as to ton in 1998 met with five heads of state (Afwerki, Kabila, Kagame, Museveni, and Zenawi) and pro- be unable to render fair judgments (box 2.2). claimed a "new Africa Renaissance sweeping the continent." Unfortunately, only two months after Clinton's hopeful declaration all five leaders were at Notes war--mostly with one another. 10. For example, there is an enormous literature on the 1. Not all unexpected occurrences teach lessons, how- measurement of poverty in India, with a large num- ever.An analogy with earthquakes might help.Earth- ber of estimates of poverty rates.The controversy quakes cannot be predicted; the lessons learned from stems from two major sources: (1) the discrepancy one are not about better prediction. But the physical between the rate of growth of personal consumption and economic damage from an earthquake can be expenditures in the national accounts and that of predicted based on its magnitude, location, and the reported expenditures in household surveys; and (2) design and construction of the affected structures. changes in the method of the surveys between the These damage prediction models can be updated in 50th and 55th rounds of India's National Sample response to events--particularly when they fail Survey (NSS). Here we use the estimates of Deaton badly, in predicting either too much or too little (2003c),which are based on the NSS,and use a plau- damage. sible technique to adjust for the changes in the recall 2. Accused of changing his views, Keynes responded period between the rounds. with a famous quip: "When the facts change, I 11. One of the less frequently mentioned is the fickleness change my mind--what do you do,sir?"(Moggridge that this induced in the opinions bandied about in 1976, 163­64). financial and international institutions. In 1996 the 3. While these data on GDP per capita are widely East Asian model was perhaps misunderstood but it accepted, they are controversial. Many analysts argue was unquestionably sought after; in early 1998 the that mismeasurement of the value of pretransition financial crisis threatening the entire region was cited output and the undercounting of the new informal as proof that the whole East Asian model was mis- sector mean that the fall in output has been less guided and that the economies needed fundamental severe than it appears (see, for example, Shleifer and reform if they were to recover from crisis.By 2000,as Treisman 2004). Everyone, however, agrees that the Korea sailed out of the crisis, that type of talk ended recession in most countries was deep,long,and hard. as abruptly as it had started. G R I S T A N D T H E M I L L F O R T H E L E S S O N S O F T H E 1 9 9 0S 53 12. See Country Note B,"Lessons from Countries That (symptom) while another had 10 empirical measures Have SustainedTheir Growth." that were sufficiently highly correlated that multi- 13. Individual national economies and the world economy collinearity caused their individual t-statistics to fall are enormously larger today than 100 years ago.Take the below some threshold level when included jointly. best possible case,in which the relevant"market size"is Then growth regressions with one symptom of each just the national economy.The U.S. economy in 1990 syndrome would give roughly the right answer, was 55 times larger than in 1870,but the growth of per while mechanical "horse races" to assess robustness capita GDP was 2.6 percent during 1870­80 and 1.8 would give the wrong answer. percent during 1980­90. Of course, the relevant vari- 20. See reviews byTemple (1999) and Pritchett (2000). able in the models is"market size."This can be defined 21. Just as in the Solow model, the growth impact of to include trade with the rest of the world,so that Mar- policy reform is a transitional effect in moving from ket Size=Domestic Economy+*(Rest of World) so one level of income to another. Chapter 8 addresses that =1 implies all countries face the same market the question of whether the impacts of policy size.But this makes the empirical point about the prob- reforms as estimated from aggregate (growth) mod- lem of historically nonaccelerating growth in the lead- els are consistent with those from microeconomic ing countries even stronger because (1) with reduced studies of gains from reform. transport costs and lower trade barriers has increased, 22. Also in chapter 7, this volume returns in depth to a and (2) the rest of the world has grown,so that the true second empirical problem:there are many economic market size growth for the United States could be models that do not predict a linear relationship much higher than the 55-fold increase in the U.S. between measures of policy outcomes or a summary domestic economy. statistic of policy actions. 14. The two standard deviation range is 3.8 percentage 23. The arguments made by the early authors have since points a year (table 2.3). been formalized in a number of theoretical papers 15. Some countries had rapid growth of the residual (Murphy,Shleifer,andVishny 1989;Hoff and Stiglitz while others had growth, when correctly measured, 2001;Rodrik 2001a) that identify the market failures at about the OECD level or less. that these interventions addressed and clarify theo- 16. This gained momentum with Barro (1991) and has retically the economic intuition on which they were been reviewed many times, perhaps most notably by based. Temple (1999). 24. Focused on economic policy, this study does not 17. Over and above the proximate determinants of address concerns about the inequities in diseases such investment in physical or human capital, which may as AIDS or malaria, nor about access to social ser- or may not be included depending on how individ- vices such as education,nor about gender equity,nor ual authors want to examine channels of causation. about specific social injustices.These may be at least 18. A syndrome is an underlying disease process that as important as the present topic. manifests itself in related symptoms.A doctor might 25. See Country Note C,"Poverty and Inequality:What be interested in which of a particular set of symp- HaveWe Learned from the 1990s?" toms (nausea,fever,pains) best predicts an underlying 26. WDR 2001 provides other examples of how eco- syndrome or differentially diagnoses one syndrome nomic incentives affect the emergence of institutions versus another.She might be interested in the under- that sustain the functioning of markets and the dif- lying biological causes behind certain syndromes but ferent coordination or risk-reducing problems they be equally interested in the impact of a syndrome on are meant to resolve. the health of the patient, no matter what its etiology. 27. Whereas at most 2 percent of the population voted 19. In the absence of some well-developed notion of a in Argentina, Brazil, or Chile at the end of the syndrome, it is not good practice to criticize the 1800s, more than 10 percent voted in the United robustness of a variable because its significance level States, where the participation rate in voting also is changed by the addition of another variable.Nor is increased much faster. Three-fourths of the U.S. deciding what are the robust correlates of growth by population owned land,whereas less than a fifth did simply throwing all available variables into a mechan- so in Argentina, and far fewer did in Brazil. Access ical procedure (Sala-i-Martin 2003). Suppose, for to education was similarly better distributed in the instance, that one syndrome had only 1 measure United States. Country Note A Economic Growth from the Very Long-Term Perspective of History T o economists, the reasons for superior to Europe's?Analyzing long-run growth countries'growth performance implies analyzing complex historical and politi- lie in the incentives created by cal processes. Scholars have proposed numerous policies and institutions. Typically, economists hypotheses to uncover the "deep" exogenous examine questions such as the following: Does forces at work. Three of these forces have taxation discourage savings and investment? Are received the most attention in economic a country's public institutions capable of enforc- research: geography, openness to foreign trade, ing property rights and delivering public goods? and institutions. Does the trade regime facilitate integration in Since Montesquieu,authors have periodically the global economy? Does the private sector considered geography as a "deep," truly exoge- have sufficient confidence in the future direction nous factor explaining economic performance. of policies? Are fiscal policies consistent with the There are many channels through which physi- long-run solvency of the public sector? Underly- cal geography affects growth: a country's geogra- ing these questions is the goal of ascertaining phy shapes its natural-resource endowments (oil, whether the country has enough incentives to minerals, diamonds) and public health environ- use existing resources and to accumulate capi- ment (disease burden), and limits or enhances tal--factors that ultimately determine growth agricultural productivity (quality of soils,amount performance. of rainfall). A striking one-third of the world's On the other hand,since the early writings of gross domestic product (GDP) is produced in the Montesquieu, Karl Marx, and Max Weber, social temperate ecological zones within 100 km of the scientists and economic historians have sought to world's navigable waterways;these zones amount uncover the deep underlying reasons for the to only 4 percent of the world's landmass.Almost wealth and poverty of nations. From their per- none of the industrialized countries are in the spective, economic growth is deeply rooted in a tropics or subtropics, or landlocked (Gallup, country's history and structural conditions, Sachs, and Mellinger 1999). which shape societies' choices and ultimately the In an authoritative study on the long-run policies and institutions they adopt. In this con- geographic determinants of development, social text, they seek to address questions such as the ecologist Jared Diamond (1997) argues that following: Was there something special about Eurasia had large geographical advantages over Western Europe that made it the birthplace of the Americas and Africa, and that these lie at the the Industrial Revolution, notwithstanding the heart of current income disparities. He argues earlier and superior scientific and technological that since plant and animal species spread most achievements of China and the Arab world, and effectively within ecological zones, the east-west Asia's superior agricultural productivity? Some- orientation of the Eurasian landmass made it eas- thing that was absent elsewhere, in the Americas, ier to diffuse early human technologies across the Africa, or Asia? Why did India's economy stag- continent. As a result, Eurasia enjoyed a larger nate for so long, despite its enormous natural diversity of plant and animal species, and thus wealth and human skills that until the 1700s were easier domestication of useful species, than did 56 F RO M T H E V E RY L O N G - T E R M P E R S P E C T I V E O F H I S TO RY 57 societies in America and Africa--continents that factor can potentially reveal valuable insights are oriented north-south. High-productivity about the true causes of countries' development agriculture led to large, dense, stratified societies, successes and failures. For instance, Western with subsequent advances in technology Europe benefited both from the geographical (weaponry, oceangoing ships) and political advantages of east-west continental orientation organization. discussed by Diamond (1997) and from being Another important causal factor widely stud- predominantly a coastal region in the temperate ied in economic history is international trade, ecozone (Gallup, Sachs, and Mellinger 1999).All and hence access to sea-based trade and proxim- of this made land scarce and valuable (Herbst ity to export markets. Economists since Adam 2000). Additionally, rugged mountainous relief Smith's An Inquiry into the Nature and Causes of effectively separated Western Europe into a sys- the Wealth of Nations have argued that foreign tem of "competing jurisdictions of decentralized trade helps economic growth because it encour- power," constantly warring with one another, ages the division of labor and specialization, the none being able to completely defeat and control locomotives of human development throughout the others (Landes 1998).These factors raised history. Numerous empirical studies have shown returns to innovation, discovery, and adoption of that openness and growth proceed "hand-in- new warfare techniques, which later gave Euro- hand" (Balassa 1978; Sachs and Warner 1995; peans first-mover advantage over other parts of Ben-David 1993; Krueger 1997; Frankel and the world. Rose 1996). Historically, two remarkable exam- Another perspective on the "European eco- ples of early industrialization are the island nomic miracle" (Jones 1981) emphasizes the key nations of Britain and Japan,one being the cradle role of risk, uncertainty, and predation in the for- of the industrial revolution, and the other the mation of European socioeconomic and political only successful industrialization in Asia until the institutions, issues that have long been familiar to second half of the 20th century. economists.Two fundamental forces were at work. Finally, recent analyses of long-term eco- First, the distribution of income in Europe was nomic growth emphasize the crucial role played unusually equal relative to,for instance,Asia."Late by institutions, that is, the formal and informal Manchu China with a population of some 400 rules and norms that govern personal and social million supported a two percent of the population behavior; including the socioeconomic arrange- elite which consumed in the late 1800 one fourth ments that constrain predation by the state and of the national product.Whereas a relatively large individuals (North 1990). Institutions were a share of the population of Europe had risen above central focus of attention for the Latin American subsistence level before the Industrial Revolution, Structuralist school in the 1950s and 1960s; but the vast majority of the population in China, today they are seen by some as the most signifi- Northern India,Mesopotamia and Egypt hovered cant factor in long-term development. Recent slightly above or below the threshold of survival." econometric and case studies have shown that Smaller inequalities allowed for a social environ- even when controlling for historical endogene- ment able to keep in check predation by both ity, institutions remain "deep" causal factors, individuals as well as organized groups, including while openness and geography operates at best the state."Europe alone managed the politically through them (Acemoglu, Johnson, and Robin- remarkable feat of curtailing arbitrary power, thus son 2001; Rodrik 2003b; Rodrik, Subramanian, reducing risk and uncertainty, encouraging more andTrebbi 2002). productive investment, and promoting growth." Of course, hypotheses explaining economic Second, in Europe's natural environment, adverse growth are difficult to verify empirically. Coun- shocks (in the form of floods, earthquakes, epi- tries are not amenable to controlled experiments, demics) were less common than in, for example, and, in reality, complex relationships are at work China or India and contributed to reducing the among geography, institutions, and trade. Each uncertainty associated with investment decisions, 58 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s thereby increasing the incentives for individuals and firms to invest and take risks (Jones 1981). Economic development is deeply embedded in countries' history and structural conditions, and understanding these is essential for the design of effective growth strategies. The rest of this report highlights the widely differing strategies adopted by successful countries, and that the art of formulating effective growth strategies lies in careful consideration of country-specific factors, opportunities, and constraints. Chapter 3 Something Special about the 1990s? D EVELOPING COUNTRIES' growth in the 1990s was FIGURE 3.1 higher than in the 1980s but Growth in Developed and Developing Countries, 1963­99 lower than expected. Per capita income in the 6.0 median developing country continued to grow more slowly than that in the median Organisa- tion for Economic Co-operation and Develop- 4.0 ment (OECD) country, and except in East and SouthAsia the growth of developing countries as 2.0 a group continued to trail that of OECD (figure 3.1). This encouraged perceptions that the 0.0 reforms of the 1990s had a disappointing payoff 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 for growth. Was developing countries' growth really dis- Developed OECD countries (21) Developing countries (69) appointing during the decade? If so, could it be that an adverse external environment held them Source: World Bank, World Development Indicators 2004. back? What about these countries' own policies? Note: Median of GDP per capita growth rates, five-year moving averages; 69 develop- Did policies improve over the decade as much as ing countries and 21 OECD countries for which complete series exist for 1961­2002. The Republic of Korea is included in the group of developing countries. commonly believed? Section 1 of this chapter reviews measures of growth performance over the 1990s, and section 2 examines global eco- the median OECD country, the per capita nomic trends, finding that the external environ- income of the average citizen in the developing ment was not unfavorable for growth. Section 3 world has grown at least as fast as that of his or analyzes the extent of economic reforms during her counterpart in OECD countries (table 3.1). the decade,showing that they were extensive and This difference highlights the several possible significant. ways in which developing countries'growth per- formance can be measured, and that the appro- priate gauge to use depends on the purpose.One 1. Developing Countries' measure is the weighted average of per capita Growth during the 1990s income growth, where the weights are develop- ing countries' population or gross domestic While per capita income in the median develop- product (GDP).This approach provides an indi- ing country has grown more slowly than that in cator of aggregate performance but does not cap- 59 60 ECONOMIC GROWTH IN THE 1990s TABLE 3.1 Growth in Developed and Developing Countries, 1960s­1990s 1960s 1970s 1980s 1990s 1990­2002 Median GDP per capita growth Developing countries (69) 2.0 1.8 ­0.5 1.3 1.0 Developing countries (78) ... 1.9 ­0.3 1.0 1.0 Developing countries (93) ... ... ­0.2 1.0 0.9 Developing countries (116) ... ... ... 0.6 0.8 Developed OECD countries (21) 3.8 2.7 2.0 1.9 1.8 Import demand growth High income ... 3.4 6.3 7.1 6.0 Developing countries 15.2 4.7 1.8 7.9 7.2 GDP per capita growth High-income countries 4.2 2.6 2.5 1.8 1.7 Developing countries 3.6 2.9 0.7 1.8 1.8 GDP growth High-income countries 5.4 3.8 3.1 2.5 2.3 Developing countries 5.8 5.1 2.6 3.5 3.4 Source: World Bank, World Development Indicators 2004. Note: Average cumulative GDP per capita growth rates for the decade (for example, 1990s are 1991­2000 growth rates). Except for the median, all statistics refer to the sum of GDPs for the relevant group of countries, divided by the total population of that group. ture variations across developing countries. For In the discussion in this chapter, the growth example, if China grows fast but Malawi does performance of a group of countries is measured not, the aggregate measure will largely reflect using median statistics,a standard consistent with China's performance simply because of China's the focus of this report on policies and on coun- size. Even if many smaller countries parallel try performance--for which the country is the Malawi's level of accomplishment, the weighted most suitable unit of analysis. average will still not capture their growth per- Using the median as a summary statistic is not formance. without its own problems, however. First, if the An alternative, the unweighted average, cor- number of small countries is sufficiently large, rects for this deficiency by treating each country the median will represent their performance.But as an observation, but it introduces another bias it is not always proper to give the same weight in in that it may give an undue weight to outliers. the analysis to China as to Uruguay,to India as to For example, the discovery of oil in Equatorial Estonia, and gauging developing countries' Guinea enabled that country to grow at rates growth with medians might bias results in the about 15 percent a year in the 1990s. Because of "many small countries" picture just as using Equatorial Guinea's high growth rate,measuring aggregate data biases them in the "China-India" Africa's performance through an unweighted direction.Thus in the analysis that follows we average would give the impression that the con- report aggregate data and medians together tinent grew faster than it did.Bosnia and Herze- whenever possible. govina is another example.Following the war,its A second problem with using medians as a per capita GDP grew by 80 percent.This obser- summary statistic is that they can easily be vation alone would skew the estimate of the regionally biased. Comparing developing coun- average for all developing countries upward by tries' growth across decades using means requires 0.5 percent. The higher the variance, the less us to follow a fixed sample of countries that have representative the arithmetic average will be. a complete GDP per capita series for the period SOMETHING SPECIAL ABOUT THE 1990S? 61 in question.But out of the 69 countries that meet alone grew from 9 to 17 percent.) If one excludes this criterion from the 1960s on, nearly three- China's contribution of 1.1 percent (13 percent fourths are from only two regions:Latin America of all developing countries' GDP in 1995, grow- and the Caribbean (21) and Sub-Saharan Africa ing at 9 percent over the 1990s) the developing (28). Extending the sample of countries and, of countries' growth rate drops to 0.7 percent. If necessity, limiting the timeframe to the later one excludes India (7 percent, growing at 3.6 decades, we find a smaller pick-up in the median percent) it drops further to 0.6 percent. Because country growth rate over the 1990s.The magni- of the size of these two economies, and ignoring tude of the apparent growth contraction in the that the concept applies to economies, not to 1980s also shrinks depending on the number of people, it has even been argued that these trends countries included in the sample. Extending the demonstrated absolute convergence (Fischer 1990s to include 2001 and 2002 reduces median 2003). per capita growth significantly from 1.3 percent A regional perspective on growth during the to 1 percent for the sample of 69 countries 1990s suggests that the change in performance (mostly as a reflection of Latin America's per- during the decade is the result of changes in per- formance), and reduces it marginally for the formance in three regions: the Middle East and larger samples (table 3.1). North Africa and Latin America and the As noted above, estimating growth on the Caribbean, where performance improved, and basis of the performance of developing countries Eastern Europe and Central Asia, where perform- as a group, where the per capita GDP is the ance deteriorated (table 3.2). aggregate GDP for all developing countries divided by the total population, shows a much higher per capita growth of 1.8 percent a year in 2. Global Economic Trends in the 1990s.This figure reflects the above-average the 1990s growth performance of China and India, two countries whose combined GDP increased dur- During the decade a tremendous increase took ing the 1990s from one-seventh of all developing place in global integration in goods,services,and countries' GDP to one-fourth. (China's GDP investment flows (table 3.3). TABLE 3.2 Developing Countries' Growth, 1990s: Regional Perspectives (growth of GDP per capita for median country in each region) 1960s 1970s 1980s 1990s Sub-Saharan Africa (28­41) 1.4 0.6 ­ 0.9 ­1.0 ­ ­0.9 ­0.4 ­ ­0.2 South Asia (5­6) 1.7 0.7 3.1 ­ 3.3 3.0 ­ 3.1 Middle East and North Africa (6­12) 2.4 3.6 ­0.2 ­ 0.7 0.5 ­ 1.0 Latin America and the Caribbean (21) 1.8 2.6 -0.7 1.8 Europe and Central Asia (1­24) 6.2 4.3 1.5 ­1.8 ­ 1.0 East Asia and Pacific (8­12) 2.1 5.5 1.6 ­ 2.6 2.9 Source: World Bank, World Development Indicators 2004. Note: This table gives a regional perspective on developing countries' growth. As indicated in the text, median estimates depend crucially on the choice of countries. denotes the estimate for the group of countries with complete GDP per capita data over four decades, 1960­2002. For exam- ple, for 1970 Sub-Saharan Africa, 0.6 percent is the median growth for the group of countries with complete GDP series over these four decades. If one includes in this group the countries for which GDP data are available starting in 1970 through the 1990s, the growth rate rises to 0.9 per- cent. Figures in parentheses represent the number of countries with complete GDP per capita series starting in 1960 "" as well as the number of countries for which data are available for the 1990s. For example, in Europe and Central Asia indicates that only one country (Turkey) has a com- plete GDP per capita series that starts in the 1960s, but 24 countries have data available for the 1990s. 62 ECONOMIC GROWTH IN THE 1990s TABLE 3.3 Global Integration, 1980­2000 Exports and imports of goods and services FDI flows as a share of GDP as a share of GDP (in current US$) (in percent) 1980­85 avg. 1990 2000 1980 1990 2000 Developing countries 41.0 39.2 55.3 0.3 0.8 2.6 Developed countries 40.7 39.1 46.4 0.6 1.0 5.1 Sources: Trade and FDI (foreign direct investment) flow figures from IMF Balance of Payments Statistics; GDP from the World Bank's World Develop- ment Indicators. Note: A nominal measure is used here because of the difficulty in obtaining price deflators for services trade. Regardless of whether a real or a nom- inal measure is used, there was still a large increase in trade integration on the "goods" side in the 1990s. The analysis in the rest of this chapter regarding goods trade uses "real" measures, with nominal values deflated by the relevant price indexes. Notwithstanding this improvement in inte- ronment over a period of time, the Loayza and gration, some empirical studies suggest that neg- Easterly studies provide an intuitively appealing ative external shocks reduced developing explanation as to why the reforms of the 1990s countries' growth in the 1990s below their did not generate the results expected: surely the potential, offsetting the positive impact of eco- reason must be a negative external shock. nomic reforms. Easterly (2001) finds that the slowdown in Analyzing policy reforms in developing economic growth of developing countries' countries, Loayza, Fajnzylber, and Calderón OECD trading partners provides a potential (2002) and Easterly (2001) find that countries explanation for the decade shift dummies: that reformed their policies--improving macro- economic management, opening the economy, In contrast, the effect of OECD trading liberalizing the financial sector,and so on--grew partner growth on LDCs' home country faster in the 1990s than countries that did not growth is huge (if anything, implausibly take such steps. At the same time, both of these large) . . . one less percentage point of studies find that an unexplained negative shock OECD trading partner growth is associ- affected all countries not only in the 1990s but ated with 2.1 less percentage points of also in the 1980s. In their studies, dummy vari- home country growth. ables for the 1980s and 1990s are large and statis- tically significant,implying that,other things held Comparing the two 20-year periods 1960­70 equal,developing countries grew nearly two per- and 1980­90, Easterly highlights that growth in centage points more slowly in the 1980s and high-income and developing countries moved 1990s than in the 1960s or the 1970s. closely together (figure 3.2), and that growth Both of these studies have come under criti- slowdowns in industrialized countries over the cism. Loayza, Fajnzylber, and Calderón's study 1970s and 1980s mirrored and preceded those in has been criticized because it uses outcome vari- developing countries during the "lost decade" of ables,such as a country's ratio of trade to GDP,or the 1980s. But when he examines the 1990s sep- financial depth or inflation rates, to represent the arately, he finds that the relationship between extent of economic reforms, and thus is unable developed and developing countries growth is to attribute changes in economic growth to not that strong. actual changes in policies, or to claim that Latin There is little doubt that that the 1980s--the America's reforms added significantly to that "lost" decade--had its share of negative shocks, region's economic growth (Rodrik 2003a). including declines in primary commodity prices, Ignoring these specification problems, if a collapse in oil prices, a sharp hike in U.S. inter- period dummies truly capture the external envi- est rates,debt crises,a sudden stop in capital flows SOMETHING SPECIAL ABOUT THE 1990S? 63 TABLE 3.4 FIGURE 3.2 Unprecedented Growth of World Trade, 1990s Growth Slowdowns in Developed and (average annual cumulative growth rates, percent) Developing Countries 1960s 1970s 1980s 1990s 6.0 World GDP growth 5.5 3.7 3.0 2.6 4.0 World trade growth 7.9 5.7 4.9 6.6 Developing countries 2.0 Export growth 6.1 5.2 5.2 7.4 Import growth 6.4 7.7 1.6 5.3 0.0 High-income countries Export growth 8.3 6.3 5.0 6.6 1961 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 Developed countries Import growth 8.7 4.9 5.3 6.8 Developing countries Source: World Bank, Global Economic Prospects. Source: World Bank, World Development Indicators 2004. Note: Imports and exports are of goods and nonfactor services. Note: GDP per capita growth rates (weighted averages), 1961­2002. an annual rate of 8.5 percent from less than 2 percent in the 1980s. to developing countries,and a collapse in import demand from developing countries. Increased Integration through Trade But it is harder to tell the same story about Taking export shares in GDP as a measure of the 1990s. Certainly the financial shocks of the globalization shows that developing countries are decade--notably inAsia,the Russian Federation, now more integrated with the world economy andTurkey--may suggest that the external envi- than are high-income countries (figure 3.3). ronment was inimical to developing countries' Between 1990 and 2000, developing countries' growth. But trade and capital flows, which are export revenues doubled as a share of GDP,rising two major channels whereby economic per- from 12.5 to almost 25 percent. formance in industrialized countries affects The exports-to-GDP ratio for the median developing countries, were both expanding dra- developing country rose during the 1990s from matically during the 1990s.Analyzing indicator- 24 to 27 percent (figure 3.4). by-indicator the variables underlying the The exports-to-GDP ratio alone is not a suf- external environment suggests indeed that the ficient measure of trade integration, because 1990s was not unfavorable to developing coun- changes in relative prices will alter it even if there tries' growth. are no changes in real flows. But given that the median developing country sustained a stable real Unprecedented Expansion of International exchange rate over the 1990s (figure 3.5),the ris- ing ratio does suggest increasing integration into Trade world trade over the decade. World trade boomed in the 1990s.The overall Services trade rose during the decade,though volume of trade grew 2.5 times faster than world goods trade integration dominated the globaliza- GDP, compared to the average of 1.5 times over tion scene (table 3.5). the period since World War II. Import demand expanded at an accelerating pace in industrial- Diversification into Higher-Value Products ized countries and also recovered in developing The composition of developing countries' countries (table 3.4). exports shifted dramatically from agricultural and The merchandise export growth of develop- resource exports into manufactures, which now ing countries quadrupled in the 1990s, rising to constitute nearly 80 percent of exports from all 64 ECONOMIC GROWTH IN THE 1990s FIGURE 3.3 FIGURE 3.5 Export Shares of GDP, 1980­2002 More Competitive Real Exchange Rates (percent) 200 25.0 22.5 150 20.0 17.5 100 15.0 12.5 50 10.0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 Source: World Bank, World Development Indicators 2004. High-income countries Developing countries Note: Median real effective exchange rate index. Sample of 52 developing countries, 1995 = 100. Source: World Bank, Global Economic Prospects 2003. income group of 1980 also raised the share of FIGURE 3.4 manufactures in their exports, but somewhat less Faster Integration into World Trade during the 1990s rapidly, to nearly 70 percent.1 60 Not All Developing Countries Benefited The rising tide of exports did not lift all boats, 40 however. Forty-three countries achieved no increase, on average, in their merchandise exports 20 between 1980 and 2000 (World Bank 2003b), and their share in world exports declined.A group 0 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 of least-developed and Sub-Saharan countries 197 197 197 197 197 198 198 198 198 198 199 199 199 199 199 200 remained highly dependent on primary com- modity trade, and often on just one or two com- Developing countries (68) modities. Many in this group were plagued by Developed OECD countries (23) civil conflict and engaged in politically motivated Source: World Bank, World Development Indicators 2004. trade embargoes.Both of these factors were often Note: Median exports of goods and services as a percentage of GDP; 68 developing and complicated by inept governance. 23 OECD countries, 1970­2000. Balancing the sample of developing countries over dif- ferent decades--that is, including all countries for which data are available--does not International Terms of Trade: No Clear Trend significantly alter these results. Developing countries suffered no large terms- of-trade shocks during the decade (figure 3.8). developing countries (figure 3.6 and table 3.6). Although primary commodity prices had Many countries successfully diversified into declined in the 1980s relative to those of manu- medium- and high-technology products. factures, they were stable over the 1990s, fluctu- Countries whose incomes were low in 1980 ating without a clear trend (figure 3.9).Analyzing managed to raise their exports of manufactures the impact of commodity-price cycles during from roughly 20 percent of their total exports to the 1990s, a number of studies have concluded more than 80 percent (figure 3.7). As a result, that overall these cycles were more effectively many grew quickly and entered the ranks of managed than in previous decades and should today's middle-income countries.The middle- not have adversely affected the prospects for SOMETHING SPECIAL ABOUT THE 1990S? 65 TABLE 3.5 Exports and Imports of Goods and Services as Shares of GDP, 1980­2000 (current US$) Export and import Developed countries Developing countries shares of GDP 1980 1990 2000 1980 1990 2000 Goods 34.2 31.0 36.8 37.1 31.6 45.9 Services 7.7 8.1 9.6 7.9 7.6 9.4 Goods and services 41.8 39.2 46.4 45.1 39.2 55.3 Sources: Trade figures from IMF Balance of Payments Statistics; GDP from the World Bank's World Development Indicators. developing countries' growth (World Bank, FIGURE 3.6 Global Economic Prospects 2001). Even for non- Developing Countries Diversified into Manufactures, oil-exporting Sub-Saharan countries, changes in 1960s­1990s real incomes were generally small and the policy environment was much better than in the previ- 80 ous decades. Oil prices were generally lower in the 1990s 60 than in the 1970s and 1980s, and since most developing countries are oil importers rather 40 than oil producers, this drop has meant a gener- ally favorable trend (figure 3.10). 20 Expansion of International Capital Flows 0 The largest increase in integration in the 1990s 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 was in investment flows. Globally, FDI as a share Manufactures Agriculture Metals Fuel Food of GDP more than quadrupled between 1990 and 2000. Source: World Bank, World Development Indicators 2004. After reaching a nadir in the late 1980s, capi- Note: Median merchandise export shares, three-year moving average; 60 developing tal flows to developing countries expanded rap- countries with complete data, 1962­2002. For some countries a few missing observa- tions were filled in by simple linear interpolation. Technically, because the median idly during the 1990s (figures 3.11 and 3.12). could be a different country in each year, the sum of the shares is generally smaller Capital flows to all regions were unambiguously than 100. Thus, they have been renormalized. stronger on average than in the 1970s and in the 1980s. Private capital flows boomed, rising from 1 ment in all countries, mature and developing.A percent of developing countries' GDP in the growing consensus holds that push factors 1980s to more than 4 percent in the 1990s, and explain much of the acceleration. Servén,Albu- displacing official flows as the principal source of querque, and Loayza (2003) find that global fac- finance. Official flows declined slightly, to less tors became progressively more important in than half a percent of developing-country GDP FDI flows over the 1990s, and by the end of the by the end of the decade. decade could explain (in a statistical sense) nearly A combination of pull factors (liberalization half of these flows.The flood of credit was facili- of capital accounts and domestic financial mar- tated by advances in financial technology, such as kets) along with a range of push factors (regula- vehicles for risk pooling (mutual funds), the tory changes in mature market economies) development of new financial instruments resulted in an explosion of international invest- (derivatives and securitization), and a decline in 66 ECONOMIC GROWTH IN THE 1990s TABLE 3.6 Diversification Took Place before the 1990s Manufactured export shares Regions (number of countries in sample) 1970 1980 1990 2000 Sub-Saharan Africa (24) 7.0 8.1 18.5 15.9 South Asia (5) -- 53.2 76.1 77.3 Middle East and North Africa (12) 9.6 13.2 26.9 26.5 Latin America and the Caribbean (21) 15.4 25.8 36.1 46.6 Europe and Central Asia (4) -- 60.7 64.3 79.2 East Asia and Pacific (6) -- 3.5 39.2 58.2 Source: World Bank, World Development Indicators 2004. Note: Median share of manufactures in overall merchandise exports; sample of 72 developing countries. Numbers of sample countries are shown in parentheses. For some countries, a few missing observations were filled in by simple linear interpolation. Average for given and two preceding years used--for instance, "1970" is an average of 1968­70 export shares. --. Not available. FIGURE 3.7 FIGURE 3.8 Developing Countries' Exports of Manufactures, No Large Terms-of-Trade Shocks for 1981­2001 Developing Countries, 1990s In middle-income countries, manufactures make up 2.0 70 percent of exports . . . world 80 Manufacturing 1.0 of 70 exports (%) GDP of 0.0 60 share (percent) 50 ­1.0 2002­ 40 Percent Resources ­2.0 countries 30 exports (%) 20 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 10 Agricultural 60 (87) 80 (96) 90 (104) exports (%) exports,1981 0 Source: World Bank, World Development Indicators 2004. Middle-income 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 1900 2001 Note: Median terms-of-trade shocks as a percentage of GDP, . . . and in low-income countries, manufactures make up 1970­2000. Numbers of developing countries in the sample in 80 percent of exports parentheses. 90 world Manufacturing of 80 exports (%) 70 (percent) Resources share nominal and real interest rates. The London 60 exports (%) 50 interbank offered rate (LIBOR, both nominal 2002­ 40 and real) was much lower in the 1990s (figure Agricultural countries 30 exports (%) 3.13), and one result was an energetic search for 20 yield. 10 Conditions in developing and emerging mar- exports,1981 0 ket countries clearly influenced the geographical Low-income 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 1900 2001 destination of capital flows. Growth in many developing countries accelerated in the 1990s; Source: World Bank, Global Economic Prospects 2004. many had completed external debt restructur- SOMETHING SPECIAL ABOUT THE 1990S? 67 FIGURE 3.9 FIGURE 3.11 Decline in Nonenergy Commodity Prices Capital Flows to Developing Countries Expanded in the 1990s 300 6 200 5 100 4 GDP of 3 0 2 1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 Percent Agriculture Metals and minerals 1 Fertilizers Nonenergy commodities 0 Source: World Bank staff calculations. Note: Constant dollar indexes deflated by manufactures unit value 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 (MUV) index with 1990=100, 1960­2003. Official Private Source: World Bank, Global Development Finance 2004. Note: Aggregate net resource flows to developing countries as a percentage of GDP. FIGURE 3.10 Oil Prices Were Lower in the 1990s than in the 1970s and 1980s tries' main creditors were not bondholders but 80 commercial banks. In the 1970s and 1980s, as commercial banks recycled oil surpluses from oil 60 producers to other developing countries' banks, banks accounted for about 90 percent of devel- 40 oping countries' public external debt to private creditors.Developing countries'debt expanded at 20 double-digit annual rates in the 1970s.The debt crises of the 1980s slowed the growth of bank 0 financing,and by the end of the 1980s,the banks' 1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 share had declined to nearly 30 percent. The Nominal oil price, spot current $/bbl Brady plan restored market confidence in inter- Real oil price, 1990 constant US$ national lending to developing countries, and Source: Economist Intelligence Unit. debt flows increased again in the 1990s. Success- ful macro-stabilization in many emerging economies, the opening of their capital markets, ings, successfully stabilized their economies, and technological innovation contributed to the embarked on substantial liberalization of their rapid growth of bond finance (see figure 3.12, financial sectors, and, as discussed in section 3 right side). below on policy reforms, had undertaken signif- In the shift from debt to equity,FDI played an icant privatization. important role, as nonfinancial corporations Underlying the expansion in capital flows to increased their exposure to developing countries. developing countries were two important shifts Developing countries accounted for about 30 in the structure of these flows: first, from bank percent of global FDI flows--a share that lending to bonds and portfolio financing and, remained stable over the decade.While North- second,a shift from debt to equity.Up to the time South FDI flows declined, South-South FDI of the Brady plan in 1989, the developing coun- increased substantially over the 1990s, from less 68 ECONOMIC GROWTH IN THE 1990s FIGURE 3.12 Capital Flows Were Driven by a Surge in FDI and Portfolio Equity Flows 4 0.9 3 0.6 GDP GDP of of 2 0.3 Percent Percent 1 0.0 0 ­0.3 0 3 6 9 2 5 8 1 4 7 0 197 197 197 197 198 198 198 199 199 199 200 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 Source: World Bank, Global Development Finance 2004. Left: Net FDI flows to developing economies as a percentage of GDP. Solid line is the flows to the top 10 recipients (in descending order): China, Brazil, Mex- ico, Argentina, Poland, Chile, Malaysia, Thailand, Czech Republic, and R. B. de Venezuela. Right: Portfolio flows to developing economies as a percentage of GDP. Solid line is the flows to the top 10 recipients (in descending order): Argentina, Mex- ico, Brazil, Turkey, Russian Federation, Philippines, India, China, Colombia, and Malaysia. Such investment choices were usually motivated FIGURE 3.13 by trade costs, such as for transport and tariffs, Nominal and Real Interest Rate, 1980­2002 though in the 1990s they were also related to the (London interbank offer rate) privatization of utilities and the entry of foreign banks (Shatz andVenables 2000). 20 What distinguished the 1990s was the rapid growth in international vertical FDI, that is, 15 year investment by firms that break up the production of final goods geographically into discrete stages, per 10 typically choosing the location for each stage on the basis of factor abundance. Along with tech- Percent 5 nological progress and policy reforms that liber- alized trade and finance, this investment led to 0 the creation of global production networks that 9801 9821 9841 9861 9881 9901 9921 9941 9961 9981 0002 0022 locate each stage of production in the country where it can be accomplished at the least cost. In Nominal LIBOR, US$6/month developing countries as a group, parts and com- Real LIBOR (deflated by the U.S. CPI) ponents exports (a proxy for participation in Source: Bloomberg. global networks) grew faster during the 1990s than in the 1980s,and much faster than in indus- than 10 percent of the total at the beginning of trial countries (figure 3.14).As a result, develop- the decade to nearly 40 percent at the end. ing countries' share of global parts and Empirical studies suggest that FDI flows to components exports increased from a mere 7 developing countries were largely horizontal in percent in the early 1990s to 21 percent in 2000. the past, as when multiplant firms chose to repli- The possibilities for participating in interna- cate roughly the same activities in many coun- tional trade expanded during the 1990s as the tries to serve the local markets in those countries. global division of labor changed and more SOMETHING SPECIAL ABOUT THE 1990S? 69 resources were shifted into labor-intensive activi- ties, in which developing countries have a com- FIGURE 3.14 parative advantage (World Bank 2002b).2 At the Developing Countries' Parts and Components Exports Grew Faster in the 1990s same time, however, global production networks (percent per year) were concentrated in just a few countries: the top five emerging-market exporters of parts and com- 25 ponents (China, Mexico, Korea, Malaysia, and Thailand) accounted for 78 percent of the total. 20 Such concentration on just a few large devel- oping countries was the defining feature of pri- 15 vate capital flows over the 1990s. At the end of 10 the decade, the top 10 recipients of FDI received more than 70 percent of all net inflows, and for 5 the top 10 portfolio equity recipients, the pro- portion was even higher.While it is true that the 0 largest recipients are also the largest emerging High-income countries Developing countries economies, the amounts of financing that they 1981­1990 1990­2000 attracted accounted for a large share of their GDP and exports (World Bank 2002b). Sources: UN Comtrade and World Bank. For the median developing country, by con- trast, portfolio capital flows were effectively zero throughout the 1970­90 period. Further, the prerequisite for borrowing internationally), even spreads at which economies borrowed were high after adjusting for the emergence of former during the decade, reflecting creditors' focus on Soviet Union republics as new sovereign bor- risk, not just returns. Moreover, many countries rowers.The result of such diversity of investors experienced sudden stops--abrupt and extremely and instruments, as well as much better under- disruptive reversals in the flow of capital--and standing of emerging markets as an "investment quite a few were struck by financial crises. class,"was that international capital flows became Should one then conclude that the external a much more resilient and viable form of devel- financial environment for developing countries in opment financing.This coming of age could be the 1990s was largely negative?The answer is no. seen at the end of the 1990s: when debt flows First, the supply of funds available to the dried up, equity flows remained significant-- developing countries was far greater in the 1990s unlike in the 1980s,when the drying-up of bank than in any previous decade.Figure 3.15 suggests lending led to a protracted debt crisis. that capital flows to the median country at least Fifth,though the spreads at which economies did not decline, and those to a few large borrowed were high during the decade, reflect- economies rose sharply. ing creditors' concern with risk, they cannot be Second, significant deregulation and liberal- compared with those of previous decades; as ization, as well as financial innovations, greatly noted above,bond financing emerged only in the expanded the choice of investment vehicles. early 1990s after the completion of Brady Third, the group of investors was growing, to restructuring. include banks, nonfinancial institutions, mutual Sixth, the financial and banking crises that and pension funds, and individual investors. rocked the 1990s were not a new phenomenon Fourth, considerable progress was made over the (see chapter 2 and Kindleberger 1984,2000),and decade in improving the transparency, data shar- their severity was largely a reflection of the ing, and circulation of information between underlying fragility of the economies affected investors and emerging markets. By the end of (Reinhart and Kaminsky 1999), rather than of the 1990s, many more countries were rated (a adversity in the international environment. 70 ECONOMIC GROWTH IN THE 1990s FIGURE 3.15 FIGURE 3.16 Capital Flows to a "Median" Developing Country as a Developing Countries Paid Less Interest Percentage of GDP, 1970­2002 On External Debt in the 1990s 10 4 8 3 GDP of 2 6 1 4 Percentage 0 2 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 1970­2002 (70) 1990­2002 (93) 0 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 Source: World Bank, Global Development Finance 2004. 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 Note: Median interest payments on the external debt as a per- Official (73) Private (73) centage of gross national income; sample sizes in parentheses. Total (81) Total (114) Source: World Bank, Global Development Finance 2003. nal environment, it is difficult to conclude that adverse external conditions explain why devel- Seventh, developing countries' interest pay- oping countries' growth was below expectations ments on external debt were lower in the 1990s in the 1990s. International trade expanded rap- (figure 3.16). Throughout the decade, heavily idly and so did capital flows to developing coun- indebted poor countries continued to receive an tries in the 1990s. The context was one of unprecedented amount of debt relief.The Highly relatively stable terms of trade, reasonably low oil Indebted Poor Countries' debt relief initiative of prices, and lower interest payments on develop- the International Monetary Fund (IMF), the ing country debt. World Bank, and other multilateral and bilateral True, the averages mask vast divergences in creditors had committed US$40 billion for debt experiences and growth outcomes: some coun- relief to 26 countries by the end of the decade. tries indeed suffered from real exchange rate Moreover, large positive net transfers from the appreciation; some were adversely affected by a International Development Association (IDA) decline in their terms of trade;while for others the and bilateral concessional sources offset negative prospects of growth were frustrated by the contin- net transfers for the International Bank for ued import protection in developed countries;and Reconstruction and Development (IBRD),IMF, so on.Yet overall it is still difficult to argue for neg- and private sources--in effect becoming another ative shocks in the external environment that form of debt relief by replacing concessional debt could account for the negative common shocks with nonconcessional lending containing a large postulated in Loayza, Fajnzylber, and Calderón grant element. (2002) and Easterly (2001). Econometrically, lower OECD growth Conclusion: No Unusual Adversities in explains away the negative dummies for the 1990s, as in Easterly (2002).3 Yet the channels through the External Environment which OECD countries influenced developing From this brief review of the features that are countries' performance are unclear. Easterly him- most commonly thought to represent the exter- self indicates that SOMETHING SPECIAL ABOUT THE 1990S? 71 The OECD slowdown may have caused the ket premiums disappeared, and real exchange LDC slowdown . . .However,I am not able rates became more competitive. to demonstrate a clear mechanism by which these external shocks translated into lower Trade Liberalization growth for the developing world.A variable that interacts OECD growth with the share On balance,developing countries made convinc- of OECD trade in the economy is insignif- ing progress in opening their economies to inter- icant,for example [emphasis added]. national competition. Since the mid-1980s, they have nearly halved their average tariff rates from It is always possible to find some adverse exter- 28 percent (1980­85) to 15 percent nal shock affecting some countries.But economet- (1995­2000). (See figure 3.17.) ric testing is always a joint hypothesis test of the Further, all of the developing regions have at economic theory and the empirical model. least halved the incidence of nontariff barriers Unable to determine precisely what aspects of the (table 3.7), and in most countries tariff-rate dis- external environment produced such an adverse persion has declined significantly. shock, or what are the exact channels through Progress has varied by regions, countries, and which slower OECD growth affected developing policy instruments. Countries in South Asia, Latin countries' performance, Easterly (2001) suggests a America, and East Asia achieved the most impres- possibility that the type of growth regressions used sive gains in opening their economies, but there in the empirical analysis might be misspecified, was little progress in the Middle East and North given that stationary economic growth is regressed Africa and only moderate successes inAfrica.Over on nonstationary, upward-trending, indicators of the span of a few years, Latin American countries policy performance: became more open to international trade and cap- Alternatively, I have shown that the sig- ital than East Asian countries did over decades. nificance of the 1980s and 1990s decade South Asia remained the most protectionist region dummies in regressions omitting OECD with the highest tariff rates, even after a decade of growth reflects in part mis-specification reforms involving the deepest tariff cuts and a sharp rather than shocks [emphasis added]. reduction in tariff dispersion (table 3.8).Africa's moderate reduction of tariffs masks drastic reforms In short,it is difficult to conclude that adverse in a number of individual countries. Kenya external conditions explain why growth was reduced its import tariffs from 41 percent in the below expectations in developing countries over late 1980s to 14 percent in 1999. Guinea did the the 1990s, and Easterly's suggestion that the same, from 76.4 percent to 10.8 percent. In some dummies result from econometric misspecifica- Middle Eastern and North African countries, the tion needs to be taken seriously. signing of trade agreements with the European Union in the late 1990s eventually started a grad- 3. Policy Reforms in the 1990s ual process of opening up (World Bank 2003j). Developing countries made significant policy Financial Sector Liberalization changes during the 1990s. The reform agenda of the 1990s ranged from The liberalization of finance has been at least as financial deregulation and privatization to impressive as that of trade (figure 3.18).The main upgrading labor codes and mounting anticor- goal of financial reform was to grant greater ruption campaigns, but the most impressive operating freedom to market intermediaries and results during the decade came from opening up, at the same time to strengthen prudential regula- or further opening up, economies to interna- tion and oversight.As a result, at the end of the tional trade and capital flows.Tariffs were cut,the decade only South Asia's financial systems coverage of nontariff barriers shrank, black mar- remained "partly repressed" according to the 72 ECONOMIC GROWTH IN THE 1990s FIGURE 3.17 Reduction in Tariffs in Developing Countries, 1980­2000 80 70 60 50 40 30 20 10 0 Sub-Saharan South Asia Middle East and Latin America Europe and East Asia Africa North Africa and the Caribbean Central Asia and Pacific 1980­1985 1986­1990 1991­1995 1995­2000 Sources: World Trade Organization, World Bank, UNCTAD. Note: Median average tariff in percent, based on unweighted averages for all goods in ad valorem rates, or applied rates, or most-favored-nation rates, whichever data are available for a longer period. Abiad-Mody measure, and economies in Latin requirements. Less rapid successes were achieved America and Sub-Saharan Africa had moved to in privatization and the liberalization of entry "partly liberalized" status.4 The Abiad-Mody barriers such as licensing requirements and limits measure is complete only up to 1996, but signif- on the participation of foreign banks. icant reforms had already been instituted by then. Although there are exceptions within each Although financial reforms were largely region,countries within regions tended to liberal- implemented in packages, they tended to focus ize their financial sectors at roughly the same time most closely on steps to eliminate interest con- and in roughly the same way. Latin American trols and credit controls, such as directed credit countries carried out drastic and rapid reforms in schemes, sectoral credit ceilings, and high reserve the late 1980s and early 1990s. By contrast, East Asian countries implemented financial liberaliza- TABLE 3.7 tion gradually,starting in the early 1980s,opening Reduction in Nontariff Barriers in Developing up in small policy steps with the whole process Countries, 1990s stretching over the 1990s. South Asian countries reformed only in the early to mid-1990s. South Country 1989­94 1995­98 Africa in 1980 and the Arab Republic of Egypt in Sub-Saharan Africa (12) 26.0 10.4 1987 followed a "big bang" approach to financial South Asia (4) 57.0 58.3 liberalization, while others including Ghana, Middle East and North Africa (4) 43.8 16.6 Morocco, and Zimbabwe followed a rather grad- Latin America and the Caribbean (13) 18.3 8.0 ual approach (Abiad and Mody 2002). East Asia and Pacific (7) 30.1 16.3 Source: Michalopoulos 1999. Liberalization of the International Note: Average number of commodities subject to nontariff measures as a per- Financial System centage of total. Figures in parentheses are the number of countries in each region for which data are available. In the case of South Asia, significant Equally significant were the measures taken to lift reductions have taken place since. restrictions on the international movement of cap- SOMETHING SPECIAL ABOUT THE 1990S? 73 TABLE 3.8 Tariff Dispersion Decline in the 1990s FIGURE 3.18 Financial Sector Liberalization, 1973­96 Region 1990­1994 1995­1998 1999­2002 20 South Asia Bangladesh 114.0 14.6 13.6 15 India 39.4 12.7 12.4 Sri Lanka 18.1 15.4 9.3 index Africa 10 South Africa 11.3 7.2 11.7 Malawi 15.5 11.6 10.5 Zimbabwe 6.4 17.8 18.6 Abiad-Mody 5 East Asia Philippines 28.2 10.2 7.3 Thailand 25.0 8.9 14.3 0 Indonesia 16.1 16.6 10.8 73 75 77 79 81 83 85 87 89 91 93 95 19 19 19 19 19 19 19 19 19 19 19 19 China 29.9 13.0 10 Latin America High-income Latin America and South Asia (5) countries (9) the Caribbean (7) Argentina 5.0 6.9 7.2 East Asia (8) Middle East and Brazil 17.3 7.3 12.9 Africa* (7) Colombia 8.3 6.2 6.2 Source: Abiad and Mody 2003. The database covers financial sector policy changes in Mexico 4.4 13.5 9.3 35 countries over the 24 years from 1973 to 1996. Source: World Bank, World Development Indicators 1998, Note: The Middle East and Africa region also includes Turkey. Figures in parentheses are 2000, 2003; WTO Trade Policy Reviews, various issues. the number of countries in each region for which the index has been calculated. See note 6. Note: Country observations are for one year within the time periods noted above. tion. Countries of Central and Eastern Europe ital. The numbers of countries using multiple and the former Soviet Union designed new tax exchange rates and requiring compulsory surren- codes. Liberalization of trade altered tax struc- der of export receipts declined, and several coun- tures by sharply reducing the share of trade taxes tries moved slowly to liberalize their current and in total tax revenues. capital account transactions (table 3.9 and figure The fiscal reforms were implemented in a 3.19).However,financial crises in the late 1990s,in much more stable macroeconomic environment East Asia in particular, forced policy makers to than that of the 1980s. Fiscal balances improved reevaluate the conventional wisdom that opening in most regions and inflation declined. Real the capital account as soon as possible is the right exchange rates that had been overvalued in the policy to follow (see chapter 3). 1960s and 1970s were devalued. Tax Reforms 4. Conclusions Although not as extensive as trade and financial liberalization, tax reforms were a significant area The economic environment of the 1990s has of reform during the 1990s (Lora 2001a; IDB often been seen as unstable,volatile,and unforgiv- 1997).To increase revenue and to reduce the effi- ing for economic growth. In reality, however, it ciency cost of taxation, many developing coun- was quite favorable for developing countries.Dur- tries lowered their marginal tax rates, simplified ing the decade,practically all developing countries and rationalized their tax systems, introduced embarked on ambitious market-oriented reforms. value added taxes, and strengthened tax collec- Since the mid-1980s most of these countries have 74 ECONOMIC GROWTH IN THE 1990s TABLE 3.9 Capital Account Restrictions Were Progressively Dismantled, 1970­97 1970 1980 1990 1997 Multiple exchange rate practices Sub-Saharan Africa (46) 20 19 18 14 South Asia (7) 50 16 14 0 Middle East and North Africa (16) 33 33 18 23 Latin America and the Caribbean (31) 24 39 42 19 East Asia and Pacific (16) 11 2 13 3 High-income countries (21) 17 10 0 0 Current account restrictions Sub-Saharan Africa (46) 88 73 73 82 South Asia (7) 100 79 86 86 Middle East and North Africa (16) 70 43 52 47 Latin America and the Caribbean (31) 48 44 58 48 East Asia and Pacific (16) 45 28 31 62 High-income countries (21) 38 26 20 10 Capital account restrictions Sub-Saharan Africa (46) 97 95 97 85 South Asia (7) 100 84 86 86 Middle East and North Africa (16) 70 50 52 47 Latin America and the Caribbean (31) 66 71 84 53 East Asia and Pacific (16) 79 64 63 68 High-income countries (21) 85 70 52 0 Surrender of export proceeds Sub-Saharan Africa (46) 97 95 96 77 South Asia (7) 100 100 100 57 Middle East and North Africa (16) 68 50 52 38 Latin America and the Caribbean (31) 79 85 97 50 East Asia and Pacific (16) 93 82 63 56 High-income countries (21) 68 55 40 0 Source: International Monetary Fund. Note: The data show percentage of the countries imposing restrictions according to IMF methodology. Years are three-year averages, except for 1997, which is the 1996­97 average. Figures in parentheses are the number of countries in each region for which the index has been calculated. succeeded in reducing tariffs, liberalizing their rates, high in the 1980s, came down in the 1990s, financial sectors, privatizing their public enter- and so did oil prices.Large numbers of poor coun- prises, and reducing their deficits. Driven by tries received unprecedented amounts of debt exports of manufactures, world trade grew much relief.By most common indicators of macroinsta- faster in the 1990s than in any previous decade. bility, the 1990s were less volatile than previous Aggregate financial flows recovered rapidly in the decades. On the macroeconomic front, inflation 1990s after reaching a nadir in the late 1980s, and declined,real exchange rates significantly depreci- an average developing country experienced no ated, and black market premiums disappeared. significant decline in capital inflows. For the From any perspective, the positive changes wit- median country, capital flows regained their aver- nessed during the decade were quite remarkable age level of the 1970s in 1997, with portfolio and and created a reasonable expectation of higher FDI flows growing particularly fast. Real interest growth. SOMETHING SPECIAL ABOUT THE 1990S? 75 Notes FIGURE 3.19 Effects of Liberalizing the Financial 1. Notwithstanding the rising U.S. growth rate in the Sector, Developing Countries 1990s, high-income countries as a group grew more slowly than in previous decades first and foremost 60 because of the slowdown in Japan, which accounts 50 for 20 percent of industrialized countries'GDP.Japan grew at just 1.4 percent in the 1990s,far below its his- OFE 40 toric average of 6.3 percent over the previous two 30 decades.The United States (32 percent of industrial- above ized countries'GDP) grew only slightly faster during 20 the roaring 1990s than in the 1980s: 3.3. versus 3.2 10 percent annually. Percent 2. These changes were not just due to declines in the 0 prices of agricultural and resource commodities rel- ­10 ative to manufactures--the strong shift in the com- 60 63 66 69 72 75 78 81 84 87 90 93 96 99 position of exports shows up even when price 19 19 19 19 19 19 19 19 19 19 19 19 19 19 changes are removed. Further, it was not just due to Median Geometric average a few large, high-growth exporters such as China and India. Excluding China and India, the share of Sources: Levine and Renelt 1995; World Currency Yearbook; Wood manufactures in developing-country exports grew 1988; World Bank, Global Development Finance and World Develop- ment Indicators. from one-tenth in 1980 to almost two-thirds in 2001. It increased sharply, but not equally, in all Note: Median black market premium and its geometric average, regions.The laggards included Sub-Saharan Africa 1960­2000; 103 developing countries. and the Middle East and North Africa, which have OFE official exchange rate. yet to reach 30 percent. Many countries, particularly the poorest,remain dependent on exports of agricul- tural and resource commodities. 3. For instance,only 4 percent of U.S.affiliates'produc- tion in the European Union is sold back to the United States, whereas for developing countries the figure is 18 percent and for Mexico it is more than 40 percent (Shatz andVenables 2000). 4. It should be kept in mind that there are questions as to whether participation in global production-sharing actually leads to higher productivity, to faster growth in value added or employment, or to any other posi- tive spillovers. 5. The slowdown in developed countries' growth over the 1990s was a mixed experience--neither univer- sal nor particularly sharp.Three-fourths of the slow- down was a result of the prolonged recession in Japan; the United States grew no more slowly, and the European Union only moderately more slowly. 6. The Abiad-Mody index of financial liberalization is an aggregate of six components of financial sector policy: credit controls, interest rate controls, entry barriers,regulations and securities markets,privatiza- tion in the financial sector, and international finan- cial transactions. Country Note B Lessons from CountriesThat Have SustainedTheir Growth A t the beginning of the 1990s, it chapter 1,at different points in time one function was broadly agreed that coun- is more binding on growth than another, and tries needed "to get their poli- each can be fulfilled in different ways.The suc- cies right"to achieve growth and overcome what cessful countries provide illustrations of "func- for many, particularly in Latin America and tional equivalents" (Rodrik 2002b),showing that Africa, had been the "lost decade" of the 1980s. function does not define form,that there are sev- Getting "policies right" had a well-defined eral ways of fulfilling the same function, and that meaning. On the macro front, it meant reducing income convergence does not imply conver- fiscal deficits, moving away from foreign gence of policies and institutions.For economists exchange rationing and multiple exchange rate and the policy makers they advise, this is perhaps systems, lowering inflation, freeing interest rates, the central realization of the 1990s.The implica- and increasing the independence of monetary tion is that there are no best-practice policies that policies. On the structural front, it meant reduc- will always yield the same positive result--there ing the scope for state intervention and discre- is no unique way to succeed. Sustained growth tion through privatization and rationalization of depends less on whether policies conform to government agencies, freeing external trade and some ideal than on whether they identify bind- replacing restrictive trade regimes by more uni- ing constraints accurately and address them effec- form and lower tariffs, and liberalizing the finan- tively. Successful growth strategies address cial sector.As discussed in several chapters of this specific, binding constraints on--for example-- report, the extremely varied results that emerged faster accumulation of capital or higher produc- from this experience--some exceeding the most tivity growth by experimenting and by adjusting optimistic forecasts and others below expecta- policies and institutional arrangements to tions--made it evident that the issues were more changes in economic, institutional, and political complex than was thought at the beginning of conditions.Similar conclusions were first reached the decade. in a 1993 study, The East Asia Miracle (World While chapters 4­7 review experience with Bank 1993) and this note suggests that they can the implementation of specific policies across be generalized to a wider set of countries.1 countries, this note focuses on the experience of individual countries that have sustained growth Defining Successful Growth Experiences during the 1990s. Defining a successful growth experience is itself not straightforward.The first The frequency of growth episodes stands in sharp section of this note identifies a list of countries contrast to the few cases in which growth has judged "successful" from a growth perspective in been sustained over time (chapter 2). In the last the 1990s on the basis of arbitrary yet reasonable 50 years,most countries have experienced at least criteria. one, and often more than one, period of several The second section discusses what can be years of growth.But few countries have sustained learned from these country experiences. Growth growth over decades (Easterly et al. 1993; Haus- requires four functions to be fulfilled.As noted in mann, Rodrik, and Pritchett 2004). Sustaining 78 LESSONS FROM COUNTRIES THAT HAVE SUSTAINED THEIR GROWTH 79 growth for long periods is what enables develop- The two criteria yield a list of 18 countries ing economies to reach the income levels of that account for about 60 percent of the world's industrialized economies, as have Hong Kong population and are extremely diverse economi- (China), the Republic of Korea, and Singapore. cally,politically,and historically.The results throw Therefore, in selecting the successful growth up a number of surprises, including countries experiences of the 1990s, care must be taken not such as Egypt or Nepal, known for highly dis- to include episodes of growth that are not part of torted policy environments and governance a long-term trend, and not to exclude relatively weaknesses, which have nonetheless succeeded modest growth rates sustained over the long run. from a growth--and also social development-- There is no foolproof method.A "successful" perspective. The group includes resource-rich growth experience in the 1990s is defined here (Botswana, Indonesia) and resource-poor as one meeting two criteria: catching up with economies (Bangladesh,Vietnam); well-estab- advanced economies over the 1990s,and sustain- lished democracies (Botswana, India); recently ing this growth over time. The first criterion democratized or democratizing countries meant selecting countries with a rate of per (Korea, Bangladesh) as well as one-party states capita income growth over the 1990s sufficient (China, Vietnam, Egypt); and landlocked to narrow the per capita income gap with the (Botswana, Lao PDR) and island economies United States: that is, per capita income growth (Mauritius, Sri Lanka) as well as continental of at least 1.7 percent a year during the 1990s. economies (China,India).Some of the successful Out of 117 developing countries with popula- countries have had a relatively recent colonial tions of more than half a million, 42 countries history (Mauritius,Indonesia),others a quite dis- grew faster than the United States in the 1990s tant one (Chile), and some have never been col- (table B.1). onized (Bhutan, Nepal, Thailand). In some, Many of these countries, however, were corruption is pervasive (Indonesia, Bangladesh), recovering in the 1990s from output collapses in and in others it is no longer a significant issue the 1980s stemming from external shocks,macro- (Korea, Malaysia). The public sector in Chile, economic crises, civil conflicts, or other adverse notwithstanding its ownership of the copper sec- circumstances.Thus,to avoid the inclusion of pos- tor,plays a much smaller role than its counterpart sibly transitory recoveries and to narrow the def- in China or India,and concentrates on formulat- inition of successes to countries more likely to be ing appropriate policy and regulatory frame- on a sustained growth path, the second criterion works and delivering essential social and meant selecting countries with per capita income infrastructure services. Chile has privatized growth of at least 1 percent a year during the extensively,not only enterprises but also its social 1980s.This eliminates 24 countries such as El Sal- security system, whereas Egypt has relied on vador, the Islamic Republic of Iran, Lebanon, public investments and state-owned enterprises, Mozambique, Peru, and Sudan (table B.1).The 1 which, as is the case in China and Vietnam, percent threshold chosen for the 1980s is obvi- account for a large share of the economy.The ously arbitrary.Applying the same criterion over banking system has been freer and sounder in two decades, that is, choosing countries that nar- Malaysia and Mauritius than in Bangladesh, rowed the per capita income gap with the United China, or India. Sri Lanka and Chile started to States both in the 1980s and in the 1990s, would open their economies in the 1970s, whereas eliminate countries such as Bangladesh and China and India started significantly reducing Tunisia, which have given all indications of being trade barriers only in the 1990s.Macroeconomic on a sustained growth path. And applying the prudence has taken on different meanings in dif- same criterion over the past four decades would ferent countries. In some countries, it has meant be too restrictive, since it would limit the list to keeping fiscal deficits low. In India it has meant a only six economies:Botswana,theArab Republic structure of public debt with long maturities of Egypt, Korea, Lesotho, Malaysia, andThailand. mostly denominated in local currency.India's fis- TABLE B.1 Growth Successes in the 1990s 1990 1980 1980­2002 GDP per capita in 1980 Population (millions) China 8.6 7.7 8.2 167 1,262 Vietnam* 5.7 1.9* 4.6* 185* 78 Korea, Rep. of** 5.0 7.4 6.1 4,098 47 Lebanon 4.4 n.a. n.a. n.a. 4 Chile 4.3 2.1 3.3 2,665 15 Mozambique 4.3 ­1.4 1.6 160 18 Mauritius 4.1 4.9 4.4 1,745 1 Sudan 3.9 ­0.1 2.1 227 31 Malaysia 3.7 3.1 3.4 2,297 23 Dominican Rep. 3.7 0.4 2.2 1327 8 Lao PDR* 3.6 1.4* 2.9* 284* 5 India 3.6 3.6 3.6 228 1,016 Thailand 3.4 6.0 4.6 1,116 61 Bhutan 3.4 5.4 4.3 232 1 Uganda* 3.2 0.7* 2.2* 236 22 Sri Lanka 3.1 3.1 3.1 455 18 Poland 3.1 -- -- -- 39 Bangladesh 3.0 1.1 2.1 249 131 Tunisia 2.9 1.1 2.1 1,641 10 Iran, Islamic Rep. of 2.7 ­0.7 1.2 1,380 64 Botswana 2.7 7.2 4.7 1,538 2 Guyana 2.7 ­3.0 0.1 927 1 Indonesia 2.6 4.4 3.5 503 206 Cambodia 2.6 -- -- -- 12 Panama 2.6 ­0.7 1.1 3,042 3 Trinidad and Tobago 2.4 ­1.2 0.8 4,612 1 Costa Rica 2.4 ­0.5 1.1 3,097 4 Burkina Faso 2.3 0.8 1.6 181 11 Greece 2.2 0.3 1.2 10,702 11 Egypt, Arab Rep. of 2.1 2.9 2.5 731 64 El Salvador 2.1 ­1.5 0.5 1,595 6 Nepal 2.1 2.4 2.2 148 23 Albania 2.0 ­0.8 0.7 910 3 Lesotho 2.0 2.4 2.2 360 2 Peru 2.0 ­3.0 ­0.3 2,569 26 Benin 1.9 ­0.5 0.8 362 6 Namibia 1.9 ­2.4 ­0.1 2,469 2 Ghana 1.9 ­1.3 0.4 394 19 Syrian Arab Rep. 1.9 ­1.1 0.5 719 16 Mali 1.8 ­1.9 0.1 305 11 Fiji 1.8 0.2 1.1 2,311 1 Ethiopia* 1.8 ­1.7* 0.3* 117* 64 Source: World Bank, World Development Indicators (WDI) 2003. Note: Successful countries (shaded in blue) that grew by more than 1.7 percent a year, which was the growth rate of GDP per capita in the 1990­2002 period. * Indicates that the GDP per capita data series starts later than 1980 (Vietnam in 1984, Lao PDR in 1984, Uganda in 1982, and Ethiopia in 1981). Population in millions as of 2000. GDP per capita is in 1995 U.S. dollars. ** The Republic of Korea has graduated into the ranks of developed countries, but is included in this table because during the 1980s it was still considered part of the developing world. --. Not available. 80 LESSONS FROM COUNTRIES THAT HAVE SUSTAINED THEIR GROWTH 81 cal deficits (10 percent of GDP) have risen while external indebtedness has declined. FIGURE B.1 What is common among these countries has Investment as a Share of GDP been their persistent ability to grow over time.In 40 low-income countries, positive shocks often become growth episodes. Examples of such 30 shocks include the adoption of new agricultural technologies, investments in infrastructure, cent 20 increases in commodity prices, or new industrial erP investments.The challenge of development is to 10 transform growth episodes into sustained growth.Albeit with different degrees of success, 0 the 18 countries have been able to meet this 1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 challenge. Year In seeking to learn from the experience of Success cases (16) the successful countries, three issues need to be Other countries (42) kept in mind. First, one or two decades is not long enough to lift a low-income developing Source: World Bank, World Development Indicators 2004. country to the income levels of industrialized Note: Only 16 success cases (out of 18) are listed owing to incom- economies. China, at its current exceptionally plete data. high levels of growth,would still need 35 years to catch up with Korea's current per capita income. (figure B.1). Why have governments in these Second, growth over one or two decades does countries been able to invest more than the typ- not ensure that it will be sustained in the future. ical country? Why have entrepreneurs in these Brazil experienced growth for almost a century, countries been more willing to take on risks and followed since the early 1980s by persistent stag- invest, often in new activities? Typical explana- nation of its per capita income.Argentina had a tions range from the existence of an entrepre- similar experience of long periods of growth fol- neurial class and institutions ensuring that lowed by prolonged stagnation.Third, there are investors can appropriate their returns,to policies large differences in performance among the 18 that reduce uncertainty on returns and encour- countries: China's per capita income grew at 8 age risk taking through privileged access to percent a year over two decades, whereas Nepal's credit, imports, or other inputs. grew at just over 2 percent. Four features seem to differentiate these 18 countries from less successful developing coun- tries. Diversity of Experiences: East Asia First, growth overrode other social and economic Lessons Relearned and Generalized objectives. Growth was a central objective of poli- Key functions to be fulfilled in sustained growth cies not only rhetorically but in practice.While processes are the accumulation of capital, alloca- each country followed its own specific growth tive efficiency, technological progress, and the strategy,a common element was a focused deter- sharing of the benefits of growth.The discussion mination to adjust policies and institutions prag- below illustrates the different ways in which matically whenever growth started to falter. Key these functions have been fulfilled. in this process were the use of unambiguous indi- cators of performance (such as exports), institu- Accumulation of Capital tions with some degree of accountability,and exit The 18 countries have accumulated capital faster strategies when results were below expectations. than other economies.At times the difference has Downturns or adverse shocks have been taken as been as large as 10 percentage points of GDP opportunities for decisive reforms that strength- 82 ECONOMIC GROWTH IN THE 1990s ened economic foundations, rather than as pened in Sri Lanka--was a New Economic Pol- excuses for inaction. icy aimed at sharing wealth equitably for all The oil shock of the 1970s, for example, was Malaysians through growth. In China, when an opportunity for Korea to open its economy growth faltered in the late 1990s, the government and expand exports (a sustainable longer-term expanded public investment and rationalized the strategy) whereas Brazil responded by raising tar- export regime. More recently, the Asian financial iffs and introducing a second phase of import- crisis provided the opportunity to reform banking substitution policies (an approach that had and corporate governance in Indonesia, Korea, negative growth consequences in the long run) and Malaysia. In these three countries, the recent (chapter 5). Both India and Sri Lanka faced bal- economic reforms have been accompanied by ance of payments crises in the later 1980s and political change fostering democracy and early 1990s; India responded with stabilization accountability at the highest levels of government. measures and a comprehensive reform strategy Second,and partly because growth was such a central that liberalized and opened its economy, and Sri objective, the 18 countries show remarkably narrow fluc- Lanka adopted a more flexible exchange rate tuations in their growth rates over time. A record of policy,renewed its emphasis on privatization,and steady growth is of central importance because it introduced further trade reforms. Tunisia's reduces the uncertainty associated with invest- response to debt problems in the 1980s was to ment decisions.As noted above,developing coun- shift to a more competitive exchange rate,a grad- tries seldom sustain their growth: low average ual opening of the trade regime, and limitations growth typically results from volatile growth rates, on external borrowing--all of which remained rather than absence of episodes of rapid growth. key features of its policies throughout the 1990s. What distinguishes countries such as Botswana, Chile established the credibility of its socially ori- Chile,China,India,Indonesia,Korea,and Malaysia ented new administration during the democratic is less that they achieved high levels of growth in transition in the early 1990s by strengthening fis- some years than the fact that they have systemati- cal policies while expanding social spending. cally avoided episodes of slow growth. By and Korea's response to the 1998 financial crisis was large, developing countries experience a year of to relax restrictions on foreign direct investment negative per capita growth roughly once every and improve corporate and banking governance. three years,whereas in EastAsia,the average is half Botswana responded with effective stabilization that rate and in Organisation for Economic Co- policies to each of the terms-of-trade shocks it operation and Development (OECD) countries suffered, which were generally severe because of one-third that rate (table B.2).Korea has had only the limited diversification of the economy. three years of negative per capita growth since Similarly in Indonesia, each of the external 1961.Ability to avoid downturns and periods of shocks the country faced provided a stimulus to low growth is what explains East Asia's "miracle" strengthen the policy regime.With the fading of growth relative to other developing countries as the second oil boom in the early 1980s,Indonesia well as the 18 countries' above-average perform- introduced two devaluations and microeconomic ance. reforms that diversified exports and strengthened Third,the ability to reduce the volatility of growth is productivity growth. Even though these reforms the result not only of decisive responses to shocks, but did not resolve deep-seated institutional problems also of macroeconomic policies that reduced vulnerabili- related to corruption and cronyism at the top,and ties and hence the costs of shocks. The 18 countries to a weak judiciary,they were sufficient to put the have had less recourse to external debt than other country on a growth path that was sustained for developing countries (figure B.2).While access to nearly two decades. In multiethnic Malaysia, the external capital helps to increase the pool of sav- response to the racial riots of 1969--which could ings so that an economy can grow faster, it also have destabilized the country for decades and can be misused and weaken macroeconomic dis- reduced growth far below its potential, as hap- cipline (see Country Note F on financial crises). LESSONS FROM COUNTRIES THAT HAVE SUSTAINED THEIR GROWTH 83 TABLE B.2 Economic Successes: Steady Growth, 1960­2002 Years in which growth rate was Negative Below 1% Below 2% Above 2% All developing countries 14 19 24 18 Sub-Saharan Africa (28) 18 22 27 15 Botswana 2 3 4 38 Lesotho 10 15 16 26 South Asia (5) 8 11 17 25 Bangladesh 11 15 21 21 India 8 10 14 28 Nepal 10 18 22 20 Sri Lanka 4 6 14 28 Middle East and North Africa (6) 15 18 22 21 Egypt, Arab Rep. of 4 10 15 27 Latin American and the Caribbean (21) 12 19 25 17 Chile 7 11 18 24 East Asia and Pacific (7) 7 8 10 32 China 5 6 7 35 Indonesia 7 8 10 32 Malaysia 5 5 7 35 Thailand 2 2 6 36 High-income OECD (22) 5 8 16 27 Korea, Rep. of 3 3 4 38 Source: WDI 2003. Note: The table shows evidence for the 89 countries for which growth data are available for the four decades since 1961. Regional aggregates are medians. The table is calculated for countries for which complete 1960­2002 GDP per capita series are available. Thus it excludes Bhutan, Lao PDR, Nepal, Tunisia, and Vietnam. The Republic of Korea "graduated" into a high-income category in the early 1990s, and thus is classified here in the high-income OECD group rather than in East Asia and Pacific. The 18 countries have kept inflation low and sta- industrial policies indicates that these policies ble (table B.3).Above all,their exchange rates have played an important role in their growth strate- been much less volatile than those of other devel- gies, regardless of the ability to measure such a oping countries (table B.4). role (Hoff and Stiglitz 2001). Fourth, the role of activist industrial policies is still controversial but is likely to have been important.It has Efficiency in Resource Allocation been well documented that governments of East The 18 successful economies have had, and in Asian countries took an activist role in the some cases continue to have, various degrees of process of industrialization and that this sup- distortions that weaken efficient resource alloca- ported constructive risk taking by both the pub- tion and cause significant economic waste. In lic and the private sector. Whether East Asia's Bangladesh,for example,the poor governance of exceptional growth has taken place because of or banks was until recently the source of impaired in spite of these industrial policies is controver- financial intermediation. In India until the sec- sial. Some studies suggest that had policies and ond half of the 1990s, trade restrictions were the institutions converged to "best practice," growth source of significant economic efficiency losses. would have been faster, while others conclude Yet all 18 economies have gradually and that the fact that all the East Asian "miracles," persistently improved their policy regimes on a except Hong Kong (China), adopted activist wide range of fronts: macroeconomic manage- 84 ECONOMIC GROWTH IN THE 1990s TABLE B.4 FIGURE B.2 Real Exchange Rate Volatility External Debt as a Share of Gross National (median of the variance of the real effective exchange Income rate, per year) 100 Success countries (15), 1980­2002 44.4 (17), 1988­2002 47.5 80 Other developing countries (66), 1980­2002 112.4 (81), 1992­2002 117.5 60 cent Source: IMF 2004a. erP 40 Note: Figures in parentheses denote the number of countries in the sample. 20 important to note that efficiency in allocation is 0 induced not only by sound policies--such as 1976 1979 1982 1985 1988 1991 1994 1997 2000 competitive exchange rates and an open trade Year regime--but also by institutions, for example, Success cases (18) that enforce contracts and enable markets to Other developing countries (61) function (World Bank, World Development Report Source: World Bank, World Development Indicators 2004. 2001, 2005a). In some instances, institutions and political realities help to define the set of feasible policies, as policy makers such as Russia's former TABLE B.3 Minister of Finance Yegor Gaidar are quick to Inflation Volatility acknowledge (World Bank 2005b). (median of the variance of the inflation rate, consumer prices, per year) Technological Catch-Up Success countries (7), 1961­2002 21.1 Innovation and technological progress play an (10), 1970­2002 19.3 important role in economic growth and available (13), 1980­2002 17.5 estimates suggest that productivity growth,to vary- (17), 1990­2002 17.5 ing degrees, has been an important factor in the Other developing countries (35), 1961­2002 103.0 growth experience of the 18 successful countries. (53). 1970­2002 89.3 Productivity growth is a common characteristic of (61), 1981­2002 103.0 all sustained growth processes.Almost half a cen- (74), 1991­2002 107.3 tury ago,pioneering studies by Abramovitz (1956) Source: World Bank, World Development Indicators 2004; own and Solow (1956,1957) found that increases in fac- calculations. tors of production accounted for less than a third Note: Figures in parentheses denote the number of countries of U.S.economic growth,and that the bulk of the in the sample. growth came from working smarter.While factor accumulation is essential,it is no guarantee of suc- ment,external trade,public sector enterprise and cess by itself.In most of the Middle East and North utility regulation, and finance. Progress on the Africa, for example, as discussed in Country Note policy front is reviewed in chapters 4­7. Of note D, capital accumulation has been insufficient to is the increase in the role of trade in the 18 generate rapid growth. economies, which points to an increase in their Total factor productivity (TFP) measures the efficiency of resource allocation (figure B.3).The use of better technology and improvements in difference between the 18 countries and other the quality of labor and capital. Several studies developing countries has been widening over find that TFP explains between half and three- time, reaching almost 20 percentage points of quarters of economic growth, and that differ- GDP toward the end of the 1990s. It is also ences inTFP account for most of the differences LESSONS FROM COUNTRIES THAT HAVE SUSTAINED THEIR GROWTH 85 successful countries than in the median develop- FIGURE B.3 ing country (table B.5). Integration with the World Economy, Productivity gains result from complex forces 1970­2000 in society. Organizational improvements that (median of merchandise goods and services trade as a enable quicker turnaround times for ships in ports percentage of GDP) may raise productivity as surely as does more nar- 100 rowly defined technological change. Institutional 80 innovations--such as the creation in the 18th century of public limited liability companies, 60 which enabled greater risk taking by firms; the cent letter of credit a few centuries earlier; and the erP 40 introduction of patent protection for innova- tion--are just as important sources of productiv- 20 ity gains as are breakthroughs in science or 0 technology. For example, while their contribu- tion is difficult to measure, legal reforms in India 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 over the last decade have facilitated the trans- Year parency and security of transactions in capital Success cases (14) markets, and probably account for a part of the Other developing countries (49) gains in productivity of the last decade.Competi- Source: World Bank, World Development Indicators 2004. tion is also important, providing both the incen- Note: For lack of data, not all the successful countries are included tive and the means to acquire better technologies. in this figure. However, expanding the sample (by effectively Openness is crucial; it does not imply uniform ignoring earlier decades for which no data were available--in the trade tariffs or the absence of protection but it case, for example, of Mauritius and Vietnam) does not alter the results. does require access to key inputs at international prices (Harberger 2004,23).Certain types of pro- in output growth rates among countries. For duction technologies are embedded in imported example,looking at the growth rates of 74 coun- machinery (or seeds and other inputs). Foreign tries over three decades, the 1998/99 World investment,one manifestation of economic open- Development Report: Knowledge for Development ness,has been important in many of the 18 coun- attributes three-quarters of the differences in tries and has often been a major channel for growth rates to differences inTFP. modernizing technology. Foreign direct invest- Data on productivity are not always reliable ment in Malaysia averaged more than 5 percent and important variables tend to be mismeasured: of GDP for several years, and the percentage is capital stock aggregates investments of different even higher in the most rapidly growing vintages and hence quality; depreciation adjust- provinces in China.This is not an exhaustive list ments ignore obsolescence; human capital is of the forces behind productivity; others include measured through education inputs, not value, government and firms spending on research and and so forth. And although the old Cambridge development (R&D), tertiary education, and capital controversy questioning the meaning of additions to human capital or better infrastruc- capital in an aggregate production function has ture, all of which can improve the functioning of been largely forgotten, a recent retrospective the economy. concludes that it has not been resolved (Cohen The challenge of technological catch-up is and Harcourt 2003). In China, TFP has about expanding a country's production possibility accounted for about half of the GDP growth rate frontier.This challenge is country-specific because in the last two decades, while in the Philippines the frontier is defined by technology that includes TFP is negative (Yusuf and Evenett 2002).2 Pro- the organizational and institutional settings in ductivity has generally grown faster in the 18 which people and firms operate. Less distorting 86 ECONOMIC GROWTH IN THE 1990s TABLE B.5 Total Factor Productivity Growth of Successful Countries, 1960­2000 TFP growth 1960s 1970s 1980s 1990s China 0.5 0.7 4.2 5.1 Vietnam* -- -- -- -- Korea, Rep. of** 2.4 ­0.7 2.4 0.9 Chile 0.9 0.1 0.7 2.1 Mauritius 0.0 1.5 3.0 2.3 Malaysia 1.0 1.1 0.3 0.9 Lao PDR* -- -- -- -- India 0.7 ­0.3 2.5 1.3 Thailand 1.7 0.8 2.4 0.1 Bhutan -- -- -- -- Sri Lanka 1.9 0.2 ­0.3 1.5 Bangladesh 1.0 ­0.7 1.6 0.6 Tunisia 2.2 2.0 ­0.2 1.1 Botswana -- -- -- -- Indonesia 1.3 1.8 0.3 ­0.9 Egypt, Arab Rep. of 0.2 1.5 0.0 0.9 Nepal -- -- -- -- Lesotho -- -- -- -- Development "successes" (12) 1.0 0.8 1.2 1.0 Other developing countries (46) 1.0 0.3 ­1.7 ­0.1 Source: Bosworth and Collins 2003. * Indicates that the GDP per capita data series starts later than 1980 (Vietnam in 1984, Lao PDR in 1984, Uganda in 1982, and Ethiopia in 1981). Population in millions as of 2000. GDP per capita is in 1995 U.S. dollars. ** The Republic of Korea has graduated into the ranks of developed countries, but is included in this table because during the 1980s it was still considered part of the developing world. --. Not available. policies move an economy to its own frontier; but incentives have been effective. Further, interven- shifts of the frontier itself through technological tions to improve technology and productivity are progress or factor accumulation are the essence of hard to evaluate and replicate, and successful ini- growth processes. tiatives have coexisted with less successful ones. Governments have played a role in techno- In Brazil, for example, the government correctly logical catch-up in each of the 18 countries, but identified the potential of computers in the each in their own way.Korea used bank loans for 1970s,and set up a publicly funded research cen- indirect funding of private firms engaged in ter and protected domestic producers from for- modernizing their production methods. Singa- eign competition. A large domestic computer pore taxed labor to discourage low-skilled jobs. industry had developed by the mid-1980s, but Malaysia funded vocational training,provided tax Brazilian computers were costlier than the better breaks, and established special economic zones. computers that were available abroad, and the Many have intervened by imposing local content domestic computer industry did not withstand restrictions. In India, an important contribution external competition when Brazil liberalized its of government has been to stimulate the green trade in the 1990s.At the same time, Brazil suc- revolution. ceeded in developing commercial airplanes Since economic growth puts a halo on all of (Embraer), which have won a significant share of a country's policies,it is difficult to discern which the world market. In Indonesia, an attempt to LESSONS FROM COUNTRIES THAT HAVE SUSTAINED THEIR GROWTH 87 develop an aircraft industry has been more costly infrastructure (the 8-7 program in China), social than successful: the government spent $400 mil- spending (Tunisia), policies to increase opportu- lion on R&D (not just on aviation) and invested nities for economically underprivileged groups some $3 billion in a showcase aircraft factory (affirmative actions for bumiputra in Malaysia or without success (World Bank, World Development scholarships in Bangladesh for girls' secondary Report 2005). By contrast, the garment industry education) or poverty-targeted programs (food in Bali, Indonesia, was "accidentally" industrial- stamps in Sri Lanka or employment programs in ized in the 1980s, after foreign tourists (mainly India and Bangladesh). Soeharto's Indonesia surfers) saw the commercial potential in Balinese developed the concept of"economic democracy," indigenous designs and became marketing inter- which advocated reliance on the free market for mediaries connecting local producers with for- growth. In Malaysia, the New Economic Policy eign retail outlets.This success happened despite formally articulated a consensus strategy to elim- the skepticism of the country's then­research and inate the identification of race with economic technology minister for Indonesians' becoming function. In Chile, distributive programs, and, "tailors to the world." since the country's return to democracy, increas- To sum up, the improvements in productivity ing social spending, have been central objectives in the 18 successful countries can be best under- of policies. In India, reduction of famines and stood as resulting from the mix of general poli- improvement in the living standards of the popu- cies and institutions that created the conditions lation have guided policies since Independence. for adoption of technology, together with direct Sri Lanka has maintained a long tradition of government interventions toward this goal, not inclusion, even though mistargeting and politi- all of which succeeded.The exact contribution cization of access to benefits have been serious of each element is extremely difficult to measure. issues.Egypt andTunisia have made efforts to raise the consumption levels of low-income groups, Shared Growth through Opportunities, Pub- for example, through housing programs or subsi- lic Expenditure, and Distributive Programs dies for items of popular consumption. In East Asia, the emphasis on growth led the Growth-oriented strategies for reducing region's governments to focus on augmenting poverty and generating opportunities require productive capacity and the efficient delivery of access to public services, and all countries have social services, rather than on augmenting con- sought to expand public services with different sumption among groups that might otherwise be degrees of success.Even in Soeharto's Indonesia, left behind. Since these countries have sought to where corruption eroded the effectiveness of equalize ex ante opportunities rather than ex some of the country's key institutions, infra- post outcomes, efforts to expand opportunities structure and social services were considerably and mechanisms that facilitate upward mobility expanded to reach significant segments of the have played a more important role than distribu- population and played an important role in cre- tive programs.Elsewhere,however,direct income ating opportunities and distributing the benefits transfers or subsidization of specific commodi- of growth. The expansion of education and ties--as has been common in India, Egypt, Sri health services during the 1970s, as well as the Lanka, and Nepal--have played much more agricultural development policies in the 1960s important roles. and 1970s, were both important in this respect. With few exceptions,3 ensuring that the ben- The village grant program (Inpres Desa) is a efits of growth reached all segments of the popu- good example,as well as a case of innovation and lation was part of the growth strategy. To adaptation.This program, which started in the distribute the benefits of growth, governments 1970s as a top-down grant for centrally pre- relied on a different set of policies and programs. scribed expenditures at the village level, evolved In some cases, they redistributed assets and land, into the village improvement program (VIP), while in others they used public expenditures in which in turn inspired the Kecamatan Develop- 88 ECONOMIC GROWTH IN THE 1990s ment Program,the largest successful community FIGURE B.4 development program in the world. Average Annual Growth of per Capita Income of Perhaps just as important for rural develop- Different Income Groups, 1980 to Mid/Late 1990s ment as targeted programs, though little recog- nized as such, has been the maintenance of a 10.0 competitive exchange rate. The distributive impact of government pro- 8.0 grams is difficult to disentangle from the impact of growth. For example, it is well known that returns to education are higher in rapidly grow- 6.0 ing economies--with the result that investments in education are both capacity enhancing and 4.0 distributive, and more so at high rates of eco- Percent nomic growth. Investments in water and sanita- 2.0 tion, and other forms of infrastructure, are capacity enhancing and distributive.What seem 0.0 to have been important are pragmatic interven- tions that ensured that the incomes of the bot- ­2.0 tom 20 or 40 percent of the population grew. In Bottom 40% Middle 40% Top 20% East Asian countries, the incomes of the bottom China Malaysia Brazil 40 percent of the population have grown quite Mexico Nigeria Korea, rapidly over the last 20 years (figure B.4). Korea Indonesia Rep. of is the East Asian country with the highest Sources: World Bank, World Development Indicators 2004; Global Development Finance growth rate of incomes of the bottom 40 per- 2004. cent, even though its average aggregate growth FIGURE B.5 Ratio of Real per Capita Income of Bottom 40 Percent to That of Top 20 Percent, 1980 to Mid/Late 1990s 0.30 0.25 0.20 0.15 0.10 0.05 0.00 1980 1990 Mid-late 1990s China Malaysia Brazil Mexico Nigeria Korea, Rep. of Indonesia Sources: World Bank, World Development Indicators 2004; Global Development Finance 2004. LESSONS FROM COUNTRIES THAT HAVE SUSTAINED THEIR GROWTH 89 TABLE B.6 Progress on Social Indicators, 1980­2000 Under-five mortality rate Secondary school enrollment 1980 1990 2000 1980 1990 2000 China 120 49 40 46 49 68 Vietnam 87 53 30 42 32 67 Korea, Rep. of 18 9 5 78 90 94 Chile 98 19 12 53 74 86 Malaysia 63 21 9 48 56 69 Lao PDR 218 163 105 21 25 38 India 202 123 94 30 44 49 Thailand 102 40 29 29 30 83 Sri Lanka 100 26 20 55 74 .. Bangladesh 239 144 82 18 19 46 Indonesia 172 91 48 29 44 57 Egypt, Arab Rep. of 235 104 45 50 76 85 Nepal 234 143 91 21 33 40 Tunisia 201 52 28 27 45 78 Botswana 142 58 101 19 43 73 Mauritius 86 25 20 50 53 77 Bhutan 267 166 100 .. .. .. Lesotho 190 148 133 18 25 32 All developing countries 131 103 90 41 47 63 Source: WDI 2003. .. Not available. was slower than China's.And even though China equitable outcomes have been more sustainable itself has seen more rapid growth among the over time. richest 20 percent of its population (the urban Rapid growth was accompanied by wide middle and upper class), the growth rate of its improvements in social indicators and access to poorest citizens has still been very rapid, at more expanding public services. Primary and second- than 6 percent a year. By contrast, countries in ary education expanded massively in Botswana other regions, as represented by Mexico and and Indonesia, and health improved vastly in Nigeria in figure B.4, have mostly seen below- Egypt and Tunisia, as well as in Chile and average growth rates for their poorest 40 per- Malaysia, among other countries (table B.6). cent. Brazil is an exception, but its aggregate While there may be an element of reverse cau- growth in this period has been very slow. sation--when incomes rise, health and educa- This result does not just reflect the initial dis- tion outcomes improve--governments in all the tribution of income (figure B.5). In 1980, Korea success cases were central to the expansion of and Nigeria had roughly similar income distri- social services. butions as measured by the share of the bottom 40 percent of the population in total income.Yet the relative growth rates of income of the bot- Notes tom and top strata differed substantially between the two countries. Similarly, Malaysia and Mex- 1. The experiences of China, the Republic of Korea, Indonesia, and Malaysia were reviewed for the Shang- ico had similar initial income distributions but a hai Conference on Poverty Reduction (see different sharing of growth. It appears that the http://www.worldbank.org/wbi/reducingpoverty/ growth strategies themselves produced the dif- cases-SearchTOC. html), and others as part of the ferent distributional outcomes, and that more Global Development Network's work on growth 90 ECONOMIC GROWTH IN THE 1990s (http://www.gdnet/). This country note draws on those sources as well as onWorld Bank economic and sector work. 2. Harberger (2004) explains how recessions or the absence of Schumpeterian "creative destruction" could produce negativeTFP estimates. 3. For example Nepal,where growth has been concen- trated in the Kathmandu Valley and has had only a modest impact on poverty. Part 2 Development Controversies of the 1990s Chapter 4 Macroeconomic Stability: The More the Better? M ACROECONOMIC POLICIES a clarification of the meaning of macroeconomic improved in a majority of instability and of how to measure it empirically. developing countries in the Conceptually, macroeconomic instability refers 1990s, but the expected growth benefits failed to to phenomena that make the domestic macro- materialize, at least to the extent that many economic environment less predictable, and it is observers had forecast. In addition, a series of of concern because unpredictability hampers financial crises severely depressed growth and resource allocation decisions, investment, and worsened poverty. growth.2 Macroeconomic instability can take the What is the relationship between these devel- form of volatility of key macroeconomic variables opments? This chapter argues that both slow or of unsustainability in their behavior (which growth and multiple crises were symptoms of predicts future volatility). deficiencies in the design and execution of the To examine the evolution of macroeconomic pro-growth reform strategies that were adopted stability, we look at the behavior of macroeco- in the 1990s with macroeconomic stability as nomic outcome variables including the growth their centerpiece.1 Section 1 reviews how of real output, the rate of inflation, and the cur- macroeconomic stability evolved during the rent account deficit.It focuses on the volatility of 1990s. Section 2 evaluates this experience from the growth rate and the levels of inflation and the the perspective of promoting economic growth, current account deficit.3 Changes in the behav- examining how a policy agenda that focused on ior of these endogenous variables can reflect macroeconomic stability turned out to be associ- changes in the macroeconomic policy environ- ated with a multitude of crises. Section 3 draws ment as well as exogenous shocks.Thus to distin- lessons, which essentially concern the depth and guish the roles of these two factors we look at the breadth of the macro reform agenda,the need for behavior of fiscal, monetary, and exchange rate attention to macroeconomic vulnerabilities, and policy variables as well as at real and financial the importance of policies outside the macro- exogenous shocks to developing countries. economic sphere. Stability of Macroeconomic Outcomes 1. Macroeconomic Facts of the Developing countries have traditionally experi- 1990s enced much greater macroeconomic instability than industrial economies.This problem is widely How did macroeconomic stability evolve over perceived to have worsened,4 but in fact the the 1990s?Answering this question requires,first, volatility of developing countries' key macroeco- 93 94 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s nomic aggregates declined in the 1990s.5 For the Republic of Korea) but also countries whose example, the median standard deviation of per growth volatility declined (such as Madagascar, capita gross domestic product (GDP) growth fell which suffered a large drop in GDP in 1991; from 4 percent in the 1970s and 1980s to about 3 Mexico;and Ecuador).There is evidence that this percent in the 1990s, although it remained signif- crisis-type volatility is significantly more adverse icantly higher than the comparable figure for for growth than normal volatility (Hnatkovska industrial economies (1.5 percent) (figure 4.1).6,7 and Loayza 2004).10 The reduction in GDP volatility was widespread Inflation rates improved in the 1990s.Among but far from universal:of the 77 developing coun- middle-income countries the median annual tries for which complete information is available inflation rate declined from a peak of 16 percent for 1960­2000,about a third (27 countries) expe- in 1990 to 6 percent in 2000. Among low- rienced more volatile growth in the 1990s than in income countries, inflation peaked during the 1980s. In turn, the volatility of private con- 1994­95 in the wake of the devaluation of the sumption growth also declined relative to the pre- CFA franc, and then declined (figure 4.3).The vious decade in low-income developing incidence of high inflation among developing countries. In middle-income countries, however, countries declined sharply after peaking in 1991 consumption volatility remained virtually (figure 4.4). But over the 1990s as a whole, the unchanged at the record highs of the 1980s.8 number of developing countries experiencing The reduction in the aggregate volatility of average inflation higher than 50 percent was no GDP growth concealed the increasing role played smaller than in the 1980s. by extreme instability (figure 4.2). In the 1990s, Other things being equal, reduced aggregate large negative shocks accounted for close to one- volatility and lower inflation probably improved fourth of total growth volatility, against 14 per- the incomes of the poor.The inflation tax tends to cent in the 1960s and 1970s and 18 percent in the fall disproportionately on poorer households, 1980s.9 And the increasing incidence of growth crises affected not only countries whose growth volatility rose (such as Indonesia, Malaysia, and FIGURE 4.2 Structure of GDP Growth Volatility, 1961­2000 FIGURE 4.1 (percent, mean of 77 developing countries) GDP Growth Volatility, 1966­2000 (percent, medians by country income group) 80 70 5 60 4 50 40 3 30 20 2 10 0 1 1960­70 1971­80 1981­90 1991­2000 Normal Extreme Crisis Boom 0 Source: Author's own elaboration using data from World Bank, All countries Industrial Least Middle- Low-income (97) countries developed income countries(33) World Development Indicators, and Hnatkovska and Loayza 2004. (20) countries (77) countries (41) Note: Extreme shocks are defined as those exceeding two standard 1966­70 1971­80 1981­90 1991­2000 deviations of output growth over the respective decade. Total Sources: World Bank, World Development Indicators; Hnatkovska and Loayza 2004. volatility = Normal + Extreme; Extreme = Crisis + Boom. M AC RO E C O N O M I C S TA B I L I T Y: T H E M O R E T H E B E T T E R ? 95 FIGURE 4.3 Inflation Rates, 1991­99 (GDP deflator, medians by country income group) 30 25 20 rate 15 10 Inflation 5 0 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 Industrial countries (20) Least developed countries (77) Middlie-income countries (41) Low-income countries (33) Source: World Bank, World Development Indicators. FIGURE 4.4 High Inflation in Developing Countries, 1961­99 (relative frequency, percent) 14 12 10 8 6 4 2 0 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 Above 50% Above 80% Source: World Bank, World Development Indicators. which hold few or no financial assets to shelter the median current account deficit/GDP ratio them against rising prices, and whose wage earn- was about one percentage point lower than in the ings typically are not fully indexed to inflation. 1970s and 1980s.13 In the latter, it rose by about Through this and other channels,higher aggregate half a point in relation to the 1980s to exceed 5 volatility is empirically associated with worsening percent of GDP in the 1990s (figure 4.5). income distribution.11 The median current account deficit among Stability of Policies developing countries decreased slightly in the 1990s,although there was a contrast between mid- Conventional indicators of policy stability also dle- and low-income countries.12 In the former, improved over the 1990s. Most notably, the 96 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s Since the overall fiscal balance is affected by FIGURE 4.5 the trajectory of interest rates on public debt Current Account, 1966­2000 (which is beyond the direct control of the (percentage of GDP, medians by country income group) authorities), the primary balance likely offers a 2 more accurate measure of a country's fiscal 1 stance. Its evolution over the 1990s shows clear 0 increases in surpluses,particularly after 1995 (fig- ure 4.7). By the end of the decade, the median ­1 developing country held a primary surplus, ­2 although a much more modest one than that typ- ­3 ical of industrial countries.14 ­4 It is more difficult to gauge monetary stabil- ­5 ity,given the diversity of monetary arrangements ­6 across developing countries and over time. One All countries Industrial Least developed Middle-income Low-income (70) countries (17) countries (53) countries (32) countries (19) rough measure is the resort to seigniorage--that 1966­70 1971­80 1981­90 1991­2001 is, money financing of the deficit. Measured by Sources: World Bank, World Development Indicators; IMF, BoP4 the change in the money base relative to GDP, seigniorage collection rose in the late 1980s and Note: The countries featured are those for which data are available over the entire period shown. early 1990s, and then declined in middle- income and (more modestly) low-income overall fiscal deficit of developing countries economies (figure 4.8).The pattern is roughly shrank from a median value of 6­7 percent of similar to that of the inflation rate (figure 4.3 GDP in the early 1980s to 2 percent of GDP in above). the 1990s, before rebounding to about 3 per- The diversity of exchange rate arrangements cent by the end of the decade.The fiscal cor- across countries makes it hard to gauge trends in rection was particularly pronounced among exchange rate policy for developing countries middle-income countries (figure 4.6). as a group. One indirect approach looks at FIGURE 4.6 Developing Countries' Overall Fiscal Balance (percentage of GDP, medians by country income group) 1 0 ­1 ­2 ­3 ­4 ­5 ­6 ­7 ­8 ­9 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 Least developed countries (37) Middle-income countries (29) Low-income countries (8) Sources: World Bank, World Development Indicators; Institute of International Finance. Note: The countries featured are those for which complete data are available from the late 1970s on. The availability of consistent fiscal balance data is very limited, particularly for low-income countries. M AC RO E C O N O M I C S TA B I L I T Y: T H E M O R E T H E B E T T E R ? 97 FIGURE 4.7 Primary Fiscal Balance, 1990­2002 (percentage of GDP, medians by country income group) 5 4 3 2 1 0 ­1 ­2 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 All countries (61) Industrial countries (20) Least developed countries (41) Source: Fitch Ratings. Note: These data differ in source and coverage from those underlying Figure 4.6. Therefore the two figures are not strictly comparable. trends in real exchange rates. Real exchange rates depreciated over the 1990s in a majority of FIGURE 4.8 developing countries. For the median develop- Developing Countries: Seigniorage Revenues, 1966­2000 ing country, the volatility of the real exchange (percentage of GDP, medians by country income group) rate (as measured by the standard deviation of 3.0 the rate of change of the real exchange rate) declined from the record highs of the 1980s,but 2.5 the decline was limited to middle-income 2.0 countries, and over the 1990s developing coun- tries as a group exhibited much more volatile 1.5 real exchange rates than industrial countries (figure 4.9). 1.0 The relatively high volatility of real 0.5 exchange rates partly reflected the high inci- dence of exchange rate crises (figure 4.10).The 0.0 incidence of devaluations peaked in 1994, with 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 the devaluation of the CFA franc, and in 1998, Middle-income countries (34) Low-income countries (30) with the East Asia and Russian Federation Sources: IMF, International Finance Statistics; World Bank, World Development Indica- crises.When we look at the decade as a whole, tors. it emerges that exchange rate crises were Note: The countries featured are those for which data are available over the entire period slightly less frequent in the 1990s than in the shown. 1980s, but much more so than in the 1960s and 1970s.15 The External Environment High real exchange rate volatility and fre- quent exchange rate collapses suggest that over What role did external shocks, real or financial, the 1990s progress in achieving robust nominal play in the observed trends in macroeconomic exchange rate arrangements was limited. instability? 98 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s As to real disturbances, developing countries FIGURE 4.9 suffered only modest terms-of-trade shocks in Real Exchange Rate Volatility, 1961­2000 the 1990s (see chapter 3).The volatility of the (percent, medians, by income group) terms of trade declined in all developing regions, in most cases to levels comparable to those of the 16 1960s.The only exception was the Middle East 14 and North Africa region, whose terms of trade 12 were still less volatile than in the 1970s and 10 1980s. 8 It is more difficult to assess the volatility of the financial environment.The behavior of inter- 6 est rates in the world's major financial markets 4 captures some of this volatility, but the interest 2 rates paid by developing countries incorporate 0 risk premia that make these rates much more All Industrial Least Middle-income Low-income countries countries developed countries countries volatile than industrial-country interest rates.16 (80) (19) countries (61) (32) (26) Volatility measures based on such risk premia, or 1961­70 1971­80 1981­90 1991­2000 indeed on flows of capital to developing coun- Source: Aten, Heston, and Summers 2001. tries, are not necessarily good indicators of the Note: Figure shows the standard deviation of the rate of change in the real exchange volatility of the international financial environ- rate. The countries featured are those for which data are available over the entire ment, since they partly depend on events in the period shown. borrowing countries themselves. Figure 4.11 shows the volatility of international net capital flows as measured by their standard deviation.This measure suggests that the external FIGURE 4.10 Developing Countries: Exchange Rate Crises, 1963­2001 (relative frequency, percent) 35 30 25 20 15 10 5 0 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 Least developed countries (77) Middle-income countries (41) Low-income countries (33) Source: IMF, International Finance Statistics. Note: For this figure an exchange rate crisis is defined as in Frankel and Rose (1996): a depreciation of the (average) nominal exchange rate that (a) exceeds 25 percent, (b) exceeds the preceding year's rate of nominal depreciation by at least 10 percent, and (c) is at least three years apart from any previous cri- sis. The countries featured are those for which data are available over the entire period shown. M AC RO E C O N O M I C S TA B I L I T Y: T H E M O R E T H E B E T T E R ? 99 FIGURE 4.11 FIGURE 4.12 Volatility of Net Capital Flows, 1977­2000 Developing Countries: Sudden Stops in Net Capital (percent, medians by country income group) Inflows, 1978­2000 (relative frequency, percent) 5 50 45 4 40 35 3 30 2 25 20 1 15 10 0 5 Industrial Least Middle- Low- countries developed income income 0 (20) countries countries countries (43) (31) (7) 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 1977­80 1981­90 1991­2000 Stop > 5 percent of GDP Stop > 2.5 percent of GDP Source: IMF, International Finance Statistics. Balanced sample includes 53 countries. Source: IMF, International Finance Statistics. Note: Data for the first half of the 1970s are too limited to allow a comprehensive Note: Figure shows the standard deviation of net capital flows as analysis. Sudden stops are defined as declines in net capital inflows in excess of a given a percentage of GDP. Using instead the coefficient of variation percentage of GDP. Reversals are allowed to take place in adjacent years; using a two- leads to qualitatively similar results. The countries featured are year window leads to similar qualitative conclusions. Note that reversals could have those for which data are available over the entire period shown. been defined instead in terms of (large) changes in the current account deficit (as done, for example, by Hutchison and Noy 2002). However, when applied to a large financial environment was modestly less volatile in cross-country sample such as the one at hand, the latter criterion tends to pick up numerous current account reversals (particularly in low-income countries) owing pri- the 1990s than in the 1980s, but that capital flows marily to terms-of-trade shocks in a context of modest changes in capital flows. to developing countries remained much more volatile than those to industrial countries. Several observers have pointed out that large and nominal stability helped achieve a moderate capital flow reversals, often termed "sudden reduction in output volatility, facilitated by a stops,"can be much more damaging for develop- somewhat more stable external environment. ing economies than is general capital-flow vari- But the picture was far from rosy. Developing ability,because such abrupt stoppages force costly economies remained much less stable than indus- and disruptive real adjustments.17 Sudden stops trial ones.And extreme volatility accounted for a were not significantly more frequent in the 1990s larger share of total volatility than previously.This than in the 1980s (figure 4.12).Their incidence latter fact accords with evidence suggesting that declined in the first half of the 1990s, but then instances of currency crashes and "sudden stops" rose again in the second half, peaking about the in capital inflows did not diminish during the time of the East Asia and Russia crises.18 1990s.The picture is therefore one of dramatic policy improvements in some areas, of more moderate improvements in the stability of macroeconomic outcomes,and of persistent vul- 2. Assessing the Experience of nerability to extreme macroeconomic events. the 1990s Below we use these findings to interpret the growth performance of developing countries The brief review, above, of the macroeconomic during the 1990s.We first review the analytical facts of the 1990s shows that developing coun- links between macroeconomic stability and eco- tries achieved notable progress on fiscal consoli- nomic growth and then apply that framework to dation and inflation performance. Better fiscal the experience of the 1990s. 100 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s Links between Stability and Growth pinning, that sources of macroeconomic fragility be eliminated to the greatest possible extent, and A stable macroeconomic policy environment fea- that the authorities actively exploit the scope for tures a fiscal stance safely consistent with fiscal stabilization policy created by these two improve- solvency,a monetary policy stance consistent with ments in the macroeconomic environment. a low and stable rate of inflation, and a robust exchange rate regime that avoids both systematic How Much Macroeconomic ProgressWas currency misalignment and excessive volatility in Made in the 1990s? the real exchange rate. Policy makers can foster stable macroeconomic outcomes both directly-- As argued above, developing countries achieved by removing destabilizing policies themselves as significant stability in the traditional macroeco- sources of shocks--and indirectly--by using poli- nomic policy sense during the late 1980s and cies as stabilizing instruments in response to early 1990s.These achievements were far from exogenous destabilizing shocks, thus enhancing universal,however,and the consequence was that the stability of key outcome variables. A stable macro instability continued to impede growth in policy framework is not an end in itself:it matters some countries and allowed traditional macro only as a means to secure a more stable overall imbalances to generate crises that in many ways macroeconomic environment. resembled those of the 1980s. Neither were the Conceptually, the link between policy stabil- achievements always based on solid institutional ity and growth is quite complex. First, the direct foundations to guarantee their permanence, and contribution that policy stability can make to they frequently did not translate into more effec- growth is likely to depend on the institutional tive use of macro policies as stabilization instru- setting.What matters is not just whether policies ments. are good today, but the perceived likelihood that A useful framework for discussing these issues they will continue to be so.To have a significant is the public sector solvency condition, which impact on growth, actual gains in macroeco- requires the present value (PV) of primary sur- nomic stability need to be seen by the private pluses (T ­ G) and seigniorage revenue (dM) to sector as signs of a permanent change in the be at least as large as the government's outstand- macroeconomic policy regime. Second, the ing stock of net debt (B): potential indirect contribution of policy stability to growth--by promoting stable outcomes in the PV (T ­ G + dM) B (0). face of external shocks--is likely to depend on how vulnerable the economy is to shocks. Stability requires a monetary and fiscal policy Macroeconomic fragility--through which even stance consistent with maintaining public sector minor shocks may have large macroeconomic solvency at low levels of inflation, while leaving consequences--may make the use of stabiliza- some scope for mitigating the impact of real and tion policies too costly, for fear of potentially financial shocks on macroeconomic perform- adverse effects; here the result is policy paralysis. ance. The former requirement imposes con- Or fragility may mean that the instability straints on the size of both the primary deficit becomes so severe that no feasible policy adjust- (G ­ T) and its money financing dM, while the ments are able to counter it. latter refers to the profiles of monetary and fiscal These two points suggest that the type of policy over the business cycle.These require- macroeconomic stability likely to be most con- ments apply not only to the present but also to ducive to economic growth--durable outcomes- the future, as implied by the present-value term based stability--involves much more than just in the expression.19 moving fiscal,monetary,and exchange rate policies Reassessing developments during the 1990s in stabilizing directions. It requires that policy- in the light of the expression above, the follow- based stability be given a solid institutional under- ing observations emerge: M AC RO E C O N O M I C S TA B I L I T Y: T H E M O R E T H E B E T T E R ? 101 · Most countries have yet to convey a convinc- ing impression of fiscal solvency. FIGURE 4.13 Government Debt, 1990­2002 · Improvement in fiscal balances was often (percentage of GDP, medians by country income group) achieved either with stopgap measures that 100 were unlikely to be sustainable or in ways inimical to growth and welfare. 90 80 · In many countries, fiscal policy remains destabilizing. 70 · Lasting nominal stability remains to be credi- 60 bly established. 50 · Robust exchange rate arrangements have 40 remained elusive. 30 · The reform agenda has proved to be incom- 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 plete. Industrial Least developed Middle-income Low-income countries (24) countries (50) countries (35) countries (12) Source: World Bank, WDI; IMF, World Economic Outlook; Fitch Ratings. We discuss these observations in turn. Most Countries Have Yet to Convey a monetization, reflected in the decline in Convincing Impression of Fiscal Solvency seigniorage revenues (figure 4.8 above).Without Fiscal adjustment in the 1990s was often weak- an equally rapid correction of the primary ened by increases in debt that offset improve- deficit, debt issuance was left as the only source ments in primary surpluses. Despite the trend of financing.The debt impact of disinflation is toward lower fiscal deficits (figure 4.6 above), the confirmed by the statistically significant associa- ratio of public debt to GDP remained high in tion between disinflation and subsequent rises in most developing countries (figure 4.13).And an debt ratios over the 1990s. incipient decline in these countries' ratios In a majority of developing countries, how- through 1997 was followed by a rise, so that by ever, primary deficits did decline over the 2001­02 the debt ratio of the median developing 1990s, and other factors accounted for the country exceeded the 1990­91 level.20 The ris- lion's share of public debt accumulation. Key ing trend appeared to be particularly marked among these were the costs of banking system among low-income countries, although data are bailouts, which in several countries provided too limited to draw firm conclusions.21 the main impetus for the growth in public The persistence of high and rising debt over debt.23 Some of the banking crises of the the 1990s reflects several factors. 1990s, especially those in East Asia in 1997, had First, improvements in fiscal performance the greatest fiscal impact in history (figure were not universal. In India, for example, con- 4.14).24 Such crises also adversely affected tinuing large primary deficits, averaging close to income distribution, through their fiscal impact 4 percent of GDP in the late 1990s, were the and other channels involving implicit net trans- main factor behind persistent high debt ratios. fers from poorer households to financial system Fiscal vulnerabilities played a role in the finan- participants, in order to rescue and recapitalize cial crises in Russia in 1998, Ecuador in 1999, the failed banks.25 and Argentina in 2002.22 In many cases, the Another factor behind the rise in debt stocks pressure of weak public finances on debt accu- in the late 1990s was large real exchange rate mulation was revealed by an attempt at rapid dis- depreciations, undertaken in a context in which inflation, which implied a drop in deficit the bulk of public debt was denominated in (or 102 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s FIGURE 4.14 Total Fiscal Costs of Systemic Banking Crises as a Percentage of GDP Indonesia 1997 Thailand 1997 Turkey 2000 Korea, Rep. of 1997 Ecuador 1998 Mexico 1994 Venezuela, R. B. de 1994 Malaysia 1997 Paraguay 1995 Bulgaria 1995 Finland 1991 Hungary 1991 Sweden 1991 Argentina 1980 Chile 1982 Uruguay 1981 Senegal 1988 Spain 1977a Norway 1987 Colombia 1982 Sri Lanka 1989 0 10 20 30 40 50 60 Percent of GDP Source: Caprio and Klingebiel 2003. a. Percentage of GNP. indexed to) foreign currency. In both Argentina Thus,as to the solvency constraint introduced and Uruguay,for example,the collapse of domes- above,the bottom line is that,in many countries, tic currencies in 2002 more than doubled the increases in the observed value of the primary debt-to-GDP ratio, from 50 percent to more surplus T ­ G did not suffice to bring down the than 140 percent of GDP in Argentina and from burden of public debt. 40 percent to more than 80 percent in Uruguay. A strong indication that perceptions of solvency Across emerging markets, debt dollarization remained shaky in the 1990s is the fact that default remained pervasive: the median ratio of foreign risk premia, as measured by sovereign borrowing currency debt to total public debt rose over the spreads in international markets, remained highly late 1990s to more than 55 percent by 2001 (fig- volatile for most emerging countries (figure 4.16). ure 4.15). As noted earlier, the evidence suggests that these A further reason for the persistence of high premia depend not only on borrowers' existing debt was the high real interest rates that prevailed debt burdens but also on investors' perceptions in many countries, particularly in the late 1990s. about the quality of borrowers' policy and institu- This largely reflected the lack of credibility of tional frameworks, and medium-term economic stabilization efforts (documented below). Exces- growth prospects--a key determinant of public sive reliance on short-term debt made some sector solvency (Kraay and Nehru 2003).Thus,the countries' overall fiscal outcomes, and thus their volatility of risk premia likely reflected, among rates of public debt accumulation, highly sensi- other factors, the markets' shifting perceptions tive to changes in domestic interest rates.In some about borrowers'ability to ensure stability and sus- countries, notably Brazil, high real interest rates tain adequate growth. contributed to a rapid pileup of public debt, fur- Perceptions of high default risk are not ther weakening perceptions of solvency and merely a symptom of perceived vulnerability. macroeconomic stability. They themselves undermine macroeconomic M AC RO E C O N O M I C S TA B I L I T Y: T H E M O R E T H E B E T T E R ? 103 FIGURE 4.15 FIGURE 4.16 Developing Countries' Foreign Currency Emerging Markets Bond Index Spreads for Latin and Debt, 1997 and 2001 Non-Latin Borrowers (percentage of general government debt, medians by (basis points) country income group) 2500 70 2000 60 50 points 1500 40 30 1000 Basis 20 10 500 0 1997 2001 0 Least developed Middle-income Low-income 95 96 countries (45) countries (37) countries (8) Dec-91 Jun-92 Dec-92 Jun-93 Dec-93 Jun-94 Nov-94 May Nov-95 May Oct-96 Apr-97 Oct-97 Apr-98 Oct-98 Apr-99 Sep-99 Mar-00 Sep-00 Mar-01 Aug-01 Feb-02 Aug-02 Feb-03 Jul-03 Source: Moody's. Latin Non-Latin Source: JP Morgan. stability over business cycles. In particular, they hamper countries' ability to conduct stabilizing speeding up the extraction of exhaustible policies:when default risk is perceived to be high resources or advancing tax collection) without and highly sensitive to changes in circumstances, altering their present value, which is the relevant a country's attempts to run deficits at times of magnitude for solvency.Another popular strategy cyclical contraction may be viewed with suspi- involves one-time asset sales to finance the retire- cion and result in large jumps in risk premia (and ment of public debt, which in principle implies thus borrowing costs), in turn discouraging the no change in the government's net worth. Like- use of counter-cyclical fiscal policy.26 wise, governments have often resorted to replac- ing explicit debt with contingent liabilities (for Often Improvement in Fiscal Balances Was example, granting debt guarantees rather than Achieved Either with Stopgap Measures or subsidies to public firms). All these measures in Ways Inimical to Growth and Welfare improve conventional indicators of cash deficit In numerous instances, fiscal improvements and gross debt--the two fiscal benchmarks themselves were perceived as purely temporary, closely watched by investors and international either because the measures used to achieve financial institutions--but have no effect on sol- them were clearly transitory or because they vency.They represent illusory fiscal adjustment.27 directly compromised future growth and wel- In other instances, the appearance of fiscal fare. In terms of the solvency constraint above, adjustment may reflect a rise in revenues result- such adjustments often reduced the current ing from a temporary boom in tax bases.This deficit significantly but had little effect (or even may happen, for example, when a transitory an adverse one) on the path of future deficits. surge in capital inflows boosts consumption in an Such temporary fiscal correction was some- economy with a value added tax (VAT)-domi- times achieved through fiscal tricks designed to nated tax system.When the consumption boom meet short-term targets for deficits or debt with- ends, a major fiscal gap opens.There is evidence out making real progress toward fiscal solvency.A that this mechanism played a significant role in common such device involves changing the tim- some emerging markets in the 1990s (Talvi ing of expenditures (for example postponing 1997). them into subsequent fiscal years or accumulat- More generally, many fiscal adjustment ing payments arrears) or revenues (for example, episodes have focused more on the quantity than 104 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s on the quality of adjustment, with very limited tends to be pro-cyclical,expanding in booms and attention given to public spending composition contracting in recessions--a pattern that makes it and its implications for growth and welfare. a major source of macroeconomic instability. Sometimes the result has been adjustment at the Take,for example,the cyclical behavior of public cost of social expenditures, leaving critical social consumption. On average in developing coun- needs unmet (IMF 2003a, chapter 6). But reduc- tries, a 1 percent increase in GDP growth tends ing spending on health and education may retard to raise the growth rate of public consumption growth not just by reducing the accumulation of spending by about 0.5 percentage point.Among human capital, but also by undermining political industrial countries the corresponding figure is support for sustaining responsible macroeco- much smaller, at about 0.15 percentage point, nomic policies. Such measures defeat the ulti- and in the G-7 countries the response of public mate objective of fiscal adjustment--namely, to consumption is actually negative.30 allow the resumption of sustained growth.28 Among developing countries, fiscal pro- More often than not, productive public cyclicality peaked in the 1980s and declined expenditures,on items such as human capital for- somewhat over the 1990s, but it remained much mation and infrastructure, have also been com- higher than in more advanced countries (figure pressed in the process of fiscal adjustment.The 4.17). Pro-cyclical fiscal policy played a key role main reason is that the emphasis on cash deficits in some of the recent crises, notably in and debt discourages projects whose costs are Argentina.31 borne upfront but whose returns accrue only over time.Such projects have the same impact on Lasting Nominal Stability Remains to Be the government's short-term financing needs as Credibly Established does pure consumption or any other spending The preceding points refer to two of the three item,but their impact on solvency is quite differ- components of the public sector solvency condi- ent because, unlike consumption, they involve tion:net debt B,and the present value of the pri- creating assets that yield future tax revenues mary surplus, PV (T ­ G).The third component (either directly or by augmenting output and is the present value of seigniorage revenue, thence augmenting revenues).The conventional PV(dM). Developing countries substantially fiscal aggregates--such as the primary or the reduced the monetization of their deficits in the overall surplus that is closely monitored by inter- 1990s (figure 4.8 above), but in many of them national financial institutions and investors-- the stability of prices remains vulnerable. ignore this distinction, and the result is that fiscal A transitory reduction in dM can be achieved adjustment tends to have an anti-investment in a variety of ways, but unless durable increases bias.29 in To the extent that reduced investment lowers (T ­ G) are institutionalized, continuing pres- growth and hence future tax bases, such a bias sures on the government budget will result in can adversely affect growth and even fiscal sol- debt accumulation that will in turn create pres- vency itself. Latin America, where reductions in sures for monetization.In many countries reduc- public infrastructure spending supplied the bulk tions in dM were not accompanied by lasting of the fiscal correction achieved by some of the solutions to fiscal problems. Some countries-- region's major countries in the 1990s, provides a notablyArgentina,Brazil,Ecuador,Mexico,Rus- good example of these perverse dynamics. sia, and Turkey--reduced inflation rates as the result of exchange rate-based stabilizations. Bet- In Many Countries, Fiscal Policy Remains ter price performance allowed them to reduce Destabilizing money growth rates, but the sustainability of this The stabilizing power of fiscal policy depends achievement was questionable in all of them. In largely on its ability to mitigate cyclical fluctua- most, persistent fiscal pressures were accompa- tions. But in developing countries fiscal policy nied by real exchange rate appreciations and M AC RO E C O N O M I C S TA B I L I T Y: T H E M O R E T H E B E T T E R ? 105 FIGURE 4.17 Pro-Cyclicality of Public Consumption, 1980­2000 (rolling 15-year windows, medians by country income group) 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 ­0.1 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 ­0.2 Least developed countries (41) G7 Industrial countries non-G7 (14) Source: Authors' own elaboration using data from World Bank, World Development Indicators. Note: The figure shows the median of country-specific coefficient estimates obtained by regressing the rate of growth of public consumption on the rate of GDP growth (plus a constant). increases in real interest rates, leading to a pileup ity. If such an arrangement is to promote lasting of public debt and calling the sustainability of the price stability, the central bank must be able to stabilizations into question. In Argentina and resist pressures for monetization arising from the Ecuador, inability to enforce fiscal discipline led fiscal side.That is, it must achieve true independ- to the adoption of hard exchange rate pegs in the ence from the finance ministry. hope that these would somehow harden the gov- The establishment of truly independent and ernment budget constraint as well.Their failure effective central banks has not been a straightfor- to do so shows that such quick fixes do not ward matter.The creation of independent central achieve lasting nominal stability in the absence of banks in República Bolivariana de Venezuela in an independent commitment to responsible fis- 1989 and in Mexico in 1993, for example, did cal policies. In Brazil, Mexico, and Turkey, not prevent the emergence of the strong political exchange rate­based stabilizations relying on pressures for credit creation that contributed to "soft" pegs eventually resulted in currency crises currency crises in both countries in the first half that gave way to short bursts of accelerated infla- of the 1990s. Similar pressures were brought to tion. Likewise, the devaluation of the CFA franc bear on Argentina's central bank in 2001, on the largely reflected the failure of the CFA arrange- eve of the collapse of the hard peg. ments to enforce fiscal discipline in the face of Some observers suggest that a good indicator adverse terms-of-trade shocks (box 4.1). of de facto central bank independence is the fre- In the search for nominal stability,some coun- quency of turnover of the central bank gover- tries in the 1990s placed their reliance on inde- nor.32 Among middle-income countries, pendent central banks with a commitment to turnover was sharply lower in the 1990s than in price stability.As does a fixed nominal exchange the 1980s, and among low-income developing rate, such an arrangement works in principle by countries it was modestly lower (figure 4.18). committing the central bank to a low value of Since the rate of turnover of central bank gov- dM, thereby imposing a hard budget constraint ernors may not be a good indicator of the on the fiscal authorities and forcing the latter to expected permanence of nominal stability,33 it adjust (T ­ G) to the requirements of price stabil- may be useful to observe the behavior of the pri- 106 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s BOX 4.1 Devaluation of the CFA franc T he 14 West African countries of the CFA franc Sub-Saharan Africa. The CFA franc became substan- zone share the CFA franc as their common tially overvalued. currency. From 1948 to 1993, the CFA franc To reverse the worsening economic performance, the was pegged to the French franc, partly to minimize currency's first major devaluation was implemented in transactions costs in international trade but also to January 1994, when the official parity was changed provide a nominal anchor for these economies. from CFAF 50 to CFAF 100 = F 1. The devaluation was The common currency was reasonably effective in accompanied by measures to improve fiscal perform- maintaining financial discipline in member countries ance (broadening the tax base and reducing expendi- for an extended period. Until the mid-1980s, these tures), as well as structural reforms focused on trade countries enjoyed lower inflation and more sustained liberalization, increasing flexibility in labor markets, economic growth than other Sub-Saharan African reducing the direct role of government in production, countries. But the shortcomings of the hard peg and restructuring financial sectors. against the French franc became apparent in the mid- The results of the devaluation were quite positive. 1980s when the zone was hit by two external shocks: Inflation accelerated at first but quickly converged to a sharp deterioration in member countries' terms of single-digit levels. Consequently, the real effective trade, arising from a decline in the world prices of depreciation of the CFA franc in 1994 amounted to about their primary export commodities, and a strong 30 percent. Real GDP growth, negative in 1993, aver- appreciation of the French franc against the U.S. dol- aged 1.3 percent for the zone as a whole in 1994, and lar. These shocks placed strong pressures on fiscal accelerated subsequently. Overall fiscal deficits, which outcomes, which depended heavily on commodity had peaked at about 8 percent of GDP in 1993, had fallen revenues and trade taxes. Member countries' failure to just over 2 percent of GDP by 1996. A substantial to impose an orderly correction, partly because they increase in saving rates reduced the current account could not adjust public sector wages downward, led deficit by some 2 percent of GDP between 1993 and to sharply higher fiscal and current account deficits, 1996. Coupled with capital repatriation and renewed large increases in external debt, and deteriorating external assistance, this substantially increased the for- growth performance relative to other countries in eign exchange reserves of regional central banks. Source: Clement et al. 1996. vate sector, to try to infer what the private sector 4.19 shows,the median degree of dollarization of expects about nominal stability. bank deposits among low- and middle-income First, since agents can partly protect them- developing countries actually rose over the selves against nominal instability by denominat- 1990s.35 The contrast with richer countries is ing their assets in foreign exchange,one indicator stark: their much lower degree of deposit dollar- of the confidence that private agents in develop- ization showed little change over the same period. ing countries may have in the permanence of Second, ex post real interest rates tend to be nominal stability is the incidence of dollarization. high when actual inflation falls short of expecta- Improved confidence in nominal stability should tions and when uncertainty about inflation is result in a reduced incidence of dollarization.34 high. During the 1990s, real interest rates were Many developing countries remained heavily declining in industrial countries, but in develop- dollarized at the end of the 1990s and, as figure ing countries they remained high--and indeed M AC RO E C O N O M I C S TA B I L I T Y: T H E M O R E T H E B E T T E R ? 107 FIGURE 4.18 FIGURE 4.19 Central Bank Independence in Developing Dollarization of Deposits, 1996 and 2001 Countries, 1975­98 (foreign currency deposits as a percentage of total, medians by country (annual governor turnover, medians by country income income group) group) 50 0.30 40 0.25 0.20 30 0.15 20 0.10 10 0.05 0.00 0 Middle-income Low-income 1996 2001 countries (41) countries (22) High-income Middle-income Low-income 1975­80 1981­90 1991­98 countries (18) countries (43) countries (31) Source: Sturm and de Haan 2001. Source: IMF, International Financial Statistics. Note: For Austria, Haiti, Israel, Mexico, Macedonia, and Netherlands we take the 1997 data, and for Ghana, Italy, Norway, Tajikistan, and Uganda we take the 2000 data. High were higher at the end of the decade than at the corresponds to OECD and non-OECD countries. beginning (figure 4.20). Of course, both dollarization ratios and ex post real interest rates reflect a variety of factors have too often been adopted only after currency in addition to perceptions of nominal instability, crises. so this evidence is only suggestive.36 But other The Mexico and East Asia crises, which indicators point in the same direction. As an involved the collapse of a variety of soft pegs, extreme example, the currency premium on the prompted what came to be known as the "two Argentine peso was positive throughout the extremes" view of exchange rate regimes. In this 1990s, and it became very large at times of tur- view, only irrevocable pegs (including both cur- bulence, in spite of the supposedly irrevocable rency boards and monetary unification or dollar- peg to the dollar that was enshrined in ization) and freely floating exchange rates were fit Argentina's Convertibility Law.37 for survival in a world of increasing financial inte- gration, because only these extreme regimes Robust Exchange Rate Arrangements Have appeared to offer enough transparency to make Remained Elusive exchange rate policy easily verifiable and hence Progress toward robust exchange rate regimes credible.38 There appeared to be an incipient probably was an early casualty of the search for flight away from intermediate regimes,39 based macroeconomic stability. Many countries on the belief that monetary stability required adopted exchange rate­based stabilization strate- either institutional arrangements that took discre- gies as a supposedly quick recipe for disinflation, tion over money growth rates out of the hands of as discussed above. These strategies not only central banks, or fully independent central banks meant adopting single-currency pegs, but also with reputational stakes in low and stable infla- made such pegs very difficult to adjust,since they tion, as well as the means (legal authority, policy tied the credibility of the entire stabilization pro- instruments, human-resource capability) to gram to the stability of the peg.In effect,defend- achieve that goal. ing the peg sometimes became an end in itself, The late 1990s showed that neither dollariza- even after the peg had clearly outlived its useful- tion nor currency boards offered a speedy short- ness. More flexible exchange rate arrangements cut to fiscal orthodoxy and nominal stability. 108 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s Few countries achieved a sound domestic FIGURE 4.20 financial system in the 1990s. As a result, an Ex Post Real Interest Rates, 1990­2001 important source of macroeconomic fragility was (percent, medians by country income group) not only left in place but may, indeed, have been 14 magnified in the 1990s. Inadequate attention to 12 financial sector soundness often left the domestic economic environment rife with institutional 10 problems involving moral hazard,rendering both 8 public and private balance sheets highly vulnera- 6 ble to changes in interest rates and exchange 4 rates.These features posed big obstacles to out- 2 come-based stability in a number of major coun- tries. Ironically, under these circumstances 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 incipient progress along conventional dimen- Industrial Least developed Middle-income Low-income sions of macro stability such as disinflation may countries (22) countries (89) countries (54) countries (30) even have made financial crises more likely. For Source: IMF, International Financial Statistics; World Bank, World Development Indicators. example, the use of the exchange rate as a nom- Note: The real interest rate is measured as the (log) difference between the nominal inal anchor may have encouraged agents to interest rate and the one-period-ahead rate of GDP inflation. ignore exchange rate risk and in the case of "hard" pegs such as that of Argentina may have Argentina's experience revealed the threat to sta- made it more difficult for regulators to induce bility that was posed by inflexible exchange rates, financial institutions to factor such risk into their which made adjustment to real disturbances portfolio allocations without raising fears that the exceedingly difficult. Earlier in the decade, the peg might be abandoned. fate of the CFA franc had offered the same lesson, Partly because of this gap in the reform though it was less publicized (box 4.1 above). agenda, the incidence of systemic banking crises was even higher in the 1990s than in the 1980s The Reform Agenda Has Proved Incomplete (figure 4.21).41 The developing countries' macroeconomic A second key source of macroeconomic reform agenda of the 1990s was deficient in its fragility was increased capital mobility, which very design, in that it left in place--or, worse, made economies vulnerable to sudden shifts in created--important sources of fragility. capital flows.The combination of unsound poli- The first of these sources stemmed from lack cies in the financial sector and open capital of attention to the soundness of the financial sec- accounts helps explain many characteristics of tor. While research has shown that an efficient the crises of the 1990s. Many of these crises domestic financial system is important for involved simultaneous currency and banking growth, the experience of the last decade collapses. Often banking problems preceded a strongly suggests that a sound one is indispensable currency crash, which then fed back into a full- for macroeconomic stability.The reform agenda blown financial crisis.42 Further, many of the of the early 1990s often ignored the central role crises were not foreshadowed by standard macro- of the financial system for macro stability--even economic imbalances.Those that were hardest to though this role had been clearly revealed by the predict--especially the Mexican and Asian Southern Cone crises of the early 1980s.To the crises--occurred in a setting where the main standard prescriptions for stability--a solvent fis- vulnerabilities concerned financial, rather than cal stance, low and stable money growth, and macroeconomic, variables and took the form of robust exchange rate policies that nevertheless balance of payments runs similar to traditional allow adjustment to shocks--it is necessary to bank runs.43 The deepest of the crises involved add policies that foster a sound financial system.40 serious problems in the financial sector (Mexico, M AC RO E C O N O M I C S TA B I L I T Y: T H E M O R E T H E B E T T E R ? 109 Asia, Ecuador, and Turkey), in private sector bal- ance sheets (Asia,Argentina), or fiscal insolvency FIGURE 4.21 (Ecuador,Argentina).Where none of these prob- Incidence of Systemic Banking Crises, Developing Countries, 1981­2000 lems was present and events took the form of a r simple currency crash (as in Brazil), crisis- 22 induced economic contraction was less severe.44 yea 20 per 18 16 The Growth Payoff crisis 14 in 12 Although many developing countries achieved 10 faster growth in the 1990s than in the 1980s, this 8 achievement was only a modest one, since countries 6 growth in the 1980s was generally slow. For a of 4 majority of countries, growth rates in the 1990s 2 remained well below those of the 1960s and 0 1970s.45 Is this growth payoff commensurate Number 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 with the progress made in macroeconomic stabi- Least developed Middle-income Low-income countries (60) countries (35) countries (24) lization, or is it disappointing? It is important to keep in mind that industrial countries also grew Source: Caprio and Klingebiel 2003. much more slowly in the 1990s than in the 1960s and 1970s. But several other issues also need to policies such as adequate provision of public be taken into account. goods, as well as of social investments that might First, as already explained, the growth payoff have both increased the growth payoff and made from macro stability depends on whether stabil- stability more durable. ity is perceived as permanent. In many instances Seen in this light, some economies may well progress in stabilization was based on policy have been overstabilized.From a microeconomic changes that were not perceived as durable, or perspective, the presumed stability gains from failed to include the reform of underlying insti- further fiscal adjustments may not have justified tutions. It is these latter reforms that ultimately the costs of forgoing key social and productive determine whether policy improvements are sus- expenditures. From a macroeconomic perspec- tainable and perceived as such by the private sec- tive, the narrow focus on stability may have pre- tor. The limited progress made on this front cluded more progress toward counter-cyclical probably undermined the contribution of macro policies.The contrast between the significant fis- policy improvements--even where they might cal adjustment achieved by most developing have been sustained--to raising economic countries and the persistence of outcomes-based growth. Moreover, a vicious circle may have instability suggests that this factor may have been taken hold in some countries, in that the social important. consensus that made the policies possible, and Third, even in countries that took radical was necessary to make them sustainable, faltered steps toward macroeconomic stabilization, the in the absence of a fairly prompt growth payoff. reform agenda of the 1990s failed to address Second, the search for macro stability, nar- macroeconomic fragilities. Most notably, inap- rowly defined, may in some cases have actually propriate policies toward the domestic financial been inimical to growth. Preoccupation with sector and the capital account of the balance of reducing inflation quickly induced some coun- payments left many stabilizing economies highly tries to adopt exchange rate regimes that ulti- vulnerable to adverse shocks. Extreme macro- mately conflicted with the goal of economic volatility actually increased among outcomes-based stability. Others pursued macro developing countries during the 1990s, and the stability at the expense of growth-enhancing adverse impacts of extreme volatility on growth 110 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s appear to exceed those of normal volatility.Thus, policy, the prevention of macroeconomic fragili- the growth payoff of the macroeconomic policy ties, and complementary pro-growth policies. improvements achieved in the 1990s was limited These elements are reviewed below. not only by their weak institutional underpin- nings but also by the extreme outcomes-based Institutions for Macroeconomic Policy instability that emerged during the decade, Formulation mainly as a result of the fragilities that the reform agenda overlooked. The institutional context in which traditional Fourth, the growth payoff of macroeconomic macroeconomic policies are formulated is criti- stability may have been oversold. Macro instabil- cal to an adequate resolution of the tradeoff ity hampers investors' ability and willingness to between policy credibility and flexibility. Both respond to investment opportunities,understood credibility and flexibility are required for sus- in the broadest sense of the term, but for macro tained and sustainable stability that ultimately stability to deliver growth, those opportunities matters for economic growth. In the fiscal arena, must exist in the first place.Thus while macro- an appropriate institutional setting should ensure economic stability may facilitate growth when transparency; sustainable solvency, possibly other forces are driving the growth momentum, through the adoption of fiscal rules; flexibility; it is not enough to drive the growth process itself: and a pro-growth structure of government budg- growth depends on the policies and institutions ets.With respect to the monetary and exchange that shape opportunities and incentives to engage rate policies within the purview of the central in growth-enhancing activities.The importance bank,the most successful institutional innovation of these complementary factors may not have to emerge in the 1990s seems to be one featur- been sufficiently appreciated early in the 1990s, ing an independent central bank with a floating and gains in macroeconomic stability were often exchange rate regime and a publicly announced not accompanied by necessary growth-enhanc- inflation target.The following discussion exam- ing policies and institutional reforms in other ines these aspects of the institutional framework parts of the economy. for the formulation of traditional macroeco- In sum, there is little reason to expect a sim- nomic policies. ple, direct association between macro stability and growth. From this perspective, the limited Fiscal Policy growth payoff that emerged from the gains in Budgetary institutions and counter-cyclical fiscal poli- macroeconomic stability achieved during the cies. The critical problem of pro-cyclical fiscal 1990s may not be very surprising. policy persisted through the 1990s. The phe- nomenon arises because,in the absence of strong budgetary institutions, a "tragedy of the com- 3. Lessons mons" sets in during good times when govern- ment revenues are high: political imperatives What lessons can be drawn from the experience cause the government to spend all of its resources of the 1990s? An important lesson is that old ver- (even to borrow) in the boom,leaving little mar- ities still hold true: perceived fiscal insolvency, gin of solvency from which to finance fiscal high and unstable inflation, and severely overval- deficits when times are bad. ued real exchange rates remain reliable recipes for What is required in such situations is to make extreme instability and slow growth.But while in it politically possible for the government to run some cases slow growth and frequent crises fiscal surpluses during good times.This calls for reflected insufficient policy improvements, the the development of budgetary institutions or the evidence also highlights shortcomings in the implementation of fiscal rules that force claimants reform agenda.Three elements are critical: the on the government's resources to respect the gov- institutional framework for monetary and fiscal ernment's intertemporal budget constraint, thus M AC RO E C O N O M I C S TA B I L I T Y: T H E M O R E T H E B E T T E R ? 111 securing prudent fiscal responses to favorable bilities, including those embedded in public shocks. pension systems. Transparent fiscal rules embodied in the coun- Fiscal flexibility. The 1990s showed that fiscal try's constitution or passed into law subject to flexibility is as important as fiscal credibility, and change only by legislative supermajorities, with that to be effective, fiscal rules need to balance stipulated penalties for noncompliance, may be these two objectives.Simple rules are more trans- effective in many contexts.46 In countries where parent and hence more easily verifiable, but they government revenues depend heavily on the need to be flexible enough for fiscal policy to prices of primary commodities, institutions such react to a changing economic environment. as oil stabilization funds may need to be created Overly rigid rules are unlikely to be sustainable to save windfalls. More generally, the key objec- or credible, as shown by the recent near-demise tive is to provide scope for automatic fiscal stabi- of the European Stability Pact owing to its neg- lizers to do their job. One promising example is lect of the role of the macroeconomic cycle. Chile's Structural Surplus rule, which establishes Another lesson of the 1990s,however,is that it fiscal policy targets adjusted for the variation in is risky for governments to depart from the path growth over the cycle.An alternative proposal,yet of fiscal rectitude,even when outcomes-based sta- to be implemented, focuses on the creation of an bility would benefit from this step, because mar- independent fiscal policy council,along lines sim- kets may interpret it as a sign of fiscal lassitude. ilar to an independent central bank, to set annual The tight fiscal policies adopted by the countries deficit limits.47 Whatever institutional arrange- most heavily affected by the Asian financial crisis, ment is chosen, a basic policy step is to set fiscal immediately after the crisis and while in the grip deficit targets in cyclically adjusted terms, a prac- of severe recessions, exemplify this problem.50 If tice that could be encouraged by the international such threats to confidence were justified, the financial institutions. importance of improving fiscal institutions is Similar arguments apply to fiscal decentral- enhanced, since the role of such institutions is ization.While local provision of public goods has precisely to secure the credibility needed for gov- much to recommend it, experience has shown ernments to exercise fiscal flexibility without that fiscal decentralization is also vulnerable to a being unjustly punished by financial markets. If commons problem unless institutional remedies the threats to confidence were overstated, how- are implemented that impose hard budget con- ever, a key moral of the experience of the 1990s straints on subnational governments.One way to is that it is important not to make a fetish out of reduce the pro-cyclical bias in decentralized sys- fiscal stability as such.The need then is only to tems is to insulate resource-sharing arrangements achieve enough stability to convince the private from the effects of the cycle.48 sector that there has been a sustainable change in Another important institutional aspect of fis- regime.Once this is accomplished,the authorities cal policy is that of transparency. Uncertainty gain scope to use macroeconomic policy instru- about the state of their fiscal accounts probably ments flexibly for stabilization purposes, and strongly influenced the risk premia that devel- should exploit this to achieve outcomes-based oping-country borrowers paid in international stability. capital markets during the 1990s. Enhanced fis- Sustainable fiscal solvency and the avoidance of fis- cal transparency is an important step in reducing cal stopgaps. For a fiscal adjustment to be per- such uncertainty. There is also evidence that ceived as durable, it must be based on sustainable more transparent budgetary procedures are asso- policies, and on measures that are likely to ciated with lower deficits and debt.49 The inter- enhance growth, on both the expenditure and ests of fiscal transparency are well served by a full revenue sides of the government's budget. In accounting of the contingent liabilities of the short,the composition of fiscal adjustment matters. public sector, including those of the central With respect to sustainability, fiscal adjustments bank, and by explicit recognition of implicit lia- should be based on measures that the private sec- 112 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s tor can expect will increase the present value of Africa, andThailand. It has the important advan- future primary surpluses.Temporary fiscal stop- tages of flexibility (since the central bank is not gaps fall short of this criterion.With respect to constrained in how it attains its inflation target) as growth, some measures such as highly distor- well as of commitment (since the central bank's tionary taxes (for example, on external trade or prestige is put publicly on the line).Most impor- on domestic financial transactions), or cuts in tant, the adoption of floating exchange rates and spending on productive infrastructure or human inflation targets allows the domestic authorities capital, may raise the present value of the pri- to establish their anti-inflationary credibility by mary surplus at the expense of growth.These establishing a track record rather than by policies may even fail to raise the present value of attempting to import it through some form of future primary surpluses if their negative effects exchange rate peg.The longest running of these on economic growth have a sufficiently adverse arrangements--Chile's--was remarkably suc- impact on growth in government revenues (East- cessful in maintaining price stability throughout erly and Servén 2003). the 1990s, while avoiding severe episodes of real exchange rate volatility. More recent converts to Monetary Policy and Exchange Rate Regimes this type of nominal institutional arrangement While the evidence suggests that low and stable have also been quite successful thus far. rates of inflation are conducive to economic growth, theory suggests that what is most impor- Robustness:The Scope of the Macroeco- tant is convincing the private sector that low and nomic Reform Agenda stable inflation is here to stay. In the 1990s this proved hard to do.As does fiscal credibility, price Beyond traditional macroeconomic policies, the stability requires an appropriate institutional proliferation of crises during the 1990s has made underpinning. One lesson of the decade is that it clear that the stability agenda should encom- purely monetary arrangements are not enough to pass not just fiscal, monetary, and exchange rate ensure the credibility of monetary policy: since policies, but also policies designed to reduce not even the most rigid monetary arrangements macroeconomic--especially financial--fragility. (a currency board or de jure dollarization) pro- These include, in particular, policies directed vide a guarantee of hard government budget con- toward the domestic financial system and toward straints, fiscal credibility is necessary too. Further, the management of the country's capital account. a credible commitment to fiscal solvency is not the same thing as a credible commitment to price The Domestic Financial System stability, since fiscal solvency is in principle com- The experience of the 1990s once again under- patible with relatively high and fluctuating levels lined the importance of an appropriately regu- of seigniorage revenue.Thus there is a separate lated and supervised domestic financial system to role for monetary institutions that can credibly avoid macroeconomic vulnerability arising from preclude excessive reliance on seigniorage rev- the concentration of lending in highly risky enues. activities or the emergence of balance sheet mis- The 1990s showed that monetary credibility matches. has to be earned the hard way,through anti-infla- Although the repressed domestic financial tionary performance. In this regard, a successful sectors that prevailed in many developing coun- innovation during the last decade has been the tries during previous decades were undoubtedly institution of an independent central bank oper- inimical to economic growth, an important old ating a floating exchange rate, and with a com- lesson that was relearned in the 1990s is that nec- mitment to price stability that takes the form of essary reforms in the domestic financial sector a publicly announced inflation target. Such an are not simply synonymous with liberalization. arrangement is currently maintained by Brazil, Removing restrictions on entry,on the setting of Chile, Colombia, Korea, Mexico, Peru, South interest rates,and on the allocation of the portfo- M AC RO E C O N O M I C S TA B I L I T Y: T H E M O R E T H E B E T T E R ? 113 lios of financial institutions without simultane- The desire to avoid macroeconomic fragility ously strengthening the institutional framework makes a strong case for institutional arrangements in which the financial sector operates creates regarding the capital account that at least pre- excessive scope for moral-hazard lending.This clude the emergence of maturity mismatches in leaves financial sector balance sheets vulnerable a country's external balance sheet, since such to insolvency in response even to moderate mismatches can make the country vulnerable to macroeconomic shocks (see chapter 7). creditor runs analogous to bank runs.54 The The key lesson is that, for domestic financial question is how to preclude them.Creditors favor systems that have not already been liberalized, short maturities as a means of monitoring bor- the pace of liberalization should be modulated to rowers and controlling their behavior precisely reflect the quality of the institutional framework when asymmetric information and moral hazard governing the domestic financial sector. As has problems are serious.Under these circumstances, been widely recognized, the appropriate institu- therefore, short-maturity borrowing will be sub- tional framework has a number of ingredients: stantially less costly to borrowers than long-term clear and secure property rights, an efficient and borrowing.The problem is,of course,that volun- impartial legal system to enforce contracts, tary short-maturity loans between private parties appropriate legal protection for creditors, well- fail to take into account the social costs associ- specified accounting and disclosure standards, a ated with the risk of creditor runs. regulatory system that screens entrants while To tackle this problem, in some East Asian encouraging competition,the imposition of ade- countries, as well as Chile, the public sector has quate capital requirements and prevention of accumulated large foreign exchange reserves to excessively risky lending, and a supervisory sys- offset liquid liabilities incurred by the private sec- tem that can effectively monitor the lending tor.This approach is likely to be very expensive: practices of domestic financial institutions. holding large volumes of low-yielding, short- Improving the quality of this framework deserves term assets instead of (illiquid) long-term invest- high priority in the macroeconomic reform ments entails serious opportunity costs and even agenda. fiscal ones, because the purchase of foreign exchange reserves needs to be sterilized by the The Capital Account sale of typically higher-yielding domestic govern- With respect to the capital account, the manage- ment liabilities. Meanwhile, the incentives that ment of a country's integration into international give rise to short-term borrowing are left in place, financial markets remains a controversial part of and the costs of insuring against creditor runs are the institutional agenda. As in the case of the ultimately borne by taxpayers. domestic financial sector, enhanced integration An alternative route is to discourage the pri- with world financial markets promises many ben- vate sector from incurring short-term external efits, but when the domestic institutional struc- liabilities in the first place--by restricting short- ture is defective the costs--in the form of macro term capital inflows--or to make those liabilities risks--may outweigh those benefits. Increased effectively less liquid in times of crisis--by financial openness makes it easier for investors to restricting short-term capital outflows. Because punish countries whose macroeconomic policies both of these policies tend to raise the cost of are perceived to be off-track.51 Despite the theo- short-term loans, they effectively operate by retical arguments in favor of opening the capital internalizing the systemic costs associated with account, the international evidence is inconclu- the risk of creditor runs. sive on whether this has been conducive to Can such restrictions be designed to be min- growth.52 Moreover, the evidence suggests that, imally distortionary with respect to other types contrary to theoretical predictions, it has not of capital flows? And can they be made effective? helped to reduce macroeconomic (especially These questions have attracted considerable consumption) volatility.53 attention in recent years. As to restrictions on 114 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s inflows, the evidence is modestly reassuring. Complementarities among Pro-Growth Cross-country and country-specific studies gen- Policies erally conclude that inflow restrictions such as Much of the rest of this volume focuses on the unremunerated reserve requirements (such as the role of pro-growth policies outside the macro- Chilean encaje) tend not to affect the overall vol- economic arena.Such policies include,for exam- ume of inflows but to affect their composition, ple, the implementation of an open international reducing the share of short-term flows in the total.55 Evidence on the effects of restrictions on trade regime,the adoption of national innovation policies, well-functioning factor markets, and an outflows is much less conclusive. investor-friendly legal and regulatory environ- On balance, the available evidence suggests ment. In some cases, those policies actually facil- that restrictions on short-term capital inflows may itate the adoption of reforms aimed at have a role to play in the pursuit of outcomes- macroeconomic stability: for example, disinfla- based macroeconomic stability in developing tion or the correction of a real misalignment is countries. However, it is important to be aware easier and less costly to achieve when labor and that such restrictions entail costs to private agents, financial markets are functioning well. through their impact on the availability or price of financing.56 Policies of this type are mutually complemen- tary with policies that focus on creating and pre- In addition to maturity mismatches, external serving macroeconomic stability. An unstable borrowing aggravates the problem of currency macroeconomic environment tends to under- mismatches, to the extent that foreign lenders mine the growth benefits of such policies. Still, are less willing to accept the risk of currency what we have learned from the 1990s is that depreciation than are domestic lenders and thus macro stability alone is not enough; policies out- refuse to extend credit in the borrower's cur- side the macroeconomic arena are themselves rency.The solution here is not to restrict access indispensable to harvest the fruits of macroeco- to external borrowing. In the short run, the nomic stability in the form of sustained high rates solution is to promote the efficient distribution of economic growth. of the exchange rate risk within the domestic economy by ensuring, through regulatory means, that it is appropriately priced and there- Notes fore borne by those agents best able to bear it (typically, those holding foreign currency assets, 1. Easterly (2001) also states the view that the multiple including exporters). In the case of sovereign crises of the 1990s represent a symptom of, rather borrowing, the priority is to ensure that bor- than an "explanation" for, the slow growth of the 1990s. rowing decisions reflect the existence and 2. In recent years interest has revived, sparked by potential cost of exchange rate risk. Over the Ramey and Ramey (1995),in the adverse effects that longer term, a larger role in ameliorating the real and nominal instability can have on economic problem of currency mismatches would be growth. For a recent evaluation of the growing assumed by institutional changes that promote empirical literature on the subject., see Hnatkovska and Loayza (2004). credible nominal stability, thus mitigating 3. The level of inflation is strongly associated with its exchange rate risk.The experience of economies volatility, as well as with the volatility of relative such as South Africa that are starting to be able prices. For these reasons, and because high levels of to borrow externally in their own currencies is inflation are likely to be viewed as unsustainable, consistent with this perspective. The interna- inflation itself is commonly taken as a summary indi- tional financial institutions could help advance cator of instability. In turn, the external current account deficit is commonly viewed as a leading this process by denominating their lending in indicator of future instability, with excessively local currency, a practice that they are already large--and thus unsustainable-- deficits often pre- starting with some emerging markets. dicting a macroeconomic crisis. M AC RO E C O N O M I C S TA B I L I T Y: T H E M O R E T H E B E T T E R ? 115 4. See IDB (1995); De Ferranti et al. (2000); and East- through bankruptcies, with the corresponding erly, Islam, and Stiglitz (2001).The popular view that destruction of productive assets. instability is on the rise is documented by Rodrik 11. On the relation between macroeconomic volatility (2001b). and poverty, see Laursen and Mahajan (2004). East- 5. Here the focus is on a sample of 97 countries with erly and Fischer (2001) investigate the impact of populations greater than 500,000, for which there is inflation on the poor. complete information on real GDP growth over the 12. The availability of data on the other indicators pre- period 1960­2000.The population lower limit is set sented in the rest of this section is in general much to exclude highly volatile island economies.The total more limited than in the case of growth and inflation. sample includes 20 industrial and 77 developing For this reason,the figures below refer to the universe economies, of which three (Israel, Hong Kong of countries for which information on the variable of [China], and Singapore) are higher-income non- interest is available over the entire period shown.That OECD countries. universe varies across different variables,and therefore 6. The decline in developing-country volatility over the conclusions of the analysis have to be taken with the 1990s is documented also by Rodrik (2001b), some caution. De Ferranti et al. (2000),and Hnatkovska and Loayza 13. In part, however, this apparent improvement reflects (2004).The same result holds if volatility is measured the "sudden stop" of capital inflows to crisis-afflicted by a robust statistic such as the interquartile range emerging-market economies. instead of the standard deviation. 14. Other measures of fiscal policy stability also showed 7. The decline in volatility was statistically significant: an improvement. For example, the volatility of pub- formal tests strongly reject the hypothesis that the lic spending (as measured by the standard deviation cross-country distribution of growth volatility did of public consumption growth) declined sharply not change between the 1980s and the 1990s,as well among middle-income countries. Among lower- as the hypothesis that the changes in volatility across income economies, however, it showed little change the two decades are centered at zero. relative to previous decades. 8. The information on private consumption is available 15. In a smaller country sample (whose coverage ends in only for a slightly smaller country sample.The fact 1997),Bordo et al.(2001) also find that the frequency that consumption volatility declined less than of currency crashes declined in the 1990s compared income and output volatility in the 1990s is under- to the preceding 15 years. scored by Kose, Prasad, and Terrones (2003), and has 16. The fact that weak policies and institutions (or other been viewed as a failure of financial openness to pro- factors) can result in high default risk even at mod- vide the consumption-smoothing mechanism pre- erate levels of debt has prompted recommendations dicted by conventional theory. for extra-cautious upper bounds on debt ratios for 9. Negative extreme shocks also accounted for a larger developing economies; see Reinhart, Rogoff, and fraction of the total volatility of gross national Savastano (2003). On the other hand, the depend- income and consumption in the 1990s than in pre- ence of spreads on lenders' expectations raises the vious decades. In technical terms, the frequency dis- possibility of self-fulfilling debt crises; see, for exam- tribution of growth rates shows heavier left tails in ple, Cohen and Portes (2003). the 1990s. For both GDP and consumption growth, 17. See Calvo (1998); Calvo and Reinhart (2000); and this is confirmed by conventional skewness statistics. Mendoza (2001). However, capital flow turnarounds 10. There are good reasons why. On the one hand, do not necessarily represent exogenous shifts in with a given set of risk management mechanisms, international investors'sentiment.They reflect in part large shocks may be more difficult to absorb than the effects of developments in the destination small ones.These threshold effects of volatility have economies (resulting from, among other factors, been found to be empirically relevant for invest- changing domestic policies) as well as in interna- ment (Sarkar 2000; Servén 2003). On the other tional financial markets affecting the perceived risk hand, owing to asymmetries built into the econ- and return differentials from investing in different omy, negative shocks have qualitatively different markets. consequences than positive ones. A clear example 18. The incidence of capital flow reversals among indus- is that of buffer stocks such as bank liquidity or trial countries (not shown in figure 4.12 to avoid international reserves: large adverse shocks (or a cluttering the graph) was also fairly high in the succession of small negative ones) can exhaust 1990s, although admittedly the level of capital flows them and trigger an adjustment mechanism very was much higher among them than among develop- different from the one involved for positive distur- ing countries. bances.The same applies to firms' net worth: once 19. Indeed,one of the key dilemmas for macroeconomic it becomes negative, adjustment takes place policy making is how to assure the private sector that 116 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s future policies will abide by the requirements of sol- problems, partly driven by fiscal stringency, con- vency and low inflation,without having to surrender tributed to the failure of income distribution to the short-run stabilization capability of monetary improve in the region during the 1990s. Combined and fiscal policy. As discussed later in this section, with disappointing growth performance, some many of the achievements and disappointments of believe this outcome to have weakened popular sup- the 1990s relate to the search for lasting solutions to port in Latin America for the reform agenda of the this dilemma. past decade. 20. The same pattern is found in IMF (2003d).Among 29. See Buiter (1990, chapter 5); Easterly and Servén the 46 low- and middle-income countries in the (2003); and Blanchard and Giavazzi (2003).A recent sample underlying figure 4.13, the debt-to-GDP review of fiscal adjustment episodes (IMF 2003a) also ratio rose in 24 and fell in 22. concludes that in many cases the cuts in public 21. These debt-to-GDP ratios do not accurately reflect investment were based on overoptimistic private the debt burdens faced by low-income developing investment forecasts and turned out to be excessive. countries relative to the other groups in figure 4.13, 30. These estimates are reported inTalvi andVegh (2000) since the low-income countries tend to have a larger and Lane (2003).They are broadly consistent with share of their debt in concessional terms.The focus those displayed in figure 4.17. Public consumption, here,however,is on changes in levels of debt over time rather than the primary deficit,is used as the measure within each group of countries. because public consumption data are available for a 22. For example, the expansionary fiscal stance that much larger sample. Argentina followed during the 1995­97 boom left 31. The expansionary fiscal stance adopted by the the authorities virtually no room to adjust to the Argentine authorities during the boom of 1995­97 global real and financial slowdown after the Russian forced them to engage in a self-destructive contrac- crisis of 1998 and to the real appreciation of the peso tion in the downswing, helping precipitate the under the hard dollar peg; see Perry and Servén macroeconomic collapse of 2001­02. See, for exam- (2003). On the Russian case, see Kharas and Pinto ple, Mussa (2002) and Perry and Servén (2003). (2001). For Ecuador, see Montiel (2002). 32. Most empirical studies conclude that legal central 23. In some countries, realization of other contingent bank independence is not significantly associated liabilities, as well as recognition of hidden ones, were with lower inflation across developing countries also significant sources of debt accumulation. (Cukierman,Webb,and Neyapti 1992;Campillo and Argentina is a good example; see Mussa (2002). Miron 1997).The likely reason is that there are sub- 24. However, Bordo et al. (2001) find that the output stantial deviations between the letter of the law and cost of banking crises did not rise significantly over its application. As an exception, however, Cukier- the 1990s. man, Miller, and Neyapti (2001) find a significant 25. See Halac and Schmukler (2003) for a detailed dis- negative effect of legal central bank independence cussion. on inflation in transition economies with a suffi- 26. This is empirically confirmed by Calderón, Duncan, ciently high degree of economic liberalization. and Schmidt-Hebbel (2003).The scope for indepen- Gutiérrez (2003) suggests that constitutional sanc- dent monetary policy can also be severely limited by tion of the independence of the central bank, as well the impact of changes in monetary stance on the cost as a clear primacy of inflation among its stated objec- of public debt through the associated changes in the tives, may provide a better measure of its anti-infla- nominal exchange rate and interest rates. tionary effectiveness. 27. The bias is amply documented in both industrial and 33. Long-serving central bank governors may be sub- developing countries; see Easterly (1999). Many servient to finance ministers who place a high pre- industrial countries have engaged in similar prac- mium on the financing of fiscal deficits, and even tices, particularly in the run-up to the European independent central bank governors need not be Monetary Union; see Easterly and Servén (2003). firmly committed to price stability.Indeed,the cross- 28. Perhaps the most dramatic example of this problem country empirical association between central bank is the failure of the South African government to governor turnover and inflation performance is not address the country's alarming rate of HIV infection robust:the relation is negative only when a few high- more aggressively, an outcome that some critics have inflation observations are included in the samples;see blamed on fears of budgetary costs.This situation de Haan and Koi (2000).This might reflect reverse may not only have undermined the country's long- causality from high inflation to turnover rather than term growth through a variety of possible channels; the other way around. it has weakened support for the government's pursuit 34. Perceptions of nominal instability are not the only of macroeconomic stability as well. Similarly, Latin factor behind financial dollarization.The degree of American countries' timidity in addressing poverty real dollarization, and the perceived stability of the M AC RO E C O N O M I C S TA B I L I T Y: T H E M O R E T H E B E T T E R ? 117 real exchange rate, also matter, as do financial system 46. Perry (2003). regulations and the availability of other assets shelter- 47. SeeWyplosz (2002) for details of this proposal. ing investors from nominal instability (such as instru- 48. See Sanguinetti and Tommassi (2003) for an analyti- ments indexed to domestic inflation, as in Chile, or cal appraisal of alternative institutional arrangements. short-term interest rates, as in Brazil). For discussion, Burki, Perry, and Dillinger (1999a) review the inter- see de laTorre and Schmukler (2003); Ize and Levy- national experience with various institutional Yeyati (1998);and IMF (2002b).Thus the interpreta- arrangements in fiscally decentralized systems. tion in the text should be taken as suggestive rather 49. Stein, Talvi, and Gristani (1998); Aalt and Lassen than conclusive. (2003). 35. On the trends in dollarization, see also IMF (2002b) 50. A recent study by the IMF's Independent Evaluation and Reinhart, Rogoff, and Savastano (2003). Unit (IMF 2003b) suggests that the problem is more 36. For example, the upward drift in interest rates likely widespread. The study finds, in particular, that in reflects also the liberalization of financial systems in "capital account crisis" cases what appear in retro- many developing countries over the 1990s. spect to have been cyclically appropriate fiscal 37. Schmukler and Servén (2002). expansions were not undertaken in part out of fear 38. Frankel et al. (2001). of adverse effects on market confidence. 39. A flight out of intermediate regimes was docu- 51. Countries' misguided attempts to ride the wave of mented by Fischer (2001), for example. But whether short-term capital have also played a major role in it in fact took place has been disputed, particularly some crisis episodes.In the words of Larry Summers, because alternative exchange regime classifications referring to the role of Mexico's Tesobonos on the tend to provide sharply conflicting verdicts on eve of the Tequila crisis:" . . . the situation was not regime trends. See Masson (2001) and Frankel and one of an innocent country somehow overwhelmed Wei (2004) for further discussion. by a flood of capital from the herd of speculators,but 40. Indeed, in the wake of the crises of the 1990s the rather a situation of countries that, for domestic pol- IMF has redefined its core competencies to include icy reasons,made very,very active efforts to dine with fiscal, monetary, exchange rate, and financial sector the devil of speculators--and ended on the menu." policies. In Leading Policy Makers Speak from Experience (World 41. The increasing incidence of banking crises is also Bank 2005b), online at http://info.worldbank.org/ documented by Bordo et al. (2001). etools/bspan/PresentationView.asp?PID=1015& 42. Kaminsky and Reinhart (1999). EID=328. 43. In accordance with this, the recent analytical litera- 52. The most comprehensive empirical study is that of ture on crises continues to stress weak fundamentals Edison et al. (2002), who fail to find robust evidence as a prerequisite for the occurrence of crises, but of a significant growth impact. Prasad et al. (2003) emphasizes the key role of ingredients such as self- argue that there may be"threshold effects":countries fulfilling expectations and multiple equilibria in trig- with sound policies and institutions are more likely gering them. See Chari and Kehoe (2003) for a to derive a growth benefit from financial integration. recent example.These views assign an increasingly 53. Kose, Prasad, andTerrones (2003). important role to financial system imperfections in 54. These runs played a key role in the East Asian crisis; full-blown balance of payments crises; see, for exam- see, for example, Rodrik and Velasco (1999). Mis- ple, Krugman (1999). matches may reflect not only an inadequate borrow- 44. The Russian crisis also turned out not to be very ing strategy but also the reluctance of investors to lend severe, but probably for exogenous reasons (that is, long term in the face of a macrofinancial framework the sharp recovery in world oil prices). More gener- they deem suspect. ally, there is evidence that twin crises are usually 55. The reason is that a uniform reserve requirement is much more damaging to output than are standard more onerous for short-term transactions than for banking-only or currency-only crises; see Bordo et the rest. Montiel and Reinhart (1999) review the al. (2001). cross-country evidence on the effectiveness of inflow 45. Of course, in the short run the objectives of macro restrictions. stability and growth may conflict with each other, as 56. In the Chilean case, Forbes (2004) argues that these stabilization measures often entail an output cost over costs were substantial. Johnson and Mitton (2002) the near term. But the growth disappointment refers find that in Malaysia capital controls served to pro- to the performance over the entire 1990s. tect cronyism. Country Note C Poverty and Inequality: What HaveWe Learned from the 1990s? D uring the 1990s the number explicitly on"growth first,distribution later,"as if of poor--those living on less the two could be determined independently and than $1 in consumption per sequentially. The neglect of the distributional day--in developing countries declined from 1.2 effects of growth was consistent with the percep- billion to 1.1 billion. Globally, the proportion of tion that poverty was simply too massive to be people in poverty dropped from 28 percent to 21 reduced without significantly augmenting eco- percent.The global decline masks large variations nomic resources. Further, the facts matched the in regional poverty reduction,which mirror vari- empirical regularities first found by Kuznets ations in growth performance.Whereas poverty (1955): as income rose, inequality first increased declined rapidly in East and South Asia, it rose in and then decreased.This pattern was interpreted Sub-Saharan Africa and in Eastern Europe and by many economists (albeit not Kuznets) as Central Asia. In Latin America and the reflecting forces that could not be changed by Caribbean, poverty rates fell marginally in the policies or government interventions. 1990s, returning to near their 1981 levels. In the In the 1970s Robert McNamara's "war on Middle East and North Africa, after a significant poverty" explicitly focused on the well-being of decline in the 1980s,poverty rates rose slightly in low-income groups.This was the first time in the 1990s (Chen and Ravallion 2004). Even development policy circles that improvements within regions, there are large variations in per- in the well-being of the poor were singled out as formance. For example, poverty in the 1990s a priority, separate from economic growth and declined by almost half in Tunisia, increased in separate from improvements in the welfare of Argentina, and declined in Brazil. the population at large.This subtle shift began a How has thinking about poverty evolved in debate: Is poverty reduction the goal or is it a light of experiences, academic research, and consequence of economic growth? Can the liv- country performance in the 1990s? ing standards of the poor improve, and poverty Up to the 1970s, raising income levels by decline, independently of progress on the accelerating growth was seen as the central goal broader development front? Does distribution of development policies and the most effective of resources to the poor retard growth, or accel- way to reduce large-scale poverty.While inequal- erate it instead? These questions stimulated ity was recognized as the important issue in some renewed analysis of data on income poverty and contexts, such as in Latin America, policies inequality.2 More and better data confirmed focused on growth, with allowances made for Kuznets' intuition that the determinants of "basic needs": access to water, health, housing, inequality were more complex than reflected in sanitation, and transport. In India, the centrality Kuznets' law.World Bank research in the 1970s of growth for reducing poverty had been clear to and early 1980s, into policies and government planners since at least the 1950s and was an interventions that could change income distri- explicit objective of successive development bution, focused on cases such as the Republic of plans.1 In Brazil in the 1960s and early 1970s,the Korea's experience of "growth with equity," years of the "economic miracle," policies focused contrasted with--among others--Brazil's and 120 POVERTY AND INEQUALITY: WHAT HAVE WE LEARNED FROM THE 1990S? 121 Mexico's rapid growth but more concentrated These are discussed next. income. In the 1990s the definition and measurement Sustained Growth Is Central to Poverty of poverty, and analysis of individual country Reduction poverty reduction experiences, received consid- erable academic, empirical, and policy attention. The 1990s reconfirmed earlier beliefs and the Attempts were also made to capture the phe- messages of the World Development Report of 1990 nomenology of poverty through a variety of about the centrality of economic growth for gen- concepts to which social scientists--economists erating employment and other income-earning and noneconomists--contributed: social capital, opportunities for the poor. Unlike in industrial- pro-poor growth, empowerment, and voiceless- ized countries, where poverty is often a result of ness. Studies carried out in countries such as individuals' inability to seize opportunities Brazil, China, and India further improved and because of various social pathologies, poverty in refined poverty concepts and analysis, as did developing countries is fundamentally a matter of poverty assessments by theWorld Bank and other lack of opportunities. Countries that have sus- agencies.3 tained rapid growth for long periods have gener- These advances have recently led to the con- ated opportunities and achieved rapid poverty cept of pro-poor growth,which has been defined reduction. And countries with rapid poverty in various ways--strictly in some cases (for reduction are those that have sustained rapid example, as growth that not only raises the growth over long periods (notably China, India, incomes of the poor, but does so at a rate faster and Vietnam). Conversely, countries with large than per capita gross domestic product [GDP] decreases in income have typically seen increases growth, thus implying simultaneous improve- in poverty (as in former socialist economies, and ments in income distribution) and less strictly in parts of Africa), and countries with increases in others (for example, as growth that increases the poverty have tended to have low or negative incomes of the poor regardless of whether the growth (as in Latin America). distribution of income improves or worsens). Appreciating the difference between growth While its policy implications are not yet well episodes and sustained growth is of central importance defined, the concept of pro-poor growth has for understanding the relationship between growth elicited strong interest from development aid and poverty. Growth episodes are quite frequent, agencies and from political leaders in developing and during growth episodes lasting two, three, or countries facing pressures stemming from more years,poverty outcomes can vary widely.Sus- democratization. taining growth for several decades is a much rarer Overall, four lessons have emerged from the accomplishment (see chapter 1 and Country Note experiences of the 1990s and from conceptual B), and yet it is this accomplishment that has and empirical work over the decade: enabled developing economies such as Korea, Sin- gapore,orTaiwan,China,to reach the income lev- · Sustained growth is vital for poverty reduc- els of industrialized countries. The poverty tion. outcomes of sustained growth are much less varied · Poverty is multidimensional. than those resulting from growth episodes.Growth plays a much larger role in poverty reduction dur- · Access to social and infrastructure services is ing long growth spells than it does during short key to the poor:it improves both their oppor- spells,where changes in measured distribution play tunities and their welfare. a larger role (Kraay 2004). · Consistent with the broadening of the notion Translating the goal of pro-poor growth into of poverty, there is now more focus on the policies and practical growth strategies remains a deeper issue of equity--its meaning, and its tremendous challenge. Short growth episodes possible consequences for growth. can be accompanied by a wide range of distribu- 122 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s tional outcomes (Ravallion 2003b), and since while distributive programs are an important they have small and often transitory effects on component of growth strategies, to ensure the poverty reduction, they provide only limited sharing of the benefits of growth, they are no insights into the impact of sustained growth on substitute for growth for the purpose of rapid poverty or into the types of policies that are and significant reduction in large-scale poverty. appropriate. In recent years, development agencies' focus One reason for the interest in concepts of on the poor, when analyzing poverty or formu- pro-poor growth has been the realization that lating programs for poverty alleviation, has led some patterns of growth (such as expansion of them to ignore income gains above the poverty labor-intensive agriculture) could have a larger line.Yet such gains cannot be ignored socially, impact on the poor than do others (such as sub- empirically, theoretically, or politically. Setting sidies for capital-intensive industrialization), and poverty reduction as the only goal of economic even that some groups could lose in situations in growth and economic development ignores the which poverty was declining in aggregate (Kan- value of income gains above the poverty line, for bur 2001). example, those to the middle class. Since a A second reason is that the palpable impact of poverty line is a social convention that a society targeted programs on poverty has sometimes led adopts reflecting its own conditions, we should to the impression that they could substitute for expect poverty lines to rise with a country's per economic growth.The 1990s saw some success- capita income.4 It is not reasonable forVietnam, ful targeted programs. Mexico's Progresa/Opor- Brazil, and the United States to have the same tunidades, for example, is a large-scale, poverty line,nor is it reasonable--current growth well-targeted program that makes transfers con- rates persisting--for Vietnam in 2010 to retain ditional on investments in human capital; it the poverty line it had in 1980.Thus the impact deservedly receives a great deal of attention, of growth on poverty reduction should be eval- partly because of its well-designed, independent uated using both a low and a high poverty line, impact evaluation system. India's food distribu- often using a measure of poverty that is sensitive tion system has helped to increase consumption to the distribution of income below the poverty and alleviate the consequences of poverty in line as well. some areas of the country, but as an instrument This discussion raises the question of what to reduce national poverty it has generally been should be the upper threshold for poverty. Com- ineffectual. China's 8-7 program has succeeded monly used poverty lines of $1/day at purchasing in addressing pockets of poverty within a context power parity (PPP), or even PPP$2/day, can only of rapid growth, expanding opportunities, and be justified if human well-being is high at such rapidly falling nationwide poverty. Elsewhere, levels of income. Even at the poverty line of some programs have succeeded in alleviating the $2/day,however,physical indicators of well-being impacts of crisis (Indonesia), and in improving such as education, nutrition, and mortality all community-level services, such as Bangladesh's indicate severe deprivation. Figure C.1 shows the stipend programs to increase girls' enrollment in mortality rate of the richest fifth of the popula- secondary school. Such programs should be seen tion in several low- and middle-income coun- as vehicles either for the inclusion of marginal- tries. It shows that the mortality rates of children ized groups or for improvements in the access of in the richest fifth of families in those countries are the poor to infrastructure and social services. much higher than the mortality rates of the poor- While they play a very important role, they can- est fifth in rich industrial countries: among the not substitute for growth in achieving rapid and richest fifth of households in countries such as significant reductions in the incidence of poverty. Nepal and Nicaragua, the infant mortality rate is There is no known example of rapid, sustained, 40 per 1,000,but even among the poorest fifth in and significant reductions in poverty in the Organisation for Economic Co-operation and absence of sustained economic growth.That is, Development (OECD) countries, the rate is POVERTY AND INEQUALITY: WHAT HAVE WE LEARNED FROM THE 1990S? 123 measures of poverty according to this definition, FIGURE C.1 because some of its dimensions are not amenable Infant Mortality Rates to measurement. More research is being done, 140 but there are thus far no well-established empir- 120 ical relationships regarding broadly defined births 100 80 poverty. 60 That said, the broadening of our understand- 1,000 40 20 ing of poverty has led to a search for more Per 0 encompassing indicators, especially in relation to 1996 1991 ­93 1996 1993 1993 ­88 ­88 1991 the following. 1992 1987 1987 Brazil, Education and health indicators. The key role of (Cebu), Nepal, Africa,ire, Nicaragua, Ghana, Pakistan, Vietnam, education and health in well-being and develop- Southted'Ivo Cô ment has been emphasized since economics Philippines began.The shift in the 1990s toward a greater Source: Wagstaff 2001. recognition of the importance of education and health as development objectives has also led governments to be more realistic and pragmatic lower than 10 per 1,000.This suggests that the in their pursuit of those objectives.These mes- poverty lines currently in use underestimate the sages are summarized in WDR 2004 on service amount of poverty in the world and, possibly, the delivery. declines in poverty consistent with achieving the Risk and vulnerability. A key fact that emerged Millennium Development Goals (Pritchett in the 1990s--both from the expanded availability 2003b). of data that tracked the same households over time (panel data) and from qualitative work such as Voices of the Poor--was the role of risk and vulner- Poverty is Multidimensional ability in explaining poverty in developing coun- A key change from World Development Report tries. In industrialized countries, little of the 1990 to World Development Report 2001 was the poverty is transient because social safety nets and conceptual expansion of the definition of higher incomes, and hence higher savings, help to poverty to include indicators of education and smooth consumption. Poverty in rich countries is health, risk and vulnerability, and voicelessness best understood as an issue of social marginaliza- and powerlessness. Studies carried out in the tion resulting from culture and nurture,restraining 1990s, such as participatory poverty assessments individuals'ability to seize opportunities.In devel- and the Voices of the Poor (Narayan et al. 2000), oping countries, by contrast, most poverty is tran- endeavored to capture what poverty means for sient poverty, responding to the flow of those who experience it, and rendered it clear opportunities as individuals fall into or out of that consumption expenditure--the basis for poverty. Empirical research on the dynamics of most poverty lines--captures only one aspect of poverty, pioneered by Jalan and Ravallion (1998) being poor. From this body of work emerged a in China and followed up by many others (for broader concept of poverty that includes (1) the example, Baulch and Hoddinott 2000) has found usual measures of consumption poverty, (2) edu- a high incidence of transitions in and out of meas- cation and health, (3) risk and vulnerability, and ured poverty.5 In the most extreme case of the (4) voicelessness and powerlessness. World Devel- volatility of poverty episodes, in Pakistan more opment Report 2001 emphasized the multidimen- than half of households (55 percent) were poor sional nature of the phenomenon of poverty, either in 1986 or five years later in 1991--but only which has been increasingly emphasized since. 3 percent were poor in both periods (table C.1). While broadening the definition of poverty is "Poverty" as an empirical phenomenon a conceptual step forward, there are no empirical hence consists of many individuals and house- 124 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s holds experiencing an episode of poverty and, Work on theory and empirical research that especially in some countries, surprisingly few could shape the design of relevant programs and "poor people"who are always below the poverty practices is just beginning.There are reasons to be line.This means that many more households are cautious because, while some empirical studies vulnerable to an episode of poverty than are actu- show that programs function better with greater ally poor at any point in time.These insights have local engagement (Isham, Kelly, and Ramaswamy encouraged attempts to address risks and vulner- 2002;Galasso and Ravallion 2000),it is far from a abilities as part of an overall social protection foregone conclusion that more "local" decision strategy, in preference to an exclusive focus on making necessarily increases "voice" in a useful chronic poverty or on the poorest of the poor. way (see, for example, Platteau 2004; Bardhan Voicelessness and powerlessness.Among the shifts 2002). Much more needs to be known about the in development thinking that have affected atti- design of local decision making that effectively tudes toward poverty reduction,perhaps the most increases voice. important has been the move away from the view that the solution to all ills is a well-designed pro- Access to Services by the Poor: gram implemented by a well-ordered civil ser- Learning from Experience vice bureaucracy and centrally funded from public resources (Pritchett and Woolcock 2004). Economists tend to view poverty in terms of the That view tends to make "the poor" a distinct consumption of private goods,but other social sci- category of people who are the passive recipients entists emphasize the importance of public goods of development rather than its active agents. as well. This is because public goods can both Now, centered around themes such as "commu- improve income-earning opportunities (as in the nity based,""community driven,""bottom up," case of access to roads,markets,water,public trans- and "participatory" development;"local gover- port, education, health, rural employment pro- nance"; "empowerment"; "social capital"; and grams) and directly improve welfare (through "inclusion, cohesion, and accountability," there is access to the same)--thus blurring the distinction recognition that key elements of local develop- between growth-enhancing public policies and ment practice in poverty reduction cannot be distributive policies that are implemented through delivered ready made, but must be homegrown public services. or locally grown (Narayan et al. 2000). A broader understanding of poverty thus highlights the importance of access by the poor TABLE C.1 to goods and services that are typically publicly Surveys Tracking Individuals over Time Show Only a provided. Shortfalls in the delivery of such pub- Small Portion of Poverty Is Accounted for by People licly provided goods can lead the poor to per- Who Are Always Poor ceive that poverty has increased, even though survey data based on the consumption of private Percentage of the population that is... goods may suggest the opposite (Kanbur 2002). (by headcount ACE poverty) Delivery of social and infrastructure services Always Sometimes Never was the central theme of WDR 2004, which poor poor poor highlighted that the effective provision of services Zimbabwe 1992­96 10.6 59.6 29.8 for the poor cannot be delinked from addressing Pakistan 1986­91 3.0 55.3 41.7 the root problems of voicelessness and powerless- South Africa 1993­98 22.7 31.5 45.8 ness.As that report showed,while the provision of Russian Federation 1992­93 12.6 30.2 57.2 public services to the poor and nonpoor alike is a Ethiopia 1994­97 24.8 30.1 45.1 key developmental issue in any developing econ- Côte D'Ivoire 1987­88 25 22 53 omy, it is not a straightforward matter, nor is it a Source: Adapted from Baulch and Hoddinott 2000. matter of resources alone, but one of striking the ACE, Absolute consumption expenditure. appropriate incentives for accountability and per- POVERTY AND INEQUALITY: WHAT HAVE WE LEARNED FROM THE 1990S? 125 formance, taking into account the political econ- (4) Get the support for innovation right--and back the omy of delivering services. right leadership and management. "Scaling up" is Though the 1990s started with the convic- not simply a matter of more resources but tion that retrenching governments and expand- rather of support for innovations that lead to ing the private sector would benefit virtually all better ways of doing things,that are tailored to areas of the economy, it is now seen that these circumstances and can be scaled up once they expectations were unrealistic, and that govern- have been shown to be successful. This ment has a key role to play in at least three areas: requires a rigorous means of separating suc- education,health,and infrastructure.Further,it is cesses from failures,to scale up the former and increasingly seen that how to organize these ser- shut down the latter. vices so that they are effective and reach the poor is the key question. It Is Important to Clarify the Causes and Answering this question will require a better Consequences of Inequality understanding of the reasons behind failures and successes.The developing world is littered with Rising inequality in some industrialized and dysfunctional services--schools with no books developing countries has brought inequality to the or no teachers, or teachers who are absent; clin- fore of policy discussions. Whether inequality ics with no drugs,without functional equipment, should be a matter of public policy, and how to with staff that does not attend or provides low- address it, depends on the underlying economic, quality care; and water services that operate only social, and political forces causing inequality and sporadically. At the same time there are services their relationship to social equity. This is not that do work well. straightforward,however.Not only is the relation- As WDR 2004 emphasized,the effective pro- ship between inequality and other key economic vision of services depends on relationships of variables such as growth, or poverty reduction, accountability from the providers of services to nonlinear, but its sign probably changes over time citizens, politicians, and policy makers. Related (Timmer andTimmer 2004). to this, the Shanghai Conference of May 2004 Some evidence is emerging, for example, thus on Scaling Up Poverty Reduction, in its attempt far at the microeconomic level more than in the to clarify the reasons for success, produced four aggregate,that inequality can negatively affect the key messages:6 functioning of institutions, the efficiency of resource allocation, and collective decision-mak- (1) Get the economics right--and get the politics ing processes, leading to a negative impact on behind the economics right. Poverty reduc- growth (Acemoglu, Johnson, and Robinson tion depends on growth, which in turn 2001;Aghion 1998). needs sustained political support for imple- The interpretation of narrow indicators of mentation of growth-promoting policies income inequality is not straightforward, how- and institutions. ever.Income inequality is not a good indicator of opportunities for the poor, and its use as a focus (2) Get the focus on clients right--and keep the focus of analysis may miss--just as narrow definitions on clients. One message relevant across differ- of poverty did--key issues of the reality and ent types of activities was that without a focus social perception of fairness, opportunity, legiti- on clients, even the best-intentioned efforts macy,and equity (box C.1).WDR 2006 will take can go astray and break down in the field. up the issues of equity and development and (3) Get the implementation right--the devil is in explore them in depth. the details. There is no single recipe for a This section outlines six reasons why inequal- successful education project or health proj- ity has become such a visible issue and the ect: implementation must be adapted to the importance of conceptual clarity in interpreting circumstances. its significance. 126 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s First, growth-oriented reforms have winners jobs were legitimate (box C.1). Just as important and losers; and there is an inclination to believe as the inequality in outcomes is the perception of that the losers are the poor, that their losses con- the fairness of the process that generates the out- stitute a severe loss of welfare and an obstacle to comes.This implies that examining increases in reforms, and that if the poor could be compen- income inequality alone, without considering sated for their losses, political support for reform the equity of the process behind the increase,will would be greater.Therefore, the reforms of the not capture the true meaning of a rise in inequal- 1990s inevitably raised the issue of the distribu- ity and its policy consequences. tion of gains and costs of reforms, and how they Third, it is often asserted that inequality is affect income distribution.Over the 1990s,how- increasingly on the international agenda because ever, as it became clear that the poor are vulner- "global" inequalities are rising. Changes in able continuously, not only during episodes of inequality are not a unique feature of the 1990s; reform, the concern of policy makers began to Bourguignon and Morrison (2002) suggest that evolve away from episodic forms of compensa- inequality rose steadily from 1880 to 1980 and tion toward providing more permanent forms of stopped rising in the 1980s and 1990s (see figure social protection. Meanwhile, studies have also 1.4 in chapter 1).Yet the issue is complex empir- suggested that it may be naïve to believe that mit- ically, because much depends on what estimates igating the impacts on the poor will reduce the of inequality are used and how the data are opposition to reform. Research by Graham and weighted: whether by country, to estimate the Pettinato (2002) indicates, for instance, that peo- cross-national distribution of income;or by pop- ple who have benefited from reforms are as likely ulation, to estimate the global personal distribu- to report themselves opposed to reform as peo- tion of income; or by using national accounts or ple whose incomes have stagnated or fallen.The household data, to estimate the growth of formation of attitudes toward reforms is com- income. This has engendered debates on the plex, and clearly needs further research. extent of cross-national inequality and the reduc- Second, in some transition countries such as tion in global poverty,between Bhalla (2002) and China or the former Soviet Union, inequality Sala-i-Martin (2003) on the one hand and has increased significantly. A view has emerged Milanovic (2004) on the other; see also that inequality is on the agenda because the exchanges in The Economist (2004). A recent resentment generated by higher inequality is a review by Bourguignon (2004a) shows that even powerful political factor and creates social pres- using the same data, different measures of sures that slow reforms and create uncertainty inequality can show falling or rising inequality. about policy continuity. Resentment about ris- Fourth, measures of inequality fail to capture ing inequality needs to be disentangled,however, the significance of underlying economic and from the perception that the process by which for- poverty conditions.For instance,Ethiopia is a rel- tunes were accumulated was unfair. Perceptions atively unequal country in which the income of of process and procedural fairness are important the richest 10 percent of the population is 10 influences on an individual's attitudes toward times that of the poorest 10 percent. Should outcomes, and there is no simple association inequality be the primary policy issue in between inequality of outcomes and perceptions Ethiopia, or should the country focus on gener- of the fairness of those outcomes--perceptions ating higher average income? Currently even that go to much deeper issues of equity. In Sri "the rich" in Ethiopia (except for perhaps the Lanka,for example,there was widespread resent- very few) have incomes and indicators of physi- ment at the politicization of civil service employ- cal well-being well below those of "the poor" in ment, because the process for allocating civil industrial countries, and per capita income in service jobs was perceived as unfair.The key issue Ethiopia is 50 times lower than that in the U.S. was less whether civil servants were overpaid Fifth, while rising inequality is sometimes than whether the opportunities to access those equated with rising poverty, the poverty impact POVERTY AND INEQUALITY: WHAT HAVE WE LEARNED FROM THE 1990S? 127 BOX C.1 Perceptions of Fairness in Allocating Opportunity Are Central (Case Study: Sri Lanka) S ri Lanka's Presidential Commission on Youth deemed by youth as incompatible with the basic was appointed in late 1989 to examine the norms of fairness, equity, and merit. This was based causes of youth discontent and unrest that led on the representations made by civil society before to the Marxist rebellion of 1987­89. The Commission's the Commission suggesting an alarming degree of report highlights that meritocratic processes are crit- public dissatisfaction in this sphere of recruitment to ical for maintaining social and political stability. The the public service. Hence, the Commission stated that Commission reported a strong consensus within the the 1972 Constitution, which removed the power of country that politicization and abuse of power and appointment vested in the Public Service Commission injustice were the main causes of youth unrest in Sri and brought such appointments within the purview of Lanka. The Commission felt that the politicization of the Cabinet of Ministers, triggered a politicization employment, through the "chit system" (the practice process that had far-reaching social repercussions in of receiving a letter from one's Member of Parliament a country where the public sector accounts for the in order to find public sector employment), was bulk of formal employment. Source: Report of the Presidential Commission on Youth, 1990 (as cited in World Bank 2000h). of changes in inequality is not self-evident but From this it could be deduced that poverty has depends on the causal interpretation of events. fallen "in spite of" increasing inequality. How- Policies that improve both the growth and the ever,understanding the impact of China's policies distribution of income are "win-win," but even makes it important to differentiate"constructive" policies that worsen the distribution of income inequality (which provides the incentives needed can be "win" policies for improvement in to move resources to their most efficient use) poverty if their effects on growth are large from "destructive" inequality (which generates enough. envy and socially unproductive distribution) As detailed in Bourguignon (2004b) there is a (Timmer and Timmer 2004). Thus it may be poverty-growth-inequality triangle, and income more causally correct to say that poverty in China poverty, income distribution, and changes in went down because inequality was allowed to go mean income are linked by an "arithmetic iden- up.This would be the case if the policy changes tity"--which, however, alone does not reveal the that led to a rapid increase in average incomes underlying causal forces, including policies, that needed to allow inequality to rise in order to pro- drive both growth and its distribution.For exam- vide incentives for investment and innovation. ple,in China,as an arithmetic matter,the worsen- Discussion at an international meeting in 1987 ing inequality in the 1990s partially offset the rate with Chinese, Korean, and Indian economic pol- of growth of mean income, so that the growth- icy makers illustrates this point: elasticity of poverty declined. Poverty did not worsen, but the rate of poverty reduction was After the other delegations [from India, slower than it would have been if inequality had Korea] presented their experiences in remained constant.With an equal society and no managing a market economy,the Chinese growth there was very little poverty reduction. Vice Minister presented an outline of the With rapid growth there has been a massive Chinese reform program. At the end of reduction in poverty and an increase in inequality. this presentation,Manmohan Singh,in his 128 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s usual gentle but forceful tone, asked: opportunities of the poor to invest and hence can "Would not what you are trying to do raise both equity and growth. result in greater inequality in China?"To This is an enormous shift from the view, pre- that the Vice Minister replied, with firm vailing until the late 1970s, that concentrating conviction: "We would certainly hope income among the rich increases savings so!" (because the rich have a lower propensity to --Edwin Lim,World Bank (2005a) spend), and hence capital accumulation and hence growth, so that more inequality is good Vietnam has seen the same combination of for growth.That view has an element of truth rising inequality and spectacularly declining but it ignores some key issues that are now back poverty as China.Vietnam lifted a third of its on the theoretical and empirical agenda (see, for population out of poverty between 1993 and example, Aghion 1998; Banerjee and Duflo 2002--achieving what may be the greatest 2003): that inequality can lead to underinvest- reduction in poverty in a single decade in the ment by the poor, and that unequal income dis- history of mankind.Vietnam confirms the con- tributions tend to encourage the persistence of ventional wisdom reflected in World Development poor policies and the protection of rents. Report 1990:poverty reduction was rapid because Bourguignon (2004b) reviews theoretical it began when poverty was high, incomes were ways in which more equal income distribution equal, and growth was intensive in unskilled could raise growth, but finds that the "available labor. Inequality in consumption expenditures aggregate evidence is inconclusive."The scenar- rose modestly over this period: the Gini coeffi- ios and policies in which improved equity and cient was .34 in 1993, and .37 in 2002. growth are complementary are increasingly on Thus, while it is analytically appropriate to the table.7 say that "if the growth of consumption expendi- tures had not changed and if inequality had not increased, poverty reduction would have been Notes more rapid," this arithmetic fact is not dictated by causality in the sense that a feasible (econom- 1. http://planningcommission.nic.in/plans/planrel/ fiveyr/welcome.html; Bhagwati and Srinivasan ically, administratively, and politically) policy that (2002). would have accomplished that goal may simply 2. The early focus was on inequality because few coun- not exist. It is perhaps unlikely that the fastest tries (with the prominent exception of India) had well- reductions in poverty in history, such as those in defined poverty lines, or systematic measurements of Vietnam in the 1990s (or in Indonesia and China poverty.By many standard measures of poverty,growth, inequality, and poverty are arithmetically linked, but in the 1970s, 1980s, and 1990s) could have been the research was not explicitly about absolute measures even faster. of poverty. Sixth, there is some emerging, though as yet 3. The World Development Reports of 1990 and of 2000­01 inconclusive,evidence that more equal societies are benchmark the evolution of thinking about poverty able to be more efficient: they adopt better poli- over the decade. More recent sources of insights are cies,provide public goods more efficiently,allocate WDR 2004 on Making ServicesWork for the Poor and the Global Learning Process on Scaling up Poverty Reduc- capital more efficiently,and reduce crime and inse- tion,which culminated in the Shanghai Conference of curity.There is evidence from the United States, May 2004.Both examine modes of improving services for example, that appropriately designed social to low-income groups that were developed mostly dur- safety nets can improve both distribution and effi- ing the 1990s. ciency (Blank 2002).The new view (discussed in 4. Any poverty line is a social convention.There is no right or wrong choice, and no technocratic stan- chapter 1) is that by reducing poverty countries dard,for establishing a poverty line,but only ways to can improve the functioning of markets and that of make comparisons consistently across households, institutions in a manner that is both "pro-poor" regions, and over time for a given poverty line. For and"pro-growth"--social policies can increase the international comparisons the World Bank has con- POVERTY AND INEQUALITY: WHAT HAVE WE LEARNED FROM THE 1990S? 129 ventionally used a "one dollar a day" poverty line, strated that measurement error in income or con- derived from the official poverty lines prevailing in sumption expenditures can account for a substantial some of the poorest countries in the early 1990s, and fraction of these measured transitions. "two dollars a day" as another standard for the poverty 6. The global learning process that led up to this con- line. Each country usually uses its own poverty line, ference emphasized hearing directly from practition- and in some countries the poverty line is defined in ers from the South. It produced more than 100 case relation to average incomes and hence automatically studies of projects in all regions of the world in areas evolves with changes in average incomes.On the tech- from education to health to community develop- nical complexities of setting poverty lines, see Raval- ment to targeted programs, and sponsored field visits lion (1994) and Pradhan et al. (2001). across projects to promote South-South learning. 5. Measured is a very important caveat. Suryahadi et al. 7. These will be addressed in World Development Report (1999) and Luttmer (2001) and others have demon- 2006 on equity. Chapter 5 Trade Liberalization: Why So Much Controversy? E CONOMISTS HAVE LONG RECOG- Drawing on the experience and academic nized the gains from interna- research of the 1990s, this chapter identifies five tional trade; the study of these lessons: gains is where modern economics began. Over centuries, international trade has brought · Openness to trade has been a central element together remote parts of the world and different of successful growth strategies.In all countries civilizations, helped disseminate knowledge and that have sustained growth the share of trade ideas, and shaped the course of regions and in gross domestic product (GDP) has nations. Rapid reductions in transport and com- increased, and trade barriers have been munications costs accelerated this trend in the reduced. 19th century, and international trade reached · Trade is an opportunity, not a guarantee. unprecedented levels at the beginning of the While trade reforms can help accelerate inte- 20th century.Trade declined, however, following gration in the world economy and strengthen the two World Wars, the 1929 crisis, and the an effective growth strategy, they cannot worldwide increase in protectionism. ensure its success. Other elements that address A reversal in protectionism started afterWorld binding constraints to growth are needed, War II among the industrialized countries, and possibly including sound macroeconomic spread to the developing countries in the 1970s. management, trade-related infrastructure and Trade reforms were further expanded and con- institutions, and economywide investments in solidated in the 1980s and 1990s across the devel- human capital and infrastructure. oping world: in South Asia, East Asia, Latin America, Eastern Europe, and, to a lesser extent, · There are many possible ways to open an in Africa and the Middle East.Yet in the 1990s, economy.The challenge for policy makers is the results of trade reform have varied and some- to identify which best suits their country's times fallen short of expectations. Critics of the political economy, institutional constraints, economic and social effects of globalization have and initial conditions. As these vary from also become more vocal.Why have some trade country to country, it is not surprising that liberalizations been reversed, and why have oth- there is a striking heterogeneity in country ers brought prosperity, opportunities, and eco- experiences regarding the timing and pace of nomic diversification? Is there still a role for the reforms. Different countries have opened up protection of infant industries in growth strate- different sectors at different speeds (for exam- gies? Does trade liberalization lead to economic ple, Bangladesh and India); others have growth? Finally,does trade liberalization improve achieved partial liberalization through the or reduce poverty? establishment of export processing zones (for 131 132 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s example,China and Mauritius);and yet others trade--measured as a share of exports in GDP-- have combined unilateral trade reforms with is now larger in developing than in developed participation in regional trade agreements (for countries.Another important trend is the shift in example, Estonia). the composition of developing-country exports toward manufactures. Countries whose incomes · The distributive effects of trade liberalization were low in 1980 have managed to raise their are diverse, and not always pro-poor. Trade exports of manufactures from about 20 percent reforms were expected to increase the of their total exports to more than 80 percent.1 incomes of the unskilled in countries with a Virtually all successful economies have comparative advantage in producing increased their openness to trade.In part because unskilled-intensive goods.Yet evidence from successful trade reforms have been introduced in the 1990s suggests that even in instances conjunction with other policy initiatives,it is dif- where trade policy has reduced poverty, there ficult empirically to identify the growth effect of are still distributive issues.One important pol- trade policy alone, compared with the growth icy lesson is that countries need to help work- effect of other policy initiatives, and to disentan- ers affected move out of contracting gle whether trade causes growth or growth (import-competing) sectors into expanding causes trade.As an economy accumulates physi- (exporting) sectors.This is an issue relevant to cal and human capital, shifts its comparative both developing and industrialized countries. advantage toward more capital-intensive activi- · The preservation and expansion of the world ties, and becomes internationally competitive in trade system hinges on its ability to strike a a wider range of goods and services, it will better balance between the interests of indus- inevitably trade more. But is higher trade the trialized and developing countries. Global result or the cause of its growth? Most likely both markets are the most hostile to the products processes are at work.This section reviews the produced by the world's poor--such as agri- evidence on these questions and then argues for cultural products and textiles and apparel.The the need to pursue trade reform as part of a com- problems of escalating tariffs, tariff peaks, and prehensive growth strategy. Openness to the quota arrangements systematically deny the global economy has helped efficiency and poor market access and skew the incentives growth in many cases (East and South Asian against adding value in poor countries.These countries, Botswana, Chile, Mauritius,Tunisia), problems can be addressed through collective but it has failed to do so in many others.These action, best pursued through the Doha experiences do not necessarily imply that less Round and the World Trade Organization. trade reform would have been desirable, but that Although there is a role for nonreciprocal trade reform must be done and sequenced sensi- preferences and for reciprocal regional bly, as part of an effective growth strategy. approaches, this comes at a cost to excluded countries, is arbitrary and political, and thus is The 1990s:An Overview not first best in terms of generating the right incentives for investment. Reforms in the 1980s and 1990s were the origin of a strong expansion in international trade (box 5.1).As detailed in chapter 3, developing coun- 1. Trade Reform as a tries are now more integrated with the world Component of a Successful economy than are high-income countries. Growth Strategy The integration of labor emerged as another important issue on the globalization agenda dur- This chapter begins by reviewing key changes in ing the 1990s. In 2001, developing countries trade policy, trade volumes, and the composition received some US$71 billion in migrants' remit- of trade in the 1990s. One striking fact is that tances--a sum that was nearly 40 percent more T R A D E L I B E R A L I Z AT I O N : W H Y S O M U C H C O N T ROV E R S Y ? 133 BOX 5.1 Trade Policy over the Centuries P rotection of domestic industries has a long the initial phases of industrialization. Third, growth history. In the 12th century, for example, to strategies based on import substitution proved diffi- maintain the competitive edge of their textile cult to implement in practice, and the practical and industries, Flanders and England restricted the move- political aspects of implementation often negated ment of experienced weavers. In the 13th century, most of the expected gains (Balassa 1971; Little, Sci- England enacted laws restricting the types and origin tovsky, and Scott 1970). High nominal tariffs often of fabrics certain individuals could wear. In 16th and provided negative protection to emerging activities 17th century France, the state promoted selected and protection to activities with negative value industries, through import protection, direct owner- added, and contributed to misallocation and under- ship, or subsidies, as did Japan later during the Meiji utilization of capital in capital-scarce economies. period. While the protection of domestic industries Overvaluation of the exchange rate resulting from took various forms--such as subsidized capital, or import restrictions discouraged exports and penal- monopoly or monopsony rights--protection from ized agriculture--further reducing the size of the imports was the most widely used and became partic- market for import-competing industries. ularly important after the start of the industrial revo- As a result, during the 1980s and 1990s virtually lution. During the 1800s and the first half of the all developing countries followed the examples set by 1900s, tariffs on imports in industrial countries were Singapore, Hong Kong (China), Korea, and Taiwan as high as 30­50 percent (World Bank, World Develop- (China): encouraging exports and reducing levels of ment Report 1991). protection. Industrialization based on import protec- Many developing countries pursued import substi- tion was gradually discredited and, starting in the tution industrialization strategies in the three mid-1980s, most developing countries sought to decades that followed World War II, but by the mid- reduce levels of import protection and liberalize 1980s, most developing countries were seeking to trade. Chile and Sri Lanka were among the first liber- reduce their import protection and liberalize trade. alizers, having started already in the 1970s. Three developments had raised doubts about the Argentina and Uruguay followed shortly thereafter. long-run effectiveness of strategies based on import By the early 1990s, researchers and policy makers protection. First, in the 1960s, the Republic of Korea generally accepted the superiority of outward orien- and Taiwan (China) had begun adopting export-ori- tation over import substitution as a development ented growth strategies that not only yielded supe- strategy.a Trade liberalization expanded in the 1990s, rior economic performance, but also helped these two leading to increased integration of developing economies to withstand the severe interest rate and economies in world trade. The fall of communism in oil price shocks of the 1970s. Second, high tariffs, Central and Eastern Europe, together with the col- administrative restrictions, and rationing of foreign lapse of the former Soviet Union, reinforced this exchange and of import licenses created high returns view. Countries that had not already embarked on to rent seeking, reinforcing vested interests and an liberalization began to do so now, while others scaled environment that stimulated corruption and weak- up their efforts. They included hitherto very highly ened national institutions. The results, including protected and inward-looking economies such as state capture by vested interests and the misuse of India, and countries in Sub-Saharan Africa that government discretion, discredited import substitu- looked to integration with the world economy as a tion strategies even among economists who believed key instrument for reversing hitherto dismal growth in the strategic importance of import substitution in performance. (Box continues on the following page.) 134 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s advantage lies in the export of medium- and low- skilled, labor-intensive services. BOX 5.1 (continued) Trade Reform, Exports, and While some of the reforms were unilateral, others Economic Growth were accomplished in the context of multilateral trade agreements such as the Uruguay Round. Important com- For decades, researchers have been debating the ponents of those reforms included large tariff reductions merits of economic openness and its association and the elimination of quotas, as well as the relaxation with growth. Academic debates on whether of restrictions on foreign investment. Looking at the openness to trade causes higher growth are rid- improvement in market access for the developing coun- dled with problems of measurement,reverse cau- tries, tariff cuts in industrial countries accounted for sation (faster growing countries tend to open about a third of the improvement and tariff cuts in the their markets more quickly), and omitted vari- developing countries themselves accounted for two- able bias (countries that successfully lower tariffs thirds (World Bank, Global Economic Prospects 2004). also adopt other complementary policies).4 Notwithstanding difficulties in interpreting a country experiences during the1990s, almost all See Krueger (1997) and Baldwin (2003) for expositions on the evo- lution of economic thinking over this issue during the second half of economists agree that liberal trade is important the 20th century. for growth over the long run (box 5.2). Research that focuses on the relationship between trade reforms and economic growth in the 1990s also finds that trade reforms are associ- than all official development assistance and sig- ated with higher growth, although the strength nificantly more than net debt flows to develop- of the association varies across different studies.5 ing countries in that year.2 However, such Yet trade liberalization by itself is not enough for remittances went to only a few developing coun- economic growth.Studies show that trade policy tries, and their importance for developing coun- is most likely to be associated with positive out- tries as a group declined over the 1990s, from comes when it is conducted in a favorable eco- slightly above 4 percent of all foreign exchange receipts to slightly below.3 Remittances would provide a much larger FIGURE 5.1 share of foreign exchange receipts for developing Temporary Labor Mobility, Underused countries were it not for industrial-country Mode of Trade in Services restrictions on labor migration. If rich countries (value of world trade in services by mode; percent) were to permit the temporary immigration of up to 3 percent of their total labor force, developing Mode 4 countries would gain as much as $160 billion a (movement of natural persons) 1% year (Walmsley andWinters 2003). Mode 1 Virtually all commitments under the General (cross-border Agreement on Trade in Services have focused supply) Mode 3 mainly on the first three modes of international 28% (commercial service delivery rather than on mode 4, the presence) 57% "movement of natural persons," which involves Mode 2 the temporary movement of labor to provide ser- (consumption vices. Mode 4 accounts for only 1.4 percent of abroad) services trade (figure 5.1).The lack of liberaliza- 14% tion in labor services has been particularly costly Source: World Bank, Global Economic Prospects 2004. to developing countries, whose comparative T R A D E L I B E R A L I Z AT I O N : W H Y S O M U C H C O N T ROV E R S Y ? 135 BOX 5.2 The Trade and Growth Debate T he debate among economists and policy unique feature of the HH and RR analyses was their makers over the relationship between trade use of the various authors' actual data sets in test- and growth has risen to prominence during ing the robustness of their results. HH and RR criti- the last few years, owing on the one hand to the cized the empirical studies on data grounds, on mixed growth outcomes of developing countries that model-specification grounds, and on grounds of have undergone extensive trade liberalization and, econometric techniques. Data problems included, on the other hand, to differences over data, econo- among others, the use of poor measures of trade bar- metric techniques, and model specifications among riers (including the World Bank's classification of professional economists. trade regimes, which they criticized as subjective in The resurgence of interest in the 1990s among Edwards' paper), and the use of measures that are economists on the impact of trade on growth can highly correlated with other sources of bad economic be attributed to the significant improvements that performance such as poor exchange rate manage- have taken place in endogenous growth theory as ment (as in Dollar's and Sachs and Warner's papers). well as to the availability of more comprehensive Separately, Rodrik also criticized one of the more data and new econometric techniques. According recent papers on the topic, Dollar and Kraay (2001), to the new growth theory (attributed to Romer on data and model-specification grounds. The data 1986; Lucas 1987; and Grossman and Helpman problem arises from the combination of policy meas- 1992), whether import protection raises or lowers ures (tariff averages) with outcome measures the growth rate depends on the pattern of imports (imports as a share of GDP). The model specification and exports. Economists on both sides of the problem arises from regressing income on trade debate accept that as a matter of theory the rela- shares when both are endogenous (outcome vari- tionship between trade and growth is ambiguous. ables). The issue is hence an empirical one, which has become the focus of the debate in the last few · Notwithstanding these criticisms, it would be safe years. to say that most authors agree on the following: The launching of the debate can be attributed to First, that trade protection is not good for eco- Rodriguez and Rodrik (2000) (RR) and Harrison and nomic growth. Even RR themselves state in their Hanson (1999) (HH), who reviewed a number of paper that they have seen no credible evidence to empirical studies in the 1990s. While HH showed support the notion that trade protection is good that the Sachs and Warner (1995) study reflected for economic growth, at least for the post-1945 the gains from macroeconomic stability rather than period. trade reform, RR reviewed a number of studies, · Second, that trade openness by itself is not suffi- including Dollar (1992), Sachs and Warner (1995), cient for growth. RR argue in their paper that and Edwards (1998). RR expressed doubt "that there researchers and policy makers have been over- is a strong negative relationship in the data between stating the systematic evidence in favor of trade trade barriers and economic growth, at least for lev- openness, when what is really necessary is to fur- els of trade restrictions observed in practice," view- ther identify the connection between trade and ing "the search for such a relationship futile." A economic growth. 136 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s nomic environment,6 and that while lack of reg- ure 5.2 shows that tariff reductions in the 1990s ulations can undermine the growth effects of were positively and significantly associated with trade, in countries with effective regulation the developing countries' export shares.The positive effects of trade reforms are positive for growth.7 association between tariff reductions and export In developing countries that successfully inte- growth is consistent with so-called Lerner sym- grated into the global economy in the 1990s, a metry, whereby taxing imports has the same variety of factors reinforced each other: a stable effect on international trade as does taxing investment climate, greater market access, com- exports.8 This means that reducing tariffs pro- plementary macroeconomic policies, and unilat- motes exports. Cross-country regressions also eral or multilateral trade reforms. Table 5.1 suggest that in the 1990s real export growth was illustrates how the trade intensity of economies higher in countries with greater macroeconomic changed in response to reductions in tariffs. In stability, countries that reduced tariffs more, and the countries that began the 1990s with very countries that had more effective government.9 high tariffs,and reduced them the most,the share Detailed case studies reinforce these lessons on of imports plus exports in GDP rose significantly. the determinants of export activity. Studies using But in countries that began the decade with detailed plant-level data have shown that manu- more moderate tariffs and lowered them further, facturing firms that move into exporting are fre- the responses varied widely. One possibility-- quently the most productive in an economy. consistent with the evidence presented in table Consequently, policies that encourage invest- 5.1--is that at more moderate levels of protec- ments in human and physical capital, and that tion, other changes in the economy play a grow- support technological change, are likely to pro- ing role in determining changing trade shares. mote export growth.Evidence for Morocco sug- One important avenue through which tariff gests that many exporters are new enterprises, so reductions in the 1990s contributed to economic that policies that encourage new plant entry and growth is through their impact on exports. Fig- at the same time ease the exit of inefficient enter- TABLE 5.1 Tariff Reductions and Changes in Goods Trade Integration, 1990­2000 % changes in Change in integration, 1990­2000 tariffs, from late 1980s to late 1990s <1 time 1­1.5 times 1.5­2 times >2 times 40­70 reduction India Bangladesh Sudan 20­30 reduction Pakistan, Burkina Faso Benin, Ecuador, Kenya, China Peru, Thailand 10­20 reduction Egypt, Arab Rep. of, Republic of Congo, Argentina, Colombia, Philippines Iran, Mauritania, Indonesia, Turkey, Costa Rica, El Salvador, Mauritius, Zambia Uganda, Venezuela Guatemala, Nicaragua, Sri Lanka 0­10 reduction Tanzania, Paraguay, Chile, Côte d'Ivoire, Ghana, Nepal Senegal Bolivia, Jamaica, Malaysia, Nigeria, South Africa 0­2 increase Mozambique Madagascar, Trinidad and Mexico Tobago 2­10 increase Tunisia Jordan, Morocco, Oman, Saudi Arabia >10 increase Syrian Arab Rep. Source: World Bank staff calculations, available at http://sitesources.worldbank.org/INTRANETTRADE/Resources/tar2002.xls. Note: Trade integration, defined as the share of goods exports plus imports in GDP, is measured in real terms and excludes services trade. T R A D E L I B E R A L I Z AT I O N : W H Y S O M U C H C O N T ROV E R S Y ? 137 FIGURE 5.2 Changes in Export Shares of GDP and Changes in Tariffs, 1990­2000 2.5 2.0 X/GDP 1.5 real in 1.0 0.5 change 0.0 ­0.5 ­0.4 ­0.3 ­0.2 ­0.1 0 0.1 0.2 0.3 ­0.5 Percentage ­1.0 Percentage change in relative price impact of tariffs Source: World Bank staff calculations, available at http://siteresources.worldbank.org/INTRANETTRADE/Resources/tar2002.xls. Note: Changes are for the entire 10-year period 1990­2000, not annual changes. The correlation coefficient is ­0.25 and statistically significant. prises are likely to play an important role. Evi- cannot ensure that the strategy is effective. Liber- dence from Mexican and Indonesian censuses alization of trade in Argentina in the 1980s and suggests that exporters are likely to use skilled 1990s, and in Chile in the early 1980s, for exam- labor, which suggests that policies supporting the ple, was accompanied by an appreciation of the development of human capital are important. real exchange rate, which reduced the competi- Plant-level studies and anecdotal evidence also tiveness of domestic industries and incentives to point to the importance of foreign investors in export--with adverse consequences for the bal- helping developing-country exporters to break ance of payments and real economy. In many into new markets. Recent studies control for the countries of the former Soviet Union and some possibility of reverse causality,taking into account in Eastern Europe in the 1990s, trade was liberal- the fact that foreign firms may create or take over ized while property rights were not well defined the most efficient firms.10 Even if the importance and the institutional base for a market economy of foreign investment is difficult to identify in was not well developed.These, and other institu- cross-country studies, plant-level studies provide tional issues preventing the free movement of ample evidence that foreign ownership has been resources, often meant that trade reforms did not associated with export activity (box 5.3). Studies expand economic opportunities but restricted on Indonesia, Mexico, and Morocco show that them instead (Bolaky and Freund 2004). joint ventures and foreign-owned plants are sig- Trade reforms are most likely to stimulate nificantly more likely to export than other types growth when they are part of a comprehensive of enterprises. Although the mechanism is not strategy.Important elements of an effective growth completely clear, foreign firms are likely to pro- strategy can include sound macroeconomic man- vide knowledge of foreign markets and customer agement, building of trade-related infrastructure preferences, as well as access to new technology and institutions, economywide investments in and financing opportunities. physical and human capital, greater access to developed- and developing-country markets, and maintenance of a sound rule of law.Because these The Need for Effective Growth Strategies elements are often difficult to implement, there While trade integration opens new opportunities has been excessive emphasis on trade policy alone, and can strengthen an effective growth strategy, it rather than as a component of an overall growth 138 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s BOX 5.3 The Impact of Foreign Direct Investment on Growth F oreign direct investment (FDI) has been an (Alfaro et al. 2000) or is open to trade (Balasubra- important force in the global integration of manyam, Salisu, and Dapsoford 1996). national economies. Countries welcome FDI for A number of studies use micro-data to analyze the many reasons. Capital-scarce countries benefit from role of FDI in promoting technology transfer and rais- the infusion of a less volatile source of capital. ing host country wages (see, for example, World Bank, Greater investment financed by incoming FDI should Global Development Finance 2000; Aitken and Harrison also translate into higher growth. Foreign investors 1999; Haddad and Harrison 1993; Djankov and Hoek- are expected to provide employment opportunities, man 2000; Konings 2000; and Damijan et al. 2003). better wages and working conditions, and more train- They provide a mixed picture. However, they all agree ing. Many countries give foreign firms and joint ven- that affiliates of foreign firms are more productive tures special treatment in the expectation that these than indigenous firms. While part of these results firms will transfer new technology and knowledge to could reflect the fact that foreign firms acquire more domestic workers and firms. efficient domestic enterprises, anecdotal evidence also The cross-country evidence on the relationship suggests that local firms acquired by foreign investors between FDI and growth is mixed, in part because undergo restructuring and improve their performance incoming FDI as a share of GDP is typically quite as a result of the takeover. This direct effect should small. A cross-country study using data for 72 coun- not be ignored, because its magnitude may be signifi- tries for 1960­95 (Carkovic and Levine 2002) finds cant. Other evidence also suggests that foreign enter- no evidence that FDI exerts a positive impact on eco- prises pay higher wages (Aitken, Hanson, and Harrison nomic growth independent of other growth determi- 1997) and are more likely to comply with local labor nants (openness, black market premium, financial standards (Harrison and Scorse 2003). development, initial income, years of schooling). In sum, while quite a lot of evidence suggests However, Bosworth and Collins (1999) find that FDI, that FDI is positively associated with growth, there by raising total factor productivity, raises a country's is no consensus on the issue, and in particular no rate of output growth. Borenzstein et al. (1998) find consensus on the direction of causality. Regardless of that FDI adds to capital accumulation and raises the whether FDI independently contributes to growth, it efficiency of investment, but only where the host is clear that policies and institutions that are impor- country has a minimum level of human capital--an tant for growth would also be the ones that would indicator of absorptive capacity. The Borenzstein attract FDI as well as enhance the impact of FDI on study is consistent with evidence that suggests FDI growth. Therefore, countries should focus on such can promote growth if the country has complemen- policies and institutions rather than narrowly on how tary institutions such as developed financial markets to attract FDI. strategy. In addition to freeing markets and ensur- failures themselves. At the same time, however, ing the institutional foundation of a market econ- governments have learned how to structure inter- omy, governments may also need to address ventions in a manner that can reduce the risks of market failures that impede a supply response. capture and failure. Identifying which industries warrant special treat- Although many factors contributed to the ment is highly risky,and the experience of the last rise in trade integration in the 1990s,as discussed few decades is riddled with attempts to correct above, for brevity the following discussion is market failures that became more costly than the selective. It focuses on two critical complemen- T R A D E L I B E R A L I Z AT I O N : W H Y S O M U C H C O N T ROV E R S Y ? 139 tary areas: macroeconomic stability and trade- latter reflected Zambia's unsuccessful manage- related infrastructure and institutions. ment of the large declines that took place in 1995 and 1997 in the prices of copper, its main export Macroeconomic Stability (World Bank 2003n). In Malawi, too, macroeco- Macroeconomic stability is an important ele- nomic instability undermined export and growth ment in successful outcomes from trade performance.During the 1990s,high and volatile reforms.11 Macroeconomic stability entails low inflation, averaging 31 percent, resulted in an levels of inflation and a stable and competitive overvalued and highly volatile real exchange rate, exchange rate. Exchange rate volatility creates a seriously undermining domestic production, risky business environment in which future prof- investment, and exports. Malawi's manufacturing its and payments are uncertain,and these risks are sector contracted by 9 percent during 1995­96. higher in the many developing countries that These developments hindered Malawi's efforts to have not developed financial instruments for diversify its exports out of tobacco, where they hedging against foreign exchange risk. remain highly concentrated (World Bank 2003h). Successful exchange rate management requires, among other things, appropriate Trade-Related Infrastructure and Institutions sequencing of trade reforms and capital account Successful trade integration requires supportive liberalization. Experience has shown that capital infrastructure and institutions--the so-called account liberalization should follow, not precede, behind-the-border agenda.12 A comparison of the liberalization of trade, because the large Jamaica and Mauritius illustrates the importance inflows of capital that generally follow the freeing of institutions,as well as macroeconomic stability of the capital account could cause a large appreci- (box 5.4).Two other important constraints are ation of the real exchange rate, leading to large transport infrastructure and institutional capacity import surges that destabilize domestic industries for meeting product standards. Globally, and the balance of payments. improvements in transport and communications, India's appropriate sequencing of trade in conjunction with developing-country reforms, as well as its maintenance of a stable reforms,have allowed the production chain to be macroeconomic framework, contributed to its broken up into components,with some develop- impressive export and growth performance in ing countries playing a key role in global pro- the 1990s (World Bank 1994b). Before starting duction sharing, as noted in chapter 3. to liberalize trade, in the early 1990s India In many other countries, however, transport allowed a significant depreciation of the real remains a key bottleneck. Markets that are iso- exchange rate, which served to increase export lated may feature little competition and may fail incentives and cushion the impact of lower to realize economies of scale or scope.The result import barriers on domestic industry.Trade lib- is typically a vicious cycle of low productivity eralization preceded the opening of the capital and low profitability. Such constraints severely account. Since 1992, India's real effective limit the growth potential of the poorest coun- exchange rate has remained at more or less the tries, where agriculture supplies 15 to 52 per- same level, facilitating trade reforms. cent of GDP. In addition, since most of the poor In Zambia,by contrast,macroeconomic insta- reside in rural areas, these constraints have seri- bility undermined the potentially positive effects ous negative effects on poverty. For exporters in of structural reforms.Trade and other structural some developing countries, transport is the sin- reforms in the early 1990s gave Zambia one of gle most important component of cost.13 The the most liberal trade regimes in Africa, but main issues related to transport are lack of com- export performance has been lackluster. An petition and inadequate investments.Transport important reason is macroeconomic instability, costs are further raised by formal and informal with high inflation and high real interest rates, as fees and checkpoints. Poor transport particularly well as a highly volatile real exchange rate.The affects agricultural producers (mainly small- 140 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s BOX 5.4 Jamaica and Mauritius: Institutions and Macroeconomic Stability Make the Difference J amaica and Mauritius had nearly the same per macroeconomic stability. Subramanian and Roy capita GDP in 1984. But between 1984 and 2000, (2001) point to Mauritius's superior institutions real per capita GDP grew at about 4.8 percent a (democracy and strong participatory institutions), year in Mauritius, compared with only 0.7 percent in and ethnic diversity, which provided important links Jamaica. This is a dramatic difference in performance, to the rest of the world (68 percent of the population given the many similarities between the two countries. is Indian), and the need for participatory political Both countries have similar natural endowments institutions that were important for maintaining sta- and historical legacies. Both are island economies, bility, rule of law, and mediating conflict. Looking at have tropical climates, are subject to natural shocks indicators of institutional quality (Kaufmann, Kraay, (hurricanes in Jamaica and cyclones in Mauritius), and and Zoido-Lobaton 2002), Mauritius outperforms are former British colonies with English as the official Jamaica in all but one (regulatory quality): Mauritius language. Their economic structures are similar, with does better in government effectiveness, political about 6 percent of GDP from agriculture, about one- stability, rule of law, control of corruption, and voice. third from industry, and the remaining 60 percent or so The rule of law is a particular problem in Jamaica, from services. Sugarcane is widely grown in both coun- with crime and violence costing at least 4 percent of tries, and both enjoy preferential access to the Euro- GDP (excluding dynamic costs) (World Bank 2003g). pean Union and the United States for sugar exports. Unlike Mauritius, Jamaica has lacked a social/politi- Both established export-processing zones centered on cal compact; though recently the labor unions have garment manufacturing, with the primary impetus pro- agreed with the government to limit their wage vided by East Asian investors. increases in response to the grave economic situa- The disparate growth performance cannot be tion. attributed to differences in trade: between 1985 and Mauritius outperformed Jamaica in macroeco- 2000, real annual growth of exports was 3.9 percent nomic stability for the two decades from 1980 to in Mauritius and 3.6 percent in Jamaica, and by 2000, 2000, in terms of the level of inflation and the stabil- trade accounted for a larger share of GDP in Jamaica ity and competitiveness of the real exchange rate. In than in Mauritius. Jamaica has geographic advan- the 1990s, Jamaica's poor management of adverse tages for trade, being much closer to the United macroeconomic developments seemed to more than States and the European Union than Mauritius is to offset the potentially positive effects of a substantial either. Jamaica surpasses Mauritius in education trade (and capital account) liberalization. Financial enrollment indicators. And through the 1990s, crisis in the mid-1990s worsened the already deterio- Jamaica enjoyed higher FDI as a share of GDP than rating fiscal performance, and dramatically enlarged Mauritius. the ratio of debt to GDP. This has dampened private Two factors that may explain the difference in sector confidence, government investment, interest growth performance are institutional quality and rates, and growth. holder farmers and herders) who have difficulty sumer prices, and for sugar exports, regional and accessing markets both domestic and external. international transport costs add nearly 50 per- In Malawi,for example,high transport costs have cent to the ex-mill production costs (World weakened the competitiveness and profitability Bank 2003i). Lack of competition in road trans- of firms and farmers. Malawi is an efficient pro- port (where Malawi has restrictions on foreign ducer of sugar, but domestic transport costs operators) and high transport taxes add substan- account for 15 percent or more of local con- tially to transport costs. T R A D E L I B E R A L I Z AT I O N : W H Y S O M U C H C O N T ROV E R S Y ? 141 Product standards in international trade have protection, and the pros and cons of regional proliferated and become more stringent in trade agreements. recent years. Consumers in developed countries One element is common to almost all of the are demanding stricter food standards, while success stories:despite the diversity of approaches major food retailers, food manufacturers, and to trade reform, all successful liberalizations restaurant chains have been adopting codes of either explicitly or implicitly promoted export practice, standards, and other forms of supply- growth. Exporters were given incentives to chain governance as part of their commercial ensure that selling on international markets was strategies of differentiation.Increasingly,middle- as attractive as domestic sales.This required estab- income and some low-income countries are also lishing a regime that offset the anti-export bias. raising their product standards, in part through In turn, this required an effectively functioning the investments undertaken by multinational bureaucracy to implement the offsetting regula- supermarket or restaurant chains and competi- tion--as with the "indirect duty drawbacks" in tive responses by local firms. Korea.This proactive approach is not generally Prospects are dim for "special and differen- prescribed.Since most countries lack the institu- tial treatment" that would require less stringent tional capacity that is required to implement off- standards from poorer countries (Jaffee and setting regulation, classic trade Henson 2004). Developing countries need to liberalization--through low, uniform tariffs and develop and improve their food safety and agri- the elimination of quantitative restrictions--has cultural health management systems to position been the more conventional recommendation. themselves competitively and to enhance their export performance. Building such capacity is PartialTrade Liberalization: not beyond the reach of developing countries, China and India and some very poor countries are meeting exacting international standards. Examples China opted for partial trade liberalization, pur- include Peruvian exports of asparagus to the sued through a dual-track approach.Special eco- United States and the European Union, and nomic zones (SEZs)--one of the drivers in low-income African countries' exports of fish China's export and growth success--were set up products that meet EU hygiene standards. in the 1980s to give the firms established within Countries that meet strict export standards are them access to duty-free imported inputs. Firms generally those where the private sector is well outside the SEZs faced much higher tariffs on organized and the public sector well focused to imports, at 56 percent in 1982, falling to 44 per- meet exporters' needs, such as through out- cent in 1991and 16 percent in 2000 (Lardy grower programs for smallholder farmers, sys- 2002). tems of training and oversight for small and China established its first four SEZs in 1980 medium-size enterprises through associations in two coastal provinces (Guangdong and and groups, and twinning and regional net- Fujian),selected for their location.14The success working for small countries. of the initial zones led to the addition four years later of 14 coastal cities (including Shanghai) as "coastal open cities," with authority similar to 2. Different Paths to Trade that of the SEZs. By 1992, most cities along the Reform Yangtze River and the borders of China had been granted special privileges as coastal cities, This section discusses issues related to the path of with Shanghai being granted even more auton- liberalization, including the success of different omy.These developments, in turn, spurred the partial approaches to trade liberalization, manag- establishment of "development zones" in many ing the political economy of trade reform, inland cities that extended tax benefits and whether there is a limited role for infant industry autonomy to foreign and domestic investments. 142 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s In many cases,such zones were established with- Factors that were clearly important for the out the approval of the central government.15 In trade reforms adopted by China and India were 1993 China became the world's second-largest the credibility of reforms and the importance of destination for FDI, next to the United States. strong institutions. Some ways to achieve reform Compared with other regions, the SEZs enjoy credibility are discussed below. lower tax rates and greater authority in approv- ing foreign investment projects.The removal of Political Economy ofTrade Reforms administrative barriers had nearly as great an effect in spurring trade as China's tariff reduc- The success of trade reforms is not automatic. tions,which did not really begin until the 1990s. Political economy considerations need to be Exports grew at an annual average of 15 percent taken into account at the design stage if reforms in the 1980s, and at 19 percent in the 1990s.16 are to be sustainable.The key elements on the India followed a different model of partial political economy front are ensuring that the liberalization, liberalizing trade across all regions costs of adjustment arising from reforms are of the country but relaxing protection one sec- eased, and that reforms are credible. tor at a time.After piecemeal efforts at liberaliz- ing trade during the 1980s, India launched a Easing the Costs of Adjustment coherent trade reform program in 1991, with Easing the costs of adjustment is clearly important some faltering during 1997­2001.17 The to generate social and political support for reforms entailed concurrent reductions of some reforms. One way to ease adjustment costs is to of the highest tariff and nontariff barriers ensure that safety nets are adequate to compen- (NTBs) in the world.A large reduction in NTBs sate losers. But, as discussed earlier, a more effica- and the streamlining of a very complex import cious way is to design a reform program that licensing regime came early in the reform pro- minimizes adjustment costs. gram, while tariffs were reduced in a phased China and Mauritius provide good examples manner, with reductions continuing today. Cur- in this regard, by creating new profit opportuni- rently, the maximum customs tariff for nonagri- ties at the margin while leaving old opportuni- cultural goods is 30 percent, scheduled to be ties undisturbed.The upshot was that there were reduced to 20 percent or less in the near no identifiable losers. In China, few vested inter- future.18 Capital and intermediate goods ests opposed the SEZs because these were set up imports were liberalized first, and consumer outside the scope of central planning and did not goods (which were effectively banned) not until disrupt planned production and allocation. several years later. It was not until 2001 that all China's approach also maximized political sup- consumer goods imports were liberalized.19 port for the reforms as the number of winners India's sequencing of trade liberalization, grew over time. Mauritius partially liberalized which entailed earlier liberalization of capital and trade by establishing export processing zones intermediate goods than for consumer goods, (EPZs) and segmenting the labor market (Subra- and much steeper reduction in tariffs for some of manian and Roy 2001). Labor market rules were them, was intended to discourage the deferment much less stringent in the EPZs than elsewhere of investments that might occur if domestic pro- in the economy. Until the mid- to late 1980s, ducers expected further reductions in capital employers had greater flexibility in dismissing goods tariffs.20 The response was rapid: in dollar workers in the EPZ sector,and in the 1980s,EPZ terms,exports were growing by 20 percent annu- wages were about 36­40 percent lower than ally within three years of the start of the reform wages in the rest of the economy,with the differ- program.The strong export supply response pro- entials narrowing to 7­20 percent in the 1990s. vided impetus for a continued response, not least Aside from acting as a subsidy to exports,the seg- because the new export receipts alleviated the mentation of the labor market also prevented the pressures on the balance of payments. expansion of the EPZs from driving up wages in T R A D E L I B E R A L I Z AT I O N : W H Y S O M U C H C O N T ROV E R S Y ? 143 the rest of the economy and disadvantaging the signing on to regional trade arrangements that import-substituting industries. lock in reforms. Ensuring Credibility Should Emerging Industries Be Protected? At the very least, reforms should be publicly Although import substitution policies have been communicated so that economic agents are largely discredited, the need to address market aware of them and can respond accordingly. failures that prevent the development of interna- Mozambique lifted export restrictions on cashew tionally competitive industries has continued to nuts but with very little communication to those provoke debate. Suggestions have been made to directly affected by the reforms, so that few grant temporary modest levels of import protec- cashew nut farmers were aware that substantial tion where there is a demonstrated need reforms had been undertaken.21As a result,much (Williamson 2004a). Other authors have focused of the price increase that resulted from the on choosing the right form of protection, advo- reforms went to the traders, and the supply cating subsidies to the initial entrants rather than response was constrained. Had farmers been told the use of import duties (Baldwin 2003). of the reforms, they could have strengthened Another suggestion is to approach develop- their bargaining power in relation to the traders, ment as a process of "self-discovery," since the making it difficult for the latter to pay low prices. key challenge that a modernizing economy faces Public communication of reforms also dimin- is learning what it is good at producing (Haus- ishes the possibility of reform reversals, boosting mann and Rodrik 2002).The entrepreneur who their credibility. first discovers what the country should specialize Another way to boost the credibility of in can capture only a small part of the social value reforms is to undertake measures that are less easy that this knowledge generates, because other to reverse than price changes. In Mozambique, entrepreneurs will quickly emulate such discov- another reason why the supply response was poor eries.Thus this type of entrepreneurship will typ- was that cashew nut processors did not make ically be undersupplied and economic investments to improve their efficiency, in part transformation delayed.There may be a role for because they expected the reforms to be government involvement to provide incentives reversed.The overall reform program would have to induce such investments, as well as to exert been more credible had the price reforms been discipline in pruning investments that turn out accompanied by nonprice reforms, such as gov- to be costly. ernment investment in transport, better access to A key challenge for countries that choose to credit,promotion of competition in cashew mar- pursue such a strategy is to structure the right keting, and the creation of incentives to adopt combination of incentives (inducements) and dis- improved technologies for cashew growing.Such cipline (competitive pressures, resistance against nonprice interventions strengthen credibility by special interests). Some of the world's most suc- signaling to the public a government's commit- cessful economies during the last four decades ment to the reforms. (Korea andTaiwan [China] since the early 1960s; Further ways to promote credibility include China since the late 1970s) prospered by pursuing the establishment of institutions such as India's policies that gave inducements for investment and Tariff Commission, which is charged with the risk taking while expanding competitive pressures design and implementation of the trade reform that ensured efficient allocation by investors.Dur- program and has a tenure that outlasts govern- ing their industrial drives in the 1960s and the ments. Such long tenure helps to enhance the 1970s,Korea andTaiwan (China) provided export credibility of reforms, as it diminishes private subsidies contingent on export performance.This sector expectations that the reform program strategy allowed policy makers to distinguish will be reversed by successive governments. firms and sectors that were highly productive Finally, credibility can also be achieved through from those that were not.The subsidies included 144 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s supplying inputs, providing working capital, interests, but should maintain links with the pri- imposing import restrictions, and--in Taiwan's vate sector to maximize economywide gains. (China) textile industry in the 1950s--buying the This is not a prescription for creating new state resulting output. Local production grew spectac- enterprises, promoting existing activities, or giv- ularly as a result.But the government also pruned ing governments authority to expand their nonproductive firms subsequently. bureaucratic reach. Clearly, the institutional and Asia's successful experiences in this regard con- administrative requirements for success are for- trast with the generally failed experiences of Latin midable. America. Pursuing import substitution strategies in the 1960s and the 1970s, Latin American gov- A Role for Regional Agreements? ernments provided incentives without sufficient discipline, with the result that too many low-pro- Some countries have achieved greater integration ductivity firms operated alongside the high per- and strong growth by adopting unilateral or mul- formers. When trade openness and domestic tilateral trade reforms combined with participa- competition brought discipline in the 1990s, pro- tion in regional trade agreements. Signing on to ducers received too little support (Hausmann and regional trade agreements provides countries with Rodrik 2002).Without a good balance between access to the markets of fellow members, and can promotion and discipline, Latin American coun- help improve their domestic institutions. But evi- tries' industrial performance fell short of that in dence suggests that as many as half of regional East Asian countries during these decades. trade agreements are substantially trade-diverting. Chile has often been touted as a miracle of Trade and investment diversion cause significant free-market economics. In fact, public-private economic losses to the countries excluded from collaboration strategies have played a key role in the agreements. fostering structural change and stimulating non- Regional integration has yielded good results traditional activities (box 5.5).Yet identifying the for Central and Eastern European countries that conditions for successfully assisting new activities signed Europe Agreements in the 1990s with the is not easy. Rodrik and Hausmann (2003) European Union, and for Mexico, which joined emphasize the importance of creating an institu- the North American Free Trade Agreement tional architecture that resists the pull of special (NAFTA).For the Central and Eastern European interests, and the importance of political leader- countries, the institutional harmonization aspect ship from the top. Whatever institutions are of the Europe Agreements has been very impor- employed to support new activities,they must be tant for successful trade integration and growth transparent and accountable, or selective support (World Bank 2000d);agreements on harmoniza- is likely to evolve into a new mechanism for sup- tion of investment policies, regulatory rules, and porting private interests in the name of public institutions with those of EU members have gain. The promotion of new activities should encouraged export-oriented foreign direct conform to a set of design principles that include investment into the Central and Eastern Euro- the following: (1) incentives should be provided pean countries.In Mexico,NAFTA has had pos- only for new,"sunrise"activities,not sunset ones; itive effects on trade, foreign direct investment, (2) there should be clear benchmarks for success technology transfer, and growth, and is also asso- or failure;(3) support must have a predetermined ciated with productivity improvements in manu- end (a so-called sunset clause); (4) public support facturing.But although NAFTA has contributed should target activities such as worker training or to institutional harmonization between Mexico infrastructure investment, rather than sectors such and the United States in the areas that it covers-- as electronics;(5) subsidized activities should pro- in particular intellectual property rights, investor vide clear potential for externalities; and (6) protection, and environmental standards--it has agencies involved in these activities should be not helped to narrow other institutional gaps, autonomous enough to avoid capture by private especially in the areas of rule of law and corrup- T R A D E L I B E R A L I Z AT I O N : W H Y S O M U C H C O N T ROV E R S Y ? 145 BOX 5.5 Behind Chile's Success: A Less than Orthodox Approach C hile appears to be the exception among infrastructure necessary to export fruit to foreign mar- Latin American countries by striking the kets. Private investment and exports took off after right balance of inducements and discipline the reforms of the mid-1970s once uncertainties in promoting domestic industry. regarding land reform, macroeconomic stability, and Fruits and salmon, Chile's two largest export items labor militancy were resolved. These investments and after copper, have both benefited from private-public exports were further boosted by the sharp real depre- sector partnerships. The foundations of the fruit ciation of the currency in the mid-1980s. industry were laid in the early 1960s through the The salmon industry, which generates $600 mil- efforts of the Corporacion de Fomento, the University lion in annual exports and provides jobs for more of Chile, and the National Institute of Agricultural than 100,000 people in this country of 15 million, Research (INIA). INIA, established in 1964 with also benefited significantly from public interven- highly paid, skilled researchers, initiated the fruit tions. It was created single-handedly by Fundacion research program. The public sector carried out much Chile, a nonprofit institution created by the Chilean of the development of scientific personnel and knowl- Government in 1976. Fundacion Chile brought the edge to achieve technological transfer; identification technology of salmon farming to Chile, adapted it and planting of new varieties suitable for export to and made it commercially viable, formed private sec- foreign markets; improvements in orchard and tor businesses to use it, and eventually sold its par- postharvest management; and the development of the ticipation to Japanese investors at a great profit. Sources: Rodrik and Hausmann 2003; Ocampo 2004; and Washington Post, January 21, 2004. tion, which are nonetheless important for fared much worse in economic performance income convergence between the two countries compared to Estonia and Slovenia, which signed (Perry et al. 2003). such agreements in 1995 and 1996, respectively. Evidence suggests that for developing coun- Most important, regional trade agreements tries,signing on to regional trade agreements with can divert attention away from the multilateral developed countries, particularly large developed WorldTrade Organization (WTO) process, and countries, is most useful.Agreements should also result in higher costs than benefits for develop- strive to ensure that barriers that apply to nona- ing countries.23 This will be especially true if greement countries are kept low. Signing such the agreed upon protection relative to third agreements will not generate positive export and parties remains high. Recent experience with growth responses unless the countries themselves the Free Trade Area of the Americas, the Cen- also pursue other necessary economic, political, tral American Free Trade Agreement, and the and social reforms. Among the EU accession U.S.-Australia Free Trade Agreements suggests countries in the 1990s, benefits only accrued to that regionalism will not help the developing those countries that were also undertaking the countries much with their market access prior- necessary economic, political, and institutional ities: trade-distorting agricultural support in reforms to transform their economies into mar- the North, contingent protection, and liberal- ket-based ones.22 For example, Bulgaria and ization of temporary migration of labor. Fur- Romania signed Europe Agreements in 1993, in ther, the high costs of negotiating such advance of several other accession countries, but agreements divert resources away from such they lagged behind in the transition process and larger multilateral issues. 146 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s 3. Trade Liberalization, Poverty, 3.6 percent in India, while trade integration and Income Distribution (trade in goods and services in real terms as a share of GDP) rose from 23 to 46 percent of Despite expected gains for the economy in the GDP in China, and from 19 to 30 percent in longer term, trade reform generates both win- India. Over this period, both countries massively ners and losers in the short run.24 The critical reduced the incidence of poverty--from 28 to 9 question is whether the short-run costs of trade percent between 1978 and 1998 in China, and reform fall disproportionately on the poor.Econ- from 51 to 27 percent between 1977­78 and omists in the 1990s expected trade and foreign 1999­2000 in India.26 Since a large share of the investment reforms to help developing countries world's poor lives in these two countries, these reduce poverty.Trade liberalization was expected large reductions have served to reduce or miti- to increase demand for goods produced by devel- gate overall inequality in the world, even though oping countries' poor or low-skilled workers, inequality has risen within both countries leading to higher wages for unskilled workers (Ravallion 2003b; Sala-i-Martin 2003). and ameliorating poverty.Trade reforms were also Nevertheless, Harrison (2005) suggests that expected to raise the prices of the agricultural policy makers need to be cautious about expect- products produced by the poor and to reduce ing large gains in poverty reduction from trade prices of goods that the poor consume. Is the reforms.27 Many economists expected that emerging evidence from the 1990s consistent developing countries with a comparative advan- with these expectations? How much of the tage in unskilled labor would benefit from liber- decline in poverty rates and increasing within- alization of trade through increased demand for country inequality can be attributed to the trade their unskilled labor­intensive goods, which in reforms of the 1990s? turn should reduce inequality and poverty.How- ever, the evidence in this volume--which includes 15 separate studies of the links between Effects ofTrade Reform on poverty and globalization--suggests that the Aggregate Growth and Poverty story is more complex.One reason is that labor is not nearly as mobile as simple trade models Direct Effects assume. If comparative advantage is to increase If opening up to trade is associated with higher the incomes of unskilled workers, they need to growth, it may be associated with a decline in be able to move out of contracting sectors and poverty as well. This argument rests on two into expanding ones. A second reason is that assumptions: first, that opening up to trade leads developing countries have historically protected to higher growth, and second, that growth raises their unskilled-intensive sectors, so that trade the incomes of the poor as much as the incomes reforms may lead to less protection for unskilled of the rich. workers relative to skilled.A third reason is that What actually occurred? There is widespread even firms in countries with a comparative evidence that GDP growth reduces poverty.25 In advantage in producing goods that use unskilled other words, evidence suggests that growth ben- labor need to use skilled workers in order to efits those at the lower end of the income distri- compete in global markets. bution. If trade liberalization contributes to growth--as discussed earlier in this chapter--it Indirect Effects should be associated with reductions in poverty. Trade reforms can also affect poverty indirectly-- China and India, for example, have both experi- for example,by influencing (1) the job opportuni- enced tremendous increases in trade integration ties and wages of the poor,(2) the prices that poor and growth,as well as large reductions in poverty. consumers pay for the goods that they buy,(3) gov- From 1980 to 2000, real per capita GDP grew at ernment revenues and in turn social expenditures an annual average of 8.3 percent in China and that particularly affect the poor, and (4) income T R A D E L I B E R A L I Z AT I O N : W H Y S O M U C H C O N T ROV E R S Y ? 147 instability as well as workers' chances of becoming manufacturing of computer chips. In each case, poor (Winters et al. 2004). Even if aggregate substantial numbers of workers lost their jobs,and poverty falls or remains constant,many households some experienced very long periods of unem- may move into or out of poverty as a result of trade ployment or large wage losses, or both. liberalization. As emphasized by De Ferranti et al. (2001), Effects on jobs and wages. Some studies have such dislocations are transitional and do not found that trade reforms reduce employment in imply a permanent increase in the unemploy- the short run, but others have found that trade ment rate. Chile, for example, experienced dou- reforms increase employment over the long run, ble-digit rates of unemployment for several years as expanding sectors create new employment after liberalization, but from 1986 to 1997 its opportunities.Trade explains much of the decline unemployment rates were among the lowest in in Singapore's unemployment rate, from more the region. Mexico's present rate of unemploy- than 9 percent in the 1960s to close to 2 percent ment is roughly at its traditional level,despite that in the late 1990s.A study of 18 countries in Latin country's dramatic economic integration with America and the Caribbean over the period the United States. 1970­96 found that trade liberalization had a Although most studies find that the unemploy- negative, though small, direct effect on employ- ment effects of trade liberalization tend to be tem- ment.28The negative effect was greater in coun- porary, even short-term costs can be high in tries where the real exchange rate appreciated as human terms. Such costs must be addressed a result of capital inflows that followed the eco- through a variety of policy approaches, including nomic reforms. Similarly, in Brazil during stronger social safety nets, in order to ensure that 1990­97, trade liberalization slightly reduced trade reforms succeed. employment in the short run, but the more Effects on prices. An emerging literature using labor-intensive output mix that resulted over the household-level data suggests that, via changes in long run increased employment.29 Much larger factor and goods prices, trade liberalization can negative effects on output and employment have lead to poverty reduction. For instance, a recent been found in someAfrican countries.One study study of trade liberalization in Argentina using for Kenya,Tanzania, and Zimbabwe found that household survey data found that Mercosur has most firms responded to import competition benefited the average Argentine household across pressure by contracting rather than upgrading the spectrum of income distribution.31The same aggressively.30 Among the suggested reasons for study also finds that Mercosur has had a pro-poor such behavior are the firms' lack of preparation bias, benefiting poor households more than mid- for competition, absence of policies to promote dle-income households, and that its impact on technological improvement (especially among rich families is positive but not statistically signif- small and medium enterprises), and poor tech- icant. The reason behind these results is that nological and human infrastructure. Argentine trade policy protected the rich over Trade reforms of the 1990s in Latin America the poor prior to the reforms, and granted some and the Caribbean reduced employment in pre- protection to the poor after the reforms. viously protected industries and augmented it in Effects on social spending. Social spending is others (De Ferranti et al. 2001). Argentina lost another avenue through which liberalization may much of its automobile industry while seeing an affect income distribution, but there is no direct expansion in more sophisticated chemicals and evidence for such a relationship.The available capital- and labor-intensive manufactures. Brazil evidence,relating mostly to the 1980s,32 suggests lost much of its cereals industry to Argentina that many trade reforms had no revenue costs. under Mercosur, and its manufacturing industry Some of the main reasons were that temporary suffered more generally. Costa Rica lost much of tariff surcharges were introduced when quantita- its labor-intensive manufacturing to Mexico after tive restrictions were removed, and that changes NAFTA, but it also substantially increased its in the import/export base arising from the trade 148 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s reforms enhanced revenues. For example, Summarizing the Links between Trade Kenya's trade liberalization between 1989 and Reforms and Poverty 1999 (which entailed halving the simple average What lessons emerge from cross-country and import duty rate over the period and abolishing more detailed case studies using household data? import licensing requirements and foreign First, the poor are more likely to share in the exchange controls) led to increases both in duty gains from globalization where complementary as a share of imports and in import duty revenues policies in place. Case studies of India and as a share of GDP. The increase in revenues Colombia in Harrison (2005) suggest that glob- reflected the expansion of the revenue base, alization is more likely to benefit the poor if trade tighter exemption management, higher duty reforms are implemented in conjunction with rates on certain products,a shift in imports to the labor market deregulation.35 In Zambia, poor higher duty classes, and possibly also improve- farmers are only expected to benefit from greater ments in customs administration and the intro- access to export markets if they also have access duction of a preshipment program (Glenday to credit, technical know-how, and other com- 2000, cited inWinters et al. 2004). plementary inputs.36 The same volume also Even in cases in which revenues are cut,avail- points to the importance of social safety nets. In able evidence suggests that public spending Mexico, trade reforms in the 1990s hurt the important to the poor can be protected.There poorest corn farmers; without support from the are alternative sources of revenues--though cau- government, these farmers' real incomes would tion needs to be exercised to ensure that replace- have been halved.37 The same result has been ment taxes do not hurt the poor. And, with found more recently in Ethiopia.38 political will, social spending, particularly that Second, while financial crises are associated oriented toward the poor, may be shielded. with increasing poverty,reforms in trade and for- Effects on vulnerability and income volatility. eign investment in a number of countries have When Indonesia,Korea,andThailand opened up helped to reduce poverty. In Mexico, the poor in to trade in the late 1980s and early 1990s, no the most globalized regions have weathered the strong negative effects on poverty and vulnera- macroeconomic crises the best.39 In India,open- bility resulted.33 It remains an open question ing up to foreign investment was associated with whether openness made the 1997­98 Asian a decline in poverty. In Colombia, increasing financial crisis much more serious than the export activity was associated with an increase in shocks that had hit the three countries in the compliance with labor legislation and a fall in 1980s. It is clear, however, that financial crises are poverty. In Poland, unskilled workers--who are very costly to the poor. In Indonesia, the finan- the most likely to be poor--have gained from cial crisis of 1997 led to a 50 percent reduction in the country's accession to the European real wages.34 In Mexico, the peso crisis of the Union.40 mid-1990s led to a stagnation in real wages that Clearly, globalization produces both winners lasted nearly a decade.A recent study of financial and losers among the poor.Winters, McCulloch, deregulation across countries emphasizes the and McKay (2004); Ravallion and Lokshin need for complementary policies,such as the cre- (2004); and Harrison (2005) all emphasize this ation of reliable institutions and macroeconomic heterogeneity in outcomes. It should not be sur- stabilization policies (Prasad et al. 2003).While prising that the results defy easy generalization. financial crises resulting from unrestricted capital The poor can gain from one set of policy flows are associated with a higher likelihood of reforms, if those lower the prices they pay for poverty, foreign direct investment inflows are consumption goods, and lose from other trade associated with a reduction in poverty. The reforms that lower the prices of the goods they poverty-reducing effects of FDI are clearly doc- produce. Poor wage earners in exporting sectors umented in several recent studies on India and or in sectors with incoming foreign investment Mexico. gain from trade and investment reforms; con- T R A D E L I B E R A L I Z AT I O N : W H Y S O M U C H C O N T ROV E R S Y ? 149 versely, workers in previously protected sectors that are most exposed to international competi- are likely to lose. tion pay the highest wages. It is difficult to dis- This emerging evidence on the links between tinguish the impact of globalization from that of trade reforms and poverty points to the need for technical change, since the adoption of new carefully targeted social safety nets and comple- technologies could be stimulated by external mentary policies to ease the transition of workers competition via trade. In Mexico, for example, from contracting to expanding sectors. the tripling of manufactured exports during the 1990s has been associated with increased rates of adoption of modern production technologies,an Trade Liberalization and Inequality acceleration of productivity growth, a relative Though inequality has been increasing in both increase in the demand for skilled labor, and an rich and poor countries we still lack a compre- increase in inequality. hensive understanding of why.A popular expla- There is no evidence that trade liberalization nation is that technological change--which may permanently worsens income distribution. As or may not be associated with opening up to noted above, however, there is evidence that trade trade--has led employers to demand more skilled liberalization has been associated with--at times labor. This phenomenon, referred to as skill- significant and prolonged--adjustment costs in the biased technical change, has occurred in both form of employment losses.In Mexico,trade inte- developed and developing countries.Some econ- gration through NAFTA, while reducing poverty, omists argue that the demand for more skilled has also increased income inequality between workers is unrelated to trade liberalization, since regions: regions with lower per capita GDP and the same trend has been documented in services higher telephone density grew faster,while regions that are not traded on world markets, but others with high public employment grew more slowly argue that skill-biased technical change is itself an (Perry et al.2003). outcome of globalization. Governments need to help the disadvantaged One reason why trade reforms may be associ- by strengthening social safety nets and by provid- ated with increasing inequality is that many ing education and training for the unskilled. As countries--Colombia, Mexico, Morocco, and attested by the industrialized countries, it is a Poland, for example--have traditionally pro- daunting task to build up the administrative and tected the sectors that use mainly unskilled labor. institutional capacity required to design and Another possible reason is that exporters-- implement safety nets that are well targeted and who benefit from trade reforms--need to hire that avoid leakages. More innovative approaches skilled workers to succeed in world markets. A to trade reforms and trade reform assistance number of studies have shown that exporters are packages may be needed. more likely to use a high proportion of skilled workers, suggesting that as countries turn to exporting, the demand for skilled workers will 4. Issues of Differential Market rise, pushing up their wages relative to those of Access unskilled workers.41 Foreign firms in developing countries tend to hire more skilled workers than After the reforms of the 1990s the world trade sys- do domestic firms. In Mexico, increasing tem has been more supportive of development. inequality is most evident in the border region-- But it remains strongly biased against the poor. the region most affected by increasing trade with Global markets are most hostile to the products the the United States. world's poor produce--agriculture, textiles, and Nevertheless, the evidence on trade liberal- labor-intensive manufactures.Escalating tariffs,tar- ization and wage inequality remains inconclu- iff peaks, and quota arrangements maintained by sive. In Argentina, Brazil, Costa Rica, the both developed and developing countries system- Dominican Republic,and Mexico,the industries atically deny the poor market access and skew 150 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s incentives against adding value in poor countries. inequality in agriculture and favoring wealthy In both rich and poor countries, protection landowners. Developed countries impose remains heavily concentrated in the most politi- higher tariffs on agricultural imports from cally sensitive areas--textiles,clothing,other labor- developing countries than from other industrial intensive manufactures,and agriculture. countries (table 5.2). Developed countries Differential treatment by developed countries impose an average tariff of 15 percent on agri- still constrains the expansion of trade by devel- cultural imports from other industrial countries, oping countries, particularly the poorest. In but average tariffs ranging from 20 percent (for developed countries, the relatively low average Latin America) to 35 percent (for Europe and tariffs mask the sometimes high protection in the Central Asia) on agricultural imports from form of tariff peaks, tariff escalation, specific developing countries.The issue of agricultural duties, and production subsidies. protection, in particular in cotton, has risen in Developed-country protection is much prominence in multilateral trade talks, and was more pronounced in agriculture than in manu- one of the main reasons for the failure of the facturing (World Bank,Global Economic Prospects most recent round of WTO talks in Cancun in 2004). Since most of the world's poor live in September 2003. Since then, Brazil has gone to rural areas and work in agriculture, rich-coun- the WTO with charges that U.S. subsidies on try subsidies combined with trade protection to cotton are inconsistent with WTO obligations, domestic agriculture worsen world poverty. and the WTO ruling on April 2004 affirmed Farm production subsidies in the United States, Brazil's charges.42 for example, are distributed overwhelmingly to On manufactured goods,tariffs are on average the richest farmers, exacerbating income lower in developed than in developing countries, TABLE 5.2 Rich Countries Levy Higher Tariffs on Poor Countries' Exports (1997 protection rates facing exporters in each region, in percentage points) Importing region East Europe and Sub-Saharan Industrial Exporting region Asia Central Asia Latin America Middle East South Asia Africa countries Agriculture Industrial countries 33.3 43.7 20.1 65.4 16.4 24.0 15.3 East Asia 31.0 30.3 15.5 45.3 38.4 19.0 30.5 Europe and Central Asia 24.2 36.4 23.8 55.3 34.2 12.7 35.1 Latin America and the Caribbean 42.1 36.0 14.8 50.3 29.7 24.7 20.4 Middle East 23.0 43.4 14.9 76.4 31.8 18.9 23.4 South Asia 16.6 34.6 13.7 41.1 27.7 11.0 25.8 Sub-Saharan Africa 26.7 20.3 14.4 39.1 30.9 33.6 23.6 Nonagriculture Industrial countries 7.4 9.6 8.5 10.4 25.2 12.2 1.0 East Asia 8.2 13.8 15.1 12.2 28.1 14.5 5.1 Europe and Central Asia 6.4 6.4 11.4 8.6 25.8 12.8 5.9 Latin America and the Caribbean 4.3 6.7 15.4 8.9 19.4 11.9 2.1 Middle East 5.4 11.5 8.8 11.4 33.6 11.7 6.0 South Asia 7.1 11.0 13.6 10.2 19.0 17.4 8.1 Sub-Saharan Africa 4.4 6.1 11.7 6.1 27.6 20.6 4.2 Source: Weighted averages calculated using GTAP Version 5 database (www.gtap.org). Most-favored-nation rates except for major free-trade blocs such as the European Union and the North American Free Trade Area. T R A D E L I B E R A L I Z AT I O N : W H Y S O M U C H C O N T ROV E R S Y ? 151 but the types of goods exported by poor coun- 2. World Bank (Global Economic Prospects 2004, 139). tries face higher tariffs in the rich countries. For These statistics are based on remittances sent through official channels. Existing payment systems make example, while exporters of manufactures from remittances difficult and costly, especially in and to industrial countries face, on average, a tariff of 1 Africa and Central America.To many parts of the percent on their sales to other industrial coun- world, unofficial remittances far outweigh official tries, exporters from developing countries pay ones. anywhere from 2 percent if they are from Latin 3. From 1990 to 2000, income from migrant workers America (where NAFTA weighs heavily) to 8 overseas (including workers' remittances and employ- ees' compensation) as a share of foreign exchange percent if they are from South Asia. receipts (measured as exports of goods, services, and Overall, rich countries collect from develop- workers' income) fell from 4.3 to 3.8 percent for all ing countries about twice the tariff revenues per developing countries. Conceptually it makes sense to dollar of imports that they collect from other rich compare income from migrant workers with receipts countries. Protection also takes forms other than from exports of goods and services since labor could be viewed as one form of a country's service exports. tariffs--among them quotas, specific duties, and Almost all of the drop for the developing world as a contingent protection measures such as whole can be attributed to the decline in migrant antidumping duties. As with tariffs, these meas- workers' income in the Arab Republic of Egypt, ures tend to be used more frequently against which in 1990 had enjoyed the largest amount of this labor-intensive products from developing coun- income in nominal terms in the developing world. tries.Antidumping duties are on average 7 to 10 The decline in migrant workers' income in Egypt during the 1990s was related to the GulfWar.Exclud- times higher than tariffs in industrial countries, ing Egypt, the ratio fell from 3.7 to 3.6 percent over and around 5 times higher in developing coun- the decade. Countries where incomes from migrant tries. Developing countries are also hampered in workers have become quite important--ranging other critical areas, including access for their between 20 and 46 percent of total foreign exchange agricultural and textile exports, and restrictions receipts in 2000--and where such income increased significantly over the 1990s (increases ranging from 10 on international labor migration. to 46 percent) include Albania, Ecuador, Jamaica, Jor- To continue the momentum toward greater dan,Nicaragua,Sudan,and Uganda.At the same time, global integration, high-income countries must however, countries including Benin, Cape Verde, further open their markets to developing-coun- Egypt, Lesotho, and Pakistan experienced declines in try exports. Industrial countries' unfair tariff such incomes, ranging from 10 to 30 percent. treatment of developing countries must be 4. Properly identifying the causal impact of changes in trade policies on growth needs to take into account addressed in the upcoming Doha round of trade other factors associated with GDP growth, and the negotiations. possibility of reverse causality (that is,if GDP growth causes changes in trade policies).This means that the variable for trade policy should be"instrumented"or Notes represented with measures that affect trade policy but are not correlated with GDP growth.Since most 1. These changes were not just due to declines in the reforms are driven by initial protection levels, one prices of agricultural and resource commodities rel- way to get around the problem is to instrument the ative to manufactures--the strong shift in the com- changes in tariffs in the 1990s with the initial tariffs position of exports shows up even when price that prevailed during 1986­90.The initial tariffs were changes are removed. Further, it was not just due to found to explain 36 percent of the changes in tariffs a few,large high-growth exporters such as China and during the decade: countries with high tariffs in the India. Excluding China and India, the share of man- late 1980s and early 1990s reduced tariffs by a higher ufactures in developing-country exports grew from percentage, while countries with already low tariffs one-tenth in 1980 to almost two-thirds in 2001. It reduced them less.The results also control for some increased sharply, but not equally, in all regions.The other policies that affected growth in the 1990s, laggards included Sub-Saharan Africa and the Mid- including exchange rate policies, government con- dle East and North Africa, which have yet to reach sumption, and inflation. 30 percent. Many countries, particularly the poorest, 5. For example, Dollar and Kraay (2001, 2003); Lee, remain dependent on exports of agricultural and Ricci, and Rigobon (2004); and Alcala and Ciccone resource commodities. (2004) all show a positive relationship between trade 152 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s and growth, whereas Rigobon and Rodrik (2004) reform period of 1982­91 show that the correlation get mixed results.Wacziarg and Welch (2003) find a coefficient between the provincial budgetary rev- positive relationship between a composite measure enue and expenditure is 0.75, compared to 0.17 in of economic reforms and economic growth,but that the prereform period of 1970­79 (Qian 2002). relationship is not significant for the 1990s; nor do Another study (Jin, Qian, and Weingast 2001) found they isolate the role of trade policy per se,but look at that such incentives were indeed significant--for the the composite measure including exchange rate growth of employment of nonstate enterprises and reforms.Their analysis is done in a panel context, in the reform of state enterprises. since they measure the impact of changes in trade 16. Qian (2002); Jin, Qian, andWeingast (2001). policy on economic growth. 17. A result of the increasing import competition from 6. SeeWacziarg andWelch (2003) and Baldwin's (2003) East and Southeast Asian countries that devalued summary of the recent debate on the topic. their currencies in the aftermath of the Asian finan- 7. See Bolaky and Freund (2004).The authors measure cial crisis. excessive regulation using a World Bank survey on 18. These tariffs underestimate true import competition labor regulations and business entry regulations.They since there are also specific tariffs. find that the benefits of expanding trade (as meas- 19. However,imports of several agricultural goods,mak- ured by trade shares) are offset by excessive regula- ing up 40 percent of Indian agricultural GDP, con- tions in the most regulated economies in the 1990s. tinue to be controlled by state trading enterprises. 8. Lerner symmetry in the two-good case can be illus- 20. World Bank (1994b). trated as follows: Px/Pm(1 + t) = [Px/(1 + t)]/Pm, 21. See McMillan, Rodrik, andWelch (2002). where Px = price of exports; Pm = price of imports; 22. Much of the benefits came in the form of export- t = tariff. oriented FDI from the EU member countries 9. Macroeconomic stability refers to the stability of the (World Bank 2000d). real effective exchange rate, as measured by the stan- 23. This discussion is taken fromWorld Bank (2004d);see dard deviation, and average inflation. Government also Stiglitz and the Initiative for Policy Dialogue effectiveness refers to combined perceptions of the (2004). quality of public service provision, the quality of the 24. See Harrison (forthcoming);Winters, McCulloch, bureaucracy, the competence of civil servants, the and McKay (2004); Goldberg and Pavcnik (2005, independence of the civil service from political pres- forthcoming) for comprehensive surveys. sures, and the credibility of the government's com- 25. See, for example, the survey papers by Berg and mitment to policies.The government effectiveness Krueger (2003);Winters, McCulloch, and McKay indicator is taken from Kaufmann, Kraay, and Mas- (2004); and papers by Dollar and Kraay (2001, 2003). truzzi (2003), and based on 17 separate sources of The general conclusion of these papers is that growth subjective data on perceptions of governance con- increases the incomes of the poor, although whether structed by 15 different organizations. or not the effect is neutral across different incomes is 10. See Aitken, Hanson, and Harrison (1997). subject to debate. 11. See Thomas and Nash (1992); and Nash and Takacs 26. Asian Development Bank (2000), cited by Bhagwati (1998). and Srinivasan (2002). 12. Tsikata (2003).This study summarizes the findings of 27. Papers from this volume, which was commissioned diagnostic trade integration studies undertaken dur- by the National Bureau of Economic Research, can ing 2001­03 for 12 least developed countries be viewed online at www.nber.org. (Burundi, Cambodia, Ethiopia, Guinea, Lesotho, 28. Marques and Pagés (1998). Madagascar, Malawi, Mali, Mauritania, Nepal, Sene- 28. Moreira and Najberg (2000).The appreciation of the gal, and the Republic ofYemen). real exchange rate during the period contributed to 13. The discussion in this paragraph is based on Jaffee the negative employment effect by encouraging and Sutherland (2003). imports and undermining exports. 14. Information in this paragraph is from Qian (2000). 30. Lall (1999). 15. The autonomy given to local governments in China 31. Porto (2003). is a very important factor in this development.This 32. SeeWinters et al. (2004) for studies cited. autonomy is provided in the form of the "fiscal con- 33. World Bank (2003e). tracting system"introduced between 1980 and 1993, 34. SeeThomas (2004). under which provincial governments are provided 35. For the study on India, see Topalova (2005). For the incentives to build up local economies and their own study on Colombia, see Goldberg and Pavcnik revenue bases. Specifically, the incentives arise from (2005). allowing the provinces to keep the lion's share of the 36. Balat and Porto (2004). increases in revenues at the margin. Data from the 37. Ashraf, McMillan, and Peterson-Zwane (2005). T R A D E L I B E R A L I Z AT I O N : W H Y S O M U C H C O N T ROV E R S Y ? 153 38. Levinsohn and McMillan (2004). 39. Hanson (2004). 40. Goh and Smarzynska Javorcik (2004). 41. For a review of recent evidence on these links, see Hanson (2004); Goldberg and Pavcnik (2004). 42. U.S. subsidies on cotton amounted to $3.7 billion in 2002 (three times the U.S. aid to Africa). Country Note D The Middle East and North Africa: Performing below Potential E arlier parts of this report high- and the region now faces a number of serious lighted that sustained growth macroeconomic vulnerabilities. In particular, con- experiences result from four func- tingent liabilities have been building up in many tions of growth being met, and that at different countries, from sources such as pension systems, moments,different functions are more relevant than banking sectors,public enterprises,and a variety of others and their absence poses more binding con- implicit and explicit government guarantees. straints. Fulfilling only some of the functions of Although intraregional trade and financial growth results in performance well below poten- flows are small,the countries in the region are tied tial. Countries in the Middle East and Northern together by large-scale labor migration. Other Africa illustrate this point: while they have suc- common characteristics include rapid gains in ceeded in accumulating capital, both physical and social indicators over the last two decades,starting human, and have also succeeded in ensuring equi- from extremely low levels (table D.1). Poverty table distribution,they have not sufficiently empha- declined rapidly in the 1960s, 1970s, and 1980s, sized efficient allocation of resources through but changed little in the 1990s.Income inequality openness and liberalizing their domestic remained relatively low throughout the period. economies. Also, in the last two decades, countries in the The 12 developing countries1 of the Middle region experienced one of the highest labor force East and North Africa have a total population of growth rates in the developing world. Between 260 million,of which the two most populous,the 1990 and 2020, the growth of the economically Arab Republic of Egypt and the Islamic Repub- active population (ages 15­64) will exceed that of lic of Iran, account for about half.2 While the the economically dependent population by a countries of the region vary widely in their natu- much greater amount than in any other region.As ral resources,population density,and stage of eco- experience elsewhere has shown, this rapid nomic and political development, they share a growth presents an opportunity,but also poses the common history and cultural heritage.They also challenge of responding to the employment share a common approach to economic policies: expectations of an increasingly urbanized labor relatively high import protection and a large role force. of government in the economy.Their experiences For most of the last century, because of its oil highlight the role of natural endowments in and natural gas reserves, the world's largest, the growth processes, and the role of openness and region has been the focus of attention from domestic liberalization. industrialized superpowers. Possibly as a result of Countries in the region have generally fol- such strategic importance, armed conflicts have lowed prudent macroeconomic management been frequent over the past four decades.3 principles and have avoided extreme instability of the kind seen in Latin America. Inflation has been Growth Performance moderate and relatively stable, and Argentina-like fluctuations in output have been exceptional. But Though the region has escaped extreme instabil- the fundamentals have in many cases weakened ity,its high output growth of the 1960s and 1970s 156 THE MIDDLE EAST AND NORTH AFRICA: PERFORMING BELOW POTENTIAL 157 TABLE D.1 Progress on Social Indicators, Middle East and North Africa, 1980­2000 1980 1985 1990 1995 2000 Headcount poverty rate, % -- 16.9 17.2 15.9 -- Life expectancy, % 59 64 67 69 70 Infant mortality rate, % 83 -- 42 36 32 Adult literacy rate, % 42 49 57 64 72 Secondary school enrollment, % gross 33 45 49 59 69 Literate female to literate male ratio, ages 15­24 62 69 79 86 92 Source: World Bank 2004f; Adams and Page 2003. Note: Table shows medians of the regional social indicators. Headcount poverty rate is a simple arithmetic average, because the sample size is small: long enough series are available for only five countries: Egypt (1981­82, 1990­91, 1995­96, 1997, 1999­2000); Jordan (1986­87, 1992, 1997); Morocco (1984­85, 1990­91, 1998­99); Tunisia (1984­85, 1990­91, 1998­99); and the Islamic Republic of Iran (1986, 1990, 1994, 1998). Though not reflected in the table, in 2000 poverty rates continued to decline in the countries for which data are available (Egypt and the Islamic Republic of Iran), and rose only in Morocco. Data shown on adult literacy rate and literate female-to-male ratio do not include Djibouti and Lebanon, for which no data are available. --. Not available. gave way to stagnation in the 1980s (figure D.1 and table D.2). Except for Egypt and Tunisia, the FIGURE D.1 countries of the region have not grown fast Median and GDP-Weighted Economic Growth in the enough to reduce the income gap with more Middle East and North Africa, 1961­2000 advanced industrial economies (Country Note 8.0 B,Lessons from CountriesThat Have SustainedTheir 6.0 Growth). Per capita income declined over the 1980­2000 period in several countries,including 4.0 Algeria, Jordan, and virtually all of the Gulf percent 2.0 countries. Natural resources have not ensured economic 0.0 performance (World Bank 2004f). Indeed, over Growth, ­2.0 the last two decades, countries without large oil ­4.0 resources have generally performed better than those rich in oil (tables D.3 and D.4). 1961 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 Labor-abundant countries have performed Year better than those where labor is scarce. In partic- Middle East and North Africa (6) Aggregate ular, in the 1980s, when practically all oil-rich Source: World Bank, World Development Indicators (WDI). economies shrank as oil prices collapsed, the oil- Note: Median growth rates for samples of countries that have complete GDP series. poor, labor-abundant countries were able to sus- Aggregate is growth rate for the region (regional GDP over regional population). All tain growth: Egypt, Morocco, and Tunisia grew data are smoothed with three-year moving averages. Extending the sample of the annually at 2.9, 1.6, and 1.1 percent, respectively, region's countries to six more does not significantly change the results in the later decades. during that decade (table D.3). Jordan was an exception: because of its dependence on remit- tances and financial support from oil-exporting income per capita declined to the point that the countries,its growth in the 1980s dwindled,trail- country is now classified as a developing country) ing that of resource-rich countries. and Oman.The return to peace accounted for much The 1990s saw improvements in performance of the improvement in the Islamic Republic of Iran, virtually everywhere,except in SaudiArabia (whose after the Iran-Iraq war,and in Lebanon andYemen, 158 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s TABLE D.2 Economic Growth in the Middle East and North Africa, 1960­2003 (median GDP per capita growth) 1960 1970 1980 1990 1990s* 6 countries 2.4 3.6 1.3 1.4 1.5 9 countries ­0.2 0.6 0.9 12 countries 1.2 1.2 All developing countries (69) 2.0 1.8 ­0.5 1.3 1.0 Developing countries (78) -- 1.9 ­0.3 1.0 1.0 Developing countries (93) -- -- ­0.2 1.0 0.9 Source: WDI. Note: Different country groups correspond to the periods for which data exist for all countries. For example, there are only 6 countries in the Mid- dle East and Northern Africa for which GDP statistics exist since 1960, 9 for which they exist since 1980, and 12 for which they exist since 1990. 1990s* means 1990s including 2001 and 2002. 6 countries: Egypt, Arab Rep. of; Morocco; Tunisia; Algeria; Syrian Arab Rep.; Oman. 9 countries: Egypt, Arab Rep. of; Morocco; Tunisia; Jordan; Algeria; Syrian Arab Rep.; Iran, Islamic Rep. of; Oman; Saudi Arabia. 12 countries: Egypt, Arab Rep. of; Morocco; Tunisia; Jordan; Lebanon; Djibouti; Algeria; Syrian Arab Rep.; Iran, Islamic Rep. of; Yemen, Rep. of; Oman; Saudi Arabia. --. Not available. TABLE D.3 from public institutions that are able to maintain Three Major Middle Eastern and North African Country law and order and exert the state's authority.The Groups energy of the bureaucracy, however, has often Abundant labor Scarce labor focused more on controlling the allocation of resources than on supporting private sector ini- Rich in Algeria, Syrian Arab Rep., Oman, Saudi Arabia tiatives and a competitive economy. Many coun- resources Iran, Islamic Rep. of, Yemen, tries in the region have been hesitant to embrace Poor in Egypt, Arab Rep. of, economic openness and competition, and some resources Morocco, Tunisia, Jordan, Lebanon, Djibouti analysts believe that political liberalization is needed to address governance issues and create Source: World Bank 2004f. an investment climate that is more predictable and more conducive to growth. after civil conflicts. In Jordan, improvement took Middle Eastern and North African countries place as the result of ambitious reforms that opened have generally fulfilled two of the central func- and liberalized the economy in the 1980s, even tions of growth: accumulation and distribution. though political uncertainties and restrictions on the Their investment rates have been high compared country's access to external markets kept the coun- with those of other developing-country groups try's growth below its potential (Khalaf Hunaidi, in (figure D.2).The Middle East and North Africa's World Bank 2005b). "social contract," with the state dominating the economy, allowed it to mobilize significant resources for investment, particularly when oil Fulfilled and Unfulfilled Functions of prices skyrocketed in the 1970s.Throughout the Growth 1970s and early 1980s, investment rates in the RecentWorld Bank reports have highlighted the region were comparable with those in the eight importance of governance reforms,to enable the high-performing East Asian economies. Invest- countries in the region to grow faster and more ment rates were high not only in the oil-rich equitably (World Bank 2003a­d). Middle East- countries--such as Algeria with a two-decade ern and North African countries have benefited average of 38 percent of gross domestic product THE MIDDLE EAST AND NORTH AFRICA: PERFORMING BELOW POTENTIAL 159 TABLE D.4 Economic Growth in the Middle East and North Africa: Impact of Natural Resources 1990 1980 1970 1960 Resource-poor with abundant labor Egypt, Arab Rep. of 2.3 2.9 4.4 2.9 Morocco 0.4 1.6 2.7 2.0 Tunisia 3.1 1.1 5.0 2.8 Jordan 0.6 ­1.8 -- -- Lebanon 5.3 -- -- -- Djibouti ­4.0 -- -- -- Resource-rich with abundant labor Algeria ­0.3 ­0.2 2.8 1.2 Syrian Arab Rep. 2.1 ­1.1 6.4 2.0 Iran, Islamic Rep. of 2.5 ­0.7 -- -- Yemen, Republic of 1.7 -- -- -- Resource-rich, labor importing Oman 0.6 4.7 1.2 16.2 Saudi Arabia 0.0 ­5.7 7.9 -- Source: WDI. --. Not available. (GDP)--but also in Jordan (31 percent),Tunisia (28 percent), and Egypt (26 percent). From an international perspective, these investment rates are extremely high. Even after the collapse of oil prices, and ensuing declines in investment rates FIGURE D.2 in recent years, they have been comparable to Investment in the Middle East and North Africa, those in the high-performing East Asian 1960­2002 economies. 40 Middle Eastern and North African countries have also invested large amounts in human capi- a 30 tal.They have dramatically reduced infant mor- GDP as of tality,raised life expectancy,and expanded school 20 enrollment.Their literacy rates have increased significantly, including for women in countries 10 such as Algeria,the Islamic Republic of Iran,Jor- Investment percentage dan, and Tunisia, and are now above those in many developing countries at similar levels of 1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 income. Year Most governments in the Middle East and Middle East and HPEAP (8) Other developing North Africa, perhaps with the exception of North Africa (5) countries (58) Morocco,have been highly redistributive.Distrib- Source: WDI. ution took place through a variety of programs, including provision of health and education, sub- Note: HPEAP designates high-performing East Asian economies; Other developing countries (58) is the sample of 58 countries for which data are available for this sidies for housing and for consumption items period. Extending the sample of Middle East and North African countries to six more including bread and transport, scholarships, and does not significantly change the results in the later decades. even jobs in the public sector. In the 1990s, how- 160 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s ever,as part of fiscal adjustment programs,many of the subsidies were phased out. FIGURE D.3 Macroeconomic stabilizations in the late Tariffs in the Middle East and North Africa, 1980­2000 1980s reduced inflation and debt and reduced (median of unweighted average tariffs) the need for external inflows. Fiscal positions were consolidated and economies recovered 40 from recessions. But for a number of reasons 30 growth in the region has failed to regain its pre- 20 1979 levels. Median Combined with the high investment rates in 10 physical and human capital, the disappointing 0 growth performance suggests that productivity 1980­85 1986­90 1991­95 1995­2000 Year growth was negative. Negative productivity Resource-poor with abundant labor growth may be the most important reason why Resource-rich with abundant labor countries in the Middle East and North Africa Resource-rich, importing labor have performed less well than countries in, for Source: World Bank, World Integrated Trade Solution database. example, East and South Asia. There are three main reasons for the low pro- ductivity of investments in the Middle East and The low productivity of investments belies North Africa. First is the dominance of produc- several waves of reform in the region's trade tion by the state, which typically uses resources regimes. Notwithstanding the reforms, invest- less efficiently than the private sector. ments tended to be allocated inefficiently and as Second, the region's tariff barriers are among a result, the growth payoff to large investments the highest in the world.For countries with lim- was low. ited domestic markets, import substitution poli- Tunisia,Egypt,and Morocco introduced their cies quickly outlive their usefulness. In the early first trade-related reforms in the 1970s (figure 1980s, the region's tariffs were quite low com- D.4).Tunisia was able to significantly accelerate pared with those in other developing regions. its growth rate in the 1970s by creating export- But they have remained at these levels, while processing zones insulated from the rest of the those in other regions have now been dramati- economy,and it continued with gradual but per- cally reduced.Average tariff rates in Algeria, the sistent steps to liberalize the economy (Oliva Islamic Republic of Iran,Jordan,Libya,Morocco, 2000).Egypt followed the route of internal liber- Saudi Arabia, the Syrian Arab Republic, and alization first,drawing on the potential of its large Tunisia have either increased somewhat or internal market; the pace of reform was much remained constant since the late 1980s (figure slower and the approaches more sporadic. Dur- D.3) (Oliva 2000).4 ing the 1990s, trade volume rose in Tunisia but Third, Middle Eastern and North African fell in Egypt.Morocco joined the GeneralAgree- countries maintain domestic restrictions on pri- ment onTariffs andTrade in 1987 and proceeded vate investments. Domestic restrictions on pri- to liberalize its financial sector, privatize state- vate investments are not always explicit and the owned enterprises, and rationalize its tax system. lack of a vibrant and developing private sector is Openness steadily increased in both Morocco not always the result of state monopolies. Red and Tunisia, as did the share of manufactured tape,the inefficiency of the judiciary,corruption, goods in these countries' exports, but only in and state capture of government regulation all Tunisia did increased openness translate into high work to deter private investment. And the growth (figure D.5). absence of clear directions on the future evolu- In Lebanon and in labor-abundant and tion of policy creates uncertainty, which further resource-rich economies--Republic of Yemen, limits private investment. Algeria,Syria,and the Islamic Republic of Iran-- THE MIDDLE EAST AND NORTH AFRICA: PERFORMING BELOW POTENTIAL 161 FIGURE D.4 Trade Outcomes in Egypt, Morocco, and Tunisia, 1960­2000 100 250 80 180 120 180 160 160 80 200 100 60 140 140 120 80 120 60 150 100 100 40 60 80 80 40 100 60 40 60 20 20 50 40 40 20 20 20 0 0 0 0 0 0 1960 1965 1970 1975 1980 1985 1990 1995 2000 1960 1965 1970 1975 1980 1985 1990 1995 2000 1960 1965 1970 1975 1980 1985 1990 1995 2000 Trade, % gross domestic product Trade, % gross domestic product Trade, % gross domestic product Real effective exchange rate, 1990 = 100 Real effective exchange rate, 1990 = 100 Real effective exchange rate, 1990 = 100 Source: WDI. FIGURE D.5 Diversification in Egypt, Morocco, and Tunisia, 1960­2002 (manufactured exports as a percentage of exports of goods and nonfactor services) 60 80 100 80 60 40 60 40 40 20 20 20 0 0 0 1960 1966 1972 1978 1984 1990 1996 2002 1960 1966 1972 1978 1984 1990 1996 2002 1960 1966 1972 1978 1984 1990 1996 2002 Source: WDI. the approach to trade reform has been more grad- Islamic Republic of Iran's exports remain com- ual and haphazard.The Islamic Republic of Iran, pletely undiversified and the manufacturing sector after the Iran-Iraq war, implemented some inter- itself shows little sign of viability (Tabibian 2003). nal liberalization that,as in Egypt,resulted in high The sustainability of the country's liberalization output growth but not much international inte- measures is unclear (Esfahani 2002). gration.In the Islamic Republic of Iran,it is diffi- cult to disentangle the effects of liberalization Conclusion from those of postwar "reconstruction," disman- tling of wartime price controls, gradual and spo- The experience of developing countries in the radic reforms,and higher oil revenues.But despite Middle East and North Africa illustrates the decades of state support to manufacturing, the importance of maintaining a balance among the 162 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s different functions of growth, and that if a Morocco,Tunisia, Jordan, Lebanon, Oman, Djibouti, country succeeds in fulfilling one function, it Saudi Arabia. 2. The region's performance and development chal- may not achieve its full potential unless it can lenges have been analyzed in a series of World Bank fulfill the other functions as well.The achieve- reports (World Bank 2003b, 2003f, 2003j, 2003l), on ment of sustained high growth in the region which this country note draws extensively. will require more clarity on the future direc- 3. The number of conflicts is at par with that in Sub- tions of policies and considerable domestic and Saharan Africa, where the number of countries is external liberalization. three times that in the Middle East and North Africa (World Bank 2003b, 2003f, 2003j, 2003l). 4. Tariffs increased for various reasons.Those in Saudi Notes Arabia rose by between 8 and 12 percent a year in response to balance of payments pressures, and 1. Algeria, Syrian Arab Republic, Islamic Republic of those in Morocco rose as the result of quota tarif- Iran, Republic of Yemen, Arab Republic of Egypt, fication. Chapter 6 Privatization and Deregulation: A PushToo Far? G OVERNMENTS IN THE 1990S thirds of the respondents to a 2002 survey in 17 traded the commanding Latin American countries agreed that"privatiza- heights of their economies for tion of state companies has not been beneficial" more nearly free marketplaces. Despite many (up from 43 percent in 1998). Even in the successes, state-owned firms had frequently United States,some commentators ask if current become inefficient, overstaffed, and a drain on bankruptcies in airlines and telecom can be public budgets. Earlier attempts short of privati- traced to earlier deregulation. Skeptics cite the zation had failed to improve their operations. impressive economic growth of India and Governments recognized that they might have China, where the government's role in allocat- taken on a role that they could not adequately ing resources has been reduced and this change fill and that a greater reliance on markets would has been popular. be beneficial. This chapter first describes the background Privatization and deregulation were often to the deregulation movement (in section 1) parts of a broader set of reforms to improve eco- and then outlines the efforts made during the nomic efficiency, and their speed and extent 1990s to privatize state-owned firms, especially reflected individual countries' convictions and in the transition countries (section 2). Studies circumstances. More rapid and widespread in that evaluate the experience with privatization Latin America than in Africa or South Asia, the are reviewed in section 3; they all find that ben- privatization process also varied by sectors. Its efits have followed.Section 4 focuses on the pri- proximate causes differed.Some countries sought vatization of infrastructure and other utilities. It greater operating efficiency; others, fiscal rev- finds that, contrary to some perceptions, priva- enues,and many acted under pressure from inter- tizing utilities did not hurt the poor.Consumers national lenders. Utilities were sold because with access (a few of them poor) paid more capacity was fast becoming a bottleneck and gov- when prices were raised, but they benefited ernments lacked the needed funds to invest. In when service improved, as it did by any physical Eastern Europe and the former Soviet Union, measure of performance. Expectations on the privatization was central to the transition to mar- role that the private sector could play in infra- ket economies, and was part of a much wider structure clearly proved unrealistic, however. process of societal change. Section 5 analyzes recent attempts in Latin Many observers now ask if privatization and America and Eastern Europe to increase the deregulation were pushed too far.Today's dissat- private sector's involvement in the provision of isfaction is not limited to countries where, as in pensions and social security; it finds instances the Russian Federation, a few well-connected where privatization may have been pushed too people took over some large firms cheaply.Two- far. Section 6 concludes the chapter. 163 164 ECONOMIC GROWTH IN THE 1990s 1. Privatization in Market productivity increased: passenger miles doubled Economies with only half as many more employees between 1977 and 1987. Sometimes it took strikes and In the 1950s and 1960s,governments in develop- bankruptcy filings to shake off inefficient working ing countries sought the commanding heights of practices that were embedded in company cultures their economies to promote economic develop- and union agreements. Some aviation pioneers ment, and many newly independent countries, could not adapt--PanAm,TWA,and Eastern went seeking to assert their authority, nationalized out of business--but others emerged such as firms that belonged to their erstwhile colonial Southwest, with its vigorous low-cost and cus- masters.State ownership was also thought to pro- tomer-friendly culture. mote development in areas where the private This success emboldened the U.S. govern- sector was too risk-averse or myopic to see the ment to tackle AT&T. In 1984 the Department latent, untapped profit. State-owned firms coex- of Justice broke up American Telephone and isted with privately owned ones. Even in the Telegraph (or "Ma Bell") into one long-distance United States,the government had been taking a and seven regional firms (the Baby Bells) offering major role in the economy since the 1930s'Great local services. Figure 6.1 shows how customers Depression, but generally did so through regula- switched in increasing numbers to MCI and tion, not outright ownership.1 other competitors. By 1996, the price of a tele- The move to privatize stemmed more from phone call per minute was only 40 percent of its pragmatism than ideology.Attempts to improve 1984 level. failing public enterprises, for example, through When mobile telephony arrived, eliminating professional managers, independent boards of the rationale to regulate (there was no longer a directors, or performance contracts, had not suc- fixed line creating a natural monopoly), the Fed- ceeded. By the end of the 1980s, reformers had eral Communications Commission had already reached broad agreement that nothing short of learnt its regulatory lesson: it auctioned the nec- privatization would do. essary radio spectrum, enriching the U.S.Trea- sury by $9 billion, and ensured that there was adequate competition. Most of the 100 largest The (American) Regulatory Revolution metropolitan areas in the United States now have The impetus behind the privatizations of the at least five cellular telephone providers. Public 1990s began decades earlier in the United States, awareness of the benefits from deregulation where a regulatory revolution starting in the late greatly increased.2 1970s won over many economists, and visibly improved consumer prices and services in air Spread of the PrivatizationTrend travel, telecom, and other industries. The first bold step was the deregulation of the The United Kingdom followed the United airline industry, starting with the abolition of the States' lead. In the 1970s and 1980s, the govern- Civil Aeronautics Board in 1978.Real fares halved ment privatized firms in coal,steel,railroads,tele- between 1978 and 2000, and service improved in com,3 electricity, and even water, despite fierce ways the public valued,such as flight schedules and opposition from militant trade unions, and when frequency.Airlines became more efficient in a vari- the initially skeptical public was won over, the ety of unforeseen ways.They bought more fuel- Labor government that succeeded Mrs. efficient fleets and developed computer reservation Thatcher's Conservatives did not reverse course. systems and statistical models to price-discriminate The rest of Europe was slower to privatize, but among different types of passengers to fill all flights. the rules introduced by the European Union to They developed the hub-and-spoke system to bal- create a single market limited subsidies to loss- ance the public demand for frequent flights against making firms, and this led to many firms being the capacity and fuel economy of their fleets.Their sold. Italy dismantled and began selling parts of PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 165 FIGURE 6.1 Market Shares of Revenues for U.S. Long-Distance Carriers, 1984­98 (percent) 100 90 80 70 60 50 Percent 40 30 20 10 0 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 AT&T MCI Sprint Others Source: Crandall and Hazlett 2000. IRI, and France and Germany sold part of their observes that in Africa often"the principal moti- telecom stakes in the late 1990s.4 vation for privatization has been to placate the Some Latin American countries began a international financial institutions."An ideologi- deregulation and privatization process as early as cal suspicion of capitalism, perceived to be hand the 1970s. Chile removed a panoply of controls in glove with colonialism, or the influence of that had accumulated over decades,and privatized vested interests, and sometimes both, delayed utilities and even social security,in part because of beneficial changes. But as governments aban- its economic plight in the 1970s and the reform- doned trade protection and indirect subsidies (for ers' academic links to American universities. example, directed lending from banks that the Chile's subsequent economic growth emboldened state controlled), the firms' losses and inefficien- Mexico,Brazil,andArgentina to follow suit,devel- cies became more apparent. oping techniques such as debt-to-equity swaps to Thus by the start of the 1990s, privatization extinguish their accrued external debts.Countries and deregulation were well under way in Latin tended to privatize firms in competitive industries America along with a growing, if controversial, before tackling infrastructure and utilities. Chile move to do the same elsewhere.6 The trickle privatized telephones in 1990, and Mexico had became a flood with the collapse of central plan- sold 361 of some 1,200 state-owned enterprises ning in Eastern Europe and the former Soviet by 1992, thereby virtually eliminating subsidies Union. that had amounted to almost 14 percent of its gross domestic product (GDP).5 Countries in Africa have privatized much less 2. The Great Transition Sale: than Latin American or transition countries, and Haste to Avoid Waste7 more reluctantly. Between 1991 and 2001, roughly 2,300 privatizations worth $9.1 billion The scale and speed of privatization in the transi- affected fewer than 40 percent of Africa's state- tion countries of Eastern Europe and the former owned enterprises. Just four countries--Ghana, Soviet Union were vastly greater than in the mar- Côte d'Ivoire, Nigeria, and Zambia--accounted ket economies,and the process spawned new tech- for a third of these transactions. Nellis (2003a) niques. 166 ECONOMIC GROWTH IN THE 1990s When centrally planned economies collapsed and would they pass with the firm or with the and turned to markets, the need for privatization building? State ownership of the land is a mean- was self-evident, and most of the discussion cen- ingless concept if the land is automatically trans- tered on how this should be accomplished.Advo- ferred with either the enterprise or the structure, cates of rapid privatization won the day, but but if the land is not part of the package,how can critics now question the haste and point to mis- the firm be sold? takes that were made: assets were sold to cronies Restitution claims added another dimension at low prices, and the many institutions that are of complexity and confusion. Communist gov- vital to supporting the market have been slow to ernments had confiscated properties and their develop.8 Minority shareholders have few pro- former owners and heirs wanted them back. tections, and privatization sometimes resulted in Some claimants had left the country but were the new owners stripping assets and spiriting willing to return; many others had remained, them abroad rather than investing to improve some of them still working on the confiscated their working. farms and firms. The situation was even more Before judging this experience, however, it is complicated in countries such as Hungary, important to appreciate how chaotic and dys- Poland,and Ukraine where earlier rounds of con- functional state ownership had been. fiscation under Nazi occupation made it harder to determine whose claims to honor and, where national boundaries had also shifted during and Opening Pandora's Box after World War II, whose laws applied. Some- State ownership in the former Soviet Union times the records were unclear or lost.Firms were countries was far murkier and complicated than the more pressing issue because they were strug- the term suggests.There was a fundamental dif- gling for survival:GDP in transition countries was ference between selling state-owned firms in falling rapidly and people were undergoing great countries where markets function and support- hardship. Since the Soviet Union had often ing institutions exist and doing the same in coun- located firms strategically, ignoring transport tries where the state collectively owns all assets.9 costs,its division into 15 republics wreaked havoc In the latter, laws protecting private property on firms that were freed to buy or sell as they could be readily passed, but enforcing them pleased. Many found that their suppliers and/or required a daunting array of institutions that buyers had become foreigners overnight. Borders communism had destroyed.A state that is strong were erected where none existed, and tolls were enough to protect private property is also strong extracted as each republic sought revenues from enough to confiscate it, especially when its customs duty to pay for basic services such as the finances are desperate.The former East Germany police. could rely on West Germany's institutions after Firms scrambled to cope. Many managers the unification,but other transition countries did were part of the nomenklatura, and while some not have this advantage. sought political protection and favors on the While the law may state that the state owns firm's behalf, others did the reverse. Some man- something (or everything), somebody (it is never agers resisted the changes, others adapted, and clear who) controls its use, and such usufructuary some exploited the situation to their personal rights have value despite the absence of a market. advantage.Survival often required laws to be bro- In the former Yugoslavia, for example, a firm's ken, and smugglers and criminals prospered, cor- workers "clearly" owned the firms, and a portion roding values, politics, and societies. The of their wage funds financed the investments, but breakdown of credit and central plans led to the did retirees who had financed the older machines increasing use of barter, shortfalls in cash led to also have claims? The land under the firm workers being paid in shoes or commodities,and "clearly" belonged to the state, not the firm, but machinery was scavenged to produce products did the usufructuary rights transfer automatically that could actually be sold. PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 167 This then was the background against which difference between China and most countries in the discussions on privatization took place.10The Eastern Europe and the former Soviet Union situation did not allow for a fully informed was the continuing power of China's Commu- debate, and even participants who agreed about nist Party despite the internal political turmoil. what to do often disagreed over the reasons. Changes in the relations between the central and Some argued that the creation of a market to local governments and in their control over replace central planning did not necessarily enterprises often amounted to privatization in all require firms to be privatized, or at least not but name, and the process was just as turbulent immediately. Janos Kornai (1990), the eminent and nontransparent as privatization in the transi- Hungarian economist, warned that the state tion countries. should guard the wealth it was entrusted with until a more responsible owner came along, but The Process such a course was impossible in countries where the government had lost effective control. Other The "when" to privatize was quickly settled: the protagonists favored rapid privatization: some to sooner the better,although the"how"really set the prevent the asset stripping that was becoming pace. blatant, and others because it was "necessary to The decision on what to privatize was also create a market economy" in the sense that pri- perfunctory: as much as possible. No distinction vatization was expected to create a demand for was drawn between regulated industries and more market-supporting institutions.11 Some competitive ones, although a few countries who favored restructuring the firms before sell- enacted but could not properly enforce antimo- ing them to raise more revenue were opposed by nopoly laws.12 Indeed, economists' forays into those who lacked confidence in the managers. political science were unchallenged when they The experience of the Treuhand, created to claimed that windows of opportunity might sell the state-owned assets of East German soon slam shut.13 Telecom, particularly licenses firms after the fall of the Berlin Wall in 1989, for cellular telephony, was an especially lucrative provided early lessons. The Treuhand had to business to privatize, and thus many telecom cope with the economic effects of the political licenses were sold with exclusivity provisions to decision to unify the two Germanys: an over- increase revenue despite the lower welfare valued currency that reduced competitiveness implied. and inflexible labor laws and practices that Transition countries did not privatize their exacerbated the resulting unemployment.Asset banks as they did firms, not because of a grand values eroded quickly through neglect. Simple plan but more because buyers would not be and quick sales were better, and imposing addi- forthcoming until governments dealt with the tional and largely unenforceable requirements mountains of nonperforming loans and the (employment maintenance clauses and/or banks' negative net worth. Inflation reduced the investment requirements) greatly reduced stock of deposits, and the public was wary of potential buyers' interest. It was inefficient and entrusting banks with savings.As a result, banks pointless for the government to restructure a had little to lend--a condition that may have firm in hopes of a better price, as restructuring helped harden the budget constraint on firms, could often be done better by buyers. quite independently of firms'privatization per se. China's experience was often cited to argue Diverse decisions were made on how to priva- that transformation and growth could be tize. Conventional techniques were clearly inade- achieved without privatization, but the differ- quate to the task at hand.The United Kingdom ence may lie more in labels than in substance divested 20 firms in 10 years and Mexico 150 in (box 6.1), for China's experience in practice six.The transition countries required an approach underscores the importance of incentives and the that could do much more, and faster. Poland had growth of private enterprise.The fundamental 8,400 state-owned enterprises accounting for 70 168 ECONOMIC GROWTH IN THE 1990s BOX 6.1 China: Stealth Privatization B efore 1978, China's plans covered almost ucts and SOEs were permitted to have joint ventures every decision of a state-owned firm--out- with private foreign investors. put, pricing, investment, working capital, These joint ventures amounted to a form of labor use and wagesæbut unlike the Soviet system, stealth privatization. SOEs contributed productive China allowed a great deal of local autonomy in prac- assets or space in exchange for equity in a joint ven- tice. Although the state-owned enterprises (SOEs) ture; foreign investors provided funds, newer machin- were owned by the central government, many were ery, and management expertise. SOEs often hold effectively controlled by provincial or municipal gov- social assets (such as cafeterias, housing) and show ernments, because the center set highly aggregated losses that mask the joint ventures' profitability. production targets and credit flows. While many observers have credited TVEs as the Reforms began in agriculture with the "household driver of China's growth, comparative studies of responsibility system." The initial move was to allo- provinces show that (1) TVEs flourished more when cate land among the 20 constituent households of a genuine private firms were prevented from emerging, commune and allow each to sell more than the con- and (2) provinces with genuine private firms grew tracted grain-procurement quota at uncontrolled faster (Huang 2003). This "private sector" (both pri- prices, and to keep the proceeds. Deng's now famous vate firms and firms in the government's statistical remark that the color of the cat did not matter so long category that includes TVEs) emerged before legal as it caught mice assuaged fears that this small move restrictions on its existence were eased. Provincial away from collectivist orthodoxy would be crushed. governments often tolerated, and sometimes encour- Output and incomes rose, and by 1982, this system aged, its operations, notwithstanding laws to the had been adopted by 80 percent of China's rural house- contrary. The TVEs could not obtain financing from holds and had spread to manufacturing. banks, but nevertheless managed to thrive. Although Similarly in manufacturing, the "management the state-owned banks only lend to SOEs, the SOEs in responsibility system" evolved to give state enter- turn fund private suppliers and joint ventures. prises more autonomy over their operations. Firms China's experience also illustrates De Soto's (2000) negotiated their own arrangements for limited profit important distinction between de facto and de jure retention, dual-track pricing, and some investment ownership. While it would be ideal if both were con- autonomy, which gave them incentives. Within gruent, China illustrates the importance of de facto firms, the relations and authority of technical man- protection. In contrast, most East European transi- agers and party officials shifted to reflect personali- tion countries emphasized de jure protection. Perhaps ties and changes in the powers of central and the absence of laws in China did not deter entrepre- provincial governments. Municipality-owned town- neurship because the presence of laws had not pro- ship and village enterprises (TVEs) expanded their tected people during the Great Leap Forward, the production of highly profitable light industrial prod- Cultural Revolution, or other tumultuous episodes. PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 169 to 80 percent of GDP at the outset;but many spun plex. The government prints and distributes off subsidiaries,and it was hard to keep track of the vouchers (free or,to prevent their being scorned, changing numbers and sizes. for a token sum), perhaps unequally to favor Since many firms were too large for any small some groups (such as military veterans, widows). group of investors to buy, and--more impor- The vouchers are then used in lieu of cash or as tant--workers and managers could disrupt and supplements, to bid for firms being auctioned to deter unwelcome buyers, employee ownership the highest bidders (with many variants--for arose almost out of necessity in all transition example, allowing sealed bids for controlling countries. Poland's 1990 Privatization Law, for interests).The government thereby exchanges its example, required managers and the workers' equity in state firms for vouchers that are then council to agree before a firm could be sold.Dis- extinguished. putes then arose over whether diffused ownership Many commentators extolled the virtues of in general,and employee ownership in particular, mass privatization,15 noting that auctions allowed was a good idea.While most commentators con- a firm to be sold to the buyers most likely to add ceded that workers should be given some shares, value, and that vouchers allowed governments to especially if doing so would stave off unrest and separate distributive from efficiency considera- opposition, only a few wanted all firms given tions.Vouchers could be allotted to reflect the entirely to their employees--especially since this government's distributive desires, and secondary would shortchange workers in those firms facing market trading (of vouchers or of shares on the a bleak future who would, besides losing their stock exchange) would result in efficient cluster- jobs, have only worthless scrip. ing of owners and holdings of firms.It was hoped Having mutual funds as intermediaries that voucher schemes would act as the seedling seemed sensible, and it was heroically thought of the stock exchange.16 that the funds would jumpstart a stock exchange Poland was the first to consider vouchers, in that would permit holdings to be subsequently 1989,but it did not introduce them until 1995,so reshuffled, should owners so desire (Bell 1995). in practice the use of vouchers is indelibly associ- This proposal satisfied the proponents of broadly ated with the Czech Republic.Many other coun- based share distribution as well as those who tries, including Albania, Estonia, Georgia, sought incentives for committed owners.It was a Mongolia, and Russia, introduced variants of the source of pride that theWarsaw Stock Exchange, Czech voucher scheme that each differed in first established in 1871,was reopened in the for- important details.The Czech Republic allowed, mer Communist Party headquarters in 1991 to and tacitly encouraged, the creation of mutual trade the equity of five privatized firms. But funds so that the public need not have to select although the U.S. Agency for International from among the thousands of firms for sale.It was Development (USAID) and other donors lav- thought that these 20 or so funds would oversee ished technical assistance, most of the exchanges the firms, because their expertise and sizable that were established in the transition countries stakes would help ensure that firms were well run. have since atrophied.14 In contrast, Poland's government organized the All these discussions took place separately but funds itself,but critics contend that this procedure not simultaneously or identically in each country, was hardly an improvement over the ministries with everyone looking over others'shoulders to see that had overseen firms in the previous era. what was being done elsewhere. Many countries The outcomes attributable to voucher schemes adopted similar techniques with some variations. are not impressive.The redistribution of wealth that is possible through vouchers is minor and may not have been worth the effort. Benefits were not Mass Privatization throughVouchers as widely distributed as might appear: most recip- Mass privatization through vouchers is simple in ients sold their vouchers for a fraction of their face concept even though its administration is com- value,17 often to the firms' incumbent managers. 170 ECONOMIC GROWTH IN THE 1990s Privatization created value, but it generally changes requires data that are difficult to com- accrued to the controlling owner,and other own- pile. Hence reliable studies are difficult and ers received little. expensive, and lag privatization by several years. The discussion below draws on comprehen- sive literature surveys by Megginson and Netter 3. The Results: Gains, but (2001); Djankov and Murell (2002); and Nellis Controversies As Well (2002, 2003a, 2003b). One strand looks at how firms in market economies fared after privatiza- In general, large benefits followed privatization, tion and who gained and who lost. Another even though they differed across countries and strand looks less carefully (because data are less stakeholders.Workers and consumers could ben- complete and reliable) at outcomes in transition efit through higher wages and lower product costs economies. even where firms were sold too cheaply,but there is also evidence that the benefits have been greater Evidence from Market Economies when privatization has been transparent and con- ducted fairly. In a thorough study of the privatization of 12 Because many things changed simultaneously, major firms in four different countries, Galal et however, the benefits that followed privatization al. (1994) find substantial net gains (averaging 30 are not proof that privatization was their cause. percent of predivestiture sales) in all but one More studies will not resolve the debate because (table 6.1).Workers always gained, as did owners empirical work cannot disentangle the effects of and governments, and consumers benefited in each of the many changes that accompanied pri- half the cases. vatization, and theory is not decisively against Megginson, Nash, and van Randenborgh government ownership. (1998) examine a different but larger sample and Nor is it clear how the benefits arose:for exam- find substantially the same results.The postpriva- ple, did owners oversee managers better or did tization performance of 61 firms in 18 countries lenders stop financing losers?Was it because man- (6 developing and 12 industrial) showed substan- agers had the freedom (or incentive) to do their tial improvements in different measures of effi- job, or because private owners demanded (and so ciency: profit margins were higher and so were obtained) better managers,or because overstaffing inflation-adjusted sales per employee, as well as was reduced, or because soft budget constraints the ratio of capital expenditures to sales.Employ- were hardened?The reasons could differ by firm, ment increased after privatization by an average but if so, they cannot be gleaned from aggregate of 6 percent. data. Even if privately owned firms fare better Boubakri and Cosset (1998) find similar than state-owned firms, is this performance due results from the same type of study, covering to a selection bias,whereby only those firms with more firms.Their sample of 79 newly privatized a better potential were sold? firms between 1980 and 1992 included many Evaluating benefits is also a complex task from low-income countries.While profitability because the substantial transfers among different and efficiency rose significantly, they rose more stakeholders vary over time, often for reasons, in upper middle-income countries than in low- such as business cycles, that are unrelated to the income countries. sale. Owners' gains can be measured by profits, Nellis (2003b) surveys studies of Latin Ameri- dividends, and/or equity prices (which may not can privatization, one of which, dealing with move together); workers' gains can be measured Mexico, can be singled out to illustrate the typi- through wages, but one must also correct for cal findings. LaPorta and López de Silanes (1999) changing employment and skill mix, which is analyze the pre- and postprivatization perform- hard to do.Workers losing their jobs often find ance of 218 Mexican firms in 26 different indus- alternative employment, but tracking these tries that were privatized between 1983 and 1991. PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 171 TABLE 6.1 Winners and Losers from Divestiture in 12 Case Studies (percent) Domestic Foreign World Net net welfare welfare Country and enterprise Govt. Buyers Consumers Workersa Others change Buyers Consumers Others change U.K. British Telecom 2.7 3.1 4.9 0.2 ­0.1 10.8 1.2 0.0 0.0 12.0 British Airways 0.9 1.4 ­0.9 0.3 0.0 1.7 0.4 ­0.5 0.0 1.6 National Freight ­0.2 0.8 0.0 3.7 0.0 4.3 0.0 0.0 0.0 4.3 Chile CHILGENER ­1.4 2.0 0.0 0.1 0.0 0.7 1.4 0.0 0.0 2.1 ENERSIS ­1.6 7.6 2.2 3.9 ­7.4 4.6 0.6 0.0 0.0 5.2 CTC 8.0 1.0 131.0 1.0 4.0 145.0 10.0 0.0 0.0 155.0 Malaysia Malaysian Airline Systems 5.2 2.0 ­2.9 0.4 0.0 4.6 0.8 0.8 15.8 22.1 Kelang Container Terminal 37.6 11.5 6.2 7.0 ­11.9 50.4 2.9 3.1 ­3.0 53.4 Sports Toto Malaysia 13.6 10.7 0.0 0.0 ­13.0 10.9 0.0 0.0 0.0 10.9 México Teléfonos de México 13.3 11.4 ­62.0 15.6 28.3 6.6 25.1 0.0 17.9 49.5 Aeroméxico 62.3 3.9 ­14.6 2.4 ­2.3 52.9 1.8 ­6.2 0.0 48.5 Mexicana de Aviación 3.5 ­1.4 ­7.7 0.0 3.2 ­2.4 ­1.3 ­3.3 0.0 ­7.0 Source: Galal et al. 1994. Note: All figures are the annual component of the perpetuity equivalent to the welfare change, expressed as a percentage of annual sales in the last predivestiture year. a. Includes workers both in their role as wage earners and as buyers of shares. The authors find a 24 percent increase in average job to another; certainly finding alternative profitability,as measured by the ratio of operating employment that adds value is easier in a grow- income to sales, arising from increases in produc- ing economy.To find the full effects, one must tivity (57 percent), labor retrenchment (33 per- look at employment beyond the privatized firms. cent), and price increases (10 percent) (table 6.2). No study has tracked displaced workers in such a Profitability rose more in competitive sectors than manner in noncompetitive sectors; so these gains did not Nellis (2003a) surveys the studies of African arise from an increase in monopoly power. Mex- privatization and finds impressive benefits. A ico had some 1,155 state-owned firms in 1982, 2001 study commissioned by the Zambian Pri- accounting for about 14 percent of GDP;and the vatization Agency found that 235 of the 254 government sold 150,liquidated 260,and merged enterprises privatized since 1991 continued to an additional 400 firms by 1988--all before the operate. The investments in nonmining firms privatization wave in the 1990s with infrastruc- were worth more than $400 million; but the ture firms including telecom. largest deals had been in mines where improve- In Argentina, some 150,000 workers were ments did not materialize,a result that warrants a dismissed between 1987 and 1997 following pri- brief explanation.Zambia nationalized its copper vatization; 90,000 workers were dismissed when mines shortly after independence in 1964.The Brazil privatized the railroads. Not all countries mines benefited from the copper boom in the have ways to ease workers' transition from one 1970s and suffered with its crash (fiber optics had 172 ECONOMIC GROWTH IN THE 1990s TABLE 6.2 Performance Changes in Privatized Firms in Mexico Competitive vs. noncompetitive Competitive vs. noncompetitive Changes in industry- industries (according industries (according adjusted performance to prospectus) to market share) Mean s.s. Median s.s. N Mean change s.s. N Mean change s.s. N change (%) change (%) c nc (difference) (%) c nc (difference) (%) Profitability Operating income/sales 168 0.353 1 0.153 1 134 32 0.061 104 62 0.108 Net income/sales 168 0.412 1 0.211 1 134 32 ­0.146 10 103 62 ­0.026 10 Operating efficiency Cost per unit 168 ­0.183 1 ­0.152 1 134 32 0.106 1 104 62 ­0.049 Log(sales/employees) 166 0.935 1 0.896 1 134 32 0.151 106 62 0.33 5 Labor Log(# of employees) 169 ­19.05 10 ­24.47 1 136 33 ­0.273 5 107 62 ­0.069 Assets and investment Investment/sales 168 ­0.048 1 0.067 1 134 32 ­0.005 104 62 ­0.005 Output Log(sales) 170 0.489 1 0.424 1 136 33 ­0.215 105 61 0.206 Net taxes Taxes 168 26,441 5 2,161 1 135 33 ­7,024 1 106 61 1,013.6 Source: LaPorta and López de Silanes 1999. Note: N = number in sample; s.s. (%) = statistical significance to a % level; c = competitive, nc = noncompetitive. The columns that compare com- petitive versus noncompetitive show the difference in mean change (competitive ­ noncompetitive). There are two definitions of competitive: (1) according to privatization prospectus, and (2) according to market share (>10 percent is considered noncompetitive). reduced the demand for copper worldwide).18 operates but not the world prices for commodi- The state-owned Zambia Consolidated Copper ties it produces. Even so, earlier privatization Mines (ZCCM) was too poorly managed to could have prevented the mines from deteriorat- adjust and was a burden on the rest of the econ- ing to a point where massive outlays were needed omy.Early privatization may have helped save the before they could operate. The outlays were mines, but the government was loath to give up uneconomical with low copper prices, resulting a source of corruption and patronage.As a result, in the mines being closed. privatization discussions that began in 1991 Nellis also cites studies by Boubakri and Cos- (under donor prodding with a new government) set (2002) examining 16 privatizations (10 in dragged on through the decade, despite losses Morocco and 6 in Tunisia) where investments averaging US$15 million a month. A $1 billion and profitability were found to have risen. Jones, offer during 1996­97 from a consortium of Jammal, and Gokgur (1998) examine 81 privati- experienced mining firms was rejected, and zations in Côte d'Ivoire and find better perform- ZCCM was finally sold in 1999 after the assets ance with net benefits for about 25 percent of had further deteriorated.19 The consortium is predivestiture sales. Appiah-Kubi (2001) finds reported to have invested more than $350 mil- benefits from the 212 privatizations in Ghana. lion in the Konkola Copper Mine,20 but no one Andreasson (1998) finds improved performance seems to know whether the Treasury received of divested firms in Mozambique and Tanzania. any cash proceeds or what happened to them. Temu and Due (1998) find that of the 158 firms Even this sale was subsequently canceled when Tanzania divested through 1999,two-thirds were the Anglo-American consortium abandoned the sold to nationals (South African firms that buy purchase. Privatization can improve how a firm larger-value firms and breweries are resented) PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 173 and that government revenues rose and subsidies tial conditions), policies, and institutions all mat- to state-owned enterprises fell. ter for economic growth. Output recovered sooner in some countries than in others;and crit- ics make much of the fact that the swiftest sellers Evidence fromTransition Countries did not grow the fastest, but this argument Across 27 transition economies, more than ignores the role that endowments and other fac- 150,000 large enterprises and several times as tors play. many small firms have not been merely sold but One important lesson from the experience is transformed. Despite the assertions that outcomes that market-supporting institutions--well-func- would have been better had privatization been tioning courts, credit agencies, accounting done differently, the fact that it was done is a firms--did not spring up in response to the remarkable achievement, and numerous studies demand that privatization created. Advocates of show that it was beneficial. speed countered critics'arguments at the time by These studies vary in quality, however.The pointing out that such institutions would not early ones were more scorecards than evalua- emerge without a demand for them.Even if they tions, and as better studies became available they were right, however, privatization should not showed some of the early bloomers of privatiza- necessarily have been slower, given the propen- tion fading. For example, the Czech Republic, sity for asset-stripping mentioned earlier. In the long the darling of the advocates of speed, stum- Czech Republic, it was not the speed of privati- bled in 1997, and its GDP recovery was delayed zation but the government's disregard of provi- until 2000.21 Poorly managed privatization was sions to protect minority equity holders'interests held partly responsible, because firm managers that drove investors away from the Prague Stock and funds were allegedly too busy looting from Exchange and turned voucher funds into noncontrolling shareholders to focus on adding untrustworthy mutual funds. value.The Russian loans-for-shares scandal and Even the features now recognized as mistakes other shenanigans such as coercive purchases of were not viewed at the time as critical flaws.22 workers' shares by managers have also muted the Incentives operate through people; and if people initial enthusiasm for privatization. perceive unfairness or fear social turmoil, they Pohl,Anderson,and Djankov (1997) examine may not invest for the future and the gains may firm-level data, and find that countries that pri- not follow.While the incentives for privatization vatized more and faster restructured better. normally are to create and increase value, the Weiss and Nitkin (1997) find that who bought fears that ownership of the large Russian firms the firms seemed to matter. Djankov and Murell (that were unfairly purchased) would be reversed (2002) find the greatest gains arose when firms have encouraged asset-stripping and capital were sold to outsiders rather than workers.This flight.So although privatization has been benefi- pattern could be the result of adverse selection; cial when viewed as a whole, the manner of the workers were often sold firms that outsiders sale matters. thought unviable, and the findings would follow if the outsiders were correct on average. Fryd- man et al. (1999) correct for such adverse selec- 4. Utilities:Why the tion by examining a random sample of mid-sized Disappointments? manufacturing firms in the Czech Republic, Hungary, and Poland.They too find that selling Utilities is the collective label for a range of dis- to outsiders resulted in greater gains. So who parate industries (telecom, electricity, roads, rail- buys the firms does seem to matter, although we roads).Their privatization is discussed separately in do not quite understand why. this section because of the more complex regula- Moving from firms to the aggregate,the con- tory issues involved.The outcomes have often been clusions are also not clear-cut.Endowments (ini- disappointing,and this section examines why. 174 ECONOMIC GROWTH IN THE 1990s Utilities in developing countries had often management.The facilities deteriorated even as failed to invest enough to keep pace with population grew and, with it, demand. improving technology or growing populations. Most of them were short of funds because, in The Invisible Burden on the Poor turn, governments kept prices low, thinking that the poor could not pay for the services the utili- The true costs of neglecting infrastructure are ties provided. Indeed, few of the poor did pay, enormous and the financial shortfall in the utili- because few of the poor had access to the ser- ties is only the tip of the iceberg.But even the tip vices (figure 6.2). Underpriced services such as is huge:public monopolies in power,water,road- electricity and piped water were overused by the ways, and railways have annual losses of almost nonpoor.Nor did governments compensate util- $180 billion ($55 billion in technical losses and ities for the financial shortfalls that low prices $123 billion due to pricing),equivalent to almost and unpaid bills implied, because the owner (the the total infrastructure investment in all develop- state) either did not know or did not care and ing countries. sought to retain power and/or patronage. Utili- Taxpayers pay the costs of underpriced power ties operated erratically because revenues rarely and water,but people without access bear a larger covered their costs: funds would dribble in from invisible burden because alternatives are much belated price increases, erratic budget transfers, costlier. A familiar sight in many developing- and/or forced loans from government-controlled country cities is the array of water tanks on the banks when the enterprises were in danger of rooftops of homes--the costly individual shutting down.This instability played havoc with response to erratic city water supply. Investment maintenance and planning and reinforced poor in such tanks is a social waste: the storage is inef- ficient and the overall water supply is not aug- mented.The poor suffer more: lacking the funds (or land titles) to invest in storage,they pay dearly for water delivered by truck. Similarly, lighting with a paraffin or kerosene lamp costs 10 to 20 times as much as running an electric lamp; pow- FIGURE 6.2 Access to Basic Services, by Income Group--Ghana, ering radios is far more expensive with batteries Mexico, and Peru than with electricity;and so on (Brook and Irwin 2003).Almost all types of infrastructure are des- perately needed,and while the returns from such 100 (%) investments are high, the magnitude of the investments needed is daunting. 80 access Governments began to privatize utilities because they could not cope with the attendant with 60 problems and were emboldened by the euphoria over privatization in general and by the specific group 40 successes of some countries. Chile broke up and of privatized its electric utilities in 1978;the United 20 Kingdom did so in 1989,Argentina and Norway in 1991, and New Zealand in the mid-1990s. Proportion 0 In developing countries the scale of utility Ghana Mexico Peru and infrastructure privatization was immense: Water, poorest 20% Electricity, poorest 20% between 1990 and 2001 more than 2,500 private Water, richest 20% Electricity, poorest 20% infrastructure projects worth $805 billion were Source: World Bank, WDR 1994. privatized (Harris 2003). Latin America and the Caribbean accounted for almost half ($397 bil- PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 175 lion), followed by East Asia, Eastern Europe, improvements were based on transaction com- South Asia, with the Middle East and Sub-Saha- mitments, the disappointment is not surprising. ran Africa tying for the rear (table 6.3).By sector, the bulk was in telecom ($356 billion) and elec- Differences across Sectors tricity ($268 billion). These often-cited data sum up the value of The value of privatization transactions led to infrastructure transactions (actually commitments), great expectations, but since meager investments rather than physical investment.The transactions did not ease the capacity bottlenecks,disappoint- benefited developing countries' fiscal and/or bal- ments inevitably ensued. Experience varied ance of payments accounts, but additions to greatly by sectors.Although few transactions have capacity were likely far lower than the buyers' been reversed,23 many are being renegotiated: outlays to buy existing utilities. Figure 6.3 shows some 74 percent of transport and 55 percent of the breakdown by divestiture and greenfield proj- water concessions in Latin America.The exami- ects, but even these data may not accurately rep- nation below illustrates the issues involved. resent physical investments in additional capacity. Calderón, Easterly, and Servén (2002) report that Telecommunications aggregate infrastructure investments declined as a Telecommunications was a clear privatization share of GDP in LatinAmerica between 1980­84 success.While technology provided the impetus, and 1995­98: from 3.1 to 0.2 percent of GDP the size of the gains was influenced by the com- (itself often falling) in Argentina,5 to 2 percent in petition that regulation encouraged. Privatization Bolivia, 3.7 to 0.6 percent in Brazil, 3.1 to 1.7 and new entries expanded the network, increas- percent in Chile, 2.5 to 0.4 percent in Mexico, ing operating efficiency and labor productivity. and 2.0 to 0.6 percent in Peru.This aggregate The productivity gains did not reduce employ- decline is still consistent with increased invest- ment, and network expansion absorbed over- ments in the privatized firms (separate data are staffing. Figure 6.4 shows the gains from harder to compile), but if expectations of competition, among both cellular providers and TABLE 6.3 Investment in Infrastructure Projects with Private Participation in Developing Countries, 1990­2002 (2002 US$ billions, by region or sector) Region or sector 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Total East Asia and Pacific 2.7 4.3 9.7 13.7 17.1 22.2 32.1 39.0 10.6 9.8 15.0 12.4 9.7 198.4 Europe and Central Asia 0.1 0.4 1.4 1.5 4.4 9.5 12.4 16.0 13.1 10.0 23.2 7.3 9.7 109.0 Latin America and the Caribbean 14.9 12.9 16.5 19.3 19.5 20.2 29.6 55.3 77.0 39.9 40.5 34.3 17.3 397.2 Middle East and North Africa 0.0 --a 0.0 3.6 0.4 0.1 0.4 5.7 3.4 3.2 4.1 3.9 1.6 26.4 South Asia 0.4 0.8 0.1 1.4 3.4 4.2 6.6 6.8 2.8 5.0 4.2 4.6 5.5 45.8 Sub-Saharan Africa 0.1 0.0 0.1 0.0 0.8 0.9 1.6 4.8 2.7 4.8 3.4 5.0 3.5 27.8 Energy 1.3 1.3 13.1 15.9 17.2 25.4 34.2 51.6 30.5 18.0 28.4 14.9 16.5 268.3 Telecommunication 6.3 13.7 8.0 9.9 18.8 20.2 28.5 44.3 56.3 38.7 47.3 40.2 23.7 355.8 Transport 10.5 3.4 4.7 5.8 9.0 9.7 18.1 22.1 19.3 9.0 9.9 10.0 5.2 136.6 Water and sewerage --a 0.1 2.0 8.0 0.5 1.8 2.0 9.4 3.5 7.0 4.9 2.5 1.9 43.6 Total 18.0 18.5 27.7 39.6 45.6 57.1 82.8 127.5 109.6 72.7 90.5 67.6 47.3 804.5 Source: World Bank, PPI Project Database. a. No private participation in infrastructure occurred. 176 ECONOMIC GROWTH IN THE 1990s FIGURE 6.3 FIGURE 6.4 Infrastructure Projects with Private Participation, Growth in Latin American Telecom Lines 1990­2001 (percent per year) (2001 US$ billions) 25 Sub-Saharan Africa 20 Middle East and North Africa 15 South Asia 10 Europe and Central Asia 5 East Asia and Pacific 0 Public Privitized Privitized Latin America and monopoly monopoly competition the Caribbean Source: Wellenius 1997. 0 50 100 150 200 250 300 350 400 Divestitures Management and Greenfield projects operation contracts substantial gains from expansion; and (3) inequitable cross-subsidies in pricing, which Source: World Bank PPI Project Database. allowed higher overall prices,benefiting the poor through increased access and the well-off with those with fixed lines. Latin American countries new and better services.The private sector's bet- that granted monopoly rights of 6 to 10 years to ter financial, technical, and managerial resources their privatized telecommunications operators have a distinct advantage in keeping abreast of expanded their networks to 1.5 times the size this increasingly complex industry. under state ownership. Countries that retained the right to issue competing licenses did even bet- Electricity ter. Electricity restructuring and privatization are These gains were within the reach of all more complicated because unique characteristics countries,including those that had not tradition- determine how the market functions.Electricity is ally benefited from much foreign investment. the ultimate real-time product with its production Box 6.2 describes how the increase in access and consumption occurring at virtually the same was financed in Bangladesh. Cellular telephone instant. Peak-time supply is very inelastic near full costs also declined; but the potential for future capacity and demand shifts seasonally as well as gains remains immense: although the number of during the day,depending on such factors as tem- GSM (Global System for Mobile Communica- perature. tion) phone sales doubled in 2003 to almost 21 Recent technological advances affecting loca- million in India, only 7 percent of India's 1 bil- tion and hence transmission capacities have dra- lion people have access to any type of phone. matically changed the cost structure of electricity China already has 200 million cellular phone generation (figure 6.5).Technological improve- users (more than the United States'140 million). ments in gas turbines allow smaller, less-polluting Despite the recent bursting of the dot.com and generators to be built more quickly near cities. telecom bubble, the potential is still immense. Similar developments in wind generators and The gains in telecommunications were the photo-voltaics may change the competitive result of (1) technological changes that almost potential of electricity in many countries, both eliminated natural monopolies; (2) low coverage developed and developing.These developments in most developing countries, which allowed have allowed the market for generation to become PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 177 FIGURE 6.5 BOX 6.2 Optimal Size of U.S. Generating Plants Cellular Phone Operators in Rural Bangladesh $/MW 1930 A round the world, new service cost, 1950 providers have taken advantage of drastically reduced economies 1970 of scale to enter global and local markets, increasing competition and reducing generation prices. Women in Bangladesh have taken 1990 1980 the possibilities of cellular technology fur- Average ther than in most countries. Using 50 200 800 1,000 microloans of little more than US$30, Plant size, MW women in rural areas have set themselves Source: Bayless 1994. up as small-scale operators in a business that can offer a net annual income of over US$600, or more than twice the 1997 per ure 6.6).Most reforms attracted considerable pri- capita GDP. Grameen Bank, which provides vate investment in generation and distribution microcredit to small-business investment, (though much less in transmission) thus reversing financed the purchase of payphones by the underinvestment of recent decades. Electric- entrepreneurial women in villages from ity prices have fallen in many Latin American Grameen Telecom, its subsidiary. Starting countries as wholesale markets have developed in March 1997, within three years of its and new generators with lower costs have aug- first operation Grameen Telecom had pro- mented supplies. As retail prices start to reflect vided phone access to nearly 2.8 million underlying costs, cross-subsidies have been people in 1,100 villages. Access to phone reduced and in some countries eliminated. services has brought many benefits to poor Not every country is equally well positioned communities beyond the additional to gain through new technology and attendant income to operators: it reduced communi- restructuring of ownership. Brazil, for instance, cation costs (particularly transport) and has a largely hydro-based power system, and box raised farmers' income by providing infor- 6.3 illustrates why uncertain rainfall and multiuse mation on market prices that increased dams preclude substantial gains from privatiza- their bargaining power with middlemen. tion, even of its nonhydro generators. In other situations, as in Norway, privatizing hydro-based power yielded substantial benefits. Reaping the potential gains that technology competitive, and many countries have benefited now allows, however, requires countries to from allowing such competition. change the thrust and manner of regulation. Of the many countries that reformed their Electricity prices have been kept so low in most electricity sectors in the 1990s, most achieved developing countries for so long that raising good outcomes. When done properly--with them to reflect underlying costs will be politi- vertical and horizontal restructuring, privatiza- cally difficult. In several developing countries, tion, and effective regulation--reform signifi- attempts to raise tariffs in the face of acute power cantly improved operating performance: labor shortages have led to street riots that have caused productivity rose, sometimes dramatically in the increases to be reversed. Although private generation and distribution; technical and non- entrants will naturally demand a credible com- technical losses fell; and service quality rose (fig- mitment that future prices will be adequate 178 ECONOMIC GROWTH IN THE 1990s FIGURE 6.6 Postprivatization Labor Productivity in Electricity Distribution in Chile, Argentina, and the United Kingdom 16 14 from 12 years 10 employee per 8 against privatization) of 6 sales date 4 GWh (measured 2 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Year Chilectra Edenor Edesur UK RECs Source: Pollitt 2003. before investing, few countries provide such an term power-purchase agreements (PPAs). In undertaking. Indonesia, Malaysia, the Philippines, and Thai- land, rapid economic growth during the late Potential Problems in Power Markets 1980s and early 1990s increased the demand for Two additional issues deserve greater attention: electricity, but central governments were unable (1) many developing countries are too small to to finance the needed physical infrastructure benefit from competition in the power sector, investments. So the single state-owned utilities and (2) the single-buyer model (since the gains signed purchase agreements with independent are from generation) entails great risks. power producers, typically contracted in dollars Privatization is not always appropriate and its with government guarantees, since default pro- suitability depends on the country's circum- ceedings against a state-owned utility were often stances. Many developing countries are relatively not allowed. small: 60 have a system peak load below 150 At first the stratagem seemed successful: megawatts (MW), 30 have a load between 150 Southeast Asia attracted $65 billion in transac- and 500 MW,and another 20 a load between 500 tions for the private provision of infrastructure and 1,000 MW. Under the most favorable cir- between 1990 and 1997--more than half the cumstances, the opportunities for introducing total for all developing countries and substan- competition in such small systems are limited, as tially more than the other major destination, suggested by the market shares shown in table Latin America (with $45 billion). 6.4. Some of the smaller countries may benefit But theAsian crisis in 1997 caused GDP,elec- by linking their grids where possible to those of tricity demand, and the currencies' value in for- their neighbors (small, distant islands obviously eign exchange markets to fall. In Indonesia, cannot) and pursuing a regional as opposed to a electricity prices doubled in local currency under national approach to regulation. the PPAs, but state utilities did not pass the The Asian financial crisis exposed the risks increase on to final consumers. So the power involved when independent power producers purchasers could not honor the PPAs, and even (IPPs) sell to a state-owned utility through long- the government guarantee could not be honored PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 179 BOX 6.3 Brazil: No Rain or Privatization Gains B razil's experience illustrates that the specifics between water management authorities and power of each country's situation should guide dispatch. Furthermore, if hydro is used to satisfy whether to include privatization in the reform peak, as opposed to base-load, demand--and it is strategy. Brazil faced an acute power shortage during difficult to forecast how much this demand will be 2002­03 despite reforms that began in the 1990s, (especially when rainfall also fluctuates year to including privatizing some of its generation. Like many year)--it is very difficult to price power correctly. countries, Brazil had underinvested in capacity for Spot markets for electricity will not clear supply and decades: annual investment had declined steadily from demand and simultaneously provide investors with a 1982 peak of $12 billion to $3 billion by 1999 adequate returns. although demand continued to grow with the economy. These difficulties with hydro power inevitably also Brazil was heavily influenced by the British model spill over into nonhydro generation. Even if com- of separating distribution and transmission from gen- bined-cycle gas turbines were economical for peak eration, which would then be privatized. Unlike the provision, or coal-fired generators for base load, the United Kingdom's, however, Brazil's generation is presence of a large hydro system with unpredictable predominantly hydroelectric. The massive costs of rainfall would make them unremunerative without dams are incurred upfront (and are sunk), and the special payments for their role as emergency capacity running (marginal) costs are negligible. And in or reserve. If such capacity were privately owned, Brazil, even the long-run average costs of power are rules that determine such transfers would be impor- lower than even those of combined-cycle gas tur- tant and would be influenced by rainfall. bines, because the gas network and market are not Hence the private ownership of generation in well developed and the sunk costs of the dam are not Brazil brought few gains, and endless disputes with fully included. private investors. The situation would differ if dams' Multiuse dams further complicate the functioning sole use were for electricity (as in Norway), or if the of competitive, privately owned generation. Rainfall cost of thermal plants determined electricity prices fluctuations and the need to maintain adequate (as in Chile and Argentina). While dams may be pri- water reserves for irrigation and water uses other vately owned, the efficiency gains in Brazil are than electricity require basinwide coordination greater during their construction than in operation. because of the effects of the crisis on the balance reflecting cost; if anything, the currency crisis of payments and the government budget.There made the gap worse. Promoting rapid invest- were strong pressures to renege on,delay,or rene- ment in an unreformed electricity sector by gotiate PPAs,resulting in protracted and acrimo- offering independent power producers long- nious disputes, especially when allegations of term PPAs with state-owned, single-buyer util- corruption surfaced (box 6.4). ities involves substantial risks both to the Although power-purchase agreements cre- investors and to the public interest,and investors ate the same type of liabilities as foreign debts, often exact a substantial premium for the risks these liabilities were not explicitly recognized they incur. Hence such contracts are rarely even when governments guaranteed them. Fur- cheap even if they are entered into competi- thermore, such contracts did not address the tively. They are an expensive way to expand underlying problem of electricity tariffs not capacity quickly. 180 ECONOMIC GROWTH IN THE 1990s TABLE 6.4 Market Shares of the Three Largest Generation, BOX 6.4 Transmission, and Distribution Companies in Various Countries, 2000 Controversial Power Purchases in (percent) Indonesia B Country Generation Transmission Distribution efore 1990, the budget (oil rev- enues) and development assis- Argentina 30 80 50 tance funded power investments Bolivia 70 100 70 in Indonesia. The private sector played Brazil 40 60 40 only a small role. The power company Chile 67 100 50 Colombia 50 100 60 (Perusahaan Listrick Negara or PLN) was Czech Republic 71 100 49 an integrated, state-owned firm with El Salvador 83 100 88 problems typical of such entities. In the Hungary 74 100 65 mid- to late 1980s, a shortfall in oil rev- Indonesia 100 100 100 enues made the lure of private invest- Malaysia 62 100 97 ments in generation (funded by export Pakistan 95 100 100 credits) irresistible. In 1990, Indonesia Panama 82 100 100 needed an estimated 12,000 megawatts Peru 100 100 100 of additional capacity by 2000, implying a Poland 45 100 21 need for investment of about $20 billion. Thailand 100 100 100 With such needs, PLN embarked on a Source: Jamasb 2002. vigorous expansion, entering what has been termed "a gold rush for the invited elite and their foreign partners."* Between 1990 when the first IPP project Transport was solicited and 1997 when the East In many segments of the transport sector (rail, Asian crisis broke, PLN signed 26 power- ports, trucking, airlines, interurban busing), the purchase agreements and energy-sales pressures of inter- and intramodal competition contracts covering 10,800 megawatts for are sufficient in most countries to justify substan- some $13 billion. The terms appeared tial liberalization and privatization. It is difficult very favorable to the investors, many of for regulators or service providers to predict what whom sought these concessions with a are efficient and market-responsive vertical rela- well-connected local partner that opened tionships and combinations of logistical roles doors in exchange for an equity stake among various rail entities, truckers, barge oper- (that the foreign investor often ators,port operators,air carriers,warehouses,for- financed): the tariffs were in the 5.7­8 warders, and so forth. cents/kilowatt-hour range, well above the Experience confirms what theory predicts: prevailing average tariff, with "take or markets freed from excessive regulatory controls pay" clauses that shifted most risks to the find efficient and innovative ways to serve trans- purchasing utility. port needs.It is important to distinguish transport services, which are generally competitive or con- * World Bank 2003c. This report focuses more testable, from physical infrastructure facilities, on the absence of accountability and reports which may have natural monopoly characteris- more on egregious skimming of contracts by tics.The case for privatizing transport infrastruc- the well-connected than on the underlying ture is less compelling than that for services. Rail problems of the sector. track,basic and access port infrastructure,and cer- tain portions of airport facilities, where monop- PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 181 oly is unavoidable or substantial amounts of sunk gains exceeded those in nearly every other U.S. capital are involved, must be regulated or even industry. From 1981 through 2000, labor and operated by the public sector. locomotive productivity increased by 317 and Reforming rail regulation. Railroads have been 121 percent,respectively.Lower rail rates--down in decline since the early 1950s in almost all 59 percent, on average, in real terms from 1981 countries; but better regulation has revived them to 2000--have saved shippers and their cus- when and where they are economical.Figure 6.7 tomers more than $10 billion annually. After shows the decline in Eastern Europe,but the U.S. decades of steady decline, rail market share experience points to ways in which their poten- (measured in ton-miles) has trended slowly tial could be realized. upward, from 35.2 percent in 1978 to more than Railroads lost their historical dominance as 40 percent today. carriers of high-value freight because of poor Nor was the United States an isolated case: service and unreliability. Passenger traffic Japanese National Railways began its restructur- switched to roads and air. Misguided regulatory ing in 1986. In the early 1990s, British Rail was policies exacerbated the problems of the rail split vertically and horizontally (box 6.5).The industry. Pricing restrictions and cross-subsidies British government subsequently privatized the from freight to passengers accelerated the loss of freight businesses and the entire infrastructure rail's freight market-share to trucking.The com- and competitively awarded several franchises in bination of public ownership and exclusive the passenger segment. monopoly dulled incentives to control costs. In developing countries, virtually all the rail Railroad productivity has been especially poor systems were owned by the state at the beginning relative to latent technological opportunities. of the 1990s,but private operators obtained con- In the United States, regulatory reform freed cessions in most Latin American and several the industry from many arcane and ruinous rules. African countries. The 1980 Staggers Act substantially deregulated The gains from privatizing railroads depend the railroads, allowing pricing flexibility and the on the manner of regulation.There will be no abandonment of unproductive and redundant gains if private operators are as constrained as track. The effects were dramatic: productivity their state-owned predecessors. FIGURE 6.7 Railroad Cargo in the Transformation Period, 1985­97 100 80 60 (tonne-kilometers) 40 Index 1986 1988 1990 1992 1994 1996 Slovak Rep. Slovenia Romania Poland Bulgaria EC-10 Czech Rep. Hungary Baltic 3 Source: von Hirschhausen and Meinhart 2001. 182 ECONOMIC GROWTH IN THE 1990s A country can choose from a continuum of ownership and market structure reform options, BOX 6.5 but the choice should be based on many coun- Problems with Unbundling in try- and industry-specific characteristics: size, Railroads level of development, institutional capacity, den- sity of the rail network, condition of fixed rail T he United Kingdom separated the facilities,strength of intermodal competition,and ownership of infrastructure efficacy of public finances. Thus an uncritical (track) and operations (trains) to choice, especially of the extreme options permit competition among rail operators (entirely private or public,complete vertical inte- without the need for regulation; but the gration or separation), does not serve the public action gave rise to serious coordination interest. problems, loss of economies of scope, and Toll roads. Private sector participation in toll unnecessary transaction costs. roads increased dramatically during the 1990s. Many innovative rail services require About $61 billion of private investment was specific investment in infrastructure (for committed to 279 such projects between 1990 example, constructing loading and trans- and 1999 in 26 developing countries. Many of shipment facilities, building spur tracks these projects, however, encountered difficulties to reach a shipper's location), and opera- and were either renegotiated or canceled: of the tors find it difficult and inefficient to 279, 21 projects in Hungary, Indonesia, Mexico, coordinate this work with the infrastruc- andThailand, accounting for $9.5 billion in total ture owner, especially if their investment private investment, were taken over by govern- incentives are poorly designed. This is ments. what happened when the British railway More than a third of the canceled projects system was vertically unbundled and the were part of the Mexican toll-road program.The core infrastructure was transferred to a Mexican toll roads were built on the basis of very privatized company, Railtrack plc. The optimistic traffic projections--especially regard- operating companies and Railtrack fre- ing the price elasticity of users'demand,which is quently disagreed on the type, magni- relatively difficult to predict.The tolls that were tude, and timing of needed track repairs. needed to cover project costs drove much of the A frequent complaint was that Railtrack traffic on to parallel,nontoll roads with the result focused too much on commercial concerns that revenue was inadequate to service the debts and not nearly enough on engineering: so incurred (from government-owned develop- the tracks were not properly maintained, ment banks). The government therefore took leading to breakdowns that interrupted over the project and lowered the tolls. operator service, jeopardized safety, and The Mexican failure,as does the France-U.K. prompted public complaints. Channel tunnel, illustrates an inherent difficulty with some types of infrastructure projects: that the state cannot avoid certain types of risks, regardless of the clauses in the contract. Once built, the marginal cost of allowing additional users is low until the point of congestion, but if borne by taxpayers because of their sheer size.Even the toll charged is low,to allow for optimal use,it if lenders were not state-owned banks (as with cannot cover a project's average costs. Alterna- Mexican toll roads) or pension funds (with govern- tives determine the price charged--Mexican law ment standing behind them), assets of this size and requires nontoll alternative roads, and the Chan- no alternative use should not be allowed to rust and nel tunnel faces competition from ferries--and decay,and governments should open their purses to these prices result in financial losses that are often put them to their intended use. PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 183 Inaccurate projections of demand (especially has not been highly favorable to reform. Major of the cross-elasticity of demand) create serious water reforms have tended to be provoked by problems that compound the underlying diffi- public health crises,and to some extent by declin- culty of most infrastructure projects.Rather than ing real water revenues owing to inflation--fac- handing over a project entirely, it may be better tors that led to reforms in Buenos Aires, Lima, to involve the private sector in some aspects of Conakry, Santiago, and Mexico City in the late the project, such as the construction phase of 1980s and early 1990s. roads (if the quality could be specified, moni- The reforms in these cities depended on the tored, and verified) or the collection of tolls relative political power of potential winners and (where new scanning technology to pay could losers,not on social benefits.The political benefits be used). Private participation is also easier with may come from expanded service to the uncon- projects,such as road bridges or ring roads,where nected population, typically the urban poor, and substitutes are few and where less elastic demand from better service to middle-income groups.But allows full cost recovery. vested interests include those consumers who already have access and stand to be charged more, Water workers with patronage employment in publicly The scope for introducing competition in the owned utilities, and the lucrative businesses that supply of water and sewerage services is much provide services inefficiently (for instance,private more limited than that in the other network util- trucks that haul water in Indian cities).Water sec- ities. Local networks of pipes and sewers remain tor reforms have been politically most difficult to the quintessential natural monopolies.Moreover, sustain in cities where the marginal supply price unbundling is not as attractive because the bene- of water is increasing steeply and where waste- fits resulting from increased competition in sup- water creates large externalities--such as Lima ply are likely to be considerably less than in other and Mexico City.In BuenosAires,by contrast,the network utilities.The costs of producing water lower cost (and bids that reflected this) of renew- are relatively low in relation to the value added at able water resources have made it possible to the transport stage,although this relationship may reduce water prices and still generate enough vary across countries. The opportunities for return to attract private investment. introducing competition in sewage treatment,on Reform outcomes. After a few years of reform the other hand, are of greater significance. Over- (1988­93) in six cities, the initial results showed all, franchising is likely to be the most effective improvements in the coverage indicators (except way of increasing competition in the sector. in Lima), often dramatically (Abidjan and Technological change in water supply systems.A sig- Conakry). Unaccounted-for water--a measure nificant difference between the water sector and combining physical losses due to poor mainte- most other infrastructure is that technological nance and commercial losses due to poor finan- change in the past couple of decades has been cial management or illegal use--fell significantly much less dramatic or rapid and has had less in Buenos Aires, Lima, and Santiago. Financial impact on the underlying economics of supply. performance and labor productivity improved, The most significant technological innovation in and revenues exceeded costs in all cases except conventional water systems has been the wide- Mexico City (figure 6.8). spread introduction of metering at the point of Nevertheless, the well-publicized problems consumption, which permits the utility to set a with the Cochabamba concession in Bolivia tariff reflecting the marginal cost of water used underscore the unpopularity of such reforms and to bill for actual consumption. (especially price increases) and reveal the deep Difficult political economy.Part of the reason why distrust with which the private sector is viewed, the water sector is behind electricity, telecoms, or especially when there are nontransparent deals transport in restructuring or privatization is that with poorly overseen governments. Such oppo- in many countries the political economy of water sition is less likely if the public realizes that, 184 ECONOMIC GROWTH IN THE 1990s regardless of whether the provider is public or pass the laws, and politics will always play a role private, costs must be covered--if not through in what is regulated and how. prices, then through subsidies that many govern- Regulation should promote competition and ments cannot afford, and that the alternative to ensure access to bottleneck facilities, not attempt higher prices is having no piped water at all. to control a firm. Information asymmetries and regulatory capture make attempts at control counterproductive. Competition is the most Utility Regulation: Some Lessons of effective regulator,and regulatory reform implies Experience focusing on ensuring competition and access to Privatization makes good regulation both more physical infrastructure. important (because private owners care for prof- Ensuring competition implies allowing entry, its, even at the public's expense) and more diffi- a goal that unbundling services sometimes helps cult (because firms have a greater incentive and to advance, although experience shows how ability to misrepresent costs and market condi- coordination problems arise when the effort is tions).The recent experience in some countries overdone.Attempts to control prices are invari- shows that better regulation is possible though ably distorting, especially with rapidly changing not inevitable.Three lessons emerge. technology and privatization of parts of the infra- structure system. 1. Regulatory Reform Should Promote The importance of ensuring access is illus- Competition, Not Control trated in figure 6.9,which compares several Latin Improving regulations has traditionally meant American countries that opened their telecom- giving greater legal powers to regulators, perhaps munications markets to private competition. No training them and trying to make them inde- new entrants gained more than 15 percent of the pendent of politics. Absolute regulatory inde- market, even in Chile, where they operated for pendence is neither feasible nor desirable. more than 20 years, because regulators were Regulators will always report to politicians who unable to ensure access to bottleneck telecom- munications facilities (such as the local loop). Even under technically competent regulators, regulatory interventions often fail.It is ironic that FIGURE 6.8 when theWorld Bank and other multilateral insti- Labor Productivity in the Water Sector tutions helped developing countries set up regu- (year of reform = 0) latory agencies they gave considerable attention to the organization and legal independence of Guinea 40 these agencies but not to their mandate: many operate without any controlling principles.Their present structure accentuates the tendency to 30 connections expand regulatory jurisdiction, often with dys- functional consequences. Guinea India's experience illustrates the difficulties in /1,000 20 staffing a regulatory agency (box 6.6), even in a Mexico Mexico country with considerable administrative abilities 10 Ivory Coast Lima and a vast pool of competent potential staff. Ivory Coast Employees BsAs 0 Santiago 2. Regulatory Reform Should Focus on Get- ­3 ­2 ­1 0 1 2 3 4 5 ting the Underlying Economics Right Source: Shirley and Menard 2002. Understanding the source of benefits helps in structuring reform. A pricing policy that does BsAs, Buenos Aires. not allow adequate revenue cannot improve the PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 185 FIGURE 6.9 BOX 6.6 Local Exchange Carriers in Latin America India's Regulatory Capacity versus Effectiveness 18 I ndia's power sector is overseen by states, but its reg- 14 ulatory problems do not stem from center-state juris- Dominican Republic share dictional issues. As in the United States, India's 1998 Chile 10 reforms created independent electricity regulatory com- competitors Market missions in each state, with the Central Electricity Regula- of 6 tory Commission (CERC) dealing with national and Mexico Colombia 2 Peru interstate issues (for example, the National Electricity Grid Argentina Code). The State Electricity Regulatory Commission (SERC) ­2 12 19 ­2 Venezuela,5 sets tariffs, enforces licensing conditions, and monitors R. B. de Years of local services liberalization compliance. To ensure the independence of these commissions, Source: Pyramid Research 2001. their mandate to protect the public interest is clearly defined and members have reasonable job security (they cannot be dismissed unless impeached for unethical con- situation, no matter if the industry is privatized duct). With funding coming through special provisions in or an independent regulator is established. For the consolidated central and state budgets, the commis- example, as of 2000, in almost all Common- sions are freed from the Ministry of Power's direct finan- wealth of Independent States (CIS) countries, cial control. household electricity prices covered less than 50 Nevertheless, a recent review of state and central elec- percent and industrial prices less than 70 percent tricity regulators showed inadequate staffing, which of the long-run marginal costs of supply (figure sharply limited regulatory capabilities (Prayas Energy 6.10). Group 2003). CERC and most of the SERCs (Orissa and Similarly, imposing social service obligations Andhra Pradesh are possible exceptions) could not fill on only some providers will not promote effi- their specialist positions. Government pay scales are cient investment,even when institutional mecha- insufficient to attract capable professionals, and requests nisms provide a credible commitment to policy for professional and technical staff appointments are rou- stability. tinely delayed for months or years. All but two SERCs had Neglecting the underlying economics is also only three or fewer professional and technical staff when why social security reform has gone astray: there they were supposed to have 8 to 10, and 8 of the 12 SERCs are some risks that the private sector simply can- studied had no permanent professional and technical staff not handle adequately.While some governments at all. They often relied on temporary staff from the may find it difficult to socialize these risks ade- incumbent utilities they were ostensibly regulating. quately,it is a mistake to think that they could be borne by the private sector. Despite potential gains from others sharing these risks, they may have to be borne by individuals until better insti- press (which informs the electorate), an engaged tutions develop. public (demanding more from governments), academia (training regulators and encouraging 3. Regulatory Structures Should Reflect studies of problems)--determine why what is Country Differences sound regulation in one country is ineffectual in Countries differ greatly in their economic struc- another.24 tures and in their institutions.These institutional Although countries can learn from the mis- differences--including courts (where appeals are takes of others, important differences among made), legislatures (where laws are passed), the them make it hard to replicate others' successes. 186 ECONOMIC GROWTH IN THE 1990s its social security system attracted many imitators FIGURE 6.10 in Latin America and Eastern Europe, and the Cost-Coverage Ratios of Electricity Prices in World Bank provided a three-pillar model of the CIS Countries pension reform in an influential report, Averting the Old Age Crisis (1994a).The experience of the 100.0% 90.0% 12 Latin American and 8 East European coun- 80.0% tries,27 however, calls into question the universal 70.0% applicability of this model. 60.0% 50.0% Each pillar of the three-pillar model is best 40.0% described by its function, since the method of 30.0% 20.0% financing and determining benefits can differ. 10.0% The first pillar seeks to prevent old-age poverty, 0.0% usually through a defined-benefit PAYG pen- Ukraine Russia Poland Slovak Czech Hungary Republic Republic sion. All countries with social security systems have this pillar, and in many it is the only one. Household prices: % of LRMC Industry prices: % of LRMC The second pillar aims to smooth consumption Source: von Hirschhausen and Opitz 2001. over the life cycle and prevent a dramatic LRMC. Long-run marginal cost. decrease in income during retirement. Con- sumption smoothing requires retirement pay- ments linked to previous income, so this pillar Countries often adopt regulatory reforms in tends to be a fully funded,28 defined-contribu- name but not in substance,to satisfy international tion pension.Funding purports to protect against agencies that sometimes require this reform as a demographic changes and insulate recipients condition for aid or loans. Developing countries from the vagaries of budget appropriations.The looked to the experience of Canada, New third pillar augments income in old age and is Zealand, the United Kingdom, and the United essentially a voluntary savings scheme for retire- States in formulating their regulations, often ment, often with tax advantages. A need-based ignoring their big differences from those coun- cash-transfer system financed by general (not tries.25 Rash attempts to regulate can be far more wage) taxes is sometimes called a zero pillar. dangerous and costly than inaction.There is good reason to be concerned that regulations could The 1990s Reforms thwart competition, not promote it.26 Most of the social security reforms of the 1990s took place in Latin America and Eastern Europe, 5. Privatizing Social Security but the two regions differ starkly.Latin America's large young population contrasts with the higher Several countries, primarily in Latin America and proportion of the aged in Europe (figure 6.11). Eastern Europe,reformed their social security sys- Social security coverage is low in Latin Amer- tems by moving to a multipillar system and giving ica, but almost universal in Eastern Europe. Fiscal a prominent role to privately managed individual sustainability is a concern for both regions,though accounts. We describe their experiences, after particularly acute in Eastern Europe (figures 6.12 defining key terms in pensions and social security and 6.13).Reforms in the two regions erected the (box 6.7). second and third pillars. Many Latin American countries sought to phase out the first pillar com- pletely,thereby privatizing social security.But most Three Pillars for Stability? governments in Europe and Central Asia (except Countries sought to reform social security for Kazakhstan) retained at least a small first pillar, not different reasons. Chile's privatization of parts of only because of their legacy of protecting the aged PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 187 BOX 6.7 Definitions A pension is deferred compensation that greater mobility and declining ties to the extended employers pay. As employers, some govern- family and community made social security a safety ments have long provided pensions for civil net that the public valued. servants and military veterans and their widows and In most OECD countries, problems with social orphans, but very few firms did so until well into the security have arisen because of demographic 20th century. Banks were an exception, perhaps to changes. Increased longevity and declining fertility avert malfeasance (bankers making bad loans before led to an increase in the proportion of the elderly in they retired). Pension coverage in private firms rose the population, meaning that fewer workers had to in the United States during the 1930s and further dur- support more pensioners. This imbalance required ing World War II, perhaps to circumvent wage con- either a reduction in benefits or an increase in tax trols. The expansion coincided with increasingly rates (coverage was already almost universal) or a progressive income tax rates, and spreading income delay in retirement age or some combination of the over a lifetime reduced the total taxes paid. three. Alternatively, the projected shortfalls could be Pensions are of two types: a defined-benefit funded* (by moving from a PAYG scheme to generat- scheme specifies the amounts (generally related to ing a cash flow surplus). But projections are error- years of service and the last few years' salary) while prone, depending not just on demography but also a defined-contribution scheme invests employees' on economic growth, and many governments, includ- (and employers' matching) contributions and links ing those of the United States, have been reluctant the payouts to the investment value. Funds are gen- to set aside as much as was needed. Often, they spent erally set aside in both types of schemes; but employ- even those funds. ers bear the investment risk in defined-benefit Many developing countries, especially in Latin schemes and employees decide the investment allo- America, introduced social security early in the 20th cation and bear the risk in defined-contribution century, but they generally limited coverage to gov- schemes. Defined-contribution schemes are more ernment employees and the unionized formal sec- portable, allowing employees to change jobs without tor--a small fraction of the labor force. Unlike those worrying about qualifying for or losing pension eligi- of OECD countries with aging populations, develop- bility, and are displacing defined-benefit schemes in ing countries' social security problems do not arise many countries. from demography--half the population consists of Social security also grew in the 20th century. children who will soon enter the labor force--but Some governments began paying the elderly, not just from poorly defined and enforced eligibility rules and their former employees, but required the recipients to overly generous payments. The high tax rates encour- have worked and earned in order to qualify, thereby age evasion not just from coverage but from the for- giving such transfers pension-like features. Social mal sector altogether: many employees work in security payments are often based on need, capped to smaller businesses that remain in, or migrate to, the favor the nonrich, and only loosely tied to informal economy. wages/earnings, giving them safety net­type fea- The situation is different in the transition coun- tures. Social security began as a pay-as-you-go (PAYG) tries of Eastern Europe.** The demography and ben- system, either unfunded or with partial funding, with efit coverage are similar to those of OECD countries, payroll taxes usually financing payouts. Rising pros- but the massive decline in output (outlined in Coun- perity made for increasingly generous systems, and try Note E on Eastern Europe's Transition) and the (Box continues on the following page.) 188 ECONOMIC GROWTH IN THE 1990s being established.The third pillar allowed, but did not mandate, additional savings with tax BOX 6.7 advantages.In LatinAmerica,new entrants to the (continued) system were only offered the second and third difficulties in administering taxes made it difficult for pillars:thus,once workers who were already cov- governments to honor their promises to the unemployed, ered retired and died, the system would be com- the retirees, and the elderly. Their solution was to cap pletely private. benefits or build up pension arrears in the first instance Results and to then cut costs in a variety of ways: increasing con- tribution rates, raising the retirement age, or experi- Figure 6.14 shows how the reforms reduced the menting with different forms of indexing benefits. governments' projected liabilities, but the transi- tion was unexpectedly costly.29 In Argentina and Bolivia, a poorly managed switchover increased *"Funding" (by generating a surplus that is invested before payouts by more than expected because of fraud- the "demographic bulge" retires) requires a country to save in ulent claims and a lax interpretation of the rules. the aggregate, which it can do only by running a current- Figure 6.15 shows that Bolivia's budget deficit account surplus. Barr (2000) points out that many discussions rose instead of falling.The transition costs were make this fallacy of composition, and several countries seeking also high in Eastern Europe because the newly to reform social security run current account deficits that pre- created private sector evaded the taxes when rates vent funding. were high to finance both the first-pillar payouts ** China's demography differs from those of both Latin and the second pillar's funding. America and the Caribbean and Eastern Europe, and its social There were other disappointments too.Despite security system covers about 18 percent of the population. But the reforms, coverage remains low in Latin Amer- the immediate problem seems to be clarifying whose liabilities ica, notwithstanding predictions that a closer link these are: while they are being transferred from the enterprises between contributions and benefits in a funded sys- to the government, it is not clear which level of government will tem would improve incentives to participate (figure ultimately be responsible, and how intergovernmental transfers 6.16). It is unclear whether this is because workers will take place. mistrust social security or simply evade the payroll tax. Capital markets did not develop either, con- but also because of concerns about the future. trary to predictions.30 Pension funds have simply Many of these countries seek membership in the held government bonds,not commercial paper or European Union, and for labor to move freely equities. Perhaps this development should not be within the European Union, the social protection surprising since equity markets require the prin- systems of its members should be compatible.Most cipal-agent problem to be effectively overcome-- EU countries provide the first pillar,although har- which usually takes much longer than the monizing their systems does not appear to be a accumulation of funds. But placing government high priority, and bilateral agreements among EU bonds in captive funds does not create a market. members govern workers moving from one mem- Mandating private savings have led to more, and ber country to another. occasionally better, financial sector and capital Pensions and social security are long-term market regulations. Countries that run a current contracts, and any change involves an extended account surplus could invest in foreign markets; transition. In Latin American countries, covered but most developing countries run current workers were given a short period in which to account deficits. choose between a reduced first pillar (a defined- Surprisingly, private pension funds proved benefit pension in what remained of the pay-as- more expensive to administer than the previous you-go system) or a new second pillar with the state-run systems.Although private competition privately managed individual accounts that were was expected to reduce administrative costs, these FIGURE 6.11 Eastern Europe's Population Is Significantly Older than Latin America's (percentage of total population in each age cohort) Latin America: 2004 Male Female 80+ 75­79 70­74 65­69 60­64 55­59 50­54 45­49 40­44 35­39 30­34 25­29 20­24 15­10 10­14 5­9 0­4 6 5 4 3 2 1 0 0 1 2 3 4 5 6 Percent Eastern Europe and Central Asia: 2004 Male Female 80+ 75­79 70­74 65­69 60­64 55­59 50­54 45­49 40­44 35­39 30­34 25­29 20­24 15­10 10­14 5­9 0­4 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 Percent Western Europe: 2004 Male Female 80+ 75­79 70­74 65­69 60­64 55­59 50­54 45­49 40­44 35­39 30­34 25­29 20­24 15­10 10­14 5­9 0­4 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 Percent Source: U.S. Census Bureau, International Programs Center. 189 190 ECONOMIC GROWTH IN THE 1990s costs increased in Latin America as part of the FIGURE 6.12 transition from centralized to decentralized man- Coverage is Greater in Europe and Central Asia than agement in an industry with possible scale Latin America and the Caribbean, but Intraregional economies. Even in Eastern Europe, where cen- Variation is Wide tralized records were maintained for the privately 80 managed funds, charges have been comparable to 70 those in decentralized Latin American systems. Moreover, private fund managers could collude EAP 60 since they were few and entry was difficult. Reg- of 50 % ulation did not focus on ensuring competition, as 40 and rents to incumbents came at workers'expense. 30 Coverage 20 Government's Role in Pension Systems 10 Although privatization is often appealing, gov- 0 ernments have a necessary role in pension sys- a co tina tems. People can save for old age (to smooth HungaryEstoniaPolandCroatia ChileLatvi Bulgaria RomaniaMexi Rica mbia PeruaraguaBolivia Salvador Colo consumption) in a variety of different ways, Kazakhstan El Argen Costa Nic including investing in their children and buying Sources: Palacios and Pallares-Miralles (2000) for Europe and Central Asia; recent household surveys for Latin America and the Caribbean. real estate or claims on other assets, but these choices are limited in many countries. People EAP Economically active population. may be unable to diversify against catastrophes such as droughts or famines, especially if all members of their family and community are affected. Besides, community and family ties fray with urbanization and economic development. The question is whether governments or mar- kets can protect against such risks better than FIGURE 6.13 personal and/or private initiatives. Eastern Europe Has Greater Fiscal Need for Reform The government may have a role in keeping than Latin America the elderly out of poverty, especially as societies 16 prosper. Since private markets to insure against poverty may not develop because of adverse EAP 14 selection and moral hazard, the government of 12 % essentially provides this insurance through the as 10 first pillar. 8 Though insurance markets develop to pool the risk of infrequent large but predictable losses 6 spending from, say, fire and automobile accidents, other 4 losses are better managed through prevention 2 and self-insurance (Ehrlich and Becker 1972). Pension 0 Old age is predictable and self-insurance means ia saving. Governments have a clear role in enforc- Rica ico StatesEU-15Poland CroatiaLatvia Chile Hungary Bulgaria Estonia BolivialvadorPeru Mex ing contracts for savings and investment vehicles, Argentina Romania Nicaragua Sa KazakhstanCosta Colomb El United but it is less clear why the second pillar should be Sources: Palacios and Pallares-Miralles (2000). mandated, if this forces savings through a poorly developed financial system. In many developing EAP Economically active population. countries, relying on investments in a family PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 191 FIGURE 6.14 Reforms in Latin America Reduced Debt: Projected Pension Debt (Explicitly Accumulated after 2001) 400 debt, 350 300 explicit GDP of 250 200 150 percentage pension-related a as 100 50 0 Government's Reform No Reform No Reform No Reform No Reform No Reform No Reform No Reform No reform reform reform reform reform reform reform reform Chile Peru Colombia Argentina Uruguay Mexico Bolivia El Salvador 2010 2030 2050 Source: Gill, Packard, and Yermo 2004. FIGURE 6.15 FIGURE 6.16 Bolivia's Pension Reform Was No Marked Rise in Participation in Latin America Unexpectedly Costly Contributors to Retirement Security System, Cash flow gap: Bolivia percentage of EAP GDP 70 deficit of 4 60 2 50 percentage EAP a 40 0 of Pension-related as 1998 1999 2000 2001 2002 % 30 Actual Projected in 1997 Source: Gill, Packard, and Yermo 2004. 20 10 business, or in children, may be a better choice 0 for all concerned. 1980 1982 1984 1986 1988 1990 9192 1994 1996 1998 The third pillar may not be necessary but, as Argentina Chile Ecuador Nicaragua it is voluntary, it is not harmful. It should be Brazil Colombia El Salvador Uruguay viewed in conjunction with the tax system.The Bolivia Costa Rica Mexico tax break for such savings tends to move the tax system from taxing (realized) income to taxing Source: Gill, Packard, and Yermo 2004. consumption--a shift that economists since Kaldor (1955) have long favored. 192 ECONOMIC GROWTH IN THE 1990s The experiences of Latin America and East- ern Europe differ greatly, but common lessons BOX 6.8 emerge.The multipillar model may be suitable Argentina: Private Accounts Do Not for some countries, but clearly not for all. Latin Protect Workers from Government America's experience shows that the problems in Risk the first pillar--that is, a need-based pension to P the aged--cannot be solved by simply adding rivately administered mandatory additional pillars, or by switching from a pay-as- individual retirement accounts you-go to a funded system. In Argentina the were thought to protect workers default on government bonds has essentially against political interference inherent in a destroyed the second pillar, which was meant to public PAYG system; but the 2001 crisis in replace the first (box 6.8).A government cannot Argentina showed otherwise. resolve the fiscal implications arising from the Argentina's private pension system was first pillar by "privatizing" it: a better choice may vulnerable even before the 2001 crisis. be to scale back the benefits, administer it well, Since the system began in 1994, nearly and ensure that it is financially sound. half the assets were invested in govern- Whether a second pillar (mandated individual ment bonds, making pensions vulnerable accounts) is advantageous depends on a country's to the government's fiscal problems. circumstances. Financial markets are not equally Indeed, converting implicit PAYG govern- developed in all countries, and the second pillar ment liabilities into explicit debt (having may be unsuitable and administratively costly in a funded system) contributed significantly many. Government bonds will likely remain the to Argentina's declining fiscal position: main asset in such funds, meaning that the sec- the bonds benefited from high interest ond pillar, like the first, will rely on the promises rates, but this gain increased the overall of a financially strapped government.The gov- fiscal deficit. ernment will always have a role in any mandated Argentina's return to the peso and the pillar, and without competition,"private provi- default on its debts left pensioners in the sion" may merely be an expensive decoration. lurch. The events of 2001­02 so eroded Many transitions have proved unexpectedly confidence in any mandated retirement expensive because of fraud and poor administra- system that participation in any system tion of the rules.Governments should be especially will likely suffer. wary with social security and pension reforms, Source: Rofman 2002. because mistakes are difficult to undo.Those with an existing system may be better advised to keep it simple and make it sound than to add pillars on a shaky fiscal foundation. Countries without social attributes and technology of each sector and its security should be especially wary about introduc- institutional, social, and political characteristics. ing such measures,because once established,a pen- Clear gains have followed the privatization of sion system cannot be easily dismantled. nonutilities (with the possible exception of large extractive industries), suggesting that privatiza- tion of such firms would be beneficial in all 6. Conclusions countries,although the benefits are greater when institutions exist that help markets function.With The 1990s experience shows how difficult both extractive firms and utilities (or with social secu- privatization and regulation are.There is no uni- rity), the government will play a role through versally appropriate reform model. Every taxes or regulatory controls,and the benefits from restructuring and privatization program needs to their privatization will depend greatly on how consider explicitly the underlying economic effectively this is done. PRIVATIZATION AND DEREGULATION: A PUSH TOO FAR? 193 Privatization is less about finding better own- Notes ers than the government than it is about separat- ing commerce from politics. Government 1. The TennesseeValley Authority was an exception; it ownership does not blur this distinction in OECD promoted rural electrification to tackle poverty in countries as much as in developing countries the depressed Appalachian region. 2. Kahn (2004) reflects on and summarizes the deregu- where oversight over governments is weaker. Pri- lation of airline and telecommunications industries vatization helps to achieve the separation but does and the current controversies (such as the Baby Bells not automatically ensure it, because governments being allowed into the long-distance market). retain other powers they could abuse and,without 3. In 1984, the privatization of BritishTelecom was the institutions to check such conduct, still influence largest stock market flotation ever. firms they do not own.The transition countries' 4. The (West) German government sold a majority stake to the public in Volkswagen in 1961 and in experience shows that market-supporting institu- VEBA (a major energy/industrial corporation in tions do not emerge quickly in response to Germany), but later bailed out shareholders when demand.And in infrastructure utilities the com- stock prices collapsed. German banks rather than mercial cannot be separated from the political, individuals or mutual funds own most industrial regardless of the adequacy of institutions. equity. Similarly, there was modest privatization in France during the 1980s, but not on the same scale If privatization is oversold as a means of sepa- as in Britain. rating commerce from politics, restoring the link 5. Nellis (2003b) summarizes the many descriptions through regulation is underappreciated.There are and studies of Latin America's privatization. huge potential gains from privatizing infrastruc- 6. Nellis and Shirley (1991). ture, but inappropriate regulation has sometimes 7. World Development Report 2003 covers the role of the prevented these from being realized.In particular, private sector in the provision of many services that governments traditionally provide and are not dis- when prices are controlled at levels that do not cussed here.This chapter focuses primarily on priva- cover costs, owners will not invest in new capac- tization in the transition economies and the ity. Regulation must also respond to technical regulation of utilities, not changes in the various changes, which make today's sensible regulation agricultural marketing boards (primarily inAfrica) or distorting tomorrow. California's misadventures the deregulation of nonutilities in developing coun- tries. in electricity show how even sophisticated regu- 8. See Country Note E,"Eastern Europe's Transition: lators can make mistakes that leave the public Building Institutions." confused about what really happened. And in 9. Even human capital could be considered state- countries where the public already distrusts mar- owned since wages were administratively set, migra- kets, privatization bears the opprobrium. tion was restricted, and the choice of professions was Experience shows that it is possible for regu- constrained.The freeing of wages and other restric- tions implicitly meant that this capital was restituted lation to focus less on control than on ensuring to individuals. access to bottleneck facilities and encouraging 10. The specifics differed by country, but an account of competition and entry, in turn encouraging Russian privatization by Boycko,Shleifer,andVishny innovation. Redirecting regulation in this man- (1996), who were advisors to the privatization ner requires a good understanding of technology agency--with its elements of political intrigue, clash of personalities, and the need to make important and economics. It also calls for modesty, espe- decisions quickly without adequate information-- cially in settings where politics can undercut reg- would resonate with anyone who worked in transi- ulators' competence. tion countries at the time. In social security, privatization has not elimi- 11. More was done to establish stock exchanges than to nated the government's role, and administrative establish institutions protecting property rights.More costs have sometimes increased.The purpose and was done to pass new laws than to help courts func- tion better. nature of such contracts may always require gov- 12. Some of the least market-oriented countries avidly ernment involvement.Changes in these arrange- passed such laws, so the few times they were used ments have large consequences that are difficult were to shake down efficient firms that increased to reverse. their market shares. 194 ECONOMIC GROWTH IN THE 1990s 13. Kogot and Spicer (2002) describe how the small privatization program has been the most successful group of influential economists were unaware or dis- to date." missive of the work of political scientists, many of 22. Except for the loans-for-shares scheme in Russia whom were knowledgeable about the countries.Dani where major assets were practically given away to a Rodrik (2003c) makes a similar point in cabal of cronies. http://ksghome.harvard.edu/~.drodrik.academic.ksg/ 23. Some 48 contracts worth $24 billion (that is,1.9 per- Stiglitzconference notes.pdf. cent by number and 3.2 percent by value) were can- 14. Claessens, Djankov, and Klingebiel (2001) offer a celed, a third of them dealing with the Mexican toll bleak outlook for equity markets and suggest that road program.Water and sanitation had higher can- efforts are better directed at improving creditor pro- cellation rates than telecom or ports. tection. 24. Analogously, the performance of state-owned firms 15. Some did so for fallacious reasons. Lieberman and is disappointing in some countries (India, Mexico) Nellis (1994) provided a premature endorsement of but not in others (Sweden, France). Russian experiments with vouchers, arguing that 25. Sophisticated rate-of-return rules are meaningless if vouchers give the public purchasing power. firms do not keep accounts adequately, and recourse Ramachandran (1997) sought to dispel some of the to courts for remedies may not be possible. Some many fallacies that arose by pointing out that vouch- notable examples include the adoption of the quasi- ers cannot provide purchasing power or be inflation- judicial U.S. model for telecommunications and ary, and that their main advantage was to effectively energy in the Philippines, a country with a notori- reduce the"minimum value"that governments often ously weak judiciary, where reform led to regulatory placed that would have prevented their sale. failure; and the adoption of a U.S.-style Public Util- 16. Although the privatization agencies disbanded, their ity Commission in Jamaica, where without the con- employees often found work in the securities and stitutional protections and administrative due process exchange commissions that were created (and later prevalent in the United States, the result was regula- atrophied when trading volumes could not justify tory instability and the nationalization of telecom- their continuation). munications in 1975. 17. Little empirical work has been done on this.When 26. TheWorld Bank's recent Doing Business 2004:Under- valued at the secondary market price for vouchers, a standing Regulation report (World Bank 2004c) shows mere $40 million could have bought all firms in that the biggest and most common mistake is to reg- Georgia in 1994. ulate too much, and to do so poorly. 18. Copper prices fell from $7,000 a tonne in 1996 to 27. Chile pioneered the multipillar system in 1981.The $3,000 in the mid-1990s to about $1,500 in the early system was adopted (with its many variants) during 2000s. Production had fallen from its peak of about the 1990s by seven countries in Latin America and 800,000 tonnes of finished copper a year to under the Caribbean (Peru, Colombia,Argentina, Uruguay, 300,000 tonnes. Mexico,Bolivia,and El Salvador) and three in Europe 19. Anglo-American headed the consortium through its and Central Asia (Hungary, Poland, and Kazakhstan). subsidiary, Zambia Copper Investments, which held Costa Rica,the Dominican Republic,Latvia,Roma- 65 percent of the shares.ZCCM Holdings had 20 per- nia, Bulgaria, Croatia, and Estonia adopted it in the cent,IFC took a 7.5 percent stake,and the Common- 2000s, and Nicaragua and Ecuador plan to do so. wealth Development Corporation took the remaining 28. Several countries provide some form of minimum 7.5 percent, while the Government of Zambia held a pension guarantee to the funded pillars, however, golden share. adding a defined benefit component to an otherwise 20. Konkola Copper Mine's December 2003 statement defined contribution pension. to the Extractive Industries Evaluation (describing 29. This section is based on Gill et al. (2004). Linde- efforts to find alternative employment and activities mann, Rutkowski, and Sluchynskyy (2000); Kritzer for the affected miners). (2001);andWorld Bank (2003a) describe the experi- 21. World Development Report 1996: From Plan to Mar- ence in Europe and Central Asia. ket concluded that "the Czech Republic's mass 30. For further discussion, see chapter 7, section 2. Country Note E Eastern Europe'sTransition: Building Institutions T he economic and political con- that market allocation provided better incentives, sequences of the end of Com- leading to less waste, and economic prosperity munism in Eastern and Central would follow, especially with the end of the Europe and the collapse of the Soviet Union costly arms race following the demise of com- have changed the lives of 400 million people in munism,. 27 countries.The transition was without prece- Transformation was a daunting task,and some dent, and vividly illustrates how complex is the economic turmoil was anticipated,but the coun- transformation and establishment of institutions, tries' prospects were not bleak. Most advisers-- even in societies that are well endowed with including international financial institutions, human capital and rich cultural traditions.1 think tanks, and academics--advocated rapid reform, meaning open trade with low tariffs, rapid removals of controls over the economy in Legacy general, and quick privatization. As reformers Fifty years of communism left a relatively large have highlighted (notably the Russian Federa- stock of physical capital--in the form of infra- tion's former Minister of Finance Gaidar, in structure, manufacturing industries, and housing World Bank 2005b), speed was essential. The (much of it technically obsolete)--and of human political and administrative collapse that accom- capital. Literacy and health indicators were com- panied the transition virtually everywhere meant parable with those of industrialized countries,and that gradual trade liberalization, or carefully the scientific and intellectual cadres were among planned and sequenced privatization, was simply the world's best. But resources had been ineffi- not possible. ciently used because prices were administratively set and the state owned everything. Centrally Outcomes made decisions on production and pricing disre- garded preferences and scarcity. For example, Output in all countries declined much more Uzbekistan grew cotton extensively, drying up sharply than expected. The 12 Central and the rivers that fed the rapidly diminishing (inland) Southeastern European countries and the 3 Aral Sea. Energy was as wastefully consumed in Baltic countries (CSB) fared better than the 12 industry as water and fertilizer were in agricul- countries of the Commonwealth of Independent ture.2 Many industries subtracted rather than States (CIS).3 Over four years,CSB had a cumu- added value: at world prices, their inputs were lative average output decline of 22 percent,or 12 worth more than their outputs. Despite Soviet percent when weighted by population. By con- scientific achievements, machinery often embod- trast,CIS output fell 50 percent over 6.5 years,or ied obsolete technologies because prices did not by 45 percent when weighted by population. signal value and influence resource allocation. Recovery diverged more sharply. Output in All East European countries faced the same CSB now exceeds its pretransition levels,but that transition issues: to reorganize production and in the CIS is still one-third lower, or one-fifth reallocate resources better.The expectation was lower when weighted by population at the end 196 EASTERN EUROPE'S TRANSITION: BUILDING INSTITUTIONS 197 of 2003.Within the CIS, there are differences as of the military at the borders with weapons and well (figure E.1). For example, Georgia's output transport are tempted to smuggle contraband.4 declined by roughly 75 percent between 1990 Domestic borrowing was difficult because sav- and 1994, and at the end of 2003 it was no more ings fell, and foreign lenders were not always than 40 percent of its pretransition level, partly forthcoming. Printing money created inflation because of civil war. Incomes are more unequal that hurt the poor,who (unlike in Africa or Asia) now within countries (the Gini coefficient had no extended families to fall back on. almost doubled in Bulgaria,Armenia, and several Reorganizing the real sector involved more other countries). More worryingly, absolute than freeing prices. Every firm had to decide poverty has risen. how and what to produce and how much to Private enterprise overtook the state sector in charge,which workers to retain,and so forth,but most countries;but its share of output varies from neither governments as owners nor state-owned more than 80 percent in Hungary to a mere 20 banks as creditors could oversee the firms' man- percent in Belarus. Newly established private agers. Mechanisms such as hard budget con- firms account for much of the growth, although straints were needed for this purpose and to many are spinoffs from older firms. International discipline firms.Some firms were slow to restruc- trade expanded,especially with countries outside ture and continued operating as before, building the former Soviet bloc. Exports from the CSB up inventories and interenterprise arrears. Some have outpaced those from the CIS (8.8 percent governments (for example, Romania, twice) annual growth versus 3.2 percent during sought to break the logjam, thereby loosening 1993­98).This was partly because these coun- the hard budget constraint.5 Utilities posed spe- tries had ports or had good roads to connect cial problems: their finances were often sapped them but also because foreign firms invested by the unpaid bills of energy-intensive industries, there, since the countries were expected to join but shutting them down could be a death sen- the European Union.Whether it is a cause or tence for consumers in cold climates, because consequence,CSB governments have introduced workers lived in company housing. Many gov- policies that provide a better investment climate ernments made mistakes when tackling issues that fostered domestic firms and attracted inflows such as this, but some recognized and corrected of direct foreign investment: Hungary has the mistakes more quickly than others. attracted foreign investments of more than 5 per- Problems were similar in agriculture: cent of its gross domestic product (GDP) since although farmers could switch crops and use 1995, among the highest rates of foreign direct inputs more efficiently in theory, those growing investment (FDI) in the world. Russia, Poland, cash crops (for example, Uzbek cotton growers) and the Czech Republic have also attracted sub- faced problems similar to those in industry when stantial FDI, although the amounts were smaller they could not easily adapt irrigation and the fractions of their GDP. marketing infrastructure. De-collectivizing agri- The transition buffeted government finances. culture was not easy: even when restitution Revenues fell. Collecting taxes requires a differ- claims were sorted out and land was redistrib- ent administration than does commanding uted,farmers sometimes had to continue operat- resource transfers, and developing new tax ing collectively because they lacked their own administrations was difficult and slow. Mean- equipment and access to credit. while fiscal deficits ballooned. Concomitantly, Financial intermediaries did not develop government spending rose, because unemploy- quickly. Countries licensed many new banks, ment called for additional outlays and because often creating so-called commercial banks with firms requested funds (subsidies, guarantees, or the stroke of a pen, but developing a credit cul- loans) to restructure their operations. Cutting ture takes time, and in any event banks had little government spending was risky: police officers to lend.As incomes fell, so did savings rates, and easily organize protection rackets, and members few savers entrusted their savings to banks, espe- FIGURE E.1 Key Indicators for Transition Countries Ukraine Russian Federation 60 120 60 120 50 100 50 100 40 40 80 80 30 30 60 60 20 Index 20 Index Percent Percent 40 40 10 10 0 20 0 20 ­10 0 ­10 0 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Current account balance, % of GDP Current account balance, % of GDP Unemployment rate, % Unemployment rate, % Inflation, % end of period Inflation, % end of period Central government total revenue, % GDP Central government total revenue, % GDP Real GDP index, 1989 = 100 (right) Real GDP index, 1989 = 100 (left) Conflict countriesa Reform nonstartersb (GDP weighted) (GDP weighted) 60 120 60 120 50 100 50 100 40 40 80 80 30 30 20 60 60 Index 20 Index Percent 10 Percent 40 40 10 0 20 20 ­10 0 ­20 0 ­10 0 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Current account balance, % of GDP Current account balance, % of GDP Unemployment rate, % Unemployment rate, % Inflation, % end of period Inflation, % end of period Central government total revenue, % GDP Central government total revenue, % GDP Real GDP index, 1989 = 100 (right) Real GDP index, 1989 = 100 (right) Source: World Bank, WDI; EBRD 2003. a. Georgia, Armenia. b. Belarus, Uzbekistan. 198 Poland Hungary 60 140 60 120 50 120 50 100 40 100 40 80 30 30 80 20 60 20 60 Index Index Percent Percent 10 40 10 40 0 0 20 20 ­10 ­10 0 ­20 0 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Current account balance, % of GDP Current account balance, % of GDP Unemployment rate, % Unemployment rate, % Inflation, % end of period Inflation, % end of period Central government total revenue, % GDP Central government total revenue, % GDP Real GDP index, 1989 = 100 (right) Real GDP index, 1989 = 100 (right) Baltic states Romania (GDP weighted) 60 120 60 120 50 50 100 100 40 40 80 80 30 30 60 20 60 20 Percent Index Percent 10 Index 40 40 10 0 0 20 20 ­10 ­10 0 ­20 0 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Current account balance, % of GDP Current account balance, % of GDP Unemployment rate, % Unemployment rate, % Inflation, % end of period Inflation, % end of period Central government total revenue, % GDP Central government total revenue, % GDP Real GDP index, 1989 = 100 (right) Real GDP index, 1989 = 100 (right) 199 200 ECONOMIC GROWTH IN THE 1990s cially since they lost much of their deposits to evasion, most countries made it harder to qualify high inflation during the early years of the tran- for pensions,unemployment insurance,and other sition. Banking deposits now range from 3 per- benefits by raising the retirement age and other cent of GDP (Georgia) to 60 percent (Czech eligibility requirements. Improving the efficient Republic);those in most transition countries are delivery of social services will take time. in the 10­20 percent range. Some countries Many people were adversely affected by the (Albania, Romania, and Russia) suffered from changes taking place, and their perceptions now destructive pyramid schemes that added to mattered because most countries were becoming depositors' distrust and to the liabilities of gov- democracies.(The few that remained authoritar- ernments that intervened. The inexperienced ian were not among the economic reformers.) It central banks found it difficult to oversee finan- did not help when newly prosperous citizens cial systems or to resist government demands flaunted their wealth,especially when this wealth that led to inflation. had been "legally stolen" through noncompeti- Privatization was only one of the many real- tive privatization.The hardships caused many cit- sector issues that governments tackled.Allowing izens to resent the harshness of markets, but no new firms to emerge required changes at many country has reverted to communism, and more levels in the bureaucracy,including company reg- impressively, there has been no significant policy istration and tax collection. As noted above, in reversal.6 But democracy's roots have yet to countries with no tradition of voluntary tax spread, especially in the CIS; media criticism is compliance, developing tax administration new; and politicians are unseasoned. Georgians, proved difficult. Governments traditionally col- and more recently Ukrainians,took to the streets lected payroll taxes, but relying heavily on this to oust an unpopular president after a rigged method implied high tax rates that deterred new election, but their new heroes face familiar con- business formation or forced activities into the straints and cannot work miracles. untaxed informal sector. Despite the difficulties, the achievements are Thus, although social concerns were impor- impressive and the outlook is bright: by the end tant, funds for schools, health care, and pensions of 2003, real GDP in the CIS (weighted by pop- had to be cut. Spending on education declined ulation) was 50 percent higher than in 1998, and to 2­8 percent of GDP and on health care to1­6 citizens of the Balkans enjoy peace after years of percent of GDP. Many social services that were war and civil conflict. Eight transition countries formerly provided by enterprises (such as hous- have joined the European Union, and this has ing and childcare) had to be transferred to local spurred their own efforts and those of other governments, but their costs far outstripped the prospective members to improve their societies governments' ability to raise revenues. When in all their dimensions. spending on such services declined, it was diffi- cult to protect vulnerable groups, especially Lessons when there were large regional disparities and no transfer mechanisms in place.Some of the spend- Though endowments matter, so do policies.The ing cuts were accomplished by eliminating overly countries that prospered most were not those generous provisions: Poland, for example, had that many observers had predicted. For example, been spending twice as much on disability pen- the former East Germany's endowments did not sions as the Organisation for Economic Co- lead to prosperity, and the Czech Republic's operation and Development (OECD) average. progress belied expectations stemming from its Pensions and unemployment insurance were "reforms"(soft budget constraints and the result- provided centrally, and these amounted to a high ing lack of industrial restructuring precipitated 10­13 percent of GDP--double their pretransi- the Czech crisis during 1996­98). tion proportion, and unsustainable. Because tax East Germany's experience, in particular, rates on wages were already high, encouraging shows the importance of sound policies.The East EASTERN EUROPE'S TRANSITION: BUILDING INSTITUTIONS 201 was fortunate to be well located, sharing West needed for compromises and consensus building. Germany's language, culture, and history (except Even if such reforms are not reversed, they may during the communist years). East Germany not bear fruit if complementary measures are not could quickly and smoothly adopt laws and taken.As Larry Summers put it,"Well executed organizations from theWest,and it obtained con- policies that are 30 degrees off are much more siderable financial assistance: 40­60 percent of its effective than poorly executed policies that are GDP from the West as cumulative transfers dur- spot on.... the ability to do things we take for ing 1991­97 and only slightly less thereafter. granted in modern market economies is actually Nevertheless,East Germany's output fell substan- a crucial part of success [in the transition tially, and remains more depressed than that of economies]"(World Bank 2005b).Some political many transition countries whose disadvantages systems seem to enable consensus building better are greater. Unifying the German currency at an than others (World Bank 2002a, 2002c). And overvalued level, and adopting West Germany's competitive democracies (such as the Czech rigid labor laws and generous welfare benefits, Republic, Hungary, Poland, Slovenia) have sus- resulted in too much of East Germany's capital tained reforms better than noncompetitive polit- being scrapped and too few investments being ical regimes (such as Belarus, Uzbekistan, made, leading to high unemployment. Foreign Turkmenistan), though here it is hard to distin- direct investors leapt over East Germany into guish the effect of new leaders from the effect of countries such as Poland that had better eco- new political regimes. nomic policies. Third, policies and endowments, including The transition countries show no consistent institutions, are not entirely independent of each statistical association between particular types of other.Societies take many years to change,and in reform policies and growth.7 This should not be retrospect it was unrealistic to have expected surprising: even if the policy reform index cho- market-supporting institutions to emerge rapidly sen identifies and measures the right policies, the in response to demand. If policies are effective response of the economy to reform also depends only when there is a broad consensus about on many other factors including government them--rather than any particular reform meas- credibility, institutions, and social cohesiveness, ure per se, or the zeal and speed with which it is which differ greatly across countries (and per- implemented--this also depends on endow- haps also over time in each of them). ments and institutions.Assured of broad support Even so, some observations are notable. for the reform course, governments can change First, as discussed further in chapter 6, some the implementation of policies to respond to slow privatizers have fared better than fast ones, opportunities and shocks (for example,by speed- yet the issue is not speed per se,but a rapid move ing up or slowing privatization, or adjusting to a market economy. Although studies have budget deficits), with good results, as in Poland found that growth has come disproportionately and Hungary. from newly created firms, privatizing existing firms ensures that these new firms have assets with which to work (McMillan and Woodruff Notes 2002). 1. The large literature on the transition includes the Second,building a consensus is important and World Bank's World Development Report 1996 and a takes time.As noted above,many reforms perforce report examining the first 10 years of transition had to be done quickly. Prices had to be freed, (World Bank 2002c). Looking at the ingredients for and firms had to be sold rapidly, especially if the successful transition, the World Bank's 2002 report state had lost control and firms operated in limbo. emphasized the role of policies facilitating the entry of new firms and limiting the flow of resources to But other changes, such as improving accounting old industries; and the role of market institutions or a court system,take time to accomplish.Accel- that, among other things, enforced property rights erating such reforms precludes the discussions and ensured good governance. It also indicated that 202 ECONOMIC GROWTH IN THE 1990s competitive democracies (that is, those that estab- 4. In several countries,high excise taxes on tobacco and lished and protected civil liberties and political cigarettes attracted well-organized smuggling opera- rights permitting multiparty democratic elections) tions that also moved drugs, weapons, and people recovered sooner and grew faster than others.While (illegal immigrants and victims of bondage). recognizing the importance of initial conditions, 5. Condon and Ramachandran (1993) show that this the report concluded that after the first few years, buildup would plateau and that the situation would policy and institutional reforms had been more correct itself. important than those conditions in explaining per- 6. As measured by the European Bank for Reconstruc- formance. tion and Development's transition indicator. 2. For example, in 1985 the Soviet Union used 0.95 7. Selowsky and Martin (1998) use panel data to meas- ton of oil equivalent per $1,000 of GDP,nearly dou- ure the effect of reforms (using a liberalization index) ble the 0.5 ton used by OECD countries (IMF,World on growth after controlling for initial conditions and Bank, OECD, and EBRD 1991). other factors (such as dummies for war).The coeffi- 3. The 14 CSB countries are Albania, Bosnia and cient on the liberalization index is statistically signif- Herzegovina,Bulgaria,Croatia,the Czech Republic, icant. However, Heybey and Murell (1999) find that Estonia, Hungary, Latvia, Lithuania, the Former correcting for its possible endogeneity (using initial Yugoslav Republic of Macedonia, Poland, Romania, level, share of industry, and the like as instruments) the Slovak Republic, and Slovenia.The CIS com- the liberalization index does not explain growth dur- prises Armenia, Azerbaijan, Belarus, Georgia, Kaza- ing the first four years of the transition. Brown and khstan, the Kyrgyz Republic, Moldova, the Russian Earle (2004) find that interfirm reallocation of out- Federation,Tajikistan,Turkmenistan, Ukraine, and put, labor, and capital are not related to productivity, Uzbekistan. although privatization improves such reallocation. Chapter 7 Financial Liberalization: WhatWent Right,WhatWentWrong? T HE FINANCIAL LIBERALIZATION Certainly the reforms produced some gains. that took place in the develop- But the growth benefits of the financial and non- ing countries in the 1980s and financial reforms in the 1990s were less than 1990s was part of the general move toward giving expected. Financial crises raised questions of markets a greater role in development. It was also whether financial liberalization was the wrong a reaction to several factors specific to finance:the model, what had gone wrong, and the appropri- costs, corruption, and inefficiencies associated ate direction of future financial sector policy. with using finance as an instrument of populist, Overall, the 1990s is probably best considered a state-led development; a desire for more financial precursor of better things that will take some resources; citizens' demands for better finance and time to achieve. lower implicit taxes and subsidies; and the pres- Section 1 of this chapter describes why and sures exerted on repressed financial systems by how financial liberalization occurred. Section 2 greater international trade, travel, migration, and discusses the outcomes of financial liberalization better communications. during the 1990s, including the crises that The financial reforms went beyond the inter- occurred and their relation to macroeconomic est rate liberalization that had been recom- policies, financial liberalization, and the over- mended by the so-calledWashington Consensus. hangs of old economic and political systems.Sec- To varying degrees, governments also allowed tion 3 summarizes the lessons from the the use of foreign currency instruments and experience of the 1990s,and section 4 draws sug- opened up capital accounts. Domestic markets gestions for future policy. Section 5 concludes developed in central bank and government debt, the chapter. and international markets expanded in govern- ment and private bonds. Capital markets devel- oped, but less rapidly, and were most successful 1. From Financial Repression to in the larger, already rapidly growing, East and Financial Liberalization South Asian countries. State banks continued to have a major role for much of the 1990s; their The financial repression that prevailed in devel- privatization was gradual and often proved oping and transition countries in the 1970s and costly. Central banks moved away from trying to 1980s reflected a mix of state-led development, finance development; they became more inde- nationalism, populism, politics, and corruption. pendent and successfully focused on keeping The financial system was treated as an instrument inflation low, but their debt increasingly of the treasury: governments allocated credit at absorbed bank deposits. below-market interest rates, used monetary pol- 203 204 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s icy instruments and state-guaranteed external borrowing to ensure supplies of credit for them- FIGURE 7.1 selves and public sector firms, and directed part Increase in Average Deposits/GDP in Major Countries, by Regions, 1960s­90s of the resources that were left to sectors they (difference between 3-year average of bank deposits/GDP favored.State banks were considered necessary to at the end of each decade) carry out the directed credit allocations,1 as well as to reduce dependence on foreigners. Bank 30 supervisors focused on complying with the often 25 intricate requirements of directed credit rather ease than with prudential regulations. Interest rates to 20 depositors were kept low to keep the costs of Incr loans low. In some cases, low deposit and loan 15 GDP rates were also populist measures intended to fo 10 improve income distribution.2 Repressed finance was thus an implicit tax centage 5 and subsidy system through which governments erP transferred resources from depositors receiving 0 low interest rates (and from those borrowers not ­5 receiving directed credits) to borrowers paying 1960s 1970s 1980s 1990s low rates in the public sector and to favored parts South Asia (3 countries) Africa (10) of the private sector. Governments had to allo- Latin America (7) East Asia (4) cate credit because they set interest rates that Source: IMF, International Financial Statistics. generated excess demand for credits. Capital Note: Countries and regions covered are: controls were needed not (as often argued) to East Asia: Indonesia, Republic of Korea, Malaysia, the Philip- protect national saving, but to limit capital out- pines, Thailand. flows fleeing low interest rates and macroeco- Latin America: Argentina, Brazil, Chile, Colombia, Mexico, Peru, República Bolivariana de Venezuela. nomic instability,and to increase the returns from Africa: Burkina Faso, Cameroon, Côte d'Ivoire, Ghana, Kenya, the inflation tax.3 In effect, capital controls were Mali, Nigeria, Senegal, Tanzania, Uganda. a tax on those unwilling or unable to avoid them South Asia: Bangladesh, India, Pakistan. and they encouraged corruption (Hanson 1994). sector "white elephants," and unproductive pri- vate activities yielded low returns, crowded out Factors behind Financial Liberalization more efficient potential users, and encouraged Three general factors provided an impetus for wasteful use of capital. the move to financial liberalization: poor results, Financial repression also worsened income dis- high costs, and pressures from globalization.This tribution. Subsidies on directed credits were often section discusses each in turn. large, particularly in periods of high inflation, and actual allocations often went to large borrowers.4 Poor Results The low interest rates led to corruption and to the Together,the limited mobilization and inefficient diversion of credits to powerful parties. Diversions allocation of financial resources slowed economic tended to grow over time,particularly when infla- growth (McKinnon 1973; Shaw 1973). Low tion reduced real interest rates on credits,and rising interest rates discouraged the mobilization of fiscal deficits and directed credits absorbed more of finance, and bank deposit growth slowed in the the limited deposits. 1980s in the major countries (figure 7.1).Capital flight occurred despite capital controls (Dooley High Costs et al. 1986).Allocation of scarce domestic credits The repressed systems were costly. Banks, par- and external loans to government deficits, public ticularly state banks and development banks, F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 205 periodically required recapitalization and the World Bank, as the costs of financial repres- takeover of their external debts by govern- sion became clear. ments. Political pressures and corruption were · In East Asia, the major countries liberalized in widespread. Loan repayments were weak the 1980s,though at different times and to dif- because loans financed inefficient activities, ferent degrees. For example, Indonesia, which because loan collection efforts were insufficient, had liberalized capital flows in 1970, liberal- and because borrowers tended to treat loans ized interest rates in 1984,but the Republic of from the state banks simply as transfers.Typi- Korea did not liberalize interest rates formally cally, banks and other intermediaries rolled over until 1992. Low inflation generally kept East their nonperforming loans until a period of Asian interest rates reasonable in real terms, inflation wiped out depositors' claims and per- however. In most countries, connected lend- mitted a general default. Since intermediaries ing within industrial-financial conglomerates were not forced to follow reasonable prudential and government pressures on credit allocation norms or mark their portfolios to market, the remained important. losses were nontransparent, even to the govern- ments that often owned them. Inflation also · In South Asia, financial repression began in the helped to conceal the problems of commercial 1970s with the nationalization of banks in India banks through their earnings on low interest (1969) and Pakistan (1974). Interest rates and deposits.The hidden costs of the repressed sys- directed credit controls were subsequently tems became more apparent once financial lib- imposed and tightened, but for much of the eralization began. 1970s and 1980s real interest rates remained reasonable. Liberalization started in the early Pressures from Globalization 1990s with a gradual freeing of interest rates; a Perhaps most important, financial repression reduction in reserve, liquidity, and directed came under increasing pressure from the growth credit requirements;and liberalization of equity of trade, travel, and migration as well as the markets. improvement of communications.5 The · In Latin America, episodes of financial liberal- increased access to international financial mar- ization occurred in the 1970s but financial kets broke down the controls on capital outflows repression returned, continued, or even on which the supply of low-cost deposits had increased in the 1980s, with debt crises, high depended.6 Capital controls may be effective inflation, government deficits, and the growth temporarily, but over time mechanisms (such as of populism (Dornbusch and Edwards 1991). overinvoicing imports and underinvoicing In the 1990s, substantial financial liberalization exports) develop to subvert them (Arioshi et al. occurred, although the degree and timing var- 2000; Dooley 1996).These mechanisms became ied across countries. more accessible as goods and people became more internationally mobile. · In the transition economies, financial liberal- ization took place fairly rapidly in the 1990s in the context of the reaction against com- The Evolution of Financial Liberalization munism (Bokros, Fleming, and Votava 2001; The shift in policies differed in timing, content, Sherif, Borish, and Gross 2003). and speed from country to country and included many reversals. Broadly: The earliest policy changes generally focused on interest rates. In many instances governments · African countries turned to financial liberal- raised interest rates with a "stroke of the pen" to ization in the 1990s, often in the context of mobilize more of the resources needed to finance stabilization and reform programs supported budget deficits and to enable the private sector to by the International Monetary Fund and play a greater role in development.(Some interest 206 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s rate increases,designed to curb capital flight,were · Gradually, state banks were privatized. Bank- intended more for stabilization than for liberaliza- ing competition increased, as a result of the tion.) New financial instruments were introduced entry of new domestic and foreign banks and, that had freer rates and were subject to lower in some cases, nonbank intermediaries. directed credit requirements.Some countries also began admitting foreign currency deposits, to In general, however, the financial reforms of attract offshore funds and foreign currency hold- the 1990s focused on freeing interest rates and ings into the financial system as well as to allow credit allocations, and made much less effort to residents legal access to foreign currency assets improve the institutional basis of finance--a (Hanson 2002;Honohan and Shi 2003;Savastano much harder, longer task. 1992, 1996). Partial interest rate liberalizations soon gener- ated pressures for more general freeing of interest 2. Outcomes in the Financial rates (albeit in some cases after reversals of liberal- Sector during the 1990s ization).As borrowers directed funds into dereg- ulated instruments and sectors, demand for Private sector credit is a key factor in growth.10 low-cost loans increased and repayments on them Banks can intermediate funds and take risks only deteriorated.7 Unfortunately, when the macro- if private credit is not crowded out by govern- economic situation was unstable and interest rates ment debt. Over the 1990s, deposits grew faster were freed,very high real interest rates developed, than in the previous decade, but in many coun- creating corporate and banking problems that tries bank credit to the private sector from added to the overhang of weak credits that were domestic sources grew only slowly.The increase exposed by liberalization. in loanable funds was largely absorbed by the At very different speeds in different countries, public sector. interest rate liberalization came to be supple- mented by other changes:8 Deposit Growth · Central banks were made more independent. Bank deposits grew as a share of the gross domes- They abandoned their earlier developmental tic product (GDP) in the 1990s, unlike in the role to focus on limiting inflation,often in the 1980s (figure 7.1 above and Hanson 2003b). context of stabilization programs. Thus, most major countries and most regions achieved a major objective of financial liberaliza- · Reserve requirements and directed credit tion.And in India and some East Asian and Latin were eased. American countries, nonbank deposits supple- · Capital accounts were liberalized, even in mented the rapid growth in bank deposits. countries where domestic foreign currency Box 7.1 discusses the resumption of deposit instruments remained banned. Foreigners growth in India as it gradually liberalized, as well were allowed to participate in capital markets9 as the growth of India's capital market. and private corporations were allowed to raise Many factors contributed to the deposit funds offshore. growth, including the slowdown in inflation in the 1990s,11 the positive real deposit rates, and · Markets were set up for central bank debt and new deposit instruments.Another factor was the government debt. Equity markets were set up legalization of foreign currency deposits. in the transition countries and liberalized Deposits in foreign currency grew as a share of where they already existed. total deposits in many countries in the 1990s, · In some countries, pension systems added and in some cases they supplied more than half defined contribution/defined benefits ele- the total by the end of the decade (Honohan and ments,often operated by private intermediaries. Shi 2003).12 Not surprisingly, the foreign cur- F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 207 BOX 7.1 India--A Successful Liberalizer with Strong Capital Markets I ndia liberalized its financial sector gradually developing countries, providing nearly one-fourth of over the 1990s, with particular success in capi- India's corporate funding from 1992 to 1996 (Reserve tal markets, while avoiding any major crisis. In Bank of India 1998) with listings more than doubling the 1980s, India had a repressed financial system from 2,000 in 1991 to over 5,000 (Standard and (Hanson 2001, 2003a). This, plus increasing macro- Poor's 2003). The post-1997 economic slowdown led economic instability, slowed deposit growth. Finan- to a stock market fall, problems in the NBFCs (which cial liberalization was part of greater reliance on were wound down), and crises in the government the private sector after the 1991 foreign exchange development banks and mutual fund, though public crisis. Interest rates were raised and gradually freed, sector commercial banks performed surprisingly well. bank regulations and supervision were strength- Recently, large capital inflows and higher growth ened, and nonbank financial corporations (NBFCs) have led to low interest rates and better bank per- were allowed under easier regulations (Hanson formance. 2003a). After a 1991 capital market crisis, regula- Although India's approach to financial liberaliza- tions were strengthened, listings were liberalized, tion served it well, three major issues remain: (1) foreign investors were allowed in, and infrastruc- crowding out, with government debt now absorbing ture was substantially improved (Shah and Thomas more than 37 percent of bank deposits compared to 1999; Nayak 1999). about 24 percent at the end of the 1980s; (2) a weak Bank deposits of nationals and nonresident Indi- information and legal framework, which, despite ans resumed their growth and NBFC deposits grew efforts at improvement, still contributes to nonper- sharply after 1992. The stock, bond, and commercial forming loans and limits access to credit; and (3) the paper markets became among the most vibrant in still-dominant role of public sector banks. India: Money (M3 in Sept.)/GDP, Deposit Rate and Inflation 75 75 M3/GDP M3+NBFCs % of GDP (%) 60 60 ) 45 Avg. M3/GDP, 1987­92 45 inflation GDP Interest rate 30 30 fo and 1 year deposit Inflation (CPI) %( ater 15 15 M3 est 0 0 Free Inter rate ­15 ­15 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 Source: Reserve Bank of India data. CPI. Consumer price index. GDP. Gross domestic product. 208 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s rency deposits were popular with members of Access to credit expanded less than many the public, many of whom had lost their savings observers hoped after the financial reforms, and pensions in inflation and repressed financial though it improved toward the end of the1990s. systems.13 But foreign currency loans were also Panel studies had suggested that financial liberal- popular with borrowers.14 ization would make more credit available to a The reforms reduced the burden on banks, wider group of borrowers. (See, for example, widening their discretion over the allocation of Schiantarelli et al. 1994, and works cited there.) resources and lowering required reserves. Now After liberalization there was some growth, but that governments could raise resources from in practice government and central bank debt newly developing debt markets, they had less crowded out many borrowers.In some countries need to require banks to invest in government where nonbank intermediaries (henceforth, debt or to hold low-return reserves with the cen- nonbanks) grew, they did increase lending to tral bank that were invested in government debt. nontraditional borrowers. But both banks and In many of the 25 largest developing countries, nonbanks were hindered by the lack of informa- the average ratio of reserves to deposits fell over tion on borrowers and weaknesses in the legal the 1990s (Hanson 2003b). Directed credit and judicial systems in the areas of collateral and requirements were reduced, interest rates were creditors' rights. raised on remaining directed credits, and nomi- Instead of increasing private credit, the rise in nal market rates fell. bank deposits over the 1990s tended to be absorbed by government and central bank debt, and by banks strengthening their offshore posi- Credit:Absorption of Deposits by tions. In particular,in the 25 developing and tran- the Public Sector sition countries with the largest banking systems, Bank credit to the private sector grew much less the average ratio of net government debt to bank than bank deposits and other bank liabilities in deposits rose by more than 60 percent,from about the 1990s (figure 7.2). 13 percent in 1993 to about 21 percent in 2000 (Hanson 2003b).15 Similar patterns prevailed in the larger African countries.16 The main reason for the rise in government FIGURE 7.2 debt was postcrisis bank restructurings, involving Changes in the Ratios of Bank Assets and Liabilities replacement of weak private credits, particularly Plus Capital to GDP, 1990s in Brazil, Indonesia, Jamaica, Mexico, and some (averages by country group) African countries. But growing government 20 deficits also played a key role in some cases, Liabilities + capital 15 notably India andTurkey.In general,the increases a GDP in banks' holdings of government debt reflected as fo 10 rises in the stock of government debt,rather than 5 any increased attractiveness of government debt Changes centage 0 to banks, or decreased willingness of banks to per take risks.17 ­5 1990­2000 1993­2000 1993­2000 1990­2000 Banks also increased their net holdings of Latin America East Asia Transition Africa central bank debt--substantially in some coun- 7 countries 4 countries 3 countries 10 countries tries--despite falling reserve requirements. On Increased private Increased net public Increased net sector credit sector and central foreign assets average in the 25 developing countries with the bank credit largest financial systems, banks' net holdings of Source: IMF, International Financial Statistics. central bank debt rose by nearly 5 percentage Note: Countries covered are the same as in Figure 1. South Asia is excluded because points of GDP over the 1990s (Hanson 2003b). countries in that region do not report private credit or capital separately. As a monetary policy instrument, central bank F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 209 debt had advantages over the previous instru- executed, the associated collateral was usually ments of credit controls on individual banks, worth less than 30 percent of the face value of the changes in reserve requirements, and variations loans, suggesting how unproductive the growth in central bank lending, which had tended to in private sector credit had been.In the transition limit competition, to affect banks bluntly, and to countries and African countries the quality of affect weak banks heavily. But the use of central credit issues was typically more related to the bank debt had costs, in that it crowded out public sector's use of the credits.21 would-be borrowers. Central banks often sold their debt to sterilize capital inflows as well as to tighten money when capital flowed out.18 Cen- Bond and Equity Markets tral banks may also have sold debt to mop up Government and central bank bond market some of the liquidity that arose from lowered development was fairly successful in the 1990s. reserve requirements, or when they needed to By 2000, more than 40 developing countries, fund their own quasi-fiscal deficits that arose including all but one of the 25 with the largest from negative spreads between their assets and financial systems, had government bond markets liabilities. (DelValle and Ugolini 2003), and more than 20 Another reason for the slow growth of pri- had central bank debt markets.The government vate credit was that banks themselves increased bond markets allowed governments to reduce their net holdings of foreign assets for hedging their reliance on foreign borrowing.The supply purposes (see figure 7.2 above). Those in the of this debt was inelastic, but it was attractive to largest 25 developing-country markets went banks for several reasons: its interest rates had from essentially a balanced foreign position in been freed,it carried a low capital requirement,it 1990, on average, to net borrowing of nearly 1 was less risky than private debt,and it had liquid- percent of GDP in 1993 and nearly 3 percent of ity once the markets became active. GDP in 1997,before reverting to being net hold- The growth of domestic equity and bond ers of foreign assets in 2000 (Hanson 2003b). markets contributed to private sector financing After 1997, external lenders cut credit lines, in East Asia and India during the 1990s. In the banks wound down their external borrowings, major East Asian equity markets, market capital- and banks increased their external assets.19 ization exceeded $20 billion in 2000, having Given the limited growth of private sector risen 80 percent or more (except in Thailand) credit,a variable that has been shown to be linked since 1990.Turnover averaged more than 50 per- with economic growth,20 it is not surprising that cent and listings in the individual countries rose the rise in GDP growth associated with the at least 40 percent between 1990 and 2000, to financial (and general) liberalization of the 1990s the point where they all exceeded listings in was less than hoped for. every Latin American country except Brazil However, the story is more complex than the (Standard and Poor's 2003).The Indian market slow growth of private credit.In the major devel- was even more successful in providing resources oping countries, especially in East Asia, the aver- to a wide group of firms after listing regulations age growth of private sector credit (especially were eased (see box 7.1 above). Chile's market including external credits) and of GDP was much also did reasonably well, though its turnover was faster before 1996 than after.About 1995, some low because of the pension funds' buy-and-hold countries began to experience financial crises. policies. However, even in these countries, banks Much of the private credit extended by banks and remain the main source of finance. nonbanks proved to be unproductive,in the sense Elsewhere, equity markets were less success- that it became nonperforming before or during ful. On the seven largest Latin American stock the crises.During bank restructurings,these cred- exchanges, listings have declined since 1997, and its were replaced with government debt (to on five of those seven they have declined ever ensure depositors were paid); when eventually since 1990; turnover in all seven is less than 50 210 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s percent of market capitalization (Standard and ous institutional advantage.This suggests that in Poor's 2003). In transition countries, equity mar- the short run, equity market growth mainly kets were created as part of the privatization reflects general economic performance, which process,22 often with only belated recognition of attracts foreign investors willing to risk sums that the importance of regulatory frameworks. List- are small to them but large relative to the market. ings declined in most of these markets as some Simply setting up a market may not add much to privatized companies were taken off the market. growth or allow firms to raise funding.Over time, Market capitalization is less than $20 billion, however, as institutions improve, equity markets except in Poland and the Russian Federation, do seem to contribute to economic growth and turnover exceeds 50 percent only in Hun- (Levine 2003; Levine and Zervos 1998).Another gary (Standard and Poor's 2003). important element in equity market performance Several factors lie behind the slow develop- seems to be foreign investor participation ment of equity markets in developing and transi- (Bekaert, Harvey, and Lundblad 2003). tion countries.The first is that potential investors Private bond markets grew even less than are deterred by the low turnover in these markets equity markets in the 1990s.They share the prob- (usually much less than 75 percent compared lems of equity markets as well as having some of with 85 percent in even the smaller deciles of their own. Concerns about potential future traded companies on the U.S. NASDAQ) and by macroeconomic instability have led bond buyers low liquidity, which reflects the small sizes of the to demand high returns for committing funds for listed companies as well as the low turnover the long term, and deterred issues of long-term (Shah and Thomas 2003). Second, listings on bonds. Often only public sector firms issued stock exchanges have been reduced by takeovers long-term bonds, and then only in a few coun- of firms by multinational corporations, and trad- tries, notably in South Asia. Potential buyers of ing has been reduced by the migration of major private bonds were also deterred by lack of pro- firms' listings to industrial-country markets. In tection in law and in fact for bondholders'rights, 2000, companies listed offshore accounted for and by the lack of good bankruptcy legislation about 55 percent of the market capitalization in and enforcement. Nonetheless, some private 15 middle-income countries, and for 27 percent bond markets have developed, for example, in of the market capitalization in 25 low-income India, and in Mexico, recently, for securitized countries; much of the trading in these stocks housing finance. also takes place offshore (Claessens, Klingebiel, and Shmuckler 2003). Family firms that could External Finance for the Private Sector list often do not, partly because the benefits are not great, partly because these firms often have Within the private sector in developing countries, privileged access to credit through related banks, external borrowing grew faster than domestic and partly because they fear that dilution of own- borrowing in the first part of the 1990s, as large ership will reduce their control.Third and more private companies increasingly drew on external fundamentally, weak institutional factors--poor credits. For example, in 17 of the countries with information, poor treatment of minority share- the largest financial markets, the ratio of private holders, and weak regulation of market partici- sector foreign borrowing (of more than one year's pants--weaken the interest of investors, both maturity) to borrowings from domestic banks domestic and foreign, in many equity markets increased fairly steadily,from 16.5 percent in 1990 (Glaeser, Johnson, and Shleifer 2001; LaPorta, to 27 percent in 1997.23 Short-term borrowing López de Silanes, and Shleifer 2002b; Black also grew substantially. However, after 1997 these 2001). credits slowed in dollar terms. In the same 17 The better performance of the East Asian and countries, external credit to the private sector Indian markets seems to result more from supe- changed little in dollar terms. However, the ratio rior economic performance than from any obvi- of these credits to domestic credit rose by 50 per- F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 211 cent by 2000, reflecting crisis-related devaluations many foreign currency loans and led to calls on and the removal of private sector credits from the guarantees,worsening the difficulties of firms banks in restructurings in these countries. and banks. The external credits to the private sector were narrowly distributed.They went only to interna- Financial Intermediaries tionally creditworthy borrowers, and four coun- tries accounted for the bulk of private sector Most of the impact of financial reforms on the external borrowing (in dollars) in 2000: Brazil institutional structure of the financial sector was (27 percent), Mexico (12 percent), Indonesia (9 not felt until the latter half of the 1990s.This lim- percent), andThailand (7 percent). ited the gains from liberalization during the Offshore equity sales were another source of decade and contributed to crises. capital for many large private companies in the State banks, with their well-known problems 1990s.The numerous developing-country com- (LaPorta, López de Silanes, and Shleifer 2002a), panies that were listed offshore in 2000 largely decreased in importance only after 1995, and reflected issues of global depository rights and indeed still dominated many financial systems in American depository rights during the 1990s. Of 2000 (figure 7.3).The continued large state bank course,this source of capital was also narrowly dis- presence meant that credit allocations changed tributed. only slowly, despite liberalization.The problems While financial liberalization benefited large, of state banks after liberalization were most obvi- well-run companies and, indirectly, other bor- ous in the transition countries,25 where the rowers in developing countries, it raised banks' banks often simply continued to lend to tradi- risks.The best firms obtained loans and equity tional clients or were captured by politically finance offshore at less cost than in the domestic powerful groups; as a result, their loans were market, albeit with currency risk.24 This left a unproductive and their already large portfolios of larger portion of the limited domestic private nonperforming loans increased. State banks in credit available to other borrowers, but it also other countries had similar problems.The con- increased the average risk in the banks'loan port- tinued dominance of these banks, the associated folios. Moreover, banks in developing and transi- weakness in credit allocation, and the implied tion countries increased their net intermediation state guarantees that allowed them to raise of external loans up to 1997, especially in East increasing deposits despite their high incidence Asia,and they also guaranteed some direct exter- of nonperforming loans, all limited the gains nal borrowings by the corporate sector, typically from liberalization and accounted for a substan- off their balance sheets. tial part of the cost of crises in the 1990s. The external borrowings were a major factor Private banks changed gradually with liber- in the East Asian external payments crises and alization, entry of foreign banks, and fiercer were also important in other crises of the 1990s. competition, but their deficiencies also con- As external borrowings grew, lenders shortened tributed to unproductive lending and crises. maturities,creating maturity mismatches for bor- Their credit management skills did not keep rowers.Further,loans made by financial interme- pace with changes in the environment such as diaries based on their own external borrowing, the growth of foreign currency operations and though typically matched in terms of currency, the greater competition that their traditional entailed substantial risks when the borrowers borrowers were facing in the real sector. More- lacked an assured source of foreign exchange. over, many private banks in East Asia and some Eventually,lenders refused to roll over their cred- Latin American countries were parts of indus- its because they considered the risks too high. trial-financial conglomerates and continued to This generated both a banking crisis and a for- provide funding to their increasingly unprof- eign exchange crisis.The resulting sharp devalu- itable industrial partners. State banks that were ations increased debt-servicing problems on privatized to local buyers in weak institutional 212 FIGURE 7.3 State Ownership in Banking, 1998­2000 Source: Map Design Unit, World Bank. F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 213 environments often suffered similar problems and had to be renationalized. BOX 7.2 Foreign banks enlarged their role as new policies Nonbank Financial Intermediaries eased restrictions on their entry in the latter half of (NBFIs) in the 1990s the 1990s,particularly in transition countries but also in Latin America and Africa.26Their entry increased N BFIs such as finance companies, competition in banking and cut costs for bank co-op banks, and nonbank clients.They competed fiercely for the best clients financial corporations exist in and drove down profits on business with them, and many countries and in the 1990s some of they also competed in lending to small firms.27 them were an important factor in private A second approach to increasing competition, sector credit and deposit mobilization. For taken by a few countries, was to simply allow more example, India eased restrictions on non- banks, by lowering entry requirements.28 Unfortu- bank financial corporations in 1992 and by nately, many of these new banks were "pocket 1996 their deposits were equal to more banks," capturing deposits to lend to their owners' than 5 percent of broad money (box 7.1 businesses and often suffering problems. A third above). In Thailand and Malaysia, finance approach was to allow the growth of nonbank companies' deposit and credit growth financial corporations (box 7.2).These intermedi- picked up in the early 1990s. In Latin aries also often suffered from problems of risky and America, co-op banks and housing banks connected lending and were often the first to fail in Colombia have been important for some when credit tightened. time. NBFIs usually offered higher deposit Greater competition can also create problems rates and credit in different forms and to for banks by cutting their profits (Caprio and different clients than banks--for example, Summers 1996; Dooley 2003). Although this loans for construction, consumer credit, problem seems to be mainly one of adjusting to and small borrowers. NBFIs also were often competition (Demirgüç-Kunt and Detragiache subject to easier regulations on interest 1998), it does force owners to decide whether rates, reserves, and capital than were they should continue costly competition, try to banks, as well as less supervision. How- exit, or loot the bank.Thus regulation and super- ever, NBFIs had a history of periodic crises vision, particularly with regard to bank interven- in Latin America and East Asia, as, for tion and exit, are important issues when example, in Thailand in the 1980s (Sun- liberalization increases competition. dararajan and Balino 1991, 47­48). After 1997, many NBFIs in India, Malaysia, and Regulation, Supervision, and Deposit Thailand went bankrupt, depositors Insurance shifted to banks, and, to some degree, banks increased their loans to the former Banking Regulation and Supervision NBFI borrowers. Improvements in the prudential regulation and supervision of banks lagged behind the liberaliza- tions of the 1990s and contributed to crises.The oversight of banks in developing countries started from a low base in the 1990s because, during the but implementation--a political as well as a tech- period of financial repression, bank supervisors had nical issue--often lagged, even after costly crises. focused on compliance with directed credit rules. Enforcement was patchy, even of weak regula- International standards for supervision--the 25 tions on income recognition, provisioning, capi- Basel Core Principles--were not agreed upon until tal, and connected lending, and weak banks September 1997. Countries did enact their own continued in operation. International standards prudential regulations and upgraded supervision, on minimum bank capital were not set until 214 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s 1988,in the Basel agreements between industrial The actual impact of deposit insurance and countries for internationally active banks. guarantees has been mixed.The statistical evidence The issues in improving regulation and super- suggests that the gains from deposit insurance vision were not just technical but also political. depend on its particular features,its credibility,and The crises of the 1990s did engender attempts to the institutional environment (Demirgüç-Kunt improve regulation and supervision. But even and Kane 2002;Demirgüç-Kunt and Detragiache then, regulations were often not strengthened 2002). In many cases, the insurance created large immediately and forbearance was used to limit contingent guarantees,increased moral hazard,and the capital injections that governments otherwise reduced market discipline (Demirgüç-Kunt and would have had to make. For example, in East Kane 2002; Demirgüç-Kunt and Sobaci 2001; Asia, regulations on capital, income recognition, Demirgüç-Kunt and Huizinga 1999). Large and provisioning lagged behind international lender-of-last-resort support and blanket guaran- standards after the 1997 crisis (Barth,Caprio,and tees in effect provided unlimited insurance not Levine 2001),and actual capital in many Indone- only for depositors but for owners,many of whom sian andThai banks was still well below the Basel looted their banks.They were particularly ineffec- standard in 2000. tive in the context of open capital markets and By letting weak banks overexpand, the poor political and economic turmoil: for example, in oversight contributed to the crises of the 1990s.In Ecuador (IMF 2004b) and in Indonesia, liquidity developing countries, weak banks that were support to banks was almost as large after the allowed to continue operations often opted for a introduction of blanket guarantees as it was before. high-risk/high-return lending strategy or, in the One reason may be that as the likelihood increases worst case, were looted, as has also occurred in that the deposit insurance or a blanket guarantee industrial countries. Market discipline, which will be used, its cost and credibility come into might have restrained the expansion of weak question,and runs on banks and the currency may banks, was limited by poor information and increase (Dooley 2000). implicit or explicit deposit insurance. Various attempts have been made to adjust deposit insurance so as to reduce moral hazard Deposit Insurance and increase market discipline ex ante, but usu- Deposit insurance and, in crises, blanket guaran- ally depositor losses have been socialized ex post. tees, were standard recommendations of many For example, insurance limits have been placed financial advisors, and formal deposit insurance on large deposits and on deposits carrying the was initiated or improved in nearly 40 countries high rates that are often offered by weak banks, in the 1990s, mostly in transition and West but often the limits have not prevented the insur- African countries (Demirgüç-Kunt and Kane ance from extending to all depositors in a crisis. 2002; Demirgüç-Kunt and Sobaci 2001). In Another approach has been to use risk-based addition, countries such as Ecuador, Indonesia, deposit insurance premiums,in an attempt to off- Korea, Malaysia, Mexico,Thailand, and Turkey set the moral hazard and market discipline prob- introduced blanket guarantees of bank liabilities. lem, but in practice the differentials in premiums Deposit insurance and blanket guarantees are have been substantially smaller than the differen- mainly attempts to reduce the risk of bank runs. tials in bank risk (Laeven 2002a, 2000b, 2000c). Deposit insurance also has the secondary, con- This probably reflects the political power of the sumer protection benefit of protecting unsophis- local bank owners who benefit most from ticated depositors. Governments liked deposit deposit insurance. insurance as it appeared to give benefits yet had To sum up, the schemes that were introduced no costs, at least until a crisis arrived. Local pri- for the support of depositors tended to create vate banks, often politically important, liked it large contingent liabilities and to increase moral because it improved their competitiveness with hazard while reducing market discipline.They state and foreign banks. contributed to crises and volatility by encourag- F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 215 ing the funding of weak institutions after liberal- The results were not as good as anticipated. ization. Depositors and external lenders, expect- First, all systems had to cope with the change- ing to be bailed out of any problems by a over problem of paying existing pensioners government guarantee, tended to supply too while investing the contributions to the new much funding, particularly to state banks and system into assets.Without large fiscal surpluses, well-connected financial-industrial conglomer- the change-over generated a large increase in ates.Market discipline,which might have limited government debt that the new pension system this funding, was negligible, not just because of had to hold,as occurred in Argentina.As a result, weak information but also because of the implicit the demand for long-term private instruments and explicit guarantees. did not rise much. Thus the initial impact of pension reforms was simply to make the gov- Equity and Bond Market Regulation ernment's liability transparent. A second issue is Improvements in equity and bond market regula- that because capital markets typically are small, tion began in the 1990s and also proved difficult pension funds either generated price rises, as to implement. Even improving trading rules was happened even in Chile, and/or had to invest in difficult because of the difference between the bank debt,as happened in Peru.Third,costs have interests of buyers and sellers, on the one hand, been high in many of the private pension funds, and the short-run interests of market operators, reflecting set-up costs, insurance linked to the on the other.Also difficult to resolve has been the pensions, and a response to advertising that conflict between the interests of majority and encouraged excessive shifts between funds. minority shareholders.Attempts to create markets Some of these problems could have been overnight have had only limited success, not only reduced and country risk decreased for the indi- in cases of limited regulation (Czech Republic), vidual accounts by allowing the funds to invest but also where investor protection rules appeared externally, but countries have usually tried to to be reasonably good (Russia). Regulation in retain pension contributions and avoid possible Poland seems to have been relatively successful, balance of payments pressures. however (Black, Kraakmen, and Tarassova 2000; Glaesner, Johnson, and Shleifer 2001). As with Financial Sector Crises bank regulation and supervision,the issues are not merely technical but also political. Financial sector and external payments crises were features of the 1990s.29 Costly crises occurred in Mexico, the East Asian "Miracle" Pensions countries, Russia, Brazil, and some Eastern As described in chapter 6,a major change in pen- European and African countries.The new mil- sion systems occurred in the 1990s, with many lennium began with crises starting in Argentina countries shifting from pay-as-you-go systems to and Turkey and high nonperforming loans in systems in which at least part of pension income China.Africa also suffered costly financial crises is based on full funding for individual accounts. (figure 7.4). Chile was the first developing country to adopt What role did financial liberalization play in this approach,in 1981.Among the countries with the financial and currency crises of the 1990s, large financial systems, Argentina, Colombia, dubbed the"twin crises"by Kaminsky and Rein- Mexico, and Peru in Latin America, plus Hun- hart (1999)? The discussion below first assesses gary, Poland, andThailand, all adopted variants of the roles played by macroeconomic problems, this system after 1994 (Fox and Palmer 2001).The financial liberalization,and weak lending by state new systems gave individuals much better access banks and financial-industrial conglomerates and to their pensions and held the promise of gener- then outlines the difficult tradeoffs that policy ating demand for long-term financial instruments makers faced in responding to crises over a short and thereby stimulating capital markets. time horizon. 216 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s such cycles,reflecting liquidity squeezes on banks FIGURE 7.4 that have borrowed externally; problems with Selected Financial Crises, 1980­99 borrowers, especially those indebted in foreign currency; and runs on the banks to speculate on Estimated starting year and restructuring cost the currency. 60 Various events may trigger a crisis. External AR INDO 50 GDP shocks include deteriorating terms of trade, IM fo CHK 40 increases in international interest rates, and TH UR increases in risk premiums in industrial-country 30 KO estructuringr Cdl markets that automatically affect developing- BR fo centage 20 VEN BU EC MX country debt.32 Contagion in financial markets per PH TK a CZ MY 10 SG has also been cited.33 Domestically, unstable or Cost as inconsistent macroeconomic policies sooner or 0 later lead to pressures against banks and the cur- 1975 1980 1985 1990 1995 2000 rency. Political developments, such as the ouster Source: Caprio et al. 2003. of presidents Marcos in the Philippines in 1986 Note: The figure illustrates the estimated financial costs of restructuring after selected and Soeharto in Indonesia in 1998, lead con- crises, but not the losses in GDP. nected parties to liquidate their assets, putting pressure on banks and lenders with whom they did business. Macroeconomic Problems Most crisis countries had high debt and larger Financial Sector Liberalization than usual current account deficits and were pur- Financial sector liberalization seems to have been suing exchange rate­based stabilization poli- a factor in crises (Demirgüç-Kunt and Detra- cies.30 Many also had open capital accounts, but giache 1998, 2001; Kaminsky and Reinhart a causal association with crises is not clear:not all 1999). It increased capital inflows and deposits, countries with open capital accounts experi- which allowed rapid growth in credit to weak enced crises and, in the 1980s, crises had devel- public and private enterprises and the govern- oped even in countries whose capital accounts ment,as well as to real estate.Over time,the qual- were nominally closed. ity of the lending deteriorated.This may be one The combination of high debt and exchange explanation for the lags between liberalization rate­based stabilization seems to be associated and financial crises (Demirgüç-Kunt and Detra- with unsustainable booms in capital inflows, giache 2001), and between financial crises and imports, and GDP and shifts in relative prices, currency crises (Kaminsky and Reinhart followed by reversals in these variables as financ- 1999).34 Eventually, corporate bankruptcies, ing slows and maturities shorten, while interest banking problems, and runs on banks and cur- rates rise.The slowing of inflows reflects both the rencies developed, particularly when the rapid inherent characteristics of portfolio adjustment credit growth and inflows slowed, real growth and the growth of investor concerns regarding declined, and real interest rates rose.35 These debt-servicing capacity and exchange rate problems were often connected to unsustainable pegs.31 The rises in interest rates may reflect a fiscal policy and the defense of unsustainable cur- combination of smaller inflows, growing con- rency pegs with long periods of high interest cerns about the sustainability of the exchange rates. Problems in the timing and sequencing of rate peg, and attempts to defend the peg with liberalization, sometimes related to political tight money, often for long periods. Eventually, issues, also contributed to the crises (box 7.3). the exchange rate depreciates and debts need to In assessing the role of financial liberalization be restructured. Not surprisingly, the financial in the 1990s crises, an important question is why sector suffers a crisis in the downward phase of international lenders and domestic depositors F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 217 BOX 7.3 Problems with the Process of Financial Liberalization I n the late 1980s Nigeria liberalized interest rates mediating external inflows. Bank trust accounts were and bank entry but retained a multiple exchange liberalized and grew relative to bank deposits; they rate regime that was accessible only through also were allowed to take short-term commercial banks. This raised the demand for bank licenses, many paper, which had relatively free interest rates. Finally, of which went to well-connected parties who were the freeing of interest rates on consumer loans con- interested in arbitraging foreign exchange between tributed to a shift of loanable funds to these activities the multiple rates, not in banking. Though the num- and may have dampened Korea's saving rate, aug- ber of banks tripled, the ratio of deposits and credit menting the country's increased reliance on (short- to GDP fell, and, by the 1990s, banks were experienc- term) external borrowing. ing significant distress (Lewis and Stein 2002). Thailand set up its "offshore"/onshore Bangkok In Korea, the de facto rapid liberalization of short- International Banking Facility in 1993 with tax and term borrowings in the early 1990s, both internation- regulatory advantages that were justified as an ally and internally, led the heavily leveraged attempt to create a regional financial center oper- corporations to be increasingly financed by short-term ated by national banks. The facility allowed locals to inflows and through less regulated intermediaries. In deposit in foreign currency and local borrowers to the run-up to joining OECD, Korea had opened its cap- escape (albeit with short-term loans) from the gov- ital account by freeing short-term foreign borrowings, ernment's tight money policy and foreign cur- but left longer-term borrowings subject to restrictions rency­denominated loans. Its operations became a in an attempt to limit total capital inflows (Cho 2001). major factor in the expansion of Thailand's external This policy encouraged a maturity mismatch in lending debt (Bordo et al. 2001; Alba, Hernandez, and Klinge- and a currency mismatch on the part of borrowers, biel 1999). Pressure on the government from these especially since rates were much lower on foreign cur- borrowers was probably a factor in the government's rency loans than domestic ones. Although deposit lengthy, costly attempt to defend the baht even as it rates were formally liberalized in 1993, their rise was supported the borrowers, taking a monetary stance limited by moral suasion, government guidance, and inconsistent with the fixed exchange rate. After the high reserve requirements until 1996. New intermedi- devaluation, these obligations were a major factor in aries (finance companies converted to merchant the banks' problems and in the closure of many banks) sprang up to meet demands for funds by inter- finance companies. supplied so much funding. Part of the large explicit. Moreover, when liberalization led banks increases in loanable funds may have reflected a to lose franchise value and capital, weak regula- natural overshooting tendency in financial mar- tion and supervision did not prevent bank own- kets (Kindleberger 2000; Minsky 1992). But any ers from engaging in high-risk/high-return such tendency was certainly exaggerated by the lending or even looting. Nor did it limit banks' explicit and implicit guarantees that governments overexposure to related borrowers.Thus market provided to lenders. Government debt was discipline was eroded by government guarantees, directly guaranteed (although after crises it was implicit or explicit, while weak regulation and sometimes restructured). Growing private exter- supervision did not limit moral hazard. nal debt, funneled through banks or guaranteed Guarantees and their credibility may also by them, and growing deposits carried at least an explain why the crises in the 1990s seem to have implicit guarantee, which ex post often became happened relatively quickly (Dooley 2000). 218 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s According to this explanation, avoiding a crisis these debts worse because of higher real interest depends on maintaining foreign investors' and rates and lower inflation.Moreover,lower protec- depositors' perceptions that the guarantees (and tion and increased competition reduced tradi- the exchange rate peg) are credible. Events, tional borrowers' ability to service their debts. including fears of political change, can quickly However, the increased deposits and capital change these perceptions, leading to shifts into inflows associated with financial reform provided foreign exchange, curtailment of short-term new funds that enabled the banks to roll over their credits, and rollovers of maturing loans, trigger- loans again, adding to the ultimate volume of ing banking and exchange rate crises.36 nonperforming loans. For example, in the early stages of liberalization in the transition countries, Weak Lending "most state banks continued to lend as instructed A third factor in the 1990s crises was the weak or for patronage purposes" (Sherif, Borish, and lending, old and new, by the old financial inter- Gross 2003, 21). In East Asia, banks expanded mediaries, notably state banks and industrial- their lending to related industrial conglomerates, financial conglomerates. Before liberalization which were increasingly overleveraged these intermediaries had large overhangs of bad (Claessens, Djankov, and Lang 1998). In addition, debt, which had been rolled over several times to crises tended to generate a shift of deposits to state favored borrowers. Financial liberalization made banks, because of expectations of government BOX 7.4 Indonesia: Early Liberalization and Weaknesses Related to Political Connections I n Indonesia the freeing of interest rates and finance on the political regime. Capital outflow easing of capital and reserve requirements con- developed and rollovers of the large amount of short- tributed to large deposit growth, as well as a term external loans stopped as investor concerns doubling of the number of commercial banks (Han- mounted (despite the imposition of limits on cur- son 2001). By 1996, competition and the expansion rency speculation). Monetary policy was loosened to of 10 private banks had reduced state banks to about ease borrowers' problems. The November IMF program 45 percent of the system. However, all banks were brought little relief--runs on private banks and the very weak (World Bank 1996, 1997a). Despite the currency speeded up with the closure of 16 banks rules, state banks were overexposed to well-con- (small depositors did not begin to be paid until Jan- nected borrowers, and private banks to their owners. uary 1998) and the December illness of Soeharto. State banks reported low capital and their reported State banks, which benefited from shifts in deposits, nonperforming loans, though high, were under- made loans to well-connected borrowers on the basis stated, given the rollover of bad loans and other of projected exports that did not materialize. In Jan- maneuvers. At least two state banks were insolvent. uary 1998, outflows increased with the poor recep- Loans were often inflated by "commissions" to loan tion of the 1998­99 budget, panic buying of goods, officers. Private banks reported better figures but riots that frightened Indonesians of Chinese origin, weak supervision provided no check on them. Expo- and the possibility of introducing a currency board. sure limits were not enforced and many small banks The exchange rate fell to less than one-seventh of its were bankrupt. precrisis level. Massive central bank liquidity sup- The spillover from the July 1997 Thai devaluation port, often well in excess of banks' capital and in exposed these weaknesses and the dependence of some cases up to 75 percent of their assets, would F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 219 guarantees.37 This allowed further increases in state of the smaller provincial banks was exposed lending to favored clients who often used the by the spillover of Mexico's "Tequila" Crisis; the loans to buy foreign exchange and then defaulted support needed for their privatization amounted on the loans. to about half their assets (Clarke and Cull 1999). The overhang and growth of state banks'non- In Eastern Europe, the cost of the public sector performing loans, and their cleanup, were sub- banks' bad debt overhang was enormous--for stantial elements in the crises of the 1990s. A example, about 16 percent of GDP in Bulgaria notable example is Indonesia (box 7.4). In Thai- and about 18 percent of GDP in the Czech land, more than 80 percent of Bank KrungThai's Republic (Sherif, Borish, and Gross 2003). In loans became nonperforming.Brazil's BANESPA China, official estimates of the nonperforming (the state bank of São Paulo) was estimated to loans of the four largest state banks exceeded 20 have more than 90 percent nonperforming loans; percent of loans in 2003;various private estimates the estimated cost of the federal government's were much higher. 1997 cleanup, prior to privatizing the bank, was Privatization is often considered as a remedy about $20 billion or nearly 3 percent of GDP.The for the weak lending of public sector banks, but Finance Ministry has estimated that restructuring it has been costly in cases where the state has Banco do Brasil and Caixa Economica Federal retained a controlling interest or where sales have may cost $50 billion. In Argentina, the bankrupt been made to weak owners whose operations have doubled the money base had reserves not fallen developed, exacerbated by the concerns of the mid- (Kenward 2002; World Bank 2000c). Imposition of a dle-class Indonesian Chinese. As a result, the blan- blanket guarantee at the end of January temporarily ket guarantee stopped the bank runs only slowed outflows. Soeharto's reelection in March was temporarily--total liquidity support was nearly as followed by severe riots, often directed against large after the blanket guarantee was imposed as Indonesians of Chinese origin and Soeharto cronies. before, according to the figures in Enoch et al. Capital outflows rose once again, as did liquidity sup- (2001). Of the US$20 billion liquidity support that port. In May 1998 Soeharto resigned but pressures went mostly to private banks, 96 percent was unre- on banks continued. coverable and a substantial amount was diverted In sum, liberalization encouraged deposit growth into foreign exchange speculation, according to an and foreign inflows, but credit access depended not ex post study by the National Auditor. The cost of on profitability but on political connections, includ- the crisis is estimated at more than 50 percent of ing access to external loans from international banks GDP. Bank Mandiri, a merger of four state banks of (corporations did much of Indonesia's external debt which at least two were bankrupt before the crisis borrowing; the state banks' external borrowing was began, accounted for about 30 percent of the cost of limited by policy). Lenders and depositors looked at the crisis (more than 17 percent of GDP). More than connections, not at risk and corporate leverage. The 70 percent of the losses in the state banks were in easing of bank licensing and the lack of enforcement loans that had to be taken off the books. The poor of exposure limits worsened this problem. Then, quality of these loans is shown by the eventual when political concerns developed, the well-con- recovery rate of less than 30 percent, most of which nected tried to withdraw their assets and an outflow was realized four to five years after the crisis. 220 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s were poorly regulated and supervised. Mexico's Difficulties in Policy Responses to Crises 1991 privatization is perhaps the best known The crises presented difficult new policy prob- example. Soon after privatization, partly because lems. Traditional macroeconomic policies of of the currency crisis, the Mexican government tighter fiscal and monetary policy and devalua- was forced to renationalize the banks; it then tion were appropriate for reducing excess cleaned up their balance sheets at an estimated demand and current account deficits to finance- cost of more than US$70 billion and reprivatized able levels.But the financial sector problems,and them to international banks, beginning in 1998 their implications for the balance of payments, (box 7.5). raised a new set of more complicated issues and In Eastern Europe, the initial bank privatiza- tradeoffs,for which no single best practice exists. tions went poorly, particularly where govern- To deal with an individual bank's problems,the ments retained a controlling interest (Clarke,Cull standard recommended response is to provide liq- and Shirley 2003). In Africa, too, the experience uidity support at high interest rates and then inter- with bank privatization was often bad, with long vene with protection for small depositors. But delays and sales eventually made to undercapital- banks become insolvent well before they become ized owners who did not improve credit man- illiquid, and owners of insolvent banks may then agement and abandoned the banks when they choose a risky lending strategy or even attempt to lost their capital (box 7.6). loot the bank (de Juan 2002).38 Moreover, prob- BOX 7.5 Bank Privatization in Mexico M exico nationalized its commercial banks vented their purchase of the banks. The government in 1982. It decided to privatize them in renationalized the failed banks and protected the 1991, as part of liberalization and to raise depositors but taxpayers were left with a huge bill, fiscal resources. At the time, the privatization was estimated at 18 percent of GDP. acclaimed as a resounding success, fetching higher The features of the sale that were praised earlier prices than predicted. Although open only to domes- are now often criticized: the privatization for being tic purchasers, the sale was considered technically too hasty and the purchase prices as too high. Yet a well designed and executed. Bidders were first quali- sale was considered necessary to raise fiscal fied, and the auctions were transparent and quick, resources and sustain the government's commitment without scandal. to reform. The main mistakes were the exclusion of The macroeconomic (Tequila) crisis of 1995 took foreigners and the acceptance of purchases by polit- the shine off this success: loan defaults increased ically powerful but heavily leveraged buyers. The sharply with the collapse of the peso and the rise in exclusion of foreign bidders was partly a calculated interest rates. Failing banks were found to have made risk to shore up domestic support in a nationalistic poor loans under the relaxed regulatory framework, country. Even with foreign participation, highly often to politically powerful groups connected to leveraged locals might have bid more for the banks their controlling owners (Haber and Kantor 2003; through their access to loans. Thus, above all, the LaPorta and López de Silanes 2003). The connected Mexican experience illustrates not just a flawed pri- lending meant that the banks had effectively vatization but the complicated issues that bank pri- financed much of their own purchase. Taking into vatizers must juggle, including the difficult problems account loans from development banks, the buyers associated with dealing with local elites in a sector actually had minimal equity, but this had not pre- as sensitive as banking. F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 221 BOX 7.6 Bank Restructuring and Privatization in Sub-Saharan Africa A t the end of the 1980s many African banks internal controls were put in place, and new loan pro- were insolvent and illiquid. Governments cedures were gradually developed. undertook major restructuring programs over Treatment of depositors in failed banks varied the 1990s to deal with these problems. A gradual from country to country. In some, the government return to macro-stability and balanced government left the deposits in the restructured bank or reim- budgets--a prerequisite for bank restructuring-- bursed all depositors. In others, repayment of depos- occurred in programs supported by the International itors depended on the asset recovery of the failed Monetary Fund (IMF) and the World Bank. Directed institutions. Priority was given to compensation for credits were abandoned and interest rates liberalized. small depositors. Depositors incurred substantial Government arrears to the banking sector were often losses in Cameroon and the Republic of Congo, for securitized on various terms, with debt service often example. Some countries introduced deposit insur- guaranteed. Money markets were established. Many ance in the late 1990s, but it is unclear whether countries issued new banking laws, overhauled regula- these systems could handle banking crises as large as tions, and set up supervisory authorities. those of the 1990s. Bank restructurings were both organizational and Privatizations generally went to a major institu- financial and sometimes led to privatization, but the tional partner, often foreign. In some cases the for- process also often required multiple restructurings and eign banks were large and well known, with a was hesitant (World Bank 2001c). Some banks, partic- reputation to protect. African private banks that ularly public sector banks, were closed or weak operate in several countries have developed branches were turned into agencies. In Benin, the (Ecobank, Bank of Africa, Financial Bank, CBC, Stan- extreme case, all public sector banks were closed in bic) as a result of privatization involving foreign 1990, leaving the country without banks for some partners. However, in some cases, the foreign banks months until new private banks entered the market. In provided little improvement. other cases, bad assets were provisioned and losses As a result of the restructurings, African commer- were absorbed by existing shareholders (governments cial banks have become more solvent, liquid, and and the private sector); in a few cases, new capital profitable and a safer haven for deposits, but many was injected by the private sector; and in others, bad problems remain. In many countries, banks are still loans were removed from banks. Asset recovery corpo- weak in their lending and operations. Bank deposits rations were set up to manage bank liquidations have declined in some countries, probably reflecting a and/or to recover loans and reimburse mix of bank closures, discouragement of small depositors/creditors (for example, in Cameroon, Côte deposits, and civil strife. Commercial bank lending d'Ivoire, Ghana, Uganda, and Senegal) with mixed has generally been limited, reflecting crowding out of results. Management was changed, staff retrenched, government debt and bankers' selectivity in lending. lems in one bank typically indicate more wide- either intervening in weak banks, thus possibly spread problems, and closing a bank without provoking runs on other banks,39 or providing promptly compensating depositors may trigger liquidity support--loose money--that will spill runs on other banks and looting by bank owners. over into pressure on the exchange rate and As a bank problem becomes systemic, bank international reserves, especially in open runs turn into currency runs and pose severe economies (World Bank 2000c). problems for which there is no standard answer. Another choice is that of how much to sup- The government faces the unpleasant choice of port the exchange rate with reserves and tight 222 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s money (offsetting the loose money from liquidity and the use of government credit allocations to a support) versus how much to allow a deprecia- more market-based, internationally open system. tion.Tight money helps to protect the exchange Yet this shift, along with the other reforms, had rate, as in the traditional policy response, and thus less than the expected effects on growth.Access to helps borrowers in foreign currency, but it hurts financial services does not seem to have improved borrowers in local currency and it hurts banks, substantially in the 1990s, though there are indi- particularly if it is maintained for a long time.Liq- cations of improvements recently. Expectations uidity support and loose money will help borrow- may well have been too high.Another reason was ers in local currency, but put additional pressure an apparent "boom in bust[s]" (Caprio 1997), on the exchange rate that will hurt borrowers in related to macroeconomic policies but also to foreign currency. Use of reserves delays this prob- financial liberalization in the context of an over- lem, but reserves are finite and their decline can hang of weak institutions--financial intermedi- provoke a speculative attack on the currency. aries, financial markets, and informational, legal, Nontraditional policies have had only mixed and judicial frameworks. Problems in these areas success. Capital controls have not been effective in reduced the impact of liberalization and in some stopping currency runs.40A blanket guarantee may cases led to perverse results. or may not halt bank or currency runs,depending The weaknesses of institutions were not just on how it affects concerns about the credibility of a technical issue: they reflected the difficulty of the guarantee and the burden of future costs (Doo- changing the previous state-led development ley 2000).A few countries, including Argentina system and, more fundamentally, its underlying and Ecuador,have tried to stop bank runs by freez- political-economic basis within a short period, ing deposits and devaluing, but the disruption to while restraints on markets could be and were the payments chain has led to massive recessions.If quickly lifted.The overhang of these factors dur- deposits are to be written down, in parallel with ing the 1990s was an important reason behind loans, it is probably best to make a politically the following: unpalatable exchange of tradable bonds, as Argentina did in its January 1990 Bonex plan. · Credit allocation was weak and continued to Whatever is done,GDP growth is almost certain to go to the public sector, well-connected indi- slow if not decline (Frankel andWei 2004). viduals, financial-industrial conglomerates, In sum, the crises of the 1990s appear to be and traditional state bank clients. related to macroeconomic problems, but also to · Bank privatization was slow and partial priva- financial liberalization in the context of the over- tizations left control of intermediaries in the hang of old political and economic relationships, hands of government in many transition and manifested in state banks and politically powerful African economies,leading to continued pref- financial-industrial conglomerates. Government erential treatment of the traditional borrowers guarantees encouraged a rise of funding for these from state banks. intermediaries, which they channeled into weak loans.Eventually crises developed and the depos- · Privatizations and restrictions on foreign itors and external creditors were bailed out by entry in the financial sector often allowed governments. Privatization in these environ- local elites to retain or increase their eco- ments did not solve the problems, and often nomic and political power. required costly renationalizations. · Implicit and explicit guarantees of deposits and international loans supported local elites' ability to raise resources. 3. Lessons of the 1990s · Liberalization of bank licensing led to In the 1980s and 1990s the approach to finance "pocket" banks that mainly engaged in con- shifted from the repression of prices and markets nected lending or regulatory arbitrage, not F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 223 expansion of access (the record of nonbank absorbed part of the shocks. However, the capital intermediaries is somewhat better but they markets faced competition from implicitly or too were often linked to industrial-financial explicitly guaranteed deposits and external loans. conglomerates). Market development also was hindered by the inherent problems of capital markets in develop- · Development of the framework for capital ing countries and the difficulties of building up a markets--such as reasonable information, reasonable institutional framework quickly. legal and judicial treatment of bankruptcy, By the end of the 1990s, it became clear that treatment of minority shareholders, conduct much of the increased deposits and capital rules for market participants--was slow,com- inflows had gone into (1) unproductive private pounding the problems that capital markets borrowing or state enterprise debt that had to be in developing countries face in terms of con- replaced by government debt in order to bail out cerns about macroeconomic stability, high depositors and lenders,(2) deficit finance,and (3) costs, and low liquidity. central bank debt to stabilize the economy.Thus The process of liberalization and the limited it is not surprising that the financial liberaliza- nature of the results in the 1990s suggest four tions of the 1990s did not live up to the high major lessons, discussed next. expectations regarding sustained increases in growth or credit access. Finance Depends on Institutions Focusing on the poor quality of credits Perhaps the most important lesson of the 1990s exposes a common thread in the slow growth and for finance is that the financial sector's contribu- financial crises of the 1990s: the continuation of tion to development depends not just on resource preferential access, related to the overhang of old mobilization but also on attention to institutions: institutions, that was changed only slowly by the intermediaries, markets, and the informational, financial reforms. In many countries in the 1980s regulatory, legal, and judicial framework. and 1990s, public sector borrowing, with its Resources need to be allocated to those that offer implicit guarantee from future tax revenues, was the best combination of return and risk, and this excessive and eventually led to crises and slow depends on the quality of institutions. Building growth. But even in countries with smaller pub- up these institutions is not easy, takes time, and lic sectors and relatively limited fiscal problems, requires political support. such as Chile in the late 1970s and East Asian In the 1990s, the traditionally weak loans of countries in the 1990s, loans to industrial con- state banks and financial institutions linked to glomerates--made from the guaranteed deposits industrial conglomerates were further weakened in the private financial intermediaries that they by the higher interest rates that followed liberal- controlled or from state banks and international ization, as well as by increased import competi- lenders, to which they had preferential access tion and real appreciations that cut the because of the institutional setup--eventually profitability of traditional borrowers.Explicit and became nonperforming and contributed to crises. implicit guarantees allowed these financial insti- As noted earlier, the poor contribution of such tutions to obtain much of the liberalization- loans to sustained growth is shown by the low induced increase in deposits and capital inflows, value of the associated collateral when it was and to substantially expand lending to their tra- eventually sold. ditional borrowers, private and public. Regula- tion and supervision did not prevent this; their Delaying Needed Policies Is Costly weaknesses reflected not just technical but polit- Limiting the incidence and cost of financial crises ical issues. Market discipline was eroded by poor depends on resisting political pressures to pro- information and, more important, by implicit long unsustainable booms and to delay action on and explicit guarantees. Better capital market weak banks,41 as well as on avoiding socializing development could have relieved some risks and their losses.In the 1990s,governments often tried 224 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s to prolong booms and did not limit the expan- cially repressed 1980s. During the 1990s, the sion of weak banks.42 Unfortunately, such poli- growth in bank deposits (relative to GDP) speeded cies increased the ultimate volume of bad loans up in many countries.This acceleration reflected and the size of the crises. Then, after crises lower inflation, more realistic interest rates, and a occurred, governments typically responded by wider menu of financial instruments, including bailing out depositors and external investors foreign exchange­denominated instruments. In through liquidity support,expansion of whatever addition, domestic capital markets were started deposit insurance existed, and blanket guaran- and developed and private firms increased their tees, all of which generated large increases in external borrowing and external equity issues. government debt and contingent liabilities. Deposits and domestic capital markets per- Expectations that losses would be socialized, formed best where growth was already rapid, through explicit and implicit guarantees, also where there was a history of high deposit mobi- contributed to crises and volatility by encour- lization, and where investors were willing to take aging weak institutions to mobilize funds after risks to get equity shares in rapidly growing cor- liberalization. Depositors and external lenders, porations: East Asia and India. Elsewhere, deposit expecting to be bailed out of problems by a growth was less and capital market performance government guarantee,supplied funding to state was less good. Deposit growth picked up much banks and financial intermediaries that were less in Latin America, reflecting the region's his- part of financial-industrial conglomerates.The tory of inflation and government intervention. funding was well in excess of what could be Also,much of the growth was in foreign currency used productively. Market discipline, which deposits that complicated policy making. The might have limited this funding, was weakened decline in listings in equity markets in Latin by the implicit and explicit guarantees.43 The American and transition economies suggests that process was unstable, however. When a rise access to finance through equity issues did not occurred in the subjective probability that the widen much. Even where capital market per- guarantees would be called, net capital outflows formance was better,access suffered from the lack developed, as depositors and investors became of scale and liquidity in the markets;multinational concerned about how the guarantees would be takeovers of major firms; migration of listings to paid and funded.44 The capital flight was facili- less costly,more liquid industrial country markets; tated by the liquidity support to weak banks and and, more fundamentally, weak institutional the support of the exchange rate by reserve frameworks--in particular the lack of informa- sales.The combination of high initial returns, tion,regulatory protection of minority sharehold- limited losses on the funds that were taken out ers, and bankruptcy protection for bond holders. of the countries just before the crises, and the Private external borrowing and offshore equity ultimate provision of government guarantees issues did provide lower-cost funding but only to left the depositors and investors with good larger corporations in a few countries, and the returns during the 1990s.45 loans were subject to currency and rollover risk. Improvements in these policies will depend The slowdown in net private-to-private disburse- not just on new measures but also on strong ments and short-term loans was a major factor in implementation,which has been difficult even in the crises.Though it will be difficult,better devel- industrial countries. opment of domestic capital markets, even in the countries that have done relatively well, would Financial Liberalization Increases Financial reduce the impact of future crises. Resources A financially liberalized economy tends to gener- Successful Finance Depends on ate more financial resources than a repressed econ- Macroeconomic Stability omy.This is an old lesson (McKinnon 1973;Shaw Another old lesson is that successful financial 1973) that had been forgotten during the finan- liberalization and successful finance depend on F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 225 macroeconomic stability (World Bank, World became political imperatives in the 1990s. Development Report 1989). If anything, open Repressed finance had high costs and regressive capital accounts and volatile international capi- distributional effects. Over time, increasingly tal flows place a larger premium on sound politically active households have demanded pro- macroeconomic management. However, finan- tection for their savings and access to the invest- cial reforms, or at least more market-based ment opportunities that were once available only interest rates, were often put in place in the to political and economic elites.A second reason, 1990s in the midst of macroeconomic imbal- noted earlier,is that the increased access to exter- ances, complicating what was already a techni- nal financial markets brought about by the enor- cally difficult problem.46 For example,countries mous growth in trade, travel, and migration and with unsustainable fiscal policies often used by improvements in communications has made financial liberalization to continue their debt financial repression difficult.49 buildup and delay adjustment.47 Even when fis- Although macroeconomic stability, on which cal deficits were smaller than in the 1980s, the good finance depends, seems to exist in many countries that liberalized finance often had large countries, macroeconomic issues remain. First, in external and internal debt overhangs that con- today's open economies, slow policy responses or tributed to volatility. policy errors quickly translate into macroeco- Even a strong financial system has difficulty nomic instability. Second, large government debt protecting itself against default by an overhangs and/or large unfunded pension liabili- overindebted government, as the recent Argen- ties are problems in many countries.The burden tine crisis illustrates.48 Also, many countries that of these problems has been eased by low world liberalized were pursuing exchange rate­based interest rates,but rising world interest rates,as well stabilization, or had relatively fixed exchange as other shocks, may lead macroeconomic policy rates.These macroeconomic policies, and the astray. As the 1980s and 1990s show, excessive tight monetary policy and the credibility issues government debt can interact with inconsistent associated with them, often meant extended exchange rate and monetary policy to lead to periods of high real interest rates and burden- massive capital flight, large currency deprecia- some external borrowing, which eventually tions, and costly financial sector collapses.When contributed to countries' inability to service the government goes bankrupt, the financial sys- debt and to financial crises.Thus, the problems tem and the whole economy suffer. with financial liberalization, the crises, and the The financial liberalizations of the 1990s have limited results from financial liberalization in created a sounder basis for finance in at least six the 1990s often reflected macroeconomic pol- ways: icy deficiencies and the overhang of large exter- · Crises cleared away the "debris" of past non- nal debts. performing loans, although they left large holdings of government debt that created problems. 4. The Future of Finance · Intermediaries and capital markets have Looking ahead, the general pattern seems likely improved. In Eastern Europe, Latin America, to remain one of more market-based finance. In Africa, and even East Asia, many financial most countries the financial liberalizations of the intermediaries were gradually replaced by 1990s are unlikely to be reversed in their broad reputable foreign banks. Such banks have bet- aspects, barring large macroeconomic policy ter lending skills, are more able to engage in errors. A widespread return to financial repres- arms-length lending and resist government sion is probably now untenable for two reasons. pressures, and, potentially, impose fewer One reason is political: lower inflation and a demands on government for bailouts than the more market-based, more open financial system intermediaries they replaced. Capital markets 226 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s have also been set up or improved, but they of the 1990s naturally have raised concerns about still face many structural and institutional financial instability that can lead to poor growth. challenges. Governments, attempting to reduce the future costs of crises, have often tended to emphasize · Government and central bank debt markets prudence.But there is a tension between stability have developed.They allow central banks to and the ability of the financial system to carry carry out monetary policy more efficiently, out the key intermediary roles for develop- increase banks' liquidity, and allow less infla- ment--mobilizing funds from savers, allocating tionary finance of fiscal deficits.The growth of these funds to investors that will yield the best government debt markets also helps provide a combination of return and risk, reducing risk, benchmark that can make private debt mar- and shifting risk to those most willing to bear it. kets more efficient. A financial system that does these tasks well will · Access to credit is growing in some countries, contribute greatly to development. from foreign banks (Clarke et al. 2004), new Improving the tradeoff between stability and domestic banks, and bank-like intermediaries. intermediation in finance depends not just on With the closure of the old intermediaries, bad maintaining the systems of market-based interest credit no longer drove out good credit. New rates and credit allocations that arose during the intermediaries that hold the promise of a sus- 1990s, but also on the following: tainable increase in small-scale lending were able to grow. In Ecuador, for example, the col- · Reducing the crowding out of private credit lapse of the public sector intermediaries has left by the current large overhang of government room for dramatic growth in private banks' debt; small credits in the last two years. · Reducing the volatility of resource flows,par- · Information is improving.The accounting and ticularly on the upside of cycles and to weak auditing of intermediaries and borrowers is institutions; improving. So is information on small borrow- · Improving intermediaries and markets; and ers--public credit bureaus have been established in 23 (mostly transition countries) since 1994 · Widening access to credit. and the private credit bureaus that already The discussion below addresses each in turn. existed in many countries are improving (World Progress will depend heavily on countries' suc- Bank 2003d). cess in building institutions, improving their · Prudential regulation and supervision seem to informational and legal frameworks, and, ulti- be improving and, in a few cases, the combi- mately, achieving more competitive political sys- nation of regulation, supervision, and a better tems that will reduce the power of political- safety net has limited the impact of crises in economic elites. individual banks, for instance in Peru, although supervision has also missed major Reducing the Crowding out of Private Credit weaknesses in some countries, such as the Perhaps the most immediate obstacle to the abil- Dominican Republic. ity of the financial system to carry out its inter- mediary role, as well as a threat to stability, is the large overhang of government debt in many Improving Finance countries.50 It is often said that private credit is Further improvements in the contribution of currently limited by the unwillingness of banks finance to development depend on improving and markets to take risks. In fact, it is limited by the key tradeoff between safe and sound finance the large volume of inelastically supplied govern- and risk taking in the financial sector's interme- ment debt.This is because,to ensure that all gov- diation between savers and investors.The crises ernment debt is held (either by financial F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 227 intermediaries or by individuals), the spread ate--U.S. courts have often allowed individual between interest rates on government debt and creditors to seek preferential treatment. In the private debt has to be big enough to crowd out case of Elliot Associates vs. Peru, settled in 2000, enough private debt. Hence the way to expand Elliot Associates obtained a restraining order on the supply of private credit is not to try to make the payments on the restructured debt to which government debt less attractive but to leave more Peru had agreed with other creditor representa- space for funds for the private sector in the finan- tives. Peru eventually settled by paying Elliot cial system, or to make private debt and equity Associates $56 million for the debt that they had more attractive, so that more financial resources bought for $11 million in 1996. Ultimately, all can be raised in total. attempts at limiting excessive inflows depend on political will to limit a boom, while in practice, Reducing the Volatility of Flows and Its countries often have eased restrictions on capital Impact inflows in order to prolong a boom. Governments have made various efforts to reduce Internally, governments have tried to develop the volatility of flows, especially on the upside of capital markets as a shock absorber for the volatil- a boom, and to ease the impact of volatility, par- ity of external and internal flows. Funds invested ticularly by building up international reserves to in equity or long-term domestic government and offset shocks and, within banks, by externally private debt represent much less of a threat to the hedging foreign currency liabilities. economy than do volatile short-term external But much remains to be done. Some analysts capital flows.51Thus, capital market development have argued for reducing incentives to excessive could contribute to stability as well as assisting the capital inflows that can easily turn into excessive allocation of funds to promising activities. One outflows.They argue, for example, that India's problem, of course, is that investors in such success in avoiding the 1997 crisis was related to instruments demand high returns under the cur- its limits on banks' (and firms') offshore borrow- rent environment in developing countries, so ing,even as it allowed inflows into the stock mar- such instruments are often unattractive to poten- ket and liberalized direct foreign investment tial issuers.This problem adds to the structural regulations. Chile's implicit taxes on short-term problems of small size, lack of liquidity, and high inflows also appear to have had some success in costs that limit capital market development. reducing inflows, extending their maturities, and Domestic capital markets, particularly in the in limiting the impact of shocks, but at the cost larger countries,could be stimulated by improve- of reducing credit availability to the private sec- ments in institutional factors,such as better infor- tor (Edwards 1999; Forbes 2003). mation on firms, better rules on market conduct Another approach would be to reduce the and corporate treatment of minority sharehold- incentives to banks for increasing their net off- ers, and better legal and judicial treatment of shore borrowings.This would involve at least lev- bankruptcy. Generally, such improvements eling the playing field through application of the require substantial time and effort. same reserve, liquidity, directed credit require- Better market discipline is another approach ments, and premiums for "deposit insurance" as to enhancing both intermediation and stability. on domestic deposits. Here, too, little has been Market discipline means ensuring that depositors done. In the area of international bond issues, and international lenders have appropriate incen- some countries have begun to try to reduce the tives to limit their funding to weak intermedi- bias in bond buyers'beliefs that any restructuring aries,by ensuring that they stand to receive lower will favor them, by making restructurings easier returns on deposits and investments if problems in terms of lowering the percentage of bond occur. Market discipline complements govern- holders that is needed to accept a restructuring ment regulation and supervision and evidence offer (the Collective Action Clause). However, it exists that it can work in developing countries remains to be seen how this change will oper- (Martinez-Peria and Shmuckler 2001;Calomaris 228 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s and Powell 2001). Unfortunately, market disci- proved to be undesirable: they break down the pline depends on good information. Though payments chain and have contributed to large accounting and auditing are improving, much declines in output, as has happened in Argentina remains to be done. For example, regulations and Ecuador. But brief suspensions of deposit could encourage prompt dissemination of accu- withdrawals, while term deposits are replaced by rate information and impose stiff penalties for long-term, marketable instruments that involve a failure to do so. substantial discount (in present value terms), are a Perhaps more important, market discipline is possible alternative that would make current blunted by widespread implicit and explicit gov- depositors bear part of the cost of the crisis (Beck- ernment guarantees that developed in the 1990s. erman 1995;IMF 2004b).Of course,such policies To make market discipline work, governments are politically difficult to implement. But they face the difficult task of establishing credible lim- would not only limit the burden of crises that its on liquidity support, blanket guarantees, and future generations would have to pay, they might deposit insurance, so that at least the holders of also reduce the size of future crises,by strengthen- banks' large obligations consider themselves at ing market discipline. risk. One way to begin improving market disci- To limit weak lending and crises, governments pline might be to limit payoffs to large providers have also improved their banking laws and pru- of funds, especially since the latter can be dential regulation and supervision. Since the expected to have relatively good information strengthening of prudential regulation and super- about the strength of individual banks. It would vision only began in the later 1990s,not much evi- also help to prevent problems in one bank from dence has accumulated on how well it can work to contaminating the rest of the system. However, prevent crises.At the simplest level, regulators and the policy would immediately pass the problems supervisors in developing countries may lack the of a weak bank on to the central bank as lender of technical skills even to deal with loan quality and last resort--a role that also would need to be lim- provisioning, not to speak of more complicated ited, to contain costs. Deposit insurance would aspects of banking, such as evaluating complex also come into play, and would need to be truly operations in capital markets and foreign exchange, limited to small deposits. Premiums for deposit swaps and derivatives that are poorly valued in insurance would need to reflect differences in risk imperfect markets, and risk management models. in different classes of banks. Unfortunately, the Deficiencies exist in the consolidated supervision systems of risk-based premiums that have been of financial-industrial conglomerates and in the adopted have largely copied the pricing from supervision of offshore activities--important areas industrial economies and,though better than flat, in developing countries that will not be improved premiums still provide substantial subsidies to simply by giving supervisors more power. domestic private banks, probably because of the Partly these problems reflect incentives: typi- banks' political power. cally supervisors are poorly paid and have an When banks'problems have become more sys- incentive to shift into banking, especially once temic, the past responses--large lender of last they have been trained to handle tasks well.Often resort support and blanket guarantees--have supervisors are subject to lawsuits by bankers,even undermined future market discipline and been for actions in performance of their duties--which costly to future generations. In effect, they have makes them hesitant to raise issues.52 Protecting provided nearly unlimited insurance not only for supervisors completely from legal action raises depositors but also for owners who can loot their another issue--the risk that they will engage in banks.Alternative options for dealing with crises malfeasance. Hence a tribunal separate from the would need to begin with a different approach to court system is needed to deal with accusations of dividing the costs of the crises between current malfeasance by supervisors. holders of liabilities and future generations. More fundamentally, improvements in regula- Lengthy suspensions of deposit withdrawals have tion and supervision face substantial political road- F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 229 blocks, which have arisen in industrial as well as their interest in further expansion. Some banks developing countries. For example, from time to that expanded in Eastern Europe, in hopes of time, U.S. financial economists have raised con- establishing a presence before countries acceded cerns about some U.S.banks being too big to fail. to the European Union, suffered losses as com- Also in the United States, political forces and reg- petition developed.Some that expanded in Latin ulatory forbearance are often cited as a contribu- America have suffered losses from operations and tory factor in the U.S. savings and loan crisis. In from the developments in Argentina. In the many developing countries a few large banks recent re-privatizations of Indonesian banks,only dominate the system, and bankers and major bor- one bid came from a well-known global bank. rowers are often one and the same.In this context, Lesser-known banks have been expanding inter- regulatory capital does not have even the minimal nationally, but such banks can generate more incentives that it does in arms-length transactions supervision problems than local banks, because between intermediaries and borrowers. The of the problems with international supervision. industrial-financial groups are the principal entre- Moreover, without reputations to lose, such preneurs in many countries, even large ones, so banks may pull out when things go bad in the limits on connected lending are not feasible. If country or in their home market, leaving gov- problems of loan quality develop, the strength of ernments to bear the costs. the economic and political elite is likely to lead to regulatory forbearance. Even if supervisors can Improving Access to Finance identify capital insufficiencies and other regula- Increasing small clients'access to finance is a crit- tory violations, it would be difficult for them to ical issue for the financial sector in its support of stand up to monolithic political elites, particularly development. It involves the tradeoff between when the alternative is simply to ignore a problem making banks safe and sound and making sure in return for a supplement to their small salaries. they continue to intermediate.A prerequisite to Finally, the potential strength of regulation and increasing access is to reduce the absorption of supervision is limited by the still-important role of loanable resources by the government and the large state banks that carry out government poli- central bank. cies and are nontransparent almost by design. In Pressures remain great to direct low-cost sum,regulation may not be successful unless it also credit to small borrowers. Historically, however, empowers the market to monitor banks better, by these efforts have usually been unsuccessful, encouraging market discipline. undermining sustainable finance for rural and small and medium-sized enterprises, just as Improving Intermediaries occurred under financial repression. The entry of reputable foreign banks is one way A few intermediaries have successfully sus- to improve intermediation as well as to limit the tained loans to small borrowers (box 7.7).The cost of crises.Reputable foreign banks bring bet- more traditional banking operations among them ter-trained staff to the country and generally have common features that explain their success: have better systems for evaluating and managing interest rates that cover costs,good deposit mobi- credit risk than local banks. These advantages lization, containment of administrative costs, and often spill over into the local banking system, a high rate of loan collection,all backed by appro- from competitive pressures and the movement of priate internal incentives for good staff perfor- personnel. In addition, reputable foreign banks mance (Yaron,Benjamin,and Piprek 1997).Their also are likely to cover any losses on their loans or example needs to be followed.The informational operations without demanding government sup- infrastructure for small lending also improved port, so as to avoid damaging their reputations. toward the end of the 1990s with the founding Reputable international banks have entered and improvement of credit bureaus. many countries in recent years, but losses and, in Greater competition in banking services, some cases,their own lack of capital have limited through greater entry of banks and nonbanks and 230 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s intermediaries is whether politically they can be denied access to the bank safety net, or whether BOX 7.7 they should be regulated and supervised in the Extending Credit for Small Borrowers same manner as banks to protect taxpayers as well I n addition to the well-known examples as depositors. India appropriately resisted bailing of the Grameen Bank (begun in 1976) out depositors, but Thailand's attempt to offer and Bank Rakyat Indonesia after its these intermediaries access to liquidity funding 1983 reform (Robinson 2002), other suc- contributed to an easing of monetary policy that cessful lenders began to expand toward the was inconsistent with the pegged exchange rate. end of the 1990s. These included CrediFe in Another related issue is the size of the Ecuador, MiBanco in Peru, CrediAmigo in investor/depositor base: as it widens, the distinc- Brazil, and, in India, SEWAH (which uses a tion blurs between these institutions and banks, Grameen-type approach) and self-help the pressures for claims on the safety net increase, groups that use a mixture of the Grameen and the government may be drawn into supervi- approach and traditional banking. Some of sion and regulation. Such problems have these intermediaries received support from occurred in co-op banks in India and in coun- donors. The Grameen approach relies on the tries such as Indonesia, Nigeria, and Russia, social responsibility of borrowers who where banks were allowed to set up with negli- belong to a narrow group--an approach gible capital. In Indonesia, 48 of these banks, that has also been used by some banks. many run by the politically well-connected,bor- rowed from the lender-of-last-resort facility well in excess of their capital during the crisis (Ken- ward 2002), and used the funds to support for- looser regulations and supervision, is sometimes eign exchange purchases and related businesses recommended to improve access and lending in (see box 7.4 above). general. Certainly, regulations should provide Improving access, as well as the quality of room for intermediaries that take funds from credit allocation in general, depends heavily on groups of well-informed investors/depositors improving the informational, legal, and judicial and"nip at the heels"of banks,by offering better framework. The poor supply of information returns to depositors (though with greater risk), about borrowers, though improving, limits lend- along with better service and innovation in prod- ing to smaller clients. In some countries, this ucts and lending. problem has been circumvented by lending Exactly how these entities should function-- through third parties that in effect guarantee the for example, as venture capital funds or deposit loans.53 More generally,however,better informa- takers--and where the lines should be drawn tion would enhance competition for sound bor- between them and "banks," are country-specific rowers while giving borrowers an incentive to details. Such intermediaries operated in some service their loans to maintain good credit East Asian and Latin American countries and in records.Thus,the continued spread and improve- India in the 1990s, and the outcomes illustrate ment of credit bureaus will be an important their positive and negative sides. Before they fell development in improving access to credit as well victim to crises in 1997,the nonbank intermedi- as the quality of loans. Important issues that need aries increased finance for underbanked sectors to be addressed in this process are banking such as consumer durables and construction. But secrecy;how to make banks comply promptly and to some extent their success was not in competi- accurately with the requirement to provide infor- tion and innovation but based on regulatory arbi- mation;whether the credit bureau is to be private trage relative to banks, which were constrained or public; the inclusion of related information by interest rate controls (in India) or tight money such as installment purchases; and consumers' policy (in Thailand). A critical issue with such rights to challenge and amend the information. F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 231 Improvements in the legal and judicial frame- meant that resource allocation improved less rap- work, notably the definition and execution of idly. Implicit and explicit guarantees, by remov- collateral and bankruptcy laws, are also impor- ing market discipline, contributed to excessive tant in improving credit access and lending in expansion of lending for the low-productivity general. Financial intermediaries prefer not to projects of well-connected borrowers.Weak reg- execute collateral--they are mobilizers and allo- ulation and supervision reflected not just techni- cators of funds, not managers of firms--but the cal problems but also political pressures for threat of executing collateral gives an incentive regulatory forbearance.Large,generalized liquid- for prompt debt service. Good bankruptcy laws ity support during the crises often went to make the survival of viable firms easier and allow favored parties that bought foreign exchange shifts of physical capital from nonviable firms to with it. Information, which might have helped others, with creditors receiving the maximum market discipline and limited excessive lending settlement.The potential to improve credit access had guarantees been less, was not a focus of reg- through better information, contract enforce- ulation, and it suffered from the lack of trans- ment, and technology is great: in the United parency typical of many developing countries. States, the cost of processing a small loan is now Limited credit access reflected the crowding out below the price of a modest lunch. of public sector and central bank borrowing. In Good access to financial services also involves addition,it reflected a lack of information related efficient deposit and payments services--impor- not only to technical issues but also to the tant facilities given the increase in domestic and unwillingness of established intermediaries to international migration. In Africa, unfortunately, share information on their borrowers.Weak legal the strengthening of the banking system has in and judicial frameworks,designed to protect bor- some cases reduced access to deposit and pay- rowers and often responsive to economic and ments services for small transactions. In other political elites, reduced the incentives to service parts of the world, payments services are often debts and made it difficult for new borrowers to limited and uncompetitive. Post office banks-- gain access to finance by pledging collateral narrow banks, holding only government debt-- effectively. Capital markets, which might have with better technology,and banks providing only absorbed some of the shocks, grew slowly these services (for example, in Tanzania and because of the weak institutional framework and Mongolia) are examples of innovative ways to underlying structural problems. serve these needs. The lessons of the 1990s are that improving the contribution of finance to growth depends heavily on macroeconomic stability, govern- 5. Conclusion ments that are willing to take steps to limit unsus- tainable booms, a market-based approach, and While financial liberalization delivered in some the quality of institutions (financial intermedi- aspects during the 1990s, its benefits are likely to aries, information, and the quality of the legal lie in the future and to depend on further insti- and regulatory framework).The quality of insti- tutional reforms.The crises of the 1990s, and the tutions was not changed much by the stroke-of- limited contributions of liberalization to growth the-pen liberalizations of interest rates and credit and access to finance,reflect to a large degree the allocations. Improving these institutions, and continuation of the weak institutional frame- thereby improving financial intermediation, will work related to the overhang of the old financial depend on institution building, better informa- system and, more fundamentally, the persistence tional and legal frameworks,and,ultimately,more of old political and economic power centers.The competitive political systems. Success will freeing of interest rates and credit allocation depend on a mix of increased market discipline increased resource mobilization. But the persist- and limiting guarantees, better regulation and ence of the former institutional framework supervision that includes encouraging greater 232 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s market discipline of intermediaries, greater par- credits were also diverted into loans with free rates, ticipation of reputable foreign banks, and capital for example, through curb markets, or, when some deposit rates were freed, into deposits that paid market development. Government is needed to higher rates than the loan rates on directed credits. support better markets, without intervening 5. Abiad and Mody (2003) note the link between excessively in them, backed by an open political greater openness to trade and financial liberalization. process that limits the distortions of finance in 6. Capital controls, particularly in the context of favor of well-connected parties. macroeconomic imbalances, increase incentives for corruption, worsen the income distribution, and, because they fail, create disrespect for laws. Even in the 1970s, a high proportion of the massive capital Notes inflows into Latin America leaked out (Dooley et al. 1986). More recently, in China, net short-term out- 1. As Lenin cogently put it,"The big banks are the state flows of capital and errors and omissions in the bal- apparatus which we need to bring about socialism ance of payments were very large (World Bank and which we take readymade from capitalism" 1997a, 2000c). (quoted in LaPorta, López de Silanes, and Shleifer 7. For example,in Mexico after the post-1982 high infla- 2002a, 266).Thus communist, socialist, and planned tion, the limits on interest rates on agricultural loans economies nationalized domestic and foreign com- were below the rates on some deposits for a period. mercial banks. Gerschenkron (1962) was among the Rural borrowers often simply took their loan pro- first to provide academic support for the provision ceeds and deposited them, earning a positive return by government and state banks of funds for industri- on the loans with much less effort than by farming. alization and long-term credit. In addition to state 8. Abiad and Mody (2003). banks, specialized development finance intermedi- 9. Stock markets were opened to foreign investors aries, generally public, were set up to provide credits between 1986 and 1993 in the major East Asian and for small-scale industry, agriculture, housing, and Latin American countries and in India and Pakistan long-term industrial credit.They were financed by (Bekaert, Harvey, and Lundblad 2003). government-guaranteed external borrowing,includ- 10. Levine and Zervos (1998); Levine, Loayza, and Beck ing bilateral and multilateral loans; by low-cost (2000). directed credits from banks and other intermediaries; 11. The sharp fall in inflation in the 1990s made interest and by government revenues. Often these interme- rates more realistic, even with declines in nominal diaries went bankrupt, reflecting failures to collect rates; it also reduced other financial distortions asso- debt service and dependence on unhedged external ciated with inflation. Among the 25 developing borrowing. countries with the largest financial systems, those 2. For example, Brownbridge and Harvey (1998) with hyperinflation at the beginning of the 1990s describe such financial repression in Africa. reduced inflation sharply (in some cases, such as 3. Dornbusch and Edwards (1991);Alesina, Grilli, and Argentina, to single digits), while most of those with Milesi-Ferreti (1994); and Garrett (1995, 2000). initial inflation of 10­50 percent annually reduced 4. Estimates of aggregate subsidies range from 3 to 8 inflation to single digits by 2000. In Africa, inflation percent of GDP annually (World Bank, World Devel- also fell and in most transition countries,inflation fell opment Report 1989; Hanson 2001). Regarding allo- sharply from initial high levels. cations,in Costa Rica in the mid-1970s for example, 12. Foreign currency holdings also were often large rel- the public Banco Nacional's interest rate subsidy on ative to financial systems (Hanson 2002). agricultural credits was equal to about 4 percent of 13. As an example of the popularity of these measures,in GDP and 20 percent of agricultural value added. Peru after hyperinflation at the end of the 1980s, the About 80 percent of the credit went to 10 percent of 1993 Constitution (Article 64) guaranteed citizens the borrowers; the average subsidy on these loans the right to hold and use foreign exchange. More alone would have put each recipient into the upper than 50 percent of deposits are in dollars, even in the 10 percent of the income distribution (World Bank, non-Lima savings banks. World Development Report 1989). The situation in 14. The interest rates on foreign currency credits avoid other countries was similar. See Adams and Vogel the high, up-front cost of an expected depreciation (1986); Adams, Graham, and Von Pischke (1984); that may not occur for some time--the "peso prob- Gonzalez-Vega (1984); and Yaron, Benjamin, and lem"(Hanson 2002).This improves cash flows (lower Piprek (1997). Larger firms often accessed directed deficits for governments using cash accounting) and credit and on-lent it to their suppliers, capturing the increases a loan's effective maturity. Moreover, when spread between repressed and free rates. Directed a depreciation does come, the cost is spread out in F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 233 the amortization period. Not surprisingly, govern- 20. For statistical evidence on the importance of private ments borrow externally, and many countries, for sector credit in growth see Levine and Zervos example, Mexico in 1994 and Brazil and Turkey (1998); Levine, Loayza, and Beck (2000).The evi- recently, have indexed some domestic debt to for- dence of the link between savings/investment and eign currency. For private firms, there is also the financial sector liberalization is mixed (see,for exam- hope that a depreciation may lead to a government ple, Bandiera et al. 2000), but the investment ratio bailout, either by a favorable takeover of their for- does seem to have risen in the 1990s in the larger eign loans or an asymmetric conversion of domestic Asian countries, though not in the larger Latin foreign currency debts and deposits to local cur- American countries, and it actually declined in the rency, as occurred in Mexico (1982) and Argentina larger African countries.The difference between sav- (2002). However, foreign currency loans do increase ing and investment ratios may, of course, reflect dif- bank risks, even when matched with foreign cur- ferences in capital inflows. rency deposits, since the borrowers may not have 21. Crises, unproductive credits, and their links to the easy access to foreign currency earnings. Banks unreformed institutional and political framework could have adjusted the foreign and domestic cur- that remained after liberalization are discussed in the rency proportions of their balance sheets by varying section below on financial crises. interest rate differentials, but, given the demand for 22. Stock markets were reported as of 1991 in Hungary foreign currency deposits, the spread probably and Poland; in 1994 in Croatia, the Czech Republic, would have been high, creating moral hazard prob- Romania,Russia,the Slovak Republic,and Slovenia; lems in loans in domestic currency. in 1995 in Bulgaria,Latvia,Lithuania,and Mongolia; 15. These figures understate the relative growth of pub- in 1996 in the FormerYugoslav Republic of Mace- lic sector debt because they include China, where donia, Moldova, and Uzbekistan; and in 1997 in deposit growth was large and banks'accumulation of Estonia, Kazakhstan, and Ukraine (Standard and government debt was relatively small, but the accu- Poor's 2003). mulation of state enterprise debt was large. In those 23. This average is for the 17 of the 25 largest financial transition countries for which relevant data are avail- markets for which data are available on banks' able, privatization reduced borrowing by public domestic credit to the private sector. It excludes enterprises,thereby offsetting the rise in government China, India, and Korea, which do not report sepa- debt, but deposits grew only slowly and were largely rate data on private sector credits.These three coun- absorbed by increased central bank debt. tries are large external borrowers in absolute terms 16. Note that these figures understate the growth of pri- but are likely to have smaller ratios of private exter- vate credit in India and East Asia before 1997 and nal borrowings to bank credit than the average for overstate it after 1997, because of the growth and the 17 countries. decline of the nonbank sector. 24. The additional currency risk of these funds was less 17. Government debt was either injected into the banks than it might seem,as domestic credit in many coun- as part of restructurings or, in the case of deficit tries was increasingly denominated in foreign finance, sold at whatever rates would ensure its pur- exchange. chase.Thus,as a first-order approximation,the supply 25. "[The state banks'] commercialization as joint stock was inelastic (except for changes in the proportions companies was not accompanied by sufficient com- sold internally and externally).The liquidity,low risk, mercialization of their credit management, product and low capital requirements on government debt development, service levels, operational efficiency, or affected only the rate differential between the debt risk management.All this meant poor loan perform- and private credit that was needed to crowd out the ance and eventually insolvency.Many factors worked equivalent amount of private credit, rather than the against early detection of such problems--poor amount of government debt held, which was deter- accounting and auditing standards, inexperienced mined by the inelastic supply. supervisory personnel,inadequate prudential regula- 18. In some cases, the central banks also temporarily tions, decentralized and incomplete information sys- acted as large lenders of last resort. tems (often branch accounts not consolidated with 19. The increase in external assets probably reflected an headquarters accounts) and the traditional reliance attempt to hedge the risks from their foreign cur- on the government for additional funding when liq- rency liabilities,including deposits (Honahan and Shi uidity became short . . . . Management information 2003). Although banks' net external positions were systems were weak.All these factors worked against small in 2000,gross external assets and liabilities were timely and effective scrutiny of management behav- much larger than earlier (Hanson 2003b), suggesting ior" (Sherif, Borish, and Gross 2003, 21­22). that financial liberalization had increased banks' abil- 26. Western European banks entered Eastern Europe ity to diversify themselves. hoping to gain market shares before the European 234 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s Union expanded.The shares of foreign banks in the role,while lower interest rates contributed to the large number of banks and in total bank assets grew rap- capital inflows to developing countries in the early idly in Bulgaria, the Czech Republic, Hungary, and 1990s and again recently.The rise in international Poland. In Russia and Ukraine, however, foreign interest rates that started in 1993 probably con- banks represented only a small fraction of the total tributed to a gradual tightening of credit conditions number of banks, even in 2000 (Sherif, Borish, and for developing countries. Gross 2003).In LatinAmerica,Spanish banks became 33. A substantial literature has evolved over the possi- a major force by taking over state and private banks. bility that the crises in the 1990s, particularly those In Africa, foreign banks reentered and South African in East Asia, reflected contagion in financial mar- banks were playing an increasing role in southern kets, not fundamentals; see Claessens and Forbes Africa at the end of the 1990s. (2001) and works cited there. Contagion is one 27. Research suggests that in Latin America foreign explanation of "Generation II" models of crises in banks are at least as good as domestic banks at lend- which there are multiple equilibria, associated with ing to small firms (Clarke et al. 2004), and in India high and low rates of capital inflow.No doubt inter- foreign banks' lending to small and medium-size national investors exhibit some herding behavior firms has grown faster than that of state banks. for various reasons. Another explanation is that 28. Indonesia took liberalizing bank entry to an events in one country could lead external investors extreme, with almost "free banking" (box 7.4). Rus- to reevaluate the subjective risks in others and sia and Nigeria later followed a similar approach. reduce their exposures.This also would seem like Most new banks in these countries were "pocket" contagion. banks, capturing funds for their owners' firms. In 34. The lag between liberalization and crises seems fairly Indonesia,these banks were hit hard by the crisis and long (Demirgüç-Kunt and Detragiache 2001, 105). proved costly to the government when deposits were The lag may also reflect the difficulty of pinpointing guaranteed. liberalization and crises, both of which occur over 29. Caprio and Klingebiel (2002) list 117 systemic finan- time, as discussed in Eichengreen (2001). Demirgüç- cial crises (in which most of banks' capital was Kunt and Detragiache (2001) date liberalization from exhausted) in 93 countries and 51 borderline crises the removal of some interest rate controls and note in the period 1970­99. See also Sundararajan and that the estimated lag may reflect the gradualness of Balino (1991) and World Bank, World Development interest rate liberalization.Of course,the initial rise in Report 1989. deposit interest rates may also reflect part of a defense 30. Argentina, Russia, and some African countries had against a run on the currency,as,for example,in India high public sector debt compared to public revenues in 1991.The lag between financial crises and cur- (IMF 2004a). Other countries, notably in East Asia, rency crises may reflect liquidity support to weak had high private debt, including high external pri- banks at the start of financial crises,as discussed below. vate debt, relative to GDP. Variants of exchange 35. See Diaz-Alejandro (1985);the capsule discussions of rate­based stabilization were being used by country experience in Sundararajan and Balino Argentina, Chile, and Uruguay in the late 1970s and (1991,40­49);and the descriptions of financial crises by Argentina,Brazil,and Mexico in the 1990s.Other in Kindleberger (2000). countries, notably the East Asian countries and 36. Of course, this explanation is related to Generation Turkey in the 1990s, limited the flexibility of their II models of crises, discussed in footnote 33. exchange rates.The relation between the 1990s crises 37. State banks have not been closed without paying off and the current account deficits is similar to but not depositors, except in a few African countries. the same as"Generation I"models of balance of pay- 38. The United States has required intervention in weak ments crises (Krugman 1979). In the 1990s crises, banks well before capital is exhausted, and explana- the problem was not just financing the current tions if bank failures lead to deposit insurance pay- account deficit but net amortizations of long- and ments (Benston and Kaufmann 1997). It is unclear short-term loans, which could change suddenly. how well this approach would work in developing 31. Portfolio adjustments to improved investment countries. opportunities generate rapid inflows initially, fol- 39. Even if small depositors are promptly paid off, large lowed by a slowdown in inflows and net negative depositors may switch to foreign exchange. foreign exchange flows (because of interest payments 40. Arioshi et al. (2000); Dooley (1996).The Malaysian that require internal adjustment). controls are often cited as an example of effective 32. See, for example, Demirgüç-Kunt and Detragiache controls, but they were put in place after the crisis (2002); and Kaminsky and Reinhart (1999). In the was largely over (World Bank 2000c). crises of the early 1980s, high U.S. interest rates, as 41. To paraphrase William McChesney Martin, former well as the fall in petroleum prices, probably played a chairman of the U.S.Federal Reserve Board,the role F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 235 of governments is to take away the punchbowl before notwithstanding the past failures of this approach and the party gets too wild. the increases in access that are occurring. 42. Such government behavior occurs not only in devel- 50. Such debt is not completely bad--it can serve as a oping countries but also in industrial countries, for liquid asset to improve the payments system and as a example, in the U.S. savings and loan sector before its way for individual banks to deal with limited runs. crisis. However, governments' low revenue-generating 43. This weakness would have existed even if good infor- capacities make it difficult to service these debts, lead mation had been available. to cuts in public social and infrastructure spending, 44. Interestingly, additional deposits often flowed into and divert governments from developmental issues by state banks during these periods. Despite the weak- the day-to-day problem of rolling over the debt, rais- ness of their lending, the public typically considered ing the risk of a return to inflationary finance.These them to have better guarantees.These banks, in turn, potential problems have been eased by the fall in often made additional loans to weak borrowers. interest rates worldwide.However,when interest rates 45. See,for example,Klingen,Weder,and Zettelmeyer 2004. begin to rise again, and the costs of debt service cor- 46. Financial liberalizations, even gradual ones, are not easy respondingly increase, the problems may reappear. to manage. Errors in liberalization are not always tech- 51. International equity markets can also act as a shock nical; they sometimes reflect pressures by influential absorber, but only the largest and most transparent groups. firms can list in these markets. Offshore bond mar- 47. Financial liberalization also tended to increase the fiscal kets also are developing in private as well as public deficit and make it more costly to finance, as the gov- bonds; they reduce the risk of credit crunches but ernment lost seigniorage revenues and had to pay more increase currency risk. market-based interest rates on its debt. 52. Protecting supervisors completely from legal action 48. World Bank (1998a) describes the substantial strength- raises another issue, the risk that they will engage in ening of Argentina's financial system in the mid-1990s. malfeasance. Hence, a tribunal separate from the 49. Some policies and some countries will of course devi- court system is needed to deal with accusations of ate from the general trends. Some governments where malfeasance by supervisors. democracy is limited may attempt to impose capital 53. For example, making loans for scooters, cars, and controls and return to the inflation tax as a means of homes to workers in the formal sector who often capturing resources. And many countries remain con- cannot be fired; making loans to farmers that are cerned about the narrowness of credit access for their repaid by deductions from the contracts the farmers citizens, and seek ways to provide funds for rural and have with crop buyers; and lending to small and small and medium-size enterprise lending at below- medium-size enterprises either through larger firms market rates through specialized intermediaries, or by discounting their orders from such firms. Country Note F Lessons and Controversies from Financial Crises in the 1990s S ince financial markets came into rency,currency mismatches on the banks'or bor- being, financial crises have been rowers'balance sheets made the banks vulnerable their costly companions (Kindle- to devaluations. When the devaluations came, berger 1984). But the 1990s, loosely interpreted, balance sheet losses were often larger than the will be remembered for the severity of the crises banks' entire capital, and compounded the cur- that shook Mexico in 1994, East Asia in 1997, rency crisis with a banking one. Brazil and the Russian Federation in 1999, Governments and creditors in Turkey and Turkey in 2000, and Argentina and Uruguay in some countries in Latin America and in East Asia 2002.This country note looks at the origins and responded with deflationary policies that costs of these crises and at how they have increased their net exports and some write-down changed opinions on the use of capital controls, of the debt.This, together with an improvement the choice of exchange rate regimes, and in the external environment, eventually brought approaches to crisis management. the debt crisis to an end.The adjustment was tur- In the early 1980s, high real interest rates, bulent, however, and the costs were high.The declining export prices, and a global slowdown output collapses during the "lost decade" of the combined to raise the costs of debts that had 1980s were comparable to those in the 1929 cri- been contracted in the 1970s, when real interest sis: the gross domestic product (GDP) declined rates were negative and the external environment by 7 percent in Turkey during 1979­80, 15 per- was more favorable. Rising costs of debt accom- cent in Chile during 1982­83, 20 percent in panied by declining debt-service capacity and Uruguay during 1982­83, 11 percent in inadequate response to shocks produced the debt Argentina during 1981­82, and 12 percent in crisis of the 1980s. Bolivia during 1982­86.By the end of the 1980s, Many of the crises of the 1980s had a com- the per capita incomes of most countries in Latin mon origin. Countries attempted to stabilize America and in Turkey were only marginally inflation through programs anchored on a prean- higher, and in some cases lower, than at the nounced,fixed,or only slowly depreciating nom- beginning of the decade. inal exchange rate, while delaying needed fiscal Governments, banks, and economists learned adjustments. As adjustments in the nominal important lessons from the experience of the exchange rate lagged behind inflation, the real 1980s.Except inArgentina,governments reduced exchange rate appreciated even when inflation their recourse to external financing and increased declined. Attempts to defend the nominal their reliance on domestic capital markets to exchange rate while the real rate was appreciat- finance their budgets. Commercial banks in ing led to increases in current account deficits industrialized countries reduced their lending to and declines in reserves, which ultimately trig- foreign governments. Economists learned that gered external payments crises, with capital out- attempts to reduce inflationary expectations flows and large and sudden devaluations. In through reliance on a nominal exchange rate Argentina, Chile, and Uruguay, where banks anchor could not be credible if they were accom- were allowed to hold deposits in foreign cur- panied by adverse balance of payments develop- 238 LESSONS AND CONTROVERSIES FROM FINANCIAL CRISES IN THE 1990S 239 ments. They also learned, or relearned, that investors in industrialized countries, improve episodes of real exchange rate appreciation could developing countries' access to finance, and raise have devastating effects on the real economy if developing countries' investment levels and not driven by sustainable long-term increases in growth. productivity (the Balassa-Samuelson effect). Last The results challenged expectations.Financial but not least, they learned that domestic banks' crises occurred more often than in the 1980s exposure to exchange rate fluctuations could dev- (figure 4.21 in chapter 4) and the costs were,once astate the banks' balance sheets, and that this risk again,staggering,with declines in GDP similar to of currency mismatches stemmed not only from those of the 1980s.The 1990s will be remem- their own balance sheets but also from those of bered as a decade of macroeconomic crises and their borrowers. turbulence in emerging markets. Capital flows increased significantly in the late The average cost of a crisis has been put at 1980s and 1990s, reflecting the worldwide shift about 8 percent of GDP, and that of a financial to market-oriented policies,the decline in indus- crisis accompanied by a banking crisis at 18 per- trial countries' interest rates, the fall of the Soviet cent.1 Bank restructuring costs reached 50 per- Union, the spread of international standards of cent of GDP in Indonesia,and one-third of GDP banking supervision and accounting, progress in in Korea andThailand.The output collapses were, information technology, and regulatory changes for example, 6 percent of GDP in Mexico in in industrialized countries that allowed mutual 1994, 11 percent inThailand in 1997, 13 percent and pension funds, insurance companies, and in Indonesia in 1997,and 15 percent inArgentina banks to invest abroad.The resulting rise in flows in 2002. Even though in East Asian countries, to emerging markets was massive (Tirole 2002). Turkey, and Russia per capita incomes have Capital flows to developing countries, including returned to their precrisis levels and growth has the Republic of Korea, reached US$265 billion been relatively rapid, and Argentina has experi- in 1996, six times their volume at the beginning enced two years of rapid growth, the cost of the of the 1990s, and four times the peak reached crises has been simply staggering. during the 1978­82 commercial lending boom. In some countries, the crises of the 1990s Though these amounts were small in relation to were similar to those of the 1980s in several the economies of industrialized countries, they respects.As in the 1980s, large and unsustainable were extremely large in relation to those of current account deficits played an important role developing countries: 9.4 percent of Brazil's in Argentina, Brazil, Mexico, Thailand, and GDP (1992­95), 25.8 percent of Chile's Turkey, though they were not significant in the (1989­95), 9.3 percent of Korea's (1991­95), other crisis countries. As in the 1980s, these 45.8 percent of Malaysia's (1989­95), 27.1 per- deficits were often the result of stabilization pro- cent of Mexico's (1989­94), and 51.5 percent of grams anchored on a nominal exchange rate.And Thailand's (1988­95). as in the 1980s, in some countries (Argentina, The expectation was that these inflows would Indonesia, Korea,Thailand,Turkey) the currency help developing countries integrate themselves crisis triggered a banking crisis,itself the result of into the global economy, while diversifying currency and maturity mismatches either on financial risks and reducing economic fluctua- commercial banks' balance sheets, or on their tions. In the first half of the 1990s, it frequently borrowers'. became part of the International Monetary The crises of the 1990s were much more diffi- Fund's (IMF) advice to developing countries not cult to predict than the crises of the 1980s.While only to remove restrictions on their current many observers warned of impending crises in accounts (in line with the IMF's ArticleVIII) but Argentina or Mexico,few anticipated those in East also to remove restrictions on their capital Asia. In general, interest rate spreads remained accounts. It was also perceived that these devel- remarkably low in the months preceding the crises opments would help diversify the risks to (figure F.1). 240 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s FIGURE F.1 Interest Rate Spreads and Real Exchange Rates in Crisis Countries Mexico Thailand 20 140 10 120 120 100 16 8 100 80 12 80 6 60 60 8 4 40 40 4 2 20 20 0 0 0 0 1991 1992 1992 1993 1993 1994 1994 1995 1995 1996 1996 1997 1997 1994 1995 1996 1997 1998 1999 2000 2001 Dec. Jun. Dec. Jun. Dec. Jun. Dec. Jun. Dec. Jun. Dec. Jun. Dec. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. EMBI Plus spread, % REER, 1990 = 100 EMBI Global spread, % REER, 1990 = 100 Malaysia Korea, Rep. of 12 140 10 120 10 120 100 8 100 8 80 6 80 6 60 60 4 4 40 40 2 2 20 20 0 0 0 0 1994 1995 1996 1997 1998 1999 2000 2001 1994 1995 1996 1997 1998 1999 2000 2001 Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. EMBI Global spread, % REER, 1990 = 100 EMBI Global spread, % REER, 1990 = 100 Source: JP Morgan, EMBI+ database; World Bank, WDI. Note: EMBI = Emerging Markets Bond Index; REER = real effective exchange rate. The crises of the 1990s differed from those public sector savings, in several of the crises of of the 1980s in three other important respects. the 1990s they reflected negative private sector First,indebtedness by the private sector played a savings-investment balances.There is consensus more important role than in the 1980s, both in that fiscal deficits were not a serious source of terms of imbalances between saving and invest- vulnerability in Indonesia, Korea, Mexico, or ment and in terms of external debt.Whereas in Thailand (Summers 2000). But in some coun- the 1980s the current account deficits in the tries--for example, Indonesia, Korea, and balance of payments always reflected negative Turkey--fiscal accounts did not reflect the full LESSONS AND CONTROVERSIES FROM FINANCIAL CRISES IN THE 1990S 241 Russian Federation Brazil 140 80 10 120 120 100 8 60 100 80 80 6 40 60 60 4 40 40 20 2 20 20 0 0 0 0 1995 1996 1997 1998 1999 2000 2001 2002 1996 1997 1998 1999 2000 2001 2002 2003 Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. EMBI Plus spread, % REER, 1990 = 100 EMBI Plus spread, % REER, 1990 = 100 Turkey Argentina 12 160 80 250 140 10 200 120 60 8 100 150 6 80 40 100 60 4 40 20 50 2 20 0 0 0 0 1997 1998 1999 2000 2001 2002 2003 2004 1997 1998 1999 2000 2001 2002 2003 2004 Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. EMBI Global spread, % REER, 1990 = 100 EMBI Plus spread, % REER, 1990 = 100 consolidated fiscal and quasi-fiscal picture, and lending weakened the balance sheets of com- some observers have argued that prospective mercial banks,whose quasi-fiscal cost ultimately deficits--that is, those that the government increased the domestic public debt. In Turkey, a would incur if it had to bail out the banks or the significant portion of the costs of the banking private firms, or if an exchange rate collapse crisis stemmed from state banks lending to took place--were key to understanding these politically connected borrowers. crises (though on this reasoning every country Second, the 1990s made it clear that not only is vulnerable to a crisis). Politically motivated the stock but the terms of external debt mat- 242 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s tered, and especially the maturity structure. In In theory,these inflows should have been benefi- many of the crises of the 1990s, rollover risks cial, enabling countries to increase their invest- stemming from excessive short-term debt played ment rates. In practice, they were not an a more important role than did an unsustainable unmitigated blessing. The impossible trinity stock of debt or unsustainable current account posed complex choices for the authorities and deficits in the balance of payments. Russia ran often threatened the competitiveness of real surpluses in the current account of the balance of exchange rates. Even small adjustments in inter- payments both in the year of the crisis and in the national portfolio allocations to emerging mar- years preceding it. Indonesia and Korea's current kets caused swings that were very large in relation account balances were negative, but relatively to the size of these economies. In addition, the small, and under normal circumstances their availability of cheap finance encouraged external financing would not have been a problem (Sum- borrowing beyond prudent limits. Indicative of mers 2000). this, 7 of the top 10 recipients of private capital Third, twin crises (currency and banking) flows during the 1990s suffered financial crises were much more frequent than in the 1980s, (table F.1);the exceptions were Chile,China,and even though banking supervision and bank cap- India. italization were no worse than in the 1980s, and · Some let finance flow in freely to the national were much better in some cases (such as economy,allowing the nominal exchange rate Argentina's). to bear the full brunt of the adjustment.The The novel features of the crises of the 1990s advantage of this policy is that it allows the led to a reexamination of beliefs,regarding among authorities to maintain discretion over mone- others capital controls,exchange rate regimes,and tary policy.The disadvantage is that it implies crisis management.Though debates will continue, appreciation of both the nominal and the real the balance of opinion seems to be that the crises exchange rate, an increase in the relative price of the 1990s altered the conventional wisdom that of nontradables, a loss in the competitiveness prevailed in these three areas at the beginning of of exports, a reduction in the price of imports the 1990s (see Feldstein 2003 and Williamson and a consequent increase in the current 2004b for syntheses and alternative views). Essen- account deficit of the balance of payments, tially, the crises emphasized the so-called impossi- and an adverse real shock to the economy as ble trinity, whereby it is impossible for a the tradable sector loses competitiveness with government to simultaneously maintain an open imports. No country pursued this policy to capital account, an exchange rate target, and an the letter, but many let their nominal independent monetary policy. exchange rates appreciate to some extent, or lag behind inflation. Capital Account Policies · Some countries let financial flows flow in Controls on capital inflows can reduce the risk of freely but mitigated the impact on the nomi- currency crisis, but their desirability and feasibil- nal exchange rate by building up reserves.This ity remain controversial. In the world that policy has the advantage that it prevents nom- emerged from the 1929 crisis and World War II, inal appreciation of the exchange rate, but its sources of international finance were extremely cost is a loss of monetary control.In countries limited.Most countries severely restricted capital with a fixed nominal exchange rate, whether outflows. France, for example, controlled capital by law (Argentina), or in practice (China, outflows well into the 1980s. In the late 1980s Thailand), this came at the cost of a large and 1990s, countries whose policies had tradi- expansion of credit,which in turn put upward tionally focused on managing the scarcity of for- pressure on the prices of nontradables.Typi- eign exchange suddenly faced liquidity surges cally, central banks tried to sterilize increases stemming from large inflows of debt and equity. in liquidity through open market operations. LESSONS AND CONTROVERSIES FROM FINANCIAL CRISES IN THE 1990S 243 TABLE F.1 Financial Inflows and Major Financial Shocks Financial crises Rank of recipients, by absolute Private capital flows FDI flows, 1990­96, (country, year) volume of private flows, 1990­96 1990­96, % GDP (in 1996) % private capital flows Mexico 1994­95 2 33.0 42.8 Thailand 1997 6 27.1 22.7 Indonesia 1997 7 17.7 22.7 Korea, Rep. of 1997 -- -- -- Malaysia 1997 5 62.7 47.2 Russian Federation 1998 11 4.8 18.7 Brazil 1999 (2002) 3 12.6 20.7 Turkey 2000­01 10 12.1 22.1 Argentina 2001­02 4 23.9 33.4 China 1 25.2 68.2 India 8 7.6 20.6 Chile 9 39.4 37.2 Source: IMF and Bank staff estimates. --. Not available. While this reduced monetary expansion, the increased the requirement to 30 percent and resulting increase in real interest rates was, the holding period to one year, regardless of again,the cause of an adverse real shock to the the duration of the inflow. As the volume of economy. Further, it generated a vicious cycle inflows continued to grow, Chile continued of rising inflows, higher interest rates, even to gradually extend the coverage, up until more inflows, and so on.Also, because central 1998 (Edwards 2003). In 1992, Mexico lim- banks earned less on reserves than on domes- ited commercial banks'foreign liabilities to 10 tic treasury bonds, sterilization came at a high percent of their total liabilities. India, in 1994, cost--and in turn threatened fiscal stability introduced guidelines restricting issues of (Calvo 1998). equity abroad and setting annual aggregate ceilings on private borrowing abroad. · Some countries restricted capital inflows, using a variety of methods.The advantage of this policy is that it leaves the authorities in In retrospect, it is important to distinguish control of monetary policy, and reduces pres- controls on capital outflows from controls on sure for the exchange rate to appreciate. In inflows. Countries that restricted capital inflows 1990 Indonesia imposed a ceiling on total performed better than those that did not. Chile, external borrowing by domestic banks and China, and India all introduced controls on cap- public enterprises and Thailand restored a 10 ital inflows that helped them maintain some percent withholding tax on the interest paid degree of control over monetary policy and on foreign loans. Chile first introduced helped to mitigate upward pressure on the restrictions on capital inflows in 1991, in the exchange rate..Though their banking systems form of an unremunerated reserve require- were not without weaknesses, China and India ment of 20 percent of inflows for all portfolio avoided a financial crisis and also maintained inflows; for maturities of less than a year, the strong growth. Their experience is consistent reserve requirement applied for the duration with the view of some economists (Williamson of the inflow, and for longer maturities, it 1995; Bhagwati 1998; Feldstein 2003) that the applied for one year. In July 1992, Chile efficiency gains from liberalizing capital move- 244 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s ments are small in relation to the risks this liber- ments typically attempted to defend the rates alization introduces. There is nonetheless an using foreign exchange reserves. When the opposite view, holding that controls on capital reserves eventually ran out, governments had no movements are not only inefficient (Summers choice but to let their currencies go.The result- 1999, 2000), but also difficult to implement in ing exchange rates were often a small fraction of practice. In the case of Chile, some studies sug- the fixed rates: half in Thailand and Brazil, and gest that the controls were less effective than gen- one-fourth inArgentina,at least initially,before it erally believed and that they did not succeed in stabilized at one-third. increasing the average maturity of debt. Perhaps For the many companies and banks that were more important, there is no guarantee that capi- highly leveraged in foreign exchange and tal controls will work in other nations as effec- depended on domestic currency earnings to pay tively as they did in Chile.2 their liabilities, the doubling or trebling of their debt meant bankruptcy. Banks that had lent to these companies,whether in foreign exchange or Exchange Rate Policies in domestic currency, also went bankrupt.The Few economic issues were more hotly debated economic implosion that followed was stagger- in the 1990s than that of the appropriate ing. Among the countries shown in table F.1 exchange rate regime for developing countries. above,Brazil was the only one that escaped a dra- Maintaining competitive real exchange rates is matic decline in GDP.This was mostly because central for financial stability and growth, and it Brazilian companies were much less leveraged now appears that this can best be achieved than those of Indonesia, Korea, or Thailand, through flexible regimes that prevent real appre- where corporate debt-equity ratios were in the ciation from running ahead of a country's pro- range of 250­500 percent (Dornbusch 2001). ductivity gains. Flexible regimes are also more This experience suggests that likely to discourage currency mismatches at the level of firms or banks, and to provide a more . . . the existence of large amounts of pri- accurate picture of public indebtedness. vate debt denominated in dollars or other After a period of relative disfavor, fixed nomi- hard currencies is the most serious source nal exchange rates made a comeback in academic of economic hardship facing the economy. and policy circles in the late 1980s and early 1990s . . . Avoiding large amounts of dollar- (Edwards 2003).Notwithstanding Mexico's costly denominated debt,and particularly private experience with a rigid exchange rate regime in dollar debt, is probably the most useful 1994, there was a belief that in countries facing thing that a country and corporation can inflation, a fixed exchange rate could provide a do to reduce the serious consequences of nominal anchor and keep interest rates lower than a currency fall. This is true of financial they would be if there were a currency risk.The institutions as well as of non-financial four major EastAsian countries that ran into crises companies. (Feldstein 2003) in 1997 (Indonesia, Korea, Malaysia,Thailand), Russia in 1998, Brazil in 1999, and Argentina and The balance of opinions has now moved away Turkey thereafter, had all adopted fixed exchange from rigid exchange rates and, except for China, rate regimes.The fixed rates encouraged domestic most developing countries have adopted flexible companies and banks to borrow in foreign regimes. These are not clean floats. Korea, for exchange even when their revenues were in example,has accumulated reserves of US$140 bil- domestic currency. lion since the crisis, which suggests that the gov- As domestic price and external developments ernment has intervened extensively to avoid (notably the appreciation of the U.S. dollar nominal appreciation. In India, the Reserve Bank against the yen before the 1997 East Asia crisis) (RBI) has often stated that the exchange rate will made their exchange rates unsustainable,govern- be determined by market fundamentals, which it LESSONS AND CONTROVERSIES FROM FINANCIAL CRISES IN THE 1990S 245 has been careful not to define.At times, RBI used ing" (Blejer, inWorld Bank 2005b), and this cre- the stabilization of the real effective exchange rate ated another source of stress on the banks. as a guide to set nominal rates,but at times the real Second, the devaluations caused public debts rate has been only one of the factors in deciding to rise.While their direct impact on the public the nominal rate,reflecting the fact that this "con- debt was not always large,their effect was also felt structive ambiguity" is perhaps unavoidable when through two other channels (table F.2): the costs operating a managed floating exchange rate of bailing out and recapitalizing banks, and the regime.The prevailing opinion is that RBI has impact on foreign-currency-denominated debt tried to avoid nominal appreciations more force- and the compounding of real interest rates. Real fully than it has tried to avoid nominal deprecia- interest rates exceeded 100 percent in the days tions--which is one of the reasons for the large and weeks following the devaluation as the buildup of reserves. In the case of China, the sta- authorities sought to prevent the devaluation bility of the nominal rate is based on a continuous from overshooting. Such high interest rates were buildup of reserves that prevents a nominal appre- not sustained for long, but even the more mod- ciation.There is an expectation, however, that the erate rates that succeeded them for a period of Chinese currency will need to appreciate in nom- months or years had a large impact on the inal terms to reflect increases in productivity: in buildup of debt (table F.2). In Argentina, the the long run, as an economy develops, productiv- bailout of banks and the impact of the devalua- ity and real wages rise, and incomes approach tion accounted for most of the very large increase those of industrialized economies, its exchange in public debt. In Turkey, it was the recapitaliza- rate will inevitably appreciate--as did the Japanese tion of banks and the impact of real interest rates yen in the last two decades, or the European cur- that contributed most. rencies in the 1960s in relation to the U.S. dollar. Third, the devaluations led private external The large nominal devaluations that followed debt to be nationalized through a variety of the crises had four consequences of economy- channels. In principle, borrowing abroad by wide proportions.These consequences highlight domestic private firms or private banks without the risks associated with appreciation of the real government guarantees is a strictly private trans- exchange rate, even if temporarily. action. In practice, widespread bankruptcies, or First, the devaluations led to banking crises, the threat of them, inevitably involved govern- even where banks were sound.3The banks' open ment interventions and the socialization of part positions, or the borrowers', were simply too of the costs.This constitutes one of the earliest large to withstand a devaluation as large as wit- arguments advanced for controls on capital nessed in Argentina, Indonesia, or Uruguay. In inflows: individual creditors have no exact Argentina the problem was compounded after knowledge of the exposure by other creditors the devaluation by an asymmetric conversion of while every increase in exposure causes an assets and liabilities, with assets converted at a increase in the currency risk.A large share of the lower rate than liabilities,but the devaluation was increase in public debt in table F.2 refers to the so large that even a well-capitalized bank with socialization of the costs of crises, mostly in the modest exposure could not have withstood the form of bank bailouts and recapitalization. shock. Even a well-supervised banking system Fourth, in the presence of foreign debt, the cannot be sounder than the economy in which it devaluations limited the effectiveness of mone- operates.Where there were weaknesses in the tary policies. The presence of large foreign- banking system, as in Indonesia and in Turkey's denominated liabilities can reverse the effect of state banks,the problem was compounded by the monetary policies and deepen a crisis (Mishkin poor quality of the portfolio. Last but not least, 2001).Where the government's foreign-denomi- governments raised interest rates, sometimes to nated debt is large,as in Brazil,a monetary expan- extremely high levels, in order to moderate the sion weakens the government balance sheet and extent of devaluation and prevent "overshoot- thus prompts a rise in interest rate spreads,thereby 246 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s TABLE F.2 Debt Dynamics in Crisis Economies, Cumulative Change, Three Years Before . . . Korea, Russian Mexico, Indonesia, Rep. of, Malaysia, Federation, Brazil, Turkey, Argentina, 1991­93 1994­96 1994­96 1994­96 1995­97 1995­97 1997­99 1999­2001 Change in public sector debt ­22.8 ­11.6 ­2.6 ­16.6 ­8.8 6.3 14.5 21.5 Primary deficit (­ surplus) ­12.9 ­7.5 ­2.1 ­22.1 9.3 0.7 1.8 1.7 Rec. of contingent liab. (net of privatization) 0.0 0.0 0.0 ­2.0 ­9.3 4.0 ­0.7 0.7 Contribution from real GDP growth ­3.9 ­7.3 ­3.1 ­16.0 4.0 ­3.0 ­2.4 4.2 Contribution from real interest rate 3.4 3.7 0.0 7.0 ­7.4 3.6 20.7 10.7 Contribution from real exchange rate change ­5.9 ­2.2 ­0.6 ­1.2 ­29.7 ­1.6 ­0.4 2.4 Contribution from debt indexation 0.4 0.0 0.0 0.0 0.0 1.4 0.0 0.0 Residual ­4.0 1.6 3.3 17.6 24.3 1.3 ­4.5 1.9 . . . and Three Years After Korea, Russian Mexico, Indonesia, Rep. of, Malaysia, Federation, Brazil, Turkey, Argentina, 1994­96 1996­99 1997­99 1997­99 1998­2000 1998­2000 2000­03 2002­03 Change in public sector debt 28.0 68.6 30.3 14.9 4.2 15.7 19.1 83.7 Primary deficit (­ surplus) ­16.8 ­2.7 6.0 ­19.6 ­7.0 ­6.7 ­12.6 ­3.9 Rec. of contingent liab. (net of privatization) 0.0 0.0 ­0.7 6.9 ­5.9 5.0 15.4 0.3 Contribution from real GDP growth ­0.6 2.0 ­2.9 ­2.5 ­9.4 ­2.6 ­6.4 ­5.6 Contribution from real interest rate 12.5 4.3 2.3 7.8 ­4.8 17.1 21.7 1.3 Contribution from real exchange rate change 6.5 10.8 1.8 8.7 30.9 2.8 3.3 40.6 Contribution from debt indexation 4.4 0.0 0.0 0.0 0.0 4.9 12.2 0.0 Residual 22.1 54.3 23.8 13.6 0.4 ­4.9 ­14.5 51.0 Source: World Bank 2005c. Note: The residual captures the recognition of implicit liabilities, such as banking sector bailouts, implicit social security and pension debts, and so on, for which no hard data exist, and which thus are not included directly in the calculations. It also includes various cross-products assumed away with the approximations made. For Argentina, data are available for only two postcrisis years: 2000 and 2001. opening the possibility that interest rates will rise Monetary and Fiscal Policies rather than fall. Similarly, foreign-denominated debt in domestic private balance sheets makes it There is a view, though highly controversial, that more difficult for a country to recover from a the management of the crises of the 1990s relied financial crisis because expansionary monetary on excessive fiscal adjustment;that excessively high policies will likely cause a nominal depreciation real interest rates have in some instances forced of the domestic currency.This will hurt the bal- unnecessary bank closures; and that international ance sheets of firms and banks and reduce their financial institutions often forced crisis countries to net worth. In a country without foreign-denom- undertake structural reforms that were not directly inated debt issued by the private sector, expan- related to resolving the crises--which would have sionary monetary policies can help shore up been better done through countries' own national balance sheets of financial and nonfinancial firms, decision-making processes (Feldstein 2003;Stiglitz and increase their net worth. 2001;Ahluwalia 2003). LESSONS AND CONTROVERSIES FROM FINANCIAL CRISES IN THE 1990S 247 The least controversial of these statements is Notes that fiscal policies were unnecessarily tight, as acknowledged in the IMF's own evaluation 1. World Bank, Global Economic Prospects 1998. 2. Although the reasons as to why they would work (IMF 2003b).A common explanation for this is better in Chile than elsewhere are not clear (Edwards that capital account crises were dealt with as 2003). current account crises. 3. Argentina, for example, following the Mexico crisis "Weak"banks have often been blamed for the in 1994 in which many banks had failed, consider- crises. However, while poorly run banks can ably improved the financial soundness of its banking exacerbate a crisis, and perhaps even cause one, system by restructuring and shifting the ownership of banks to foreign private banks (primarily from the well-capitalized banks cannot guarantee against a United States and Spain). financial crisis. Even well-capitalized banks can quickly lose their equity in the face of large swings in exchange rates, or the collapse of the real economy. Chapter 8 Policy Reforms and Growth Performance: What HaveWe Learned? T HIS CHAPTER SYNTHESIZES THE rect, there was a tendency to believe that they lessons from the review of could only be implemented in certain ways. experience with policy reform Going forward, more emphasis is needed on in macroeconomics, trade, privatization, and common principles, along with a more plu- finance.As the preceding chapters illustrate, each ralistic approach to implementing those prin- of these areas of policy reform is complex and an ciples. attempt to draw lessons in any one of them cre- · Growth strategies, focused on initiating and ates vigorous debate. Nevertheless, three key sustaining episodes of rapid growth, are the cross-cutting lessons seem to emerge: key to reaching much higher levels of income. · Most market-oriented reforms have had pos- Such strategies focus on attacking the binding itive payoffs, though their impact on growth constraints on growth, rather than addressing was not as large as some of the exorbitant many weaknesses simultaneously. claims made both in academic and policy cir- · Creating the institutional conditions for a cles. favorable climate for investors, both large and · Experience shows the importance of creating small, is essential. Government actions and institutional constraints on the exercise of dis- their design should be scaled to match the cretion in policy implementation. Institutions country's institutional capability. "Do no and rules should be seen as a means to facili- harm" is a wonderful guide, and the potential tate the predictable, credible, and beneficial for government action to improve on market use of discretion,rather than as a substitute for outcomes needs to be balanced against the discretion. ability of existing institutions to sustain good practices. · The expectations of the various actors in the markets play a crucial role in the success or failure of reforms, and their evolution can 1. Cross-Cutting Lessons of the lead to either virtuous or vicious circles in 1990s the reform process. For each of the three cross-cutting lessons, this These lessons are discussed in section 1.They section uses a common organizational structure: create three suggestions for a way forward,exam- it diagnoses previous successes and failures, ined in section 2: reviews the conventional wisdom of the 1990s · While the basic economic principles behind that lay behind the reform efforts, and describes most of the reforms of the 1990s were cor- the lesson itself. 249 250 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s Lesson 1 too large to be "steady-state" differences, but they also seemed too large to be transitional Most market-oriented reforms have had positive payoffs, differences in adjusting to efficiency gains. though their impact on growth was not as large as some of the exorbitant claims.And the benefits of · Countries' growth rates change dramatically: reform were, in general, predicted correctly by some countries have growth rates that propel microeconomists and sectoral experts,though not them rapidly out of poverty traps while others by crude applications of the"new growth"theory. go from rapid growth to stagnation or bust. Diagnosis before the 1990s: Conflicting Conventional Wisdom in the 1990s: Interpretations of the Relationship of "New Growth" Theory and Large Gains Growth to Policy from Reform Understanding the lessons of the 1990s for econ- This inability of the standard theory of steady- omists requires a little background on the profes- state growth to explain the facts perhaps explains sional state of play in the early 1980s. At that the love affair of academic and policy-making cir- time, growth theory was still dominated by the cles with "new growth theory" models in the Solow-Swann model (Solow 1956, 1971; Swann 1980s.Advances in the modeling of noncompet- 1956). According to that model, in the steady- itive equilibria (Romer 1983, 1986) allowed the state equilibrium, long-run growth rates are development of a new set of endogenous growth completely unaffected by national policies.That models in which national policies could influence is, while national policies could affect the level of not just the level of income but also countries' income they could not permanently affect the steady-state growth rates (Grossman and Help- growth rate.1 Meanwhile, the analysis of sectoral man 1992;Aghion and Howitt 1998).These led reforms--for example, in trade, privatization, or to the conventional wisdom of the 1990s--that the financial sector--was dominated by micro- policy reform could affect economic growth-- economic models in which gains resulted from but they never made quite clear why this should policy reforms but were typically only small frac- be so. Often, authors did not make clear whether tions of the gross domestic product (GDP).2 their growth regressions were intended to iden- This match--of the unresponsiveness of tify differences in steady-state growth or, instead, long-run growth rates to national policies in to identify impacts of policy on the level of macroeconomics, and the apparently small effi- income. Such lack of clarity pervades discussions ciency gains to be had from sectoral reforms in of growth. It is useful to dispel this confusion by microeconomics--was a stable but increasingly keeping the gains in levels with "growth" as a unhappy marriage. Stable because these were transitional phenomenon and gains in "growth" both very robust features of their respective ana- in the steady state. In the end, the hope was lytical approaches. Unhappy because by the dashed that there were large policy-driven gains early 1990s this combination of approaches in steady-state growth.Even so,this does not imply clearly could not explain some basic facts about that policy reform cannot yield large growth gains the world, particularly the developing world: when it has a large impact on the level of income. Trade policy reform illustrates this point. · Since some countries are very rich and others Many growth regressions related growth in out- are very poor,differences in growth rates must put per person in the ith country over some have been sustained and substantial. Indeed, as period of n years to the lagged level of output chapter 2 showed, growth rate differences of and some indicator of trade policy during some nearly 2 percentage points a year have been period: sustained for more than 100 years. Figure 8.1 shows the path of annual growth · The differences in growth rates across coun- stemming from a hypothetical trade reform that tries over periods of a decade or more were increases the sustainable level of output. Let us POLICY REFORMS AND GROWTH PERFORMANCE: WHAT HAVE WE LEARNED? 251 gt,t i -n= yt -yt i i -n = + *yt i -n FIGURE 8.1 +*Trade Policyif(t,t-n) Simulated Impacts of Policy Reform on the Level and +Other factors+error term Growth Rate of Output assume for now that "trade policy" at any point 0.09 Adjustment to new level with exponential decay in time can be adequately represented by a single 0.08 number.The graph shows a hypothetical econ- 0.07 omy growing at a steady-state rate of 2 percent a year. In year t = 5 there is a permanent improve- 0.06 (percent) ment in trade policy from TP to TP*. If trade Level increase = 50% 0.05 year policy raises the steady-state growth rate immedi- 0.04 Steady state ately and permanently by 2 percentage points, per growth increase = 2% the measured growth rate over any five-year 0.03 Level increase = 25% Base case growth = 2% period will increase from 2 to 4 percent and will Growth 0.02 Level increase = 5% remain at that higher level.In this case the impact 0.01 of the trade reform on the level of output will be 0 5 10 15 20 25 30 infinitely large. Years, reform in year 5 Figure 8.1 also shows the impact of a trade Source: Author's elaboration. reform that affects only the level of output. In each case we assume some dynamics for illustra- in steady-state growth). Finally, if the impact of tion--that the impact on the level of output the trade reform on the equilibrium level of out- takes 10 years to be fully felt and that the adjust- put is as large as 50 percent, the impact on ment from the baseline to the higher level of observed five-year growth rates is much larger output is linear. In this case, the reform has an for a reform that "only" affects the level than it is impact on measured five-year growth rates that for a reform that has a large impact on steady- increases as output adjusts to its new level, and state growth. then decreases to zero; that is, the economy This technical excursion clarifies that over the returns to its steady-state growth path.The graph horizon of a decade or more the impact of an shows the impact on annual growth rates of a economic reform on observed growth rates does trade reform, using three possible magnitudes of not depend at all on whether the reform raises the cumulative impact of the reform on the level steady-state growth or "only" raises the long-run of output: 5 percent, 25 percent, and 50 percent level of output with no impact on steady-state (under certain assumptions about adjustment growth.What matters is the size of the gain and dynamics). If the cumulative impact of the trade the speed of adjustment.Over the medium term, reform on the equilibrium level of output is only if the effects on the level of output are small, the 5 percent, the impacts on measured growth rates effects on steady-state growth will also be small, are small and disappear quickly, compared to the and if the effects on the level of output are large, impact of a 2 percentage point increase in the effects on growth will be large. steady-state growth rates. In contrast, if the Whether growth regressions in the 1990s cumulative impact of trade reform on the level were estimating growth effects or level effects of output is 25 percent, over a 10-year horizon mattered less than how the regressions were the effect of the reform on the observed growth interpreted. In practice they were widely seen as rate is virtually indistinguishable from the effect producing estimates of gains from policy reform of an increase in the steady-state growth rate-- that were whole orders of magnitude larger than and only over long periods does it become pos- the microeconomic estimates of those gains.The sible to distinguish one (an impact on level with example of trade liberalization again illustrates transitional growth) from the other (an increase this point.The original microeconomic ("Har- 252 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s berger triangle") estimates of the welfare gains tions increase as the square of the distortion. were on the order of 1 to 5 percent of GDP for Chapter 5 emphasized that countries that made an ambitious reform of tariffs from moderately very large reductions in tariffs (Bangladesh,India, high levels.With moderate adjustment speeds, Pakistan) achieved large gains in integration, such a reform would increase growth rates tem- while those that made smaller reductions in tar- porarily by not more than half a percent a year. iffs achieved smaller gains. Similarly, chapter 4 Even when models introduced general equilib- showed that the potential gains from taming rium effects and plausible links from trade reform hyperinflation are much larger than the potential to productivity improvements,the apparent gains gains from reducing inflation from more moder- from trade reform were too small to cause sus- ate levels. And in the financial sector, the gains tained growth increases of more than 1 percent a from interest rate liberalization depend on the year, over a period as long as a decade. By con- severity of the initial financial repression:thus for trast, it was frequently claimed that growth countries with very negative real interest rates, regressions, such as those of Sachs and Warner liberalization should produce large gains, while (1995a), supported the view that trade liberaliza- for countries with moderate financial repression tion could raise the rate of economic growth by the gains would be more modest. 2 percent a year over a 30-year horizon. This These relationships would also lead one to implies a rise of 80 percent in the level of output. expect large gains in countries that are very poor Even if trade policy reform were to raise the eco- when reforms begin, since many of these coun- nomic growth rate by only 1 percent a year, sus- tries are much less productive than they could taining these effects for a very long period--as be if they had good policies and institutions.The implied by the very small adjustment coeffi- very fast growth achieved by countries such as cients--would produce gains of as much as 50 China, India, andVietnam is consistent with the percent of GDP. view that reform can have enormous effects on Interpreting the "aggregate" and "growth the level of output, which in turn lead to rapid regression" evidence concerning the impacts of growth in the course of transition to the new trade policy on output is nearly impossible, higher levels of equilibrium output. because even though many studies take growth That said, the claims that policy reforms as the variable to be explained,the interpretation would raise growth rates permanently, or by as of the magnitude of the resulting coefficient much as 1 or 2 percentage points a year, were depends entirely on how the dynamics of the almost certainly exaggerated. The disappoint- regression are specified: the same reported coef- ment with the returns to policy reform stems ficient on a variable representing trade policy partly from the fact that regressions have sug- could imply either a small or an infinitely large gested that some policy variables, such as budget effect on the level of output. deficits, outward orientation, and privatization, are associated with economic growth. If such an Lesson of the 1990s: Policy Reform Pro- empirical association represents a stable,uniform, duced Mixed, and Modest, Gains causal relationship between the policy variable The 1990s showed that the long-run impact on and growth, it is puzzling if, at least on average, output3 of most policy reform actions in the the relationship does not hold for policy reforms. areas considered--macro, trade, privatization, However, the magnitudes of the impact of the financial liberalization--was positive and roughly policy variables on immediate growth rates were as large as claimed by microeconomic or general never very clear. equilibrium studies. To sum up, the gains from more effectively In particular, from most microeconomic- and efficiently provided infrastructure services based models we would expect that the gains will not be infinite but they are important, as are would be larger, the larger the initial policy dis- the gains from better allocation of financial tortion, because the welfare gains from distor- resources.4 Finally, not everything that is called POLICY REFORMS AND GROWTH PERFORMANCE: WHAT HAVE WE LEARNED? 253 "market-friendly" reform will work to increase of three archetypical forms: objectives with output.The details do matter and it is perfectly discretion, conditional rules, and uncondi- possible to make large and costly mistakes, as tional rules. attested by some of the examples in this volume. The background institutions of policy mak- ing are the legal and political environment into Lesson 2 which the direct institutions are embedded.The Institutional limits are needed on the exercise of discre- background institutions include not just govern- tion in policy implementation. Government discre- mental organization of checks and balances on tion cannot be squeezed out of policy making, the discretion of organizations and on the gov- and the presence of government discretion ernment itself--but also rules such as the free- implies the need for a solid institutional founda- dom of the press and the ability of citizens to tion to control it. Creating effective institutions organize. that will play this role depends not just on tech- Figure 8.2 illustrates these basics.The organi- nocratic design,but also on an underlying"shared zation responsible for implementation is the mental model" (North 1990). agent,to which the principal delegates the power to take policy actions. If for simplicity we imag- Definitions ine the organization as a single agent,5 we can For purposes of this discussion we define "pol- imagine a positive model of policy actions. One icy,""organization," and "institutions" to mean such model is that the organization will take pol- very specific things. icy actions that maximize its own objective func- A policy is a mapping from states of the world tion subject to the constraints and incentives it to actions.That is, a policy is not a single action faces. In this sense the notional policy (proposed but the description of a process that produces a objectives, model, relevant facts, and proposed sequence of policy actions.The policy actions may mapping) and the background institutions are be contingent on facts: for example if a country what establish the incentives and constraints on has a fiscal policy of running a cyclically adjusted the maximization problem of the agency. surplus of 1 percent of GDP,this requires a budget Table 8.1 gives examples of how these (policy action) that is tailored to the state of the descriptive terms fit into a specific area, such as business cycle (fact). monetary policy, as a component of macroeco- To implement a policy,translating it into prac- nomic policy. Each of the policy areas discussed tice, requires an organization of policy making. in the previous chapters, from trade to financial The direct organization of policy making sector regulation, can be understood using this includes the following: same vocabulary. · The organization that has authority to take Diagnosis before the 1990s: policy action; Government Discretion Is the Problem · The range of feasible policy actions; Up to the 1990s,a prominent diagnosis of devel- opment experience had two components. First, · The process to be followed in taking policy policy mappings were seen to consist mainly of actions; multiple objectives with discretion. Hence in · The objectives of the policy; macroeconomic, trade, financial, infrastructure, and regulatory policy the organizations with · A model that determines the relevant facts (or direct responsibility for policy actions were often states of the world); and given multiple (unclear) objectives, while at the · Some indication of the policy mapping from same time they were given control over a wide facts to actions, given the objectives and the range of policy actions. Second, while discretion model.These policy mappings can take one was seen to be used well in some times and 254 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s · Corruption was rampant. FIGURE 8.2 The Elements of Policy Action Conventional Wisdom in the 1990s: Create Market-Friendly Conditions by Relevant "states Notional policy: Reducing or Eliminating Government of the world" · objectives with (empirically Model Discretion discretion contingent facts (description · conditional Given that the diagnosis was "too much discre- that are relevant of how policy rules to the desirability actions lead tion," the conventional-wisdom goal of the (a mapping from of various policy to outcomes "states of the 1990s was to reduce public sector discretion as actions--S) --M) world" to "policy much and as fast as possible. Reformers pursued Feasible policy actions" M actions this goal in three ways (table 8.3): P :S A (The set of · unconditional actions that the · First,reducing the scope of government activ- rules (always do policymaking policy action a) ities that required discretion, by removing the organization can take) government from direction over production (by divesting the public sector of productive assets),and by eliminating unnecessary regula- Organization tions. (The organizations or agencies legally authorized to · Second, in whatever regulatory or policy take actions) activity remained under the government, Policy actions reducing government discretion, by pursuing (The outcome of all of this is rules-based formulas for decision making the actual sequence of decisions and policy actions taken) based on clear "objective" criteria and by Citizen organizing capability granting autonomy to regulatory agencies. · Third,making binding international commit- Administrative Political Impact of policy action on Judicial ments that limited the scope of domestic dis- (Executive, actions of relevant agents Legislative) cretionary action, for example, multilateral or Free regional international trade agreements, and Background press institutions of agreements limiting exchange rate flexibility. Outcomes policymaking Background institutions Lesson of the 1990s: Proper Exercise of Discretion in Policy Implementation Is Key The attempts to reduce government discretion places, it was also seen to be misused, in a variety had two phases: a rationalization phase followed of ways (table 8.2): by an optimization phase. · Inadequate information led to wrong deci- The rationalization phase, which happened sions. mainly in the 1980s, was needed and beneficial. It eliminated accretions of actions, regulations, · Technical capacity was insufficient to take and decisions that had often resulted in policies correct decisions. that served nobody's best interests.6 Examples · Multiple objectives led to ineffective actions. included reducing the variability and capricious- ness of tariff rates, closing down and consolidat- · Policy actions were politicized in a way that ing many special-purpose and money-losing sacrificed effectiveness for political expedi- financial intermediaries, and selling off assets in ency. competitive industries. · Public officials had inadequate incentives to By the 1990s, reforms were moving into the be dynamic or to innovate. optimization phase.These "second generation" POLICY REFORMS AND GROWTH PERFORMANCE: WHAT HAVE WE LEARNED? 255 TABLE 8.1 Sectoral Example of Direct and Background Institutions of Policymaking Examples from monetary policy Objectives with Conditional rule-- Unconditional Unconditional rule-- Item Definition discretion inflation targeting rule no discretion(free banking) Feasible A legally authorized Money supply Money supply Money supply None policy action act by a public sector authority "Model" Specification of Money lowers interest causal chain from rates, stimulates policy action to output outcome Relevant state Relevant facts for State of the business Prices None None of the world the application cycle of policy Policy mapping A model-informed Increase money Always increase Match outstanding mapping from states when output money supply obligations to of the world to temporarily low by k percent foreign assets policy actions Direct Public sector Central bank Central bank Central bank Currency board organization of organization policymaking authorized to act Indirect Formal and informal Procedures for administrative appeal institutions of checks on policy- Courts policymaking making decisions Executive Legislature Media Interest groups Source: Author's elaboration. reforms (Naim 1995, 1999) constituted a move responsibility and interest in the soundness of to conditional rules governing the actions of a the electricity grid. wide range of policy makers--monetary The attempts to reduce government discre- authorities, regulators of banks and utilities, and tion by imposing rules-based policies had much private contractors providing public services. In less impact than was hoped. In retrospect, there essence,many of the reforms were shaped by the were two reasons why. view that institutions should play the role of First, the risk that the public sector will abuse eliminating discretion wherever possible, rather its discretion is a necessary consequence of the than facilitating effective decision making.The monopoly nature of state power over the means reforms had to grapple with the question of how of coercion. If performance is poor because the core government responsibilities were to be car- public sector has incentives to abuse discretion ried out.In a number of sectors there is a core of (whether by failing to respond to problems,mak- public responsibility that governments cannot ing mistakes, capricious enforcement and cor- avoid.For example,while there is no compelling ruption, or outright predation), it is unlikely to reason for government to own and operate be sharply improved by reforms that limit the commercial or investment banks, government scope of government. Policy conditionality can- does have a core, unavoidable responsibility and not be effective except in those rare cases in interest in the soundness of the banking sector. which the policy action is unequivocal and com- And while there is no compelling reason why pliance is easily observed. Easterly (2000) pro- government should run an electricity company, vides an insightful analysis of attempts to limit government does have a core, unavoidable fiscal deficits through the application of rules. 256 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s TABLE 8.2 Examples of Misuse of Discretion Examples of ways in which Motivation for policymaking discretionary Policy area public sector engagement power was misused Negative consequences of discretion Macro Control of the money supply Maintenance of system of Central banks forced to "print Money creation, high and variable inflation external payments money" to finance deficits Lack of fiscal discipline led to high debts (exchange rate) through high seigniorage Mismatch of monetary and exchange rate Overvalued exchange rates policy led to overvaluation with periodic maintained with preferential crises and "maxi" devaluations access to foreign exchange for government and parastatals Trade Revenue mobilization, Firms lobbied to obtain protection Industries, once protected, never grew up industrial promotion, for politicians' "pet" projects from "infant" status control of trade balance Bribery in customs to evade Discretionary controls over imports led to trade restrictions "rent seeking" in the creation and allocation of import restrictions Privatization Supply of infrastructure Governments used firms for Underpricing and lack of autonomous and patronage (for example, control over adequate cash flow led to regulation placement of executives) underinvestment in maintenance Placement of facilities was Multiple and unclear objectives (no politically motivated "bottom line") led to productive Outright corruption in the inefficiencies and technological stagnation placement of contracts Financial Private sector capital Allocation of credit to Large losses for banks sector markets could not politically preferred activities Low deposit rates (often negative in provide long-term credit Rollover of debt for favored real terms) borrowers (for instance, High borrowing rates for nonpreferred parastatals) borrowers Selective enforcement of Capital did not flow to new, repayment obligations promising industries Source: Author's elaboration. TABLE 8.3 Efforts to Limit Government Discretion Reduce the scope of "Rules not discretion" government activity with "independent" regulation Binding international agreements Macroeconomic Dollarization, currency boards, Monetary unions inflation targeting, independent central banks Trade Elimination of barriers to trade Moving to uniform tariffs; eliminating Bilateral (NAFTA), regional (EU, nontariff barriers in favor of tariffs Mercosur), multilateral (WTO) Privatization/ Privatization Using contracts as a means of engaging regulation with private sector providers Financial sector Privatization of state-owned banks; Adopting supervisory standards Allowing entry of foreign banks liberalization elimination of regulations (for example, Basel) Source: Author's elaboration. POLICY REFORMS AND GROWTH PERFORMANCE: WHAT HAVE WE LEARNED? 257 Suppose, as is not unusual, that a government cases. The only role left to the courts was to wants to overspend--specifically, to bring declare a debtor bankrupt,8 and after the judicial expenses into the present and to push the gener- declaration of bankruptcy all future jurisdiction ation of revenues into the future, hence reducing passed to the group of creditors.The result was net public assets. Then suppose some outside that in the first few high-profile bankruptcy agency wants the country to limit its fiscal deficit cases the judges did not declare bankruptcy to a level lower than the government wants.Will because they found that a "legal" debt did not a "policy" change that limits the fiscal deficit to exist. Instead they used various criteria to show some specified amount staunch the reduction in that otherwise apparently ironclad debt con- net assets? No.The government can reduce net tracts did not in fact constitute debt.The new public assets in hundreds of ways that do not law had not changed anyone's incentives.There increase the recorded fiscal deficit.7 This ability were no credible checks on the courts' findings exposes the mirage of so-called rules-based poli- and hence the exact same result--lack of a cred- cies,because by the time one has a means to pre- ible creditor threat of bankruptcy--was reached vent all the tricks by which a simple rule such as even in the face of determined attempts to "no fiscal deficits" can be subverted, one actually remove discretion from the legal process. has the institutional conditions in place for good This conceptual framing may help us to expenditure management. understand several elements of the experience of The second reason why attempts to reduce the 1990s: government discretion by imposing rules-based · Why the success of reforms differed so widely policies had less impact than hoped for is that the across countries, and the significance of new difference between "rules" and "discretion" evidence about the importance of institutions proved much murkier than supposed.The first over policies; round of the rules-versus-discretion debate gen- erally ignored the key difference between condi- · The evolution of concerns from policy tional and unconditional rules. reform to governance and institutions; The 1990s brought home that if incentives · The mixed popularity of growth reforms and remain unchanged, and there are no background importance of perceptions in the success of institutions to check the findings of fact, the use reform; and of conditional rules can produce exactly the same policy actions as the use of discretion.Conceptu- · The evolution toward policy recommenda- ally, and often in practice, the process of policy tions designed to fit specific institutional capa- actions with conditional rules can be divided bilities, as opposed to the application of into two stages:a findings-of-fact stage and a pol- universal best practices. icy action stage; as noted above, the findings of We discuss each of these elements in turn. fact dictate the policy action (or narrow range of Intercountry differences in the success of reform. actions).The scope for exercising discretion can Why did the success of reforms differ so widely then be pushed back from the policy action stage across countries?The answer may lie in the com- to the findings-of-fact stage. bination of a country's initial level of income and A telling example comes from Indonesia's its institutional capability to implement complex attempt to create bankruptcy courts.In the wake reforms. of the financial crisis, many observers felt that Many of the biggest successes of the 1990s the lack of a credible judiciary was limiting cred- were achieved by countries that were much less itors' ability to enforce their contracts or even to productive than they could be with good policies force debtors to negotiate resettlements.Because and institutions, so that modest reforms whose judicial reform is a slow process, a new bank- implementation was not institutionally demand- ruptcy law was passed that attempted to remove ing were able to produce large gains in expected all discretion from the courts in bankruptcy future income. Examples are China's liberaliza- 258 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s tion of agriculture and India's dismantling of very with the stroke of a pen and easily observed,poli- high trade barriers.9 cies are meaningless unless they are backed by The varied experience of the transition coun- controls that make the policy actors sufficiently tries illustrates the difficulty of achieving the accountable. right mix between declared policies and institu- Take the example of replacing the public pro- tional capability. A viable financial sector that vision of an infrastructure service with private channels resources to productive investments is provision by a contractor. The public agency key to a market economy. Reform efforts in this responsible for awarding the contract must direction sometimes had acceptable results--for announce the winning bidder. Most of the sec- instance, in Hungary. In Albania, by contrast, ond generation reforms in infrastructure dealt financial sector liberalization with essentially no with extremely complex services, for which the government control led to a giant Ponzi evaluation of bids inevitably involves some dis- scheme,10 and after a brief bubble, to massive cretion (one does not merely want to choose the losses that forced the government out of power. lowest bidder without prequalification,consider- In some countries of Eastern Europe, privatiza- ation of the full range of services included in the tion worked reasonably well.In others,privatiza- contract, and so forth). But the necessary discre- tion was achieved rapidly but it was followed by tion that is created by complexity can lead to a shake-out, because the institutional capability inefficiency, malfeasance, or corruption. The for regulating the basics of corporate governance same is true with the transition from concern did not exist. Another group of East European with fiscal discipline to a broader concern with countries pursued a so-called policy of privatiza- budgetary institutions.While it is easy to place tion without any credible central authority, any conditions (either via rules or outside agencies) mechanisms of public sector accountability or that govern easily observable policy variables corporate governance, or any means of legal such as the fiscal deficit, it is impossible to man- enforcement of contracts. This concentrated date that public monies be well spent. Similarly, assets in the hands of those who were able to sensible regulation of banks requires the use of operate in such an environment. considerable judgment. Because of the impor- Latin America's experience was mixed. By tance of trust between borrowers and lenders, and large,the countries of the region began with especially in environments in which the formal a base of better policies and more advanced insti- mechanisms of contract enforcement are weak, tutions, offering less "low hanging fruit" for close continuing relationships between banks reformers than in Asian and transition countries. and firms tend to be the norm.From a regulator's Most Latin American countries had to grapple perspective this makes it difficult to distinguish with institutionally intensive reforms--financial between a perfectly rational business decision to sector regulation, and regulation of privatized carry a long-term customer over a difficult spot infrastructure--in the 1990s. Not surprisingly, by rolling over loans and a bank's unwillingness therefore, some reforms worked well and were to realize and write off bad debts.The regulator's widely popular, some worked well and were problem in observing the "true" facts about any unpopular (such as the privatization of water given loan is of course compounded when a reg- utilities in Argentina), and some worked badly ulatory agency is held accountable for thousands with recriminations all around (for example, the of such decisions made every day. first round of Mexican toll roads). The mixed popularity of growth reforms and Evolution of concerns from policy reform to gover- importance of perceptions. Analyzing the institu- nance and institutions. Current discussions about tional conditions for policy implementation may the investment climate differ from 1991 discus- also help to explain why many market-oriented sions of "market-friendly" policies.The recogni- reforms--even those for which there is evidence tion today is that, except for a very few of success--have not been altogether popular.In macroeconomic policies that can be executed Latin America,for example,bringing more mar- POLICY REFORMS AND GROWTH PERFORMANCE: WHAT HAVE WE LEARNED? 259 ket forces into the provision of infrastructure has conditional rule.But the institutions of tariff pol- improved the quality of services and expanded icy could be weak.They might lack the technical their coverage, but prices have risen and "priva- capacity necessary to assess which goods are tization" is widely unpopular. A possible expla- growth-promoting and which are not.Or,if they nation is that a lack of public confidence in the were faced with discretion or a conditional rule, regulatory institutions means that the public their findings of fact might be susceptible to may perceive deals as fixed or corrupt and price political influence or outright bribery. If the increases as simply leading to high and unjusti- direct and indirect institutions of policy making fied profits for firms, which have regulators "in are weak, the optimal policy is an unconditional their pockets." This is a hard problem to deal rule of uniform tariffs,and perhaps even zero tar- with. iffs (table 8.4). The evolution toward policy recommendations More generally, if it is perceived that corrup- designed to fit institutional capability. Suppose that tion is the central problem in public sector some goods have dynamic externalities, so that action, the tendency will be to force all discre- greater domestic production of these goods raises tion out of policy implementation--for exam- a country's overall output, and that other goods ple, by removing the government from bank do not, so that their protection and greater regulation.Good regulation is better than no reg- domestic production cause overall output to be ulation. But no regulation is better than bad reg- lower.11Assuming that tariff rates can change rel- ulation, and where mechanisms are not available ative prices, a possible tariff policy would be to to control the discretion that is inherent in place a high tariff on the good with dynamic attempts to implement reasonable policies,"no externalities and no tariff on the growth-reduc- regulation" may be the appropriate choice. Sim- ing good.This policy is a conditional rule that ilarly, if the central problem is that private depends on distinguishing which good is which. investors fear predation by the state, strong pre- In practice, however, this distinction might be conditions to prevent predation are needed-- difficult to draw and to verify. Now suppose that even if their introduction sacrifices otherwise producers of the growth-reducing good offer a desirable regulations or actions. larger bribe than producers of the growth- The debate today is no longer about whether enhancing good. "the market" or "the state" is always superior, In such a situation the optimal policy depends nor is it about "the proper role of the state" in entirely on the institutions of policy making. If the abstract.12 As theorists, most prominently we define good tariff policy institutions as those Joseph Stiglitz, have shown, one can always cre- providing institutional conditions in which the ate a theoretical model in which state action can conditional rule, "high tariffs on growth-pro- improve on the free market outcome--if the moting goods," will be applied correctly, with state action is perfect. But, as Pigou pointed out good institutions the best policy to choose is a nearly a century ago, the real choices are not TABLE 8.4 Example of the Dependence of Appropriate Policy on Institutional Conditions "Bad" institutions "Good" institutions Differentiated tariffs (either Can lead to lobbying, rent seeking, corruption, Can allow trade policy instruments discretion or "conditional and mistakes and result in complex, to promote nascent industries with rules") distorting tariffs with no positive effects possible dynamic externalities Uniform tariffs Forgoes possible benefits of differentiation, but avoids losses from rent seeking Better policy Uniform/precommitment Differentiation Source: Author's elaboration. 260 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s between the best the economist can imagine and · Get the policies right and investors will "the market."The choice is between the market respond. such as it is and what will actually happen if a · Bold action upfront signals the seriousness of given policy is adopted--which in turn depends reform. on the actual policy decisions that will be taken, which in turn depend on the quality of institu- · Signaling to the market requires ambitious tions for controlling the discretion used in pol- reform agendas. icy implementation.13 This discussion points forward to the problems Lesson of the 1990s: Expectations Are Central addressed in chapters 9 and 10.If the key problem Not only does the investment climate need to is policy implementation, and the key problem improve, but also investors (small and large, with implementation is to create the conditions domestic and foreign) need to believe that the for the effective exercise of government discre- improvement in investment climate is here to tion, the organizations of the public sector are stay. The 1990s emphasized that expectations are vitally important (chapter 9) and so are the back- central, not only as regards stabilization during ground institutions of policy making, especially crises, but also as regards the supply response to the ways in which citizens are able to monitor the policy reform.We discuss these two aspects in performance of government (chapter 10). turn. Crisis management. Restoring expectations is often the single most important factor in turning Lesson 3 around a crisis. Expectations play a crucial role in the success of policy reform. And political and social legitimacy and To restore credibility [after a crisis] you continuity are important in promoting expecta- have to show that your word is your bond tions of a more stable investment climate. . . . [I]t is crucial to choose targets that can If the gains from policy reform are to be real- be and are met.This is more important ized, individuals and firms must believe that if than issuing unrealistic projections . . . . they invest in response to the opportunities cre- --Kemal Dervis, in World Bank (2005b) ated by the policy reforms, they will reap the gains of their investment. Investment is always Our strategy at the Central Bank was based about the future, and about the future there are on the view that, given the lack of refer- no certainties, only beliefs and expectations. ence for the correct exchange rate, exchange rate expectations had to be stabi- Diagnosis before the 1990s: Policies Had lized for the bank to develop a market for Put Too Much Faith in Government as the its sterilization instruments. Otherwise, the Driver of Growth interest rates needed to induce significant As detailed above, the key explanation for the demand for the new instruments would slowdown in growth in the late 1970s and early reach unreasonable levels. In other words, 1980s was that policy makers had simply been an interest rate defense and active foreign wrong in their attempt to extend the scope of exchange market intervention were com- government action beyond the government's plementary rather than substitute policies. implementation capacity. These three policies were popularly char- acterized as a Central Bank attempt to Conventional Wisdom in the 1990s: increase demand for domestic assets--and Fixing Policies Would Ignite Growth in this way stop the bank run and the cur- The conventional wisdom of the 1990s was that rency run--by inducing greed to over- fixing policies would ignite growth.The belief come panic.The bank's main consideration was threefold: was that greed (interest rate policy) cannot POLICY REFORMS AND GROWTH PERFORMANCE: WHAT HAVE WE LEARNED? 261 overcome panic unless panic is also reduced reforms, while large reforms may have little by controlling chaotic conditions in the impact if investors perceive the results as tempo- foreign exchange market through active rary.14 Though the supply response to a policy intervention. reform is limited by credibility, larger supply --Mario Blejer (World Bank 2005b) responses make for greater support for continu- ing the policy,creating a virtuous circle in which There is disagreement on two big issues.The successful reform leads to continued reform. first is that of the proper scope of a reform pro- Many problems may interpose themselves gram in the midst of the stabilization of a finan- between policy reform and the faster growth it is cial crisis. One view is that the reform should be designed to achieve (table 8.5). limited and feasible, because an overambitious The first possibility,which has received a great reform can backfire by creating expectations that deal of attention, is that policy actions may or cannot be met. The other is that the reform may not signal policy reform. For example, if the should be big, broad, and aggressive, because that budget deficit is cut from 5 percent to 2 percent, convinces the markets that the government is does this signal macro-stability or merely reluc- serious about reform. But if in fact the big broad tant compliance with external pressures? Cer- and aggressive measures are ad hoc and not insti- tainly expectations about future macroeconomic tutionalized, there is a risk that meeting the tar- stability will differ dramatically depending on gets will not create confidence, while missing which of the two is perceived to be the case. them will create damage.This is particularly true Conventional wisdom holds that part of the rea- of implementation-intensive reforms incorpo- son why policy conditionality had a disappoint- rated into crisis stabilization packages. ing impact on growth was that the conditioned The second big issue is whether expectations changes in policy actions did not change can be positively affected by tying a government's investors' expectations about the long run. As a hands. For example, in the early 1990s there was result, the 1990s saw a growing emphasis on the a view that countries should move to either fixed ownership of reforms as key to a successful or completely flexible exchange rates to show investment and supply response.Without owner- evidence of the complete removal of govern- ship, current policy actions may not signal future ment discretion. But since the Argentina crisis, policy actions and hence do not create a power- some observers believe that removing discretion ful investment response. by creating mechanisms that impose large penal- Even if a policy shift is owned by the current ties may itself undermine expectations.Velasco government, the shift may not change expecta- and Neut (2003) argue that if the world is uncer- tions if it appears likely to be reversed by either tain and there are situations in which the lack of the current or a future government.And even if discretion will cause large losses, a precommit- investors believe that current policy actions sig- ment device can actually make things worse. nal a true shift in policy, and even if they do not Achieving a supply response. The supply expect the policy to be reversed,the fact that new response to any given policy action depends on policies often call for new organizations and how credibly that policy action signals a sustained direct institutions of policy making implies that rise in the level of income. Many of the benefits investor confidence may be difficult to build. of trade liberalization, privatization and/or This can create a particularly difficult dynamic, deregulation, and financial sector reform depend particularly in the interaction between govern- on the responses of private investors.The gains ment and providers of infrastructure.This dynamic come with new export industries, new expan- is that, even if investors would invest at existing sions of industry, improvements in efficiency and profits/prices if they were confident these prices productivity (which often require investments), would persist, they fear the government may and new activities. Small reforms may have big renege on its commitment to price regulations impacts if they are seen as harbingers of future and attempt to squeeze their profits in the future. 262 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s TABLE 8.5 Policy Reform and Growth: Sources of Differential Impacts Possible slips between policy action Question Effect and growth/output response j By how much does a policy yt action raise growth? Policy actiont Does policy action change APt ,T · Policy action conditioned anticipated policy? Policy actiont · Policy action unsustainable (either economically or politically) · Policy actions not institutionalized Do changes in trajectory of policy Rt · Changes in returns not large ,T change the trajectory of distributions APt · Policy is "wrong" ,T of profitability? Do changes in trajectory of profitability kt j* · Expected profitability higher but uncertainty ,T raise desired capital stocks? Rt higher (and investors not risk neutral) ,T · Policy changes lower profitability in the short run (adjustment costs) but raise it in the long run · Complementarities Do changes in desired capital stock(s) kt j · Financial system does not accommodate lead to investment responses? kt,T j* · Other aspects of investment climate unfavorable to investment Source: Pritchett 2003a. If that is so,investors will be willing to invest only that excessively ambitious reforms that are at a large risk premium over and above profitabil- delayed in implementation can hinder the for- ity.But--particularly in a weak political and insti- mation of positive expectations. tutional climate--the likelihood that a government will renege is higher, the higher the ex post profitability.The risk of reversal alone can 2. The Way Forward block an investment response, given that the only profit rate at which investors would be willing to Taking on board the lessons of the 1990s,what is risk their capital in new investments is one at the way forward? Three guidelines are discussed which governments cannot resist public pressure in what follows: to lower prices. Hence the risk of policy reversal · Accept that there are many ways to imple- can itself create a self-fulfilling prophecy of failure. ment common principles. These dilemmas explain the continuing search for mechanisms with which to signal a · Pursue growth strategies--not just stabiliza- government's commitment to the irreversibility tion or the avoidance of problems. of reforms.The temptation has been to argue that · Create the institutional conditions for a favor- the lack of a supply response meant that reforms able investment climate. had to be pushed harder, faster, and deeper. But this is not necessarily so.If the problem is that the Common Principles--And ManyWays to reforms are not expected to be sustained because they are too aggressive, pushing them harder ImplementThem might further undermine expectations of their Perhaps the most important and difficult lesson sustainability. To sum up, acknowledging the of the 1990s is that there is no one right way to importance of expectations does not imply that achieve development. either big bang or gradualism is the right The 1990s have not proved mainstream econ- approach to policy reform, but it is a reminder omists wrong;indeed the basic principles of eco- POLICY REFORMS AND GROWTH PERFORMANCE: WHAT HAVE WE LEARNED? 263 nomics have proved remarkably resilient. In · Competition. The principle that competition countries such as Poland, the Czech Republic, from alternative suppliers promotes produc- and the Slovak Republic,where the introduction tive efficiency does not dictate that competi- of incentives proved feasible, they have worked tion has to take any particular form. China's remarkably well. In China and Vietnam, the experience with township and village enter- introduction of stronger incentives has led to the prises, which were not private enterprises in most rapid poverty reductions in history. the usual sense but created effective competi- What was wrong, and never should have tion, is instructive (see box 6.1 in chapter 6). been part of economics, was the belief that the · Macroeconomic stability. The view that this or first principles of economics had to be imple- that particular arrangement is needed in mented in a particular way (Rodrik 2002a).This order to create macroeconomic stability is point can be illustrated with regard to four eco- belied by the diversity of experience of coun- nomic principles: tries that tried the same thing, and the simi- · Expectations about future claims. Investors need larity of experience of countries that tried certainty that they will reap the gains of their different things. investment. But this stability of expectations can be sought in a variety of ways. For exam- To conclude this discussion of the different ple,do favorable investor expectations depend modalities for implementing common princi- on property rights? Do property rights rest on ples, it should be emphasized that "one size does the same definition of property and the same not fit all"should not be interpreted as"anything means of enforcing those rights as have devel- goes."A vast array of policies in the world are not oped in some particular industrial country? fundamentally sound, and are not heterodox Experience with land titling has shown that, implementations of sound orthodox principles. in some cases, holding the title to land A vast array restrict competition in order to pro- increases a farmer's incentives, but in other tect existing owners (private and public),and cre- cases the existing informal systems have pro- ate investor uncertainty through arbitrary and vided adequate security. One way of provid- capricious behavior by state officials. What is ing property rights is through a needed is not less economics but more and bet- well-functioning legal system, but many ter economics,to identify the exact set of policies countries achieved decades of rapid growth and institutional changes needed to address bind- with very little legal certainty, when stability ing constraints on growth, based on first princi- was embodied in the political system. ples in each instance. · Openness. The principle of openness to ideas, trade, and investments with the rest of the Growth Strategies world need not entail free trade. There are many ways of engaging productively in inter- If policy reform, while beneficial, does not national markets. Even the four East Asian explain the bulk of the variation in growth per- Tigers, all famed for being outward oriented, formance across countries and time, something differed widely in the extent to which their else must. As discussed in chapter 2, recent governments intervened in the economy and research has emphasized that there are large in international trade. While Hong Kong numbers of extended episodes of rapid growth-- (China), as a trading center, was always open, some sustained and some not.How these growth the Republic of Korea opened its markets to episodes are initiated and sustained is a key ques- imports only quite late in its growth process. tion.While "policies," as represented by standard While some economies invited foreign growth-regression measures,do increase the like- investors, Korea had very little direct foreign lihood of a growth episode,they are far from suf- investment. ficient to explain growth. And what causes the 264 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s start of a sustained episode of rapid growth is not is too high; perhaps investors fear macroeco- well understood. nomic instability; perhaps too few profitable In any developing country, nearly everything opportunities are discovered; or perhaps infra- is far from ideal.The 1990s have shown that to structure deficits raise costs. achieve rapid growth, countries do not need to One problem,particularly with strategies that get everything right but they do need to get the involve donors, is that governments face pressure right things right. Identifying those right things is to act on all fronts simultaneously. In creating an the purpose of devising a growth strategy, which all-encompassing document such as a poverty is a coherent set of actions designed to initiate and reduction strategy paper it is very easy to justify sustain rapid growth. anything as being important to growth--from Devising a growth strategy requires a clear low human capital, poor health conditions, judi- diagnosis of the obstacles to growth--in particu- cial insecurity, weak infrastructure, a weak civil lar, the binding constraints, which will vary service, to stagnant investments in agriculture. widely depending on countries' initial condi- Thus too often a proposed strategy becomes a tions.To illustrate, figure 8.3 from Hausmann, menu, not a meal.To be sure, all of these prob- Rodrik, and Velasco (2004) maps the possible lems will at some stage need to be addressed.But explanations of slow growth in a country in identifying the binding constraints on growth which the slow growth is associated with low and focusing on them is the essence of strategy. rates of investment and entrepreneurship.15 The figure emphasizes that starting from fundamental Institutional Conditions for a Favorable principles can lead one's search for binding con- Investment Climate straints in many directions. For example, starting from the condition that profitability must exceed For investors,the launch of any new public policy the cost of investment,perhaps the cost of capital initiative raises the question, How will policy FIGURE 8.3 Diagnosing the Problem of Low Levels of Investment and Entrepreneurship Profitability condition: paf (k, k, l, g) = r Low paf' (k, k, l, g) High r Low p Low a Low k Low l Low g Bad local Bad international finance finance Micro risk: Macro risk: Multiple equilibria, Insufficient infrastructure, property rights, financial or spillovers, coordination high transport costs, corruption, taxes fiscal crisis failure low tax base Too little bank Country risk still too high, competition, FDI conditions Lack of R&D, low entrepreneurial Poor human capital, high spreads unattractive rents, too little "self-discovery" rigid labor market Source: Adapted from Hausmann, Rodrik, and Velasco 2004. Note: p = private appropriability of investment; l = labor input; a = total factor productivity; g = government "input" (e.g., infrastructure); f = production _ function for an individual investor or firm; paf' = expected private return to investment; k = firm-level capital; r = real interest rate; k = economywide capi- tal; FDI = foreign direct investment; R&D = research and development. POLICY REFORMS AND GROWTH PERFORMANCE: WHAT HAVE WE LEARNED? 265 actions evolve with this new policy? For govern- debates about reform. For example, should one ments and societies at large, a key question going privatize when there is no regulator? Should forward is,How does one develop the institutions banks be liberalized while prudential regulation of policy that reliably lead to the (mostly) positive is weak? Particularly with the large fiscal losses in use of discretion in policy implementation?16 the financial sector in the 1990s, should reforms First,continuity in the background institutions have been more gradual, with greater attention of policy making is conducive to success in pursu- paid to prudential regulation? In some cases "the ing individual reforms. One of the problems with use of all deliberate speed" is hard to distinguish the transition in Eastern Europe and former Soviet from "never."Another school of thought argues Union countries is that investment depends on that capacity only develops in response to need, expectations of policy implementation,that policy and so if one delays the privatization of utilities implementation depends on background institu- until one has developed an adequate regulatory tions,and that when institutions are in flux no one capability one might delay forever. Indeed, it is can say with certainty what will happen.The fact hard to build experience in regulation if there is that Indonesia has had much more difficulty than nothing to regulate.17 Korea and Thailand in restoring growth after cri- A fourth area of debate about creating favor- sis is almost certainly because Indonesia's back- able expectations is the tension between attempt- ground institutions have shifted, so that no one ing to reassure specific investors and improving can predict quite where they will lead,while those the overall investment climate. Some would of Korea and Thailand have not. Often shifts in argue that since the costs of investment are so background institutions are seismic political events high, and improvement in organizations and beyond the control of any policy maker.But expe- institutions is so slow, the best way to attract rience does suggest that new governments that are investment in the short run is to nurture individ- in the midst of an institutional shift should con- ual investors, either on a deal-by-deal basis or in sider it a priority to establish credibility around a special regimes (such as for foreign investors). few key areas, rather than undertaking a broad The latter approach, bypassing the weaknesses in array of new policy initiatives whose success may the overall investment climate, is attractive depend on expectations. because initiating a new industry or endeavor Second, if the key problem is that credible often requires attracting a large investor. Cer- background institutions that can limit predation tainly this approach has been made to work, but by the state, such as an independent judiciary or it has dangers.Complex special deals can be con- electoral accountability,do not exist and the gov- spicuously opaque and a perfect vehicle for cor- ernment cannot make a credible commitment to ruption.Particularly when their negotiated terms resist predatory behavior, it is possible that no are contested, special deals can undermine the amount of institutional reform will sufficiently perception of social and political legitimacy of a reassure investors. Acemoglu, Johnson, and government's overall policy approach (the deal Robinson (2001) have argued strongly that what with Enron in the Indian state of Maharashtra is meant by"institutional quality"is not the state's and the water deal in the Bolivian city of ability to regulate transactions between individu- Cochabamba are examples). Particularly in infra- als,but rather a country's ability to limit the state's structure, the renegotiation of individual deals temptation to expropriate. Since economic elites has proven be an enormous challenge (World often benefit from controlling the state and exist- Bank 2004e). Finally, cutting deals for specific ing institutions (Hellman, Jones, and Kaufmann investors or specific classes of investors can 2000;Acemoglu, Johnson, and Robinson 2001), undermine the pressures for systemic improve- there may be little internal impetus for reform, ment for all investors. De Soto (forthcoming) is precisely when it is needed most. eloquent on the fact that most Latin American Third, the capability of the direct organiza- investors exist outside the scope of the formal tion of policy making is often a key issue in legal economy. 266 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s Notes could then defer spending on the maintenance of public assets,causing potentially the same (or an even 1. This feature, "Solow invariance" (Hall 1999), is larger) reduction in the value of net assets while robust and driven by basic features of these models. meeting the same target for the fiscal deficit plus 2. The classic example (perhaps because it was there that payables. One could then set conditions that specify the theory had been the most clearly worked out) was a cash deficit target, a limit on payables, and a limit the calculation of the welfare losses that resulted from on the reductions in maintenance. But there are still the differences between international and domestic many other ways to reduce net public assets--for prices induced by border restrictions on trade,such as example,freezing the nominal wages of public sector a tariff.The standard analysis showed that a tariff raised workers at lower than sustainable levels, or under- prices, which benefited producers and hurt con- funding future pension obligations, or authorizing sumers, but that the efficiency losses from "too little" expenditures (such as guarantees of lending) that cre- consumption caused an overall net social loss. Graph- ate a quasi-fiscal obligation. ically this loss of consumer surplus was a triangle--in 8. The new law attempted to remove every vestige of fact the estimates of the losses from price distortions judicial discretion by declaring that if any creditor were known as Harberger triangles (afterArnold Har- petitioned for a bankruptcy and a debtor was more berger 1971).The"partial equilibrium"estimates sug- than a certain number of days overdue on a contrac- gested that a move from the current level of tual payment, the judge must declare bankruptcy. restrictions to completely free trade would produce 9. For instance, in the early 1990s tariffs in India were welfare gains on the order of 1 to 5 percent of GDP. four to five times as high as in most Latin American These small estimates implied that the temporary countries. "growth" effects caused by the transition from lower 10. A Ponzi scheme refers to any investment that pays off to higher levels of (properly measured) output from initial investors unsustainably large returns not out of efficiency-improving reforms were quite limited. actual returns from investment but from flows of 3. That is, y * . funds from new investors.These depend on rapid A growth in new investors, but in the end not every investor can be paid the promised high returns. 4. Many proponents of the efficiency case for the wel- 11. This would be the case, for example, for a good that fare gains from trade (as opposed to the "growth" is an input into many other goods and is produced arguments) are strong supporters of free trade.Jagdish by a domestic monopoly. Bhagwati frequently points out that there was never 12. No one can look at the experience of Singapore or any theoretical support for growth-regression-based the Republic of Korea (and earlier Japan) without claims on behalf of trade liberalization--but that the- being convinced that purposive government action ory and evidence on the microeconomic level pro- to promote rapid development can succeed. Con- vide all the support one needs. versely, no one can review the tragic experience in 5. Of course, in reality each organization will have its many African countries and believe that purposive own "principal-agent" problems. government action (at least ostensibly) to promote 6. For instance,in trade policy an original policy would rapid development cannot fail. be set, restrictions would be added, and then excep- 13. Comparing industrial countries with poorer coun- tions granted, and then new categories created, and tries, it is noticeable that government action is much then other new restrictions added. Many countries more pervasive in industrial countries--tax rates are had reached the point where few people actually higher, and regulation is pervasive--and that the knew what the trade regulations were (in many cases, exercise of discretion is explicit,and that much of the even customs officials did not possess fully up-to- infrastructure is owned and operated by the public date copies of the tariff code) and where, taken as a sector. A frequent practice has been to attempt to set of interventions, the trade policy was "irrational." transplant more or less wholesale the policies of Similar accretions--taking over firms that had gone industrial countries--including the direct institu- bankrupt here, making a firm a parastatal in order to tions of policy making--without adequate consider- obtain official financing there--often led to govern- ation for whether the transplants could survive in ment ownership of a variety of businesses and activ- entirely different conditions. For example, every ities for which there was no coherent rationale. industrial country regulates banks. But can banks be 7. Suppose that to meet the fiscal deficit target the gov- successfully regulated without an effective legal sys- ernment simply lengthens payments to suppliers. tem that can enforce creditor rights? Without a This does not change net public assets. One could strong tradition of an autonomous civil service that imagine then putting limits on both the cash fiscal can resist political pressures? Without effective leg- deficit and the payment of suppliers.A government islative oversight? Without transparency and an POLICY REFORMS AND GROWTH PERFORMANCE: WHAT HAVE WE LEARNED? 267 aggressive free press? Without a police force that can 16. This is the main question in chapter 9,which reviews protect impartially against threats of violence? efforts in the 1990s on several fronts. 14. These observations are part of the same overall story as 17. Countries with parastatal firms had decades in which the first two common lessons in this chapter--the ques- they could have created regulatory capability--but tion of large- versus small-level effects and the impor- they did not do so, in part because it was not per- tance of institutional quality for successful policy ceived as necessary. Similarly with financial sector implementation. regulation: developing the capability for "arm's 15. If investment were high and growth slow, a different length" regulation when the government embraces diagnostic would be appropriate. the entire sector is conceivable, but difficult. Country Note G Africa's GrowthTragedy: An Institutional Perspective A frica's slow growth was unex- pected (Easterly and Levine FIGURE G.1 1997; Collier and Gunning Africa: Getting Poorer over Decades 1997). In the 1960s, most African countries were richer than their Asian counterparts, and their 10 stronger natural resource base led many to believe that Africa's economic potential was superior to overpopulated Asia's.This view was shared by renowned economists, from Gunnar 0 Myrdal in his well known Asian Drama, to Percent Andrew Kamarck, the founding director of the World Bank's economic analysis complex, who listed seven African countries that he thought ­10 could grow at annual rates of 7 percent or more 1960s 1970s 1980s 1990s (Enke 1963; Kamarck 1967). More recently, Source: World Bank, World Development Indicators 2004. many economic reports, including several by the Note: Box-and-whisker plot for decadal growth rates of 28 African World Bank, foresaw rapid growth in Africa. countries that have GDP series from the 1960s onward. The plots The continent's growth record, however, has show averages as well as the dispersion around them: median, fallen well short of expectations. Over the last first and third quartiles, and outliers. four decades, in the 28 countries that have com- plete gross domestic product (GDP) series for predictive power particularly when it comes to this period,the median growth rate has gradually breaks with past trends--which are the essence of but persistently declined (figure G.1) and 11 development processes. countries now have income levels lower than at Seeking to understand the deep forces influ- the time of their independence. encing Africa's growth performance, researchers Ranked by their growth performance since have increasingly looked into structural factors: 1960,15 of the world's 20 slowest performers and geography (Sachs and Warner 1997; Bloom and only 2 of the 20 best performers are in Africa Sachs 1998; Mellinger, Sachs, and Gallup 1999); (table G.1). Perspectives on Africa have thus ethno-linguistic polarization and inequality become much more guarded (Easterly and (Easterly and Levine 1997); and institutions. Levine 1997;Acemoglu, Johnson, and Robinson The effect of institutions on growth has been 2002). a particularly fertile area of research in the last 10 As noted earlier in this report, growth has years, bringing new analytical insights and per- been poorly predicted not only in Africa but in spectives (Acemoglu, Johnson, and Robinson the developing world in general. Growth is diffi- 2001,2002).For example,some see the origin of cult to predict because it reflects processes of Africa's institutional weaknesses in the long-last- change, and complex historical and political ing effects of European colonial rule, which had forces.Social scientists and historians have limited little incentive to develop African local institu- 270 AFRICA'S GROWTH TRAGEDY: AN INSTITUTIONAL PERSPECTIVE 271 TABLE G.1 African Growth in Context: Average Annual Growth Rates of Real per Capita GDP, 1960­2001 Best performers Worst performers Botswana 6.4 Guyana 0.5 Korea, Rep. of 5.8 Argentina 0.5 Singapore 5.6 Côte d'Ivoire 0.5 China 5.6 Bolivia 0.3 Oman 5.4 Zimbabwe 0.3 Hong Kong, China 5.2 Burundi 0.3 Thailand 4.5 Nigeria 0.2 Ireland 4.2 Rwanda 0.2 Japan 4.1 Ghana ­0.1 Malaysia 3.9 Senegal ­0.2 Portugal 3.8 Chad ­0.4 Lesotho 3.6 Venezuela, R. B. de ­0.5 Indonesia 3.5 Central African Republic ­0.7 Spain 3.3 Zambia ­1.1 Hungary 3.2 Haiti ­1.1 Greece 3.2 Sierra Leone ­1.1 Norway 3.1 Madagascar ­1.3 Egypt, Arab Rep. of 3.0 Niger ­1.5 Finland 2.9 Liberia ­3.2 Italy 2.8 Congo, Dem. Rep. of ­3.3 Source: World Bank, World Development Indicators 2005. Note: Real GDP per capita growth rates (only for countries with GDP per capita series since 1960). tions and focused instead on developing extrac- Cold War politics did not encourage the devel- tive institutions (Crawford 1994).As discussed in opment of effective state institutions and good Country Note H on natural resources, the so- governance in Africa. In many instances, the called natural resource curse has been another United States and the Soviet Union supported factor emphasized in the literature (Sachs and political regimes and leaders intent on prevent- Warner 1995b, 2001). ing such institutions from emerging (Herbst Recently, the focus has been on the African 2000). state. Scholarly research and policy-making cir- From a longer historical perspective, the cles increasingly view poorly functioning state deeper cause may be the pattern of state forma- institutions as the root cause of Africa's develop- tion in Africa (Herbst 2000). For geographical ment problems, and believe that solutions are to reasons,state power was particularly costly to con- be found within the state itself and political insti- solidate in Sub-Saharan Africa: population densi- tutions that link the state and society (Davidson ties were low and barriers to long-distance 1992; Chege 1998; Herbst 2000; van de Walle transport too numerous.Thus Africa's pattern of 2001). state formation and consolidation differed from Post­WorldWar II geopolitics played a role in those in some other parts of the world. many countries.The system of international rela- In Europe, for example, land was scarce rela- tions polarized by the Cold War, which Africa's tive to labor, and therefore incentives to exert new democracies had to face after their inde- control over land were strong, even if at the cost pendence,turned much ofAfrica into an arena of of wars. Nation-states that could efficiently per- political struggle between the two superpowers. form key functions--mobilize fiscal and human 272 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s resources, organize and finance an army, provide following decolonization, modes of governance public goods through effective administrations, rapidly shifted to "neopatrimonial" systems of and establish legitimacy, not least though their rule, characterized by "client-patron" relation- ability to deal with citizens through representa- ships (Joseph 1998). tive institutions--were able to thrive. States that Seeking a solution toAfrica's states'inability to could not, disappeared. exercise their authority across the territories they Herbst argues that this Darwinian process of are to control,Herbst (2000) suggested rethinking state selection and survival did not take place in colonially imposed borders.While this is a highly Africa,where it was labor that was scarce,not land. controversial solution, Davidson (1992) also sug- The drawing of national borders by former colo- gests that creative thinking is needed to find alter- nial powers, independent of the new states' ability natives to nation-states, that can incorporate to exert their authority over their territories,wors- indigenousAfrican forms and traditions of gover- ened the problem by enabling"weak"states to per- nance. Recent reports suggest looser political sist without requiring them to strengthen their arrangements, to enable greater autonomy in institutional foundations, effectiveness, or political divided societies (World Bank 2000g; Ndegwa legitimacy. Because their countries lacked the and Levy 2003). external threat of war or territorial conquest that While different forms of explanation--geo- had driven much of European state-building,post- graphical, political, institutional--all provide use- colonial African leaders never faced significant ful insights and perspectives,it is unlikely that any incentives to extend their power--including single approach will be able to respond to all the power related to the provision of public goods on questions that the continent's performance raises. the entirety of their territory. States that did not For example,none is able to explain the differen- have to fight to survive had no need to invest in tial growth within the continent. Why has effective administrative and fiscal institutions, to Botswana been able to grow at the world's fastest control domestic opposition, or to make political rate for the past four decades, notwithstanding concessions to their citizens.Aid and the ColdWar one of the highest rates of income inequality in accentuated this state of affairs in some countries. the world and a reliance on natural resources, Other observers have emphasized the emer- which has been a curse in many other developing gence of the African state, not as an organic evo- countries? Why has Tanzania been able to main- lution of existing societal and institutional tain corruption at relatively modest levels, and to arrangements, but as an artificial creation oblivi- create a national ethic?Why has political stability ous to those arrangements. Mamdani (1996), for been elusive in Côte d'Ivoire in the past 10 years, instance, pointed out that European colonial rule but not in Ghana? Among Africa's largest created state institutions relying on customary law economies, why have some countries been able under a regime of "decentralized despotism," to grow so much faster than others (figure G.2)? which was exerted through indigenous chiefs. As emphasized throughout this report, specificity The population was ill-prepared to participate as is important for accurate analysis of growth and citizens in the modern states that succeeded colo- for design of effective growth strategies: depend- nial rule.Hence,Mamdani argues,most ofAfrica's ing on the country,or the time,some factors may postcolonial history is to be understood as citi- be more important than others. zens' struggle for their rights. Davidson (1992) Recent improvements in policies seem to emphasized that the nation-state as a mode of account for improvements in performance start- social organization was ill-suited to African reali- ing in the second half of the 1990s, when the ties.A European creation, it ignored the checks median growth rate rose from ­0.6 to a positive and balances embedded in indigenous power 0.9 percent--a significant 1.5 percentage point structures and their evolution in the years before increase (figure G.3). Yet behind these policy colonial rule. It alienated political structures from improvements were improvements in political the lives and needs of the population.As a result, governance in some cases (Ghana, Kenya, Mali, AFRICA'S GROWTH TRAGEDY: AN INSTITUTIONAL PERSPECTIVE 273 Tanzania, Uganda) while in others there were improvements in security (Mozambique, Sudan), FIGURE G.2 making it difficult to see a stable causal relation- Africa's Seven Biggest Economies: Volatile and Unstable (indexes, 1960 = 100) ship.Although some studies (Gelb, Ngo, and Ye 250 2004) show that structural reforms and the qual- ity of their implementation track African per- 200 formance quite well (table G.2), there is in Africa 150 a strong sense that improvements in the economic fortunes of the continent will depend on its abil- 100 ity to establish effective political governance 50 structures and to ensure security--from which better policies will necessarily emerge.This per- 0 ception is confirmed by the focus of new leaders 1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 on dealing with weak institutions--in, for exam- Nigeria Congo, Dem. Rep. of South Africa Sudan ple, Ethiopia, Ghana, Mali, Mozambique,Tanza- Kenya Ghana nia, South Africa, and Uganda (World Bank Côte d'Ivoire Source: World Bank, World Development Indicators 2004. 2000g). Expectations that the improvement noted Note: African countries with the largest populations and with GDP per capita series starting in the 1960s. above indicates a break with past trends need to be balanced with the knowledge that few developing countries have been able to transform episodes of growth into sustained and prolonged growth.As FIGURE G.3 discussed in Country Note B, "Lessons from Africa: Rebounding in the Late 1990s CountriesThat Have SustainedTheir Growth,"the (median of GDP per capita growth rates) key is countries' ability continuously to adjust and 6.0 reform institutions in a manner that enables them to sustain higher levels of income and lay the basis 4.0 for further growth. 2.0 0.0 ­2.0 ­4.0 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 Africa (28) Africa (41) Aggregate Source: World Bank, World Development Indicators 2004. Note: Numbers of countries in each sample are in parentheses. "Aggregate" is total Sub-Saharan African GDP over total population. Three-year moving averages. 274 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s TABLE G.2 Annual Growth in 17 African Countries, 1975­2003 Country 1975­84 1985­89 1990­96 1997­2003 Six sustained reformers 0.3 0.9 1.5 2.2 Six later adjusters ­2.3 0.1 ­2.2 1.8 Five governance-polarized countries ­0.9 0.3 ­0.6 ­1.6 Source: World Bank, World Development Indicators 2004. Note: Median of the real GDP per capita growth. Eight (six) sustained reformers: Benin, Burkina Faso, Ghana, Malawi, Mali, Mozambique, Uganda, and Zambia (Mozambique and Uganda do not have a complete 1975­2003 GDP per capita series). Eight (six) later adjusters: Cameroon, Chad, Guinea, Madagascar, Mauritania, Niger, Senegal, Tanzania (Guinea and Tanzania do not have a com- plete 1975­2003 GDP per capita series). Five governance polarized countries: Côte d'Ivoire, Kenya, Nigeria, Togo, and Zimbabwe. This classification is from the 1994 World Bank study, Adjustment in Africa. Chapter 9 Improving Public Sector Governance: The Grand Challenge? M ANY GOVERNMENTS IN econometric techniques, that large, cross-coun- developing countries face try analyses emerged on the impact of gover- the challenge of delivering a nance institutions on investment and growth. wide range of services essential for develop- This research has shown that corruption--which ment--from infrastructure and social services to is both a symptom and cause of bad gover- the functioning of the legal system and enforce- nance--discourages private investment and, ment of property rights--all of which pose the more generally, that the quality of governance challenge of how to get governance "right." institutions has a significant impact on economic States have responded with varying degrees growth (Mauro 1995; Knack and Keefer 1995; of success. At one end of the spectrum are the Wei 1996, 2000;World Bank, World Development failed states, where governments barely exist, and Report 1997;Kaufmann,Kraay,and Zoido-Loba- where they do, provide hardly any services. At ton 1999;Kaufmann 2003;Kaufmann,Kraay,and the other extreme are a handful of countries Mastruzzi 2003; Rodrik, Subramanian, and where governments and their leaders are doing Trebbi 2002). Further, the research provides evi- well by most development measures. In between dence that corruption distorts the allocation of are weak or predatory states that "consume the resources in ways that hurt the poor (Mauro surplus they extract, encourage private actors to 1998a, 1998b;Tanzi and Davoodi 1998; Gupta, shift from productive activities to unproductive Davoodi, and Alonso-Terme 2002).2 rent seeking,and fail to provide collective goods" Combined with urbanization and the spread (Evans 1995, 24); young democracies managing of democracy, and also the extensive public simultaneous political and economic liberaliza- awareness efforts of international organizations tion with weak bureaucracies and few checks and such asTransparency International and theWorld balances; and more mature democracies where Bank,3 the empirical research gave rise to gover- governments face the same difficulties as nance reforms in developing countries. These advanced countries when it comes to political ranged from very focused technical reforms of corruption and abuse of office.Then there are budgetary and civil service systems to more the large,continent-size polities such as India and encompassing efforts such as decentralization and Brazil, within whose national boundaries can be the overhaul of legal and judicial systems. found the entire range of governance configura- This chapter reviews these reform efforts and tions. the lessons they yield. Section 1 introduces key Though extensive research had probed the governance concepts and discusses why gover- causes and impact of poor governance, and in nance reforms are particularly challenging, and particular of corruption,1 it was not until the section 2 draws emerging lessons. Recognizing mid-1990s, with improvements in data and that one size does not fit all, section 3 presents a 275 276 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s heuristic approach to identifying avenues for reforms depending on broad country character- FIGURE 9.1 istics. Section 4 concludes the chapter. Citizens and Politicians State capture 1. Understanding the Gover- nance Conundrum4 Policy makers Public sector governance refers to how the state Bureaucratic capacity acquires and exercises the authority to provide Delegation Political and and manage public goods and services. accountability voice The quality of governance (and thus the Internal accountability nature and extent of corruption) depends funda- mentally on institutions.As discussed in chapter 1, institutions are the "rules of the game" that Citizens Client Providers shape the behavior of organizations and individ- power uals in a society (North 1990, 3). Institutions can Public goods/ be formal rules, such as a country's constitution, service provision its laws and regulations, contracts, and internal procedures.5 Or they can be informal rules, such Source: Author's elaboration. as the values and norms that drive bureaucratic behavior. Scholarly research and concern with which citizens can hold politicians' feet to the institutions is not new, but a strong interest in fire. Restraints in the form of court adjudication institutions reemerged in the 1990s, largely of disputes among contracting parties, especially because the stronger macroeconomic policies of between government and the citizenry, and leg- the 1980s had not achieved more rapid progress islative oversight of executive or ruling party in development and poverty reduction. Interest decisions and actions, foster accountability moved from "getting the policies right" to "get- between elections.In well-functioning democra- ting the institutions right" and had a particular cies, these latter features are embodied in the focus--the rules of the game on which the gov- constitution and promote the rule of law: every- ernance of the public sector is grounded. one, politicians included, behaves in accordance Fundamentally, public sector governance is with agreed rules as embodied in laws and regu- about the nature and quality of three principal lations and no one is above the law. relationships: between citizens and politicians, However, many countries are not democratic between politicians as policy makers and the and even in those that are,as discussed in the next bureaucracy (those responsible for providing chapter, the conditions needed to make democ- public goods and services), and between the racy function well are very demanding. In many bureaucracy as delivery agents and the citizenry countries, the formal trappings of democracy do as clients (figure 9.1).6 not translate into accountable decision making for a variety of reasons--from the lack of a truly inde- pendent parliament or judiciary to electoral mar- Citizens and Politicians: ket imperfections. Even in countries with regular The Heart of Governance competitive elections among multiple political In an ideal world, citizens can hold politicians parties or candidates, elections often do not have accountable for their actions and for policy out- the desired effect, so that politicians can easily comes, both through elections and through renege on campaign promises and responsibili- checks and balances on the abuse of power. Peri- ties.7 In either case, the relationship between citi- odic elections provide the basic means through zens and politicians is typically governed by weak I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 277 institutions--whether political institutions (rules major features.First,it provides the means through or arrangements that define,govern,and influence which the bureaucracy can develop and enhance how leaders and other politicians are selected and its capacity to implement policies and deliver pub- replaced) or institutions of restraint (rules or lic goods and services. Capacity in this context arrangements that establish checks and balances refers to (1) the skills bureaucrats need to deliver on the abuse of power and authority). Courts are on various mandates,(2) the resources (capital and easily swayed by influential politicians; the legisla- financial) to support the needed efforts, and (3) ture rubber-stamps the narrowly focused special the processes and systems (such as the budgeting interest initiatives of the executive; high-ranking system and the procedures for using it) that enable officials abuse their authority for private gain; and large numbers of individuals to function efficiently enforcement agencies prey on the citizenry. In together. Second, the compact establishes means other words, the rule of law is weak. through which bureaucrats can be held account- When the rule of law is weak,the risk of state able to policy makers for performing their tasks. capture is high. State capture "refers to actions of That is, the bureaucrats are responsible to policy individuals, groups, or firms . . . in the public makers for accomplishing certain tasks and are and/or private sectors to influence the formation prepared to explain and face the consequences of of laws, regulations, decrees, and other govern- deficiencies or failures. Accountability mecha- ment policies to their advantage, through the nisms typically involve checks and balances inter- illicit and nontransparent provision of private nal to government agencies,such as internal audit, benefits to politicians and/or civil servants" ex post program evaluations, and ex post report- (World Bank 2000a)8 and is a serious problem in ing, as well as external restraints such as exercised many developing countries (Kaufmann 2003). by an ombudsman,supreme audit institutions,and When pervasive, it becomes the principal stum- anticorruption commissions. bling block for efforts to reform governance Adequate capacity is needed if accountability institutions. is to work. Auditing, performance evaluation, reporting, investigations, and prosecution require Politicians and Policy Makers and the information.And to produce the right informa- tion requires processes, skills, and resources to Bureaucracy:The Core Principal-Agent provide appropriate infrastructure and create Problem appropriate incentives. In the same vein, more Politicians make policy and are responsible to the effective accountability helps strengthen capacity citizenry for the policies that are promulgated because policy makers are more willing to grant and implemented, but they delegate the imple- greater flexibility and because resources make the mentation of policy to an army of bureaucrats.In bureaucracy even more effective. Figure 9.2 delegating, they establish the rules and regula- highlights the salient features of the compact. tions that govern the operations of the civil ser- The nature of the compact between policy vice. These include, for example, formal makers and bureaucrats critically determines the institutions such as the civil service code, the outcomes of policies. When the compact is budgeting system, and reporting systems as well defective, because capacity is weak or accounta- as informal institutions (such as the practice of bility is poor, administrative (or bureaucratic) amakudari in Japan). corruption typically emerges.Weak capacity and Delegation almost always gives rise to the accountability translate into numerous opportu- principal-agent problem.The principal delegates nities for soliciting or extracting bribes and other the implementation of a task to an agent but will illicit payments.9 need to monitor the agent efficiently to confirm The compact is itself partly conditioned by exactly what has been accomplished. the extent of state capture. Administrative cor- To delegate implementation, policy makers ruption differs fundamentally from state capture establish a compact with bureaucrats that has two but is inextricably linked to it. Politicians are at 278 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s that the representative will intercede with or FIGURE 9.2 pressure the bureaucracy on their behalf. If they Politicians and Policy Makers and the Bureaucracy are generally unhappy with the response, they can vote the person out of office. But they can- Statecapture not penalize or punish bureaucrats directly or officially.10 Policy Citizens acquire leverage over the bureaucracy makers if they can organize themselves into nongovern- Bureaucratic mental organizations (Rose-Ackerman 2004).The capacity Delegation capacity to organize gives citizens"voice"(the abil- Political Administrative and accountability corruption ity to monitor the performance of the bureaucracy, voice Internal generate valuable information, and pressure politi- accountability cians for action) and "client power" (the ability to engage directly with the providers of services).Both of these attributes strengthen the compact between Citizens Client Providers power politicians and the bureaucracy and thus help to improve the delivery of public goods and services.11 Public goods/ Figure 9.3 highlights the role of voice and client service provision power in the triad. Source: Author's elaboration. Citizens generally find it difficult to organize themselves, however. Collective action is costly and thus does not always emerge naturally.A weak the heart of state capture--whether as perpe- link between citizens and politicians, as when trators or as willing respondents to the cap- electoral processes are flawed, can exacerbate the tors--but since it is bureaucrats who implement problem of administrative corruption: desperate the distorted policies that result from capture, for service, citizens may ultimately offer bribes to corrupt politicians need at a minimum the acquiescence of at least some segments of the bureaucracy. In practice this implies that politi- FIGURE 9.3 cians constrain the capacity of bureaucrats or Bureaucrats and the Citizenry weaken their accountability. For instance, in many developing countries, the annual budgets State capture allocated to the ombudsman or to the supreme audit institution are inadequate and the person- Policy nel appointed to key posts are rarely the most makers qualified (Heilbrunn 2002). Moreover, politi- Bureaucratic cians will prefer to sustain clientilistic practices capacity Delegation within the bureaucracy rather than to introduce Political and accountability formal, transparent merit-based recruitment voice Internal processes (World Bank 2000b). accountability Bureaucrats and the Citizenry: Citizens Client Providers power Where the Rubber Hits the Road Most citizens'immediate contact with the state is Public goods/ through the bureaucracy. Citizens can complain service provision to their congressman or mayor about the poor Source: Author's elaboration. quality or inadequacy of some service and hope I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 279 bureaucrats or give in to bureaucratic demands tance involving foreign law professors or expert for bribes. The inability to organize can thus practitioners (Kovacic 1995; deLisle 1999).The worsen an already defective compact. foreign experts typically brought the text of their own country's laws, either as a basis for drafting a country-specific statute or simply for 2. Public Sector Governance adoption wholesale.16 Reforms in the 1990s Many of the new laws have had little or no effect on behavior. For example, several former Using the three-part analytical framework out- Soviet nations have statutes that on paper provide lined in section 1 as a guide, this section identi- corporate shareholders more protection than fies some potential lessons from the governance under French or German law, but in practice, as reforms of the 1990s. capital flows attest, investor rights in France and Germany are much more secure (Pistor, Raiser, and Gelfer 2000, 65).Albania and Romania have Enhancing Political Accountability: enacted statutes governing the posting of collat- Legal and Judicial Reforms eral for a loan and the regulation of banks and Democracies, particularly nascent ones, face securities markets, but businesses report that the dilemmas in promoting genuine political compe- statutes are ineffective (Gupta, Kleinfeld, and tition, and thus in establishing political accounta- Salinas 2002, 13­14).And although new laws in bility.When the citizenry cannot clearly attribute Bangladesh, Benin, and Pakistan require bank responsibility for poor outcomes to politicians, debtors to repay their outstanding loans immedi- elections--already blunt instruments for account- ately, most debts remain unpaid (Messick 1999). ability--become distorted, and this enables non- Experience emphasizes the need for pragma- performing or poorly performing politicians to tism in legal reform.First,a new law must reflect remain in power (see also chapter 10).12 the realities of the institutional environment Perhaps partly because of the immense diffi- within which it is to be inserted, including in culty of addressing problems in political institu- particular the state of the institutions that will tions, countries in the 1990s turned to the other enforce it--judiciary, ministry of justice, the channel of political accountability: reforming police, and regulatory agencies (box 9.1).When legal and judicial systems, which seemed more the agencies that will enforce a new law are cor- amenable to technical solutions.13 Two trends-- rupt, technically incompetent, or insufficiently the privatization thrust and the new emphasis on independent of political authorities, the law private sector development--helped to make must compensate for these deficiencies. legal and judicial reform an imperative in many Second,implementation is easier if a new law developing countries.14 Several other factors also incorporates customs, or norms that citizens are provided an impetus: globalization and trade, already observing. Hernando de Soto, for drug trafficking,human rights,immigration,pro- instance,spearheaded several law reforms in Peru tection of intellectual property, suppression of that incorporated the norms and practices of terrorism, and the consolidation of emerging street vendors, urban transit operators, and small democracies (Messick 1999). Law reform and landholders (de Soto 1989). Peru's new land law, development activities skyrocketed during the granting urban squatters property rights to their decade,encouraged by support from the interna- land, relies on customary methods of showing tional financial institutions and key donors.15 possession to establish these rights. Experience also shows that a transparent and Pragmatic Approaches Have Been More inclusive reform process reduces opposition to a Effective and Are More Likely to Succeed new law and enhances compliance with it. In its Most reforms affecting private sector law were projects, the European Bank for Reconstruction supported through short-term technical assis- and Development has included representatives 280 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s ened as the decade unfolded, and with increasing frequency a better-functioning judiciary was BOX 9.1 linked to, if not equated with, the rule of law.17 Bright Line Rules versus Standards This broader emphasis coincided with the grow- A n early lesson that law reformers ing recognition in the international community learned was the importance of that a well-performing judiciary was part of the substituting bright line rules for solution to many development problems. By the standards wherever possible. A bright line close of the 1990s, judicial reform had, in the rule specifies the exact conduct expected words of one practitioner, become "the big tent (for example, a law establishing a 45-mile- for social, economic, and political change gener- per-hour speed limit) whereas a standard ally." Because it lent itself more readily to techni- leaves the question of violation to the cal solutions, judicial reform became the entry enforcement agency (for example, a law point for addressing a problem that was funda- that makes a "reasonable speed" the mentally political in nature. limit). Bright line rules put less of a bur- The widening of objectives caused signifi- den on enforcers than standards. Deter- cant changes in judicial reform programming mining whether a driver was exceeding 45 among assistance agencies. At the World Bank, mph is straightforward, whereas deciding for example, judicial reform projects to create whether a speed was reasonable requires commercial courts and support similarly focused investigation into such factors as time of interventions gave way to much more ambitious day, weather conditions, and the presence undertakings: supporting new institutions to of pedestrians and other traffic. Bright line assume responsibility for governing the judicial rules leave enforcers with little or no dis- branch, revising rules on the selection and pro- cretion and so reduce opportunities for motion of judges, overhauling the management bribery. Since the enforcement of bright of the judiciary's human and physical resources, line rules is far easier to monitor, this can rooting out corruption,constructing new court- provide openings for self-help and other houses and equipping them with modern com- means of enforcing the law without resort munications and computer technology, training to the authorities. lawyers and judges, providing programs to reach out to women and the poor, and establishing administrative courts. Judicial reform proved to be a far greater chal- not only of government ministries, regulators, lenge than expected. As did the attempted judges, and legislators but also of businesses and reforms of other public sector institutions during other civil society groups. the 1990s, most of these interventions produced little change (Burki and Perry 1998).As experi- Poor Incentives Rather than Weak Capacity ence grew, it became clear that the roots of poor Have Been the Root Cause of Poor Judicial performance in the judicial system lay much less Performance in a lack of resources and skills than in the behav- Concern to improve the implementation of laws ior of judges, clerks, lawyers, and litigants. For gave rise to parallel reforms of the judiciary.Early example, India has created an enormous number in the 1990s, reforms focused on contract of tribunals to handle civil service, tax, land, and enforcement, influenced by seminal work on consumer cases,to reduce the burden on the reg- institutions and economic growth that empha- ular civil courts and speed up the disposition of sized the important role that contract enforce- cases, but these reforms have had little effect ment plays in economic development (North (Moog 1997).The reason is that lawyers, clerks, 1990;Weingast 1995; see also Oliver Williamson and many litigants have an interest in court delays 1985).The rationale for judicial reform broad- and thus continue to frustrate reform efforts. I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 281 State capture has compromised the proper func- tioning of formal institutions. BOX 9.2 Integrated Justice Sector Reforms: Development of Complementary Institutions The Jamaican Case Is Essential for Reform The absence of parallel institutional reforms will T he justice sector plays a critical role in enforcing tend to negate any progress in judicial reform.For the rule of law and protecting property rights example, efforts to improve the judiciary will be and the rights of citizens. Safeguarding the hindered if they are not complemented by accountability mechanisms that underpin the justice sec- reforms of the police and the public prosecutor tor requires effective collaboration between the courts offices (box 9.2). and the police. Successful violence reduction programs While most efforts at judicial reform focused substantially increase the cost of crime by combining a on improving court systems,some dealt with alter- high probability of capture (by the police) with a high native dispute resolution mechanisms, offering probability of conviction (by the judicial system). arbitration, mediation, and other dispute-resolu- Jamaica's fairly well-developed judicial system has tion methods as a way to channel disputes away been criticized because of dramatic increases in political from the courts to private fora. These projects and criminal violence. To have an effective crime pre- tended to overestimate the potential of alternative vention strategy, Jamaica would need to match the dispute resolution methods. Unlike state-spon- effectiveness of its judicial system with better police sored courts, private dispute resolution fora can- capabilities; it would need to increase police accounta- not induce the parties to appear or to comply with bility and improve police-community relations. the resulting decision.For private methods to sub- stitute effectively for resolution by the courts, the Sources: U.S. Library of Congress country study and study by parties must have some incentive to submit their the World Bank Group Jamaican country team, 2004. dispute and to be bound by the outcome. At the same time, the projects largely ignored that many disputes could be avoided altogether if more information were available on the credit- electorate holds government to account. In worthiness and reliability of potential contracting doing so, the legislature can use several means, parties. The importance of credit bureaus and including approval of budgets,the questioning of other devices for sharing information is beginning senior government officials, the review and con- to be recognized. For instance, recent research has firmation of executive appointments, impeach- shown the significance of information availability ment and/or the power to dismiss the for deepening credit markets (World Bank, World government, the establishment of parliamentary Development Report 2003, 101­103). committees, and the formation of commissions Improving information sharing is a relatively of inquiry. apolitical approach to improving the rule of law, Relatively few attempts were made to reform but by itself it cannot solve the problems of credi- legislatures during the 1990s (Manning and ble commitment that are inherent in contracting Stapenhurst 2002). The reasons included the among private parties. The question that then more political and controversial nature of such arises is what else a country can do to make an reforms, donors' lack of experience with such investment environment more predictable, while reforms, and the nascent state of many legisla- waiting for the courts-based rule of law to emerge. tures (Lippman and Emmert 1997). Nevertheless, the spread of democracy cre- Reforms to the Legislature Deserve ated the space for legislatures to evolve into inde- Greater Attention pendent political institutions that could oversee In most countries, the legislature is the constitu- the executive and, with or without foreign assis- tionally mandated institution through which the tance, legislatures ventured slowly to build their 282 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s oversight capacity. Numerous organizations, compact between politicians as policy makers including bilateral donors, multinational organi- and civil servants as implementers of policy and zations, and international financial institutions, providers of services.With the fiscal crunch aris- supported these efforts.Their assistance ranged ing from the debt crisis of the 1980s, efforts to from supplying office and other equipment to prune and rationalize the role of the state led to helping establish legislative budget offices and privatization of state-owned enterprises. Budget strengthening committees. and financial management reforms were initi- The decade saw a trend toward legislative ated, and even challenging and controversial budget activism in developing and transition New Public Management reforms were under- countries, reflecting the process of democratiza- taken in a great number of developing coun- tion and the opening up of possibilities for leg- tries.18 islative involvement in what were previously The public sector reforms had essentially two closed budgetary systems. In Brazil, for example, thrusts.The first was to build the capacity of the where Congress had historically played no signif- public sector--personnel skills, systems, and icant role in the budget process, constitutional processes--to formulate and implement policies. changes gave Congress powers to modify the The second, whose emphasis increased during budget. In Africa, too, changes occurred. For the 1990s,was to instill clearer and more binding example, South Africa and Uganda passed finan- accountabilities in civil servants to policy makers cial administration or budget acts that give more and politicians. Underpinning both of these influence to the legislature during the budget for- trends was the move toward greater democratiza- mulation and approval processes. tion of politics,which sought to strengthen polit- The experience to date suggests that develop- ical accountability.This often took the form of a ing-country legislatures have a long way to go major decentralization of financial and functional before they can adequately perform their oversight authority to local governments,to bring govern- functions, especially over national budgets. Build- ments closer to the people. ing oversight capacity involves, among other Three caveats apply to the lessons offered things, strengthening "money committees," estab- below.First,given the slow pace at which institu- lishing dedicated research staff,enhancing comple- tions change, and the fact that most of these mentary institutions such as national audit offices, reform efforts are rather recent, it is too early to and encouraging public input at the various stages reach definitive conclusions about successes or of the budget cycle. failures.Second,very little systematic research has In assistance for democracy, it is support to evaluated results and outcomes. Third, since legislatures that most often falls short of its goals, reforms imply changes in formal and informal and legislative strengthening efforts should be rules,both of which are deeply rooted in a coun- seen as complements to related improvements in try or organization's culture and history, it is not governance. In the case of money committees, obvious that successful reform efforts in one this means dovetailing reform activities with organization can be transplanted into other cir- broader efforts to enhance government account- cumstances. ability and strengthen public financial oversight, and ensuring that training activities include par- Fiscal Management ticipants from other stakeholder organizations, Governments faced with populist pressures can such as the ministry of finance, the auditor gen- strengthen fiscal discipline by tying their own hands. eral's office,and representatives from civil society. During the 1990s, governments attempted to improve fiscal discipline in a number of ways.As noted in chapter 4, the most high-profile efforts Strengthening the Compact granted independence to central banks,as a means For more than two decades now, public sector to clarify the banks' direct responsibility for con- reforms have focused strongly on improving the trolling inflation and to grant them the flexibility I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 283 to meet this mandate. Recent research suggests erables and accountabilities of both central and that the central bank's ability to play this role line ministries.21 Part of an MTEF's value comes depends heavily on the nature and efficacy of from the discussion,debate,and agreement that it checks and balances in the larger environment generates among different parties engaged in the (Keefer and Stasavage 2003, 407). budget process--in particular the legislature-- On the public expenditure side, the experi- about the tradeoffs that need to be made among ence of the 1990s suggests that institutions define programs, activities, and projects.To be effective, and constrain the political bargaining that affects an MTEF needs to be integrated into the budget fiscal outcomes.19 Well-defined antideficit rules, process and budget documents. especially when coupled with credible limits on Among Organisation for Economic Co- government borrowing, induce smaller deficits operation and Development (OECD) countries, and more rapid adjustment of taxes and spending few have adopted the full plethora of features of to unexpected fiscal shortfalls.20 In recent years, a MTEFs, whether by choice or by circum- number of developing countries and subnational stance.22 Their adoption of multiyear features governments have passed fiscal responsibility acts evolved gradually, usually as technical improve- to strengthen fiscal management. These laws ments to the budget process.Their experience enhance transparency,as well as the accountability suggests that an MTEF may be better used as a of the executive to the legislature and the account- conceptual framework for thinking how ele- ability of both the executive and the legislature to ments of budget reform fit together, rather than the citizenry, by stimulating discussion of fiscal as a reform in and of itself. Pursuing the key fea- policies and their implications before and after tures of an MTEF individually appears to be a budget deliberations.Although no studies have yet more effective approach than wholesale adoption assessed their effects,the new laws are expected to of the full gamut of features. strengthen fiscal discipline. Largely through donor initiatives, developing Fiscal transparency is essential for maintaining fis- countries have introduced the MTEF as a process cal discipline. The Asian financial crisis highlighted technically superior to annual budgeting.InAfrica, the importance of budget transparency,since hid- implementation has been slow and MTEFs have den contingent liabilities can destroy the fabric not produced the expected results.As of 2002, 19 of fiscal discipline. Greater budget transparency countries had MTEFs at some stage of implemen- provides a basis for informed debate about budg- tation, but only five had integrated MTEFs in a etary policy among the public and the legisla- meaningful way into their budget processes. In ture, and within the executive; it also increases most cases,MTEFs have operated in parallel to the the chances that fiscal risks will be identified and general budget process (Le Houerou andTaliercio policy responses put in place.But the adoption of 2002).Only four countries had submitted MTEFs more transparent practices has often been slow, to both cabinet and parliament and,in some cases, partly because of capacity constraints and partly the MTEF has remained strictly a technical docu- for political reasons (box 9.3). ment of the ministry of finance (Le Houerou and A medium-term expenditure framework (MTEF) Taliercio 2002,13). will work only if politicians embrace it. Many coun- Just as is the budget, an MTEF is fundamen- tries adopted MTEFs during the 1990s, as a way tally political.Its effectiveness depends heavily on to increase the transparency of budgeting and the willingness of politicians to embrace it as a enhance the predictability of agency budgets. reformed budget process and to accept the disci- The primary function of an MTEF is to infuse a pline it brings. multiyear perspective into budgeting, enabling policy makers to recognize the implications of Organizational and Human Resource current budgetary decisions for future govern- Management ment finances and creating a more disciplined, Capacity constraints are binding: strategic incremental- sequenced budget process that clarifies the deliv- ism may be the only option for many developing coun- 284 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s BOX 9.3 Fiscal Transparency and Developing Countries T he International Monetary Fund introduced fiscal activities outside the budget. As of end-2003, the Code of Good Practices on Fiscal Trans- ROSCs had been completed in 63 countries; 58 of parency in 1998 as a response to the finan- these reports had been posted on the IMF Web site. cial crises of the late 1990s, and updated it in 2001. Countries electing to have a fiscal ROSC are often The Code is based on the following objectives: roles those already engaged in fiscal management reform, and responsibilities in government should be clear; yet in developing countries the reports have identi- information on government activities should be pro- fied pervasive problems of data quality (unreconciled vided to the public; budget preparation, execution, accounts, lack of clarity in accounting policies, weak and reporting should be undertaken in an open man- external audit); excessive discretion in tax adminis- ner; and fiscal information should attain widely tration and poor enforcement; unrealistic budgeting; accepted standards of data quality and be subject to weak internal controls; significant payment arrears; independent assurances of integrity. and lack of clear responsibilities at various levels of The IMF helps countries to implement the fiscal government. In a number of developing economies, transparency code through participating voluntarily the scope and extent of off-budget activities (includ- in fiscal modules of reports on the observance of ing quasi-fiscal activities) is an issue to be standards and codes (ROSCs). These assess the avail- addressed. Further, except for a few industrial coun- ability and quality of fiscal data and evaluate the fis- tries, most countries do not quantify tax expendi- cal management framework, including relations tures. Many of these problems reflect underlying between levels of government and accountability for institutional problems. Source: IMF 2003e. tries. World Development Report (WDR) 1997 recruitment and selection systems, performance argued that the state should match its role to its evaluation procedures, promotion procedures, capacity, since taking on too much makes the salary-setting rules and procedures,wage bill con- state less effective.This was certainly evident from trols, and due process protections, and ensuring the various reforms that were pursued in public that they function as planned.23The implication is expenditure management, but it is perhaps most that these core systems and practices are precondi- salient in the attempts of many developing coun- tions for effective performance management. tries to adopt New Public Management (NPM) Performance-based budgeting emerged as a approaches. corollary reform of NPM, and over the past 20 NPM reforms are a challenge even in coun- years OECD countries have gradually shifted the tries with strong capacity. In environments where focus of their public sector budgeting and man- the basics are very weak, resort to NPM-style agement from inputs to outputs. Arguably this performance management techniques has been development has enhanced public sector man- associated with poorer performance, as measured agement and increased the effectiveness and effi- by increases in administrative corruption (Ander- ciency of governments. But, as has the multiyear son, Reid, and Ryterman 2003; Schick 1998). expenditure framework, it has taken many years These authors found that the most significant fac- to mature.And the experience has debunked the tor contributing to better performing public earlier belief that performance measurement can organizations is the creation of merit-based per- often be an effective catalyst for organizational sonnel management practices: putting in place change (Schick 2003). An important lesson for I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 285 developing countries is the fact that OECD gapore have given greater autonomy to their tax countries already had basic personnel manage- administrations, although other countries in the ment systems that functioned very well, many in region are beginning to consider this.24 In gen- theWeberian tradition,before they launched into eral, an autonomous revenue agency has more performance management and budgeting. flexibility in managing personnel and finances Capacity constraints and political imperatives can and more control over corporate governance severely impede civil service reform. In many devel- than does the typical government agency.At the oping countries, establishing basic personnel same time, it also has clearer and more transpar- management systems requires overhauling the ent accountabilities. The introduction of such civil service.A malfunctioning civil service cre- agencies implied that collections would increase ates disillusionment among the citizenry, and a and service would be more taxpayer friendly. depoliticized, reasonably well-compensated, and In practice, the performance record of these skilled civil service can serve as a credible com- agencies has been mixed.25 Performance prob- mitment mechanism signaling that better public lems have resulted mainly from lack of political services are on the way (Shepherd 2003).Thus support, tensions between the autonomous rev- one might expect politicians to be interested in enue agency and the ministry of finance, and improving the internal incentive mechanisms poor organizational design,including weaknesses that affect the civil service.However,many devel- in the new accountability regime. Nonetheless, oping countries attempted civil service reforms on the whole the record suggests that, with under conditions where clientelist politics of one enough political push and proper design, these form or another were already deeply entrenched. agencies can improve tax administration and be Hence the reforms met formidable opposition sustainable (box 9.4).26 and,not surprisingly,their results have been quite Given the immense difficulty of overhauling discouraging (World Bank 2000a; Levy and the whole of government, for many developing Kpundeh 2004). A key part of the difficulty is countries enclave reforms may be the only game that such efforts must inevitably transfer some in town.The important lesson is that enclaving authority from the political echelons (legislators, must be strategic if it is not to constrain and/or ministers, and their political appointees) to a distort the capacity-building efforts of govern- cadre of depoliticized officials (civil servants) ment (box 9.5).27 Part of the reason why (Manning and Parison 2003). In many develop- autonomous revenue agencies have been mod- ing countries, the political history and environ- estly successful in some countries is that revenue ment make this a gargantuan task. collections dictate government budget Enclaving is a potential path to sustained reforms envelopes. of the civil service. The challenges of politics and In Africa, narrowly focused strategic inter- capacity constraints have led some countries to ventions tended to be more successful than experiment with enclave approaches to civil ser- broader reforms over the long run, and among vice reform, spinning off selected government the more narrowly focused reforms, the capac- entities from central government ministries. ity-building initiatives that focused on improving Increased autonomy for revenue collection expenditure accountability were considerably agencies became a key feature of governance more successful than those that focused on reforms in Latin America in the early 1990s, human resource management,according to Eng- starting with Peru in 1991, followed by berg-Pedersen and Levy (2004).These authors República Bolivariana de Venezuela, Mexico, conclude that the expenditure accountability Bolivia, Guatemala, Argentina, and Colombia. reforms, given their more technical nature, may Such reforms became popular in Africa in the have been less threatening to vested interests than late 1990s, with 12 countries experimenting the more politically sensitive administrative with revenue boards or semi-autonomous agen- reforms, and, perhaps more important, that cies. In Asia, by contrast, only Malaysia and Sin- expenditure reforms produce more readily 286 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s BOX 9.4 BOX 9.5 SUNAT in Peru: A Modest Success The Enclave Conundrum in Uganda I T n the late 1980s, Peru decided to set up a semi- he Government of Uganda now autonomous revenue authority (SUNAT), since tax has to decide the fate of more administration was riddled with corruption and on than 70 enclave units, most of the verge of collapse, with revenue collections dropping which are outgrown project implementa- to a record low of 9 percent of GDP in 1988. A compre- tion units and semiautonomous organiza- hensive staff screening and replacement program was ini- tions called secretariats. These were set tiated, and less than one-third of the original tax up at the behest, or with support, of administration staff was rehired by the new organization. donors who doubted line ministries' A modern human resource management system was intro- capacity and were dissatisfied with min- duced and SUNAT was allowed to operate under private istries' procurement practices. The enclave sector labor laws and without undue political interfer- units were allowed to set their own wage ence. Salaries were adjusted to private sector levels. A rates, above ministry levels, based on the new financing mechanism provided financial stability. By premise that the necessary skills would 1997 tax revenues had recovered to 14 percent of GDP. not be available at the public service's low After its successful start, however, SUNAT suffered salary levels. from decreasing political support for efficient revenue Ironically, with these higher salaries, collection; a decline in the quality of the tax policy the enclaved units recruited substantial framework, which made fair and efficient tax collection numbers of skilled personnel from within more difficult; and increased interference by the Ministry the public service, depleting the capacity of Finance. As a consequence, SUNAT has suffered a loss of ministries and demoralizing competent of standing in public opinion. Despite these problems, ministry staff. however, its creation has permitted the launch of far- Enclaving has also undermined the reaching efficiency- and integrity-enhancing reforms, budget process. First, although budgeted which are having an ongoing impact. Revenue collection, under development expenditures, two- at 12 percent of GDP in 2001, remains far above the pre- thirds (71 percent) of the enclave units' reform level. expenditure is of a recurrent nature. These recurrent expenditures come from the government's 10 percent contribution to donor-funded projects. While this may observable results that generate the general pub- seem a nominal proportion of any pro- lic support needed to counter vested interests. ject's total budget, the total constitutes Values, commitment, and pride in public service nearly one-third (28 percent) of the gov- matter as much as controls and compliance. NPM ernment's own generated funds. This sig- reforms sought to introduce stronger market- nificant share is spent without the rigor based incentives as a means of reforming govern- that is associated with allocating other ment bureaucracies. Emulating the experiences government expenditures. of developed countries such as New Zealand and the United Kingdom, developing-country gov- Source: Draft report, "Affording Uganda's Pub- ernments adopted performance management lic Administration Sector" (December 2003), techniques that grew out of reforms in the pri- World Bank. The report is currently being vate corporate sector and sought to enhance the reviewed by the Government of Uganda. autonomy and accountability of public sector managers and staff. Such reforms have arguably led to improved service and performance in I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 287 developed countries, but have had little success The effectiveness of information and communica- in developing countries. tions technologies (ICT) depends on the reengineering Industrial countries'experience with industrial of underlying processes and the proactive use of change performance and workplace transformation shows management. To increase transparency and effi- that workers'dedication to the job is an important ciency in the delivery of front-line services,many explanation for improvements in performance countries in the 1990s began adopting ICT in (Tendler 1997). Recognition of this relationship such varied areas as tax collection, customs valu- has caused firms that perform well to pay close ation, procurement, treasury and cash manage- attention to innovative practices that increase ment, issuance of licenses, land registration and worker dedication.Tendler contrasts this with the titling,passport issuance,and other focused front- development literature, which has been rife with line public services.This trend, which has come suspicion that "civil servants are self-interested, to be known as "e-government," involved signif- rent-seeking, and venal, unless proven otherwise." icant experimentation with the application of Her research in Ceará, Brazil, demonstrates that ICT to internal processes (box 9.7).28 the creation of a sense of calling and ownership In a wide range of countries,e-government has around public service by a committed leadership, been a powerful tool for enhancing the efficiency a dedicated work force, and an informed and and effectiveness of public services (table 9.1).29 engaged civil society can increase acceptance of Experience nonetheless suggests that e-gov- reform and improve service delivery (box 9.6). ernment is not a panacea for basic problems.First, BOX 9.6 Building a Sense of Calling and Commitment in Public Service Delivery: Ceará, Brazil C eará has substantially improved its preven- servants by regularly announcing successes and tive health services, as reflected in indica- openly rewarding good performance, building a sense tors such as infant mortality and vaccination of calling and pride around the workers and creating coverage. Agriculture programs have raised output and a sense of chosen elite in the public service. Public- productivity. Spillover effects from procurement reform ity also increased citizen awareness and public mon- programs have resulted in local economic develop- itoring of civil servants and local governments, and ment, in addition to increasing output and productiv- created a new constituency that would help leaders ity for small-firm suppliers. Public works construction and agencies overcome political opposition. has created more jobs in the economy than usual, Third, innovative organization of tasks for work- resulting in a greater share of public expenditure being ers in the public programs resulted in often voluntary allocated to labor. ownership of varied and multiple tasks. These multi- Four closely intertwined explanations can be ple tasks often coalesced into client-centric, prob- offered for this improved performance: lem-solving approaches to service delivery. First, civil servants in these programs showed And fourth, repeated messages from the govern- high dedication to their jobs. Either they were dedi- ment and reorganization of tasks kept rent-seeking cated entrants into civil service, and work conditions behavior under control by creating a sense of pride, perpetuated that dedication, or the circumstances of ownership, and recognition around public service, in the jobs elicited their commitment. the eyes of society and civil servants alike. Second, the government repeatedly and publicly demonstrated admiration and respect for the civil Source: Tendler 1997, 135­65. 288 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s BOX 9.7 Mexico's e-SAT Program for More Efficient Tax Administration I nspired by the early successes of Argentina, taxpayers to (1) obtain a personal form of electronic Brazil, and Chile in using ICT to improve the identification, (2) obtain a corporate tax ID for new delivery of government services, the Mexican gov- entities via the Internet, (3) submit a tax declaration ernment (as part of its OECD-based program of regu- and other relevant forms, (4) enquire into the status latory and administrative reform), began moving of a taxpayer account, and (5) schedule an appoint- government information and services online. Mexico's ment with a SAT tax counselor. In August 2002, SAT federal tax administration (Servicio de Administración promulgated a new regulation requiring the electronic Tributaria, or SAT) embarked on e-SAT, a program to submission of most individual and corporate taxpayer offer tax services online. Starting in 1995, e-SAT has declarations (through SAT's portal) and payments evolved gradually. By 1998, SAT had established an (through the portal of the taxpayer's bank). SAT is interactive Web page that contained basic informa- now working to develop and make operational online tion on tax laws and procedures and permitted tax- systems for the receipt and processing of credit card payers to file their annual declarations electronically. payments and payment of tax refunds. e-SAT has The main beneficiaries of the first phase of this new greatly reduced the amount of paperwork previously system were large corporations. During 2000­03, SAT managed by Mexico's tax registry, declaration, and expanded the nucleus of online tax services to allow collection units. Source: Kossick 2003. for real gains to be made, processes need to be is moving to a paperless office and electronic simplified and automated in ways that reduce the workflow. In the Republic of Korea, the success discretion of government officials. For example, a of the OPEN project has been largely due to the study of the effectiveness of efforts to introduce support of the mayor of Seoul, who took a mul- computerized integrated financial management tipronged approach to curbing corruption.31 In systems in Africa concludes that "technology can Gujarat,India,with strong support from the lead- only add value in the context of an underlying ership,computerized check posts trebled the fines commitment to disciplined decision making, and collected from overloaded trucks, but when this internal management systems geared to monitor- support waned, the verve to reduce corruption ing compliance" (Dorotinsky and Floyd 2004). weakened,and new forms of extortion hampered Second, most e-government projects have faced the overall effectiveness of the program. substantial resistance from public servants who tend to see such projects as a threat to their jobs; External Restraints publication and easier access to information dilute To combat corruption, prevention may be more effective their control and diminish their responsibility as than investigation and prosecution. With the spiral- information brokers. Hence, reform projects can ing concern over corruption in the public sector, stall unless adequate change management developing countries began to establish institu- processes are adopted.30 Third, political support tions of restraint that focused mainly on enhanc- from top leadership is critical.In Andhra Pradesh, ing internal accountability: supreme audit India,for example,top leadership publicly pushed agencies, the ombudsman function, and anticor- for e-government and allocated resources accord- ruption agencies. ingly for an ambitious program. Several services Among these, anticorruption agencies gained are now delivered online and the state secretariat most currency, in countries from Latin America I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 289 TABLE 9.1 Examples of Efficiency Gains from ICT Country, region Type of government application Time to process before application Time to process after application Brazil Registration of 29 documents Several days 20­30 minutes per document, one day for business licenses Chile Taxes online 25 days 12 hours China Online application for 32 2­3 months for business license 10­15 days for business license business services Several visits to multiple offices Several seconds for routine filing for filings for companies India, Valuation of property Few days 10 minutes Andhra Pradesh India, Andhra Pradesh Land registration 7­15 days 5 minutes India, Karnataka Updating land registration 1­2 years 30 days for approval, request completed on demand India, Karnataka Obtaining land title certificate 3­30 days 5­30 minutes India, Gujarat Interstate check posts for trucks 30 minutes 2 minutes India, Statutory certificates on caste 20­30 days 15 minutes Andhra Pradesh Jamaica Customs online 2­3 days for brokers to process entry 3­4 hours Philippines Customs online 8 days to release cargo 4 hours ­ 2 days to release cargo Singapore Issue of tax assessments 12­18 months 3­5 months Source: Bhatnagar and Deane 2003. to Africa.These agencies have had mixed results. illicit enrichment. For instance, reports from the The successes of Singapore's CPIB and Hong auditor general are submitted to the president, Kong's ICAC are widely known. More often, who then determines when and how to release however, such agencies have been seriously the information.Too often, the information is impeded by resource constraints, weaknesses in not released. complementary institutions such as the judiciary Focusing on cure rather than prevention of and the police, and multiple goals (Meagher corruption has rarely worked in developing- 2002). Some have been set up to satisfy a politi- country contexts, at least not in the short to cal need--such as an outcry from a corruption medium term. As the 1990s ended, it became scandal or loan conditionality--but lack enough clear that the gestation period for such reforms resources or political backing to actually do their would be much longer than had been expected work.They have served to deflect demands for and that, over the medium term, better results action against corruption while the authorities could probably be obtained from preventive fail to undertake any real responses (Heilbrunn measures that achieve the following: 2002). Box 9.8 outlines the demanding require- ments for an effective anticorruption agency. · Reduce the likely benefits from corruption. Pro- Supreme audit agencies are crucial for enforc- moting competition in the private sector, ing the financial accountability of the govern- through lowering barriers to entry and ment, and during the 1990s such agencies were reforming regulations where there are natural established or restructured in many developing monopolies, serves to reduce rents and rent countries. Their effectiveness has been highly seeking. Ensuring competition in procure- variable (Heilbrunn 2002). In many countries, ment through nationwide advertising and the auditor general's independence is compro- efforts to prevent collusion can also greatly mised by an executive that seeks to prevent the reduce corruption. More radical measures opposition from learning of possible cases of include calling for a referendum for the pub- 290 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s BOX 9.8 What It Takes to Create a Successful Anticorruption Agency E stablishment: Carefully situate the agency ers from the agency's placement and line of respon- within a set of well-defined supports: a com- sibility, the appointment and removal procedures for prehensive anticorruption strategy, careful top officials, or some forms of fiscal autonomy. The planning and performance measurement, realistic most important sign of independence is the absence expectations, and strong enough political backing of political intrusion into the agency's operations. (across class and party lines) to make it effective De facto autonomy enables anticorruption agencies regardless of political and personal consequences. to operate on a consistent and professional basis Agencies that score high on these measures are those with relatively little partisan intrusion. In most in Hong Kong (China), Singapore, Malaysia, Uganda, environments, this mode of operation is important and Australia's New South Wales. to success. Focus: Define the agency's focus in a way that will Powers: A successful anticorruption agency will maximize its effectiveness. For example, an agency have strong research and prevention capabilities, could focus on prevention and monitoring govern- along with the authority to do the following: access ment implementation of anticorruption policy, forgo- documents and witnesses, freeze assets and seize ing a comprehensive mandate (as in Korea); or its passports, protect informants, monitor income and jurisdiction could be mainly prospective, with only assets, propose administrative and legislative limited concern with past cases (as in Hong Kong, reforms, and exercise jurisdiction over the head of China); or it could choose cases selectively, based on state. Many agencies have most or all of these pow- clear standards (as in Argentina and New South ers on paper but frequently cannot put them into Wales); or it could deal only with the probity and effect owing to lack of coordination, weak capacity reputation of the public service (as in the United in cooperating institutions, and political factors. States and India). Clarity of focus seems to be con- Resources: Agencies in this field, as in others, sistently associated with success, except where mas- depend on well-trained personnel, including suffi- sive resources are available. cient numbers with highly specialized skills. Staff Accountability: Promote the agency's accountability should also be well compensated, subject to integrity through such factors as the application of legal stan- reviews and quick removal, and endowed with a dards, the availability of judicial review, systems of strong ethic of professionalism, integrity, and high public complaint and oversight, a requirement that the morale. Also important are sufficient funds, adequate agency answer to all branches of government and the facilities and assets, high-level information sharing, public, and precise and comprehensive expenditure and coordination with other government bodies. accountability. Some commentators also suggest mini- Complementary institutions: Anticorruption agen- mizing the agency's size, as well as the "free" support cies do not succeed without the basic features of the given by aid donors. Accountability is not uniformly rule of law: functioning courts, free and active media, associated with success. Many successful anticorrup- an active community of nongovernmental organiza- tion agencies are strongly accountable, but this is prob- tions (NGOs) and public interest groups, and other ably an outgrowth of the rule of law, which seems to be capable institutions such as supreme audit agencies more consistently associated with success (see below). and central banks. Civic factors such as free media Independence: Independence arises in some and capable nongovernmental watchdogs are not as cases simply from outside accountability, and in oth- clearly associated with their success. Source: Meagher 2002. I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 291 lic to approve large projects and the tax and agency) is himself or herself not corrupt,and that expenditure allocation choices these imply. he or she has an interest in ensuring that his or her agents are not corrupt. But most principals, · Reduce the number of transactions that create oppor- that is, politicians, even if they themselves are tunities for graft--for example, by liberalizing honest, need money to stay in power, fight elec- imports, removing price controls, removing tion battles, or buy off opposition.Those who industrial and trade licensing requirements, or contribute such money can then influence the making such licensing automatic. For exam- way power is exercised. If politicians are ple, when India liberalized industrial licenses beholden to special interests or are captured by in the early 1990s, a large industry aimed at various interest groups, fighting corruption obtaining licenses disappeared, along with the becomes rather difficult (see the discussion in corruption associated with it. Similarly, when section 3 on reform strategies in captured states). Indonesia liberalized its trade regime in the 1980s it radically reduced corruption in import licensing.Streamlining bureaucracy by NurturingVoice and Client Power reducing the number of approvals required for By the mid-1990s, it was clearly seen that civil particular transactions, reducing bureaucratic society organizations--such as NGOs and reli- instructions, simplifying rules, improving ser- gious organizations--could play an important vice standards, and decentralizing services all role in inducing better performance from gov- have this effect. ernment. In many developing countries, the · Increase information, transparency, and public over- introduction of elections opened the way for cit- sight. Corruption often occurs because of lack izens to hold politicians accountable for the per- of information. Governments lack informa- formance of the public sector. Homegrown tion on what their agents are doing on the experiments sprouted throughout the develop- ground. Consumers of government services ing world, showing that active civil society par- are not aware of what rules or charges are ticipation in reform can potentially lead to much legitimate. Clarifying rules and increasing needed improvements in the compact.32 For transparency helps to reduce opportunities for example: corruption, as does involving beneficiaries in · Citizen report card surveys, which originated the oversight of government programs. in Bangalore (box 9.9), have spread to other Indonesia's Kecamatan Development Pro- parts of India and are now being tried in other gram is a good example of this (World Bank developing countries including Peru and the 2003c). Philippines. This instrument has channeled · Establish time-tested, basic systems of personnel the collective voice of citizens seeking better management. Merit-based recruitment systems public services. and long-term career path arrangements-- · Different forms of participatory budgeting essential attributes of a Weberian system-- have emerged in parts of the developing significantly reduce bureaucratic corruption world including Brazil and South Africa. At (Evans and Rauch 2000), and there is evi- its core, participatory budgeting engages citi- dence that smaller pay differentials between zens and their elected representatives, such as the public and private sectors lead to lower the mayor or governor, in a partnership to levels of bureaucratic corruption (Van determine the priorities of the community Rijkeghem and Weder 2001). and what projects will be funded from the Much current thinking about fighting cor- local budget during the coming fiscal year. ruption is influenced by the principal-agent Typically, this process has been launched by problem.The above lessons assume that the prin- forward-looking local politicians as a means cipal (the politician or the head of a government of locking future politicians into a transparent 292 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s BOX 9.9 The Report Card Survey in Bangalore, India: Stimulating Administrative Reforms T he 10-year experience of the Public Affairs organizations. Citizens showed themselves generally Centre (PAC) in Bangalore, India, illustrates dissatisfied with the delivery of public services. The the potential of using client surveys as a results of the survey were published in a leading lever to induce upstream public management reforms newspaper, raising their visibility and leading public and an improvement in service delivery. Established officials in a number of agencies to discuss their in 1994 to improve public sector governance in India, agencies' problems with PAC and citizens' groups. PAC PAC's primary focus and strength lies in assisting cit- offered to help these agencies address some of the izen groups in "using knowledge as a basis for collec- problems. Further surveys, undertaken in 1999 and tive action" (Paul 1995). In 1994, the PAC conducted 2003, show that over 10 years the public has become its first report card survey, effectively an opinion poll, much more satisfied with the service delivery across of citizens in Bangalore on their perceptions of the all eight agencies. The lesson: demand pressures can quality of services provided by eight key government lead to needed public sector reforms. Citizen satisfaction with various agencies across three report cards 100 94 92 96 85 78 80 73 73 67 60 41 41 42 32 32 satisfied 40 19 16 16 % 20 10 13 0 2 0 0 BMP BESCOM BWSSB BSNL POLICE BDA BMTC Agencies 1994 1999 2003 Source: Paul 2003. budget process for which they can be held age citizen. Changes in quality, quantity, and accountable. access to services affect everyday lives and thus make citizens more prone to support, if not to · NGO advocacy for greater transparency and seek, reforms in governance. fairness in public procurement and monitor- Interventions to improve service delivery are ing of procurement processes has become a potential entry point for broader governance more common as well. So-called integrity reforms (box 9.10). In the Middle East and pacts developed by Transparency Interna- North Africa, there are few democracies and the tional have now been used in several coun- governance gap is significantly wider than in tries as part of huge procurement contracts, other developing regions, largely as the result of for example, in Colombia and Mexico.33 weaknesses in public accountability (World Bank Most demand-driven reforms have taken 2003c). Though there appears to be much less place in the delivery of public services, because governance reform activity than in other regions, service delivery is where the state meets the aver- civil society efforts to address poor service deliv- I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 293 BOX 9.10 BOX 9.11 Service Delivery and Civil Society in Procurement Watch: Working with the West Bank and Gaza the "Enemy" A I 1998 review of service delivery in n the Philippines, Procurement Watch the West Bank and Gaza asked a Inc (PWI) was born in part out of the representative sample of benefici- need to harness citizens' collective aries about the provision of health and anger and frustration about widespread education services by NGOs, the private corruption. It was organized as a nonprofit sector, and the Palestinian Authority. institution with encouragement from the Meanwhile, specialists carried out institu- government Department of Budget and tional reviews of education and health Management (DBM) and seed funding from facilities. The findings show that benefici- the World Bank's Asian Emergency Trust aries often select a particular provider Fund (ASEM) grant facility, along with pri- because its service is better or easier to vate contributions. To establish its credi- access--areas in which NGOs and the pri- bility with citizen groups and legislators, vate providers ranked higher than the gov- PWI quickly developed technical expertise ernment. The findings were presented at in the area of procurement. With well- dissemination workshops that were trained and experienced staff and an attended by ministers, senior civil ser- active board of directors, it drew attention vants, and senior representatives from from the media to the need for a procure- NGOs and private organizations. The minis- ment reform bill. It secured advocacy sup- ters of health and education reacted by port from different citizen groups using the findings to improve the quality throughout the country, was invited to of health and education services across the participate actively in the three technical board and to improve coordination among working groups formed to formulate the the government, the NGOs, and the private law--one in the executive, another in the sector. House of Representatives, and the third in the Senate--and was sought after for Source: World Bank 1999. advice by proponents of the bill in both the House and the Senate. With technical assistance from donors and advocacy sup- port coordinated and managed by PWI, DBM secured the enactment of the Govern- ery appear to be sprouting at the community ment Procurement Reform Act of 2002. The level.All regimes care about their legitimacy, and President signed the bill into law on Janu- one way that nondemocratic regimes can main- ary 10, 2003, almost four years after the tain their legitimacy is by providing adequate effort was initiated. public services.34 To do this, they need informa- tion about service delivery problems. Civil soci- ety organization efforts such as report card surveys provide such information. If demand-raising efforts are to be effective, and understanding needed to move a reform public officials and civil society groups must be agenda within the government. Likewise, willing to work closely together (box 9.11). reform-minded politicians and civil servants Civil society organizations need partners inside need civil society organizations to galvanize the public sector, since they have the knowledge public support. 294 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s Decentralization mitted to responding to local needs, the intended services may not be forthcoming Many developing countries embarked on without appropriate inputs by the bureaucracy wide-ranging decentralization efforts in the (Schroeder 2002). 1990s (WDR 2000; Litvack, Ahmad, and Bird A local government should be just as 1998; Burki, Perry, and Dillinger 1999b; Ebel accountable for funds transferred to it as for funds and Yilmaz 2002). In developing countries on collected directly from local taxpayers. Instru- average, the share of subnational governments ments for ensuring financial accountability are in total government spending increased by 20 often in place, but their implementation is often percent from the 1980s to the 1990s.35 poor.Financial audits of local accounts,for exam- Although some of the decentralization efforts ple, tend to be delayed for long periods, and as a may have been indirectly stimulated by fiscal result, they have not become effective instru- crisis, as in Indonesia, most were inspired by ments of accountability. Experience has varied changes in the political landscape: the collapse widely across countries, but emphasizes the of long-standing highly centralized regimes importance of imposing credible, hard budget and the emergence of strong global pressures constraints on local governments (Rodden,Eske- for democratization.In many countries,includ- land, and Litvack. 2003). Otherwise local gov- ing Thailand, pent-up distaste among the citi- ernments may borrow recklessly to fund local zenry for tight, unchecked central control and initiatives or find other ways to transfer liabilities a greater desire to hold political leaders and potentially expose the national government accountable provided the impetus. to unwanted fiscal risks. Decentralization is a political choice, whose design Throughout the 1990s, decentralization and implementation may not improve service delivery. efforts featured intergovernmental fiscal reforms. Designed well, decentralization can move deci- Argentina's experience highlights the challenges sion making closer to the people, enhance the in the design and evolution of a good intergov- efficiency and responsiveness of service delivery ernmental fiscal system (box 9.12). (Faguet 1997; Kahkonen and Lanyi 2001; Bard- The administrative aspects of decentralization are han and Mookherjee 2000), support economic as important as the fiscal aspects. The details of growth, and offer a potentially powerful tool for implementation arrangements ultimately deter- alleviating poverty. But designed inappropriately, mine outcomes. The challenges that have or introduced without strong local participation emerged fall into three categories: (1) adoption and accountability (of local officials to local citi- of a more systematic view of decentralization zens), it can lead to macroeconomic instability, (Bahl 2000); (2) balancing of responsibility with declining service levels (Martinez-Vazquez and resources, capacity, and accountability; and (3) Boex 2001), heightened regional disparities or creating incentives for implementation to match conflicts (Smoke 2001), and increased corrup- formal decentralization arrangements. Experi- tion (Brueckner 1999). ence shows the value of pragmatism in imple- Because the ultimate objective of devolved menting decentralization (Litvack, Ahmad, and arrangements is to provide needed public ser- Bird 1998; Bahl 2000; Bahl and Smoke 2003).A vices in an equitable and technically efficient pragmatic strategy would be unique to each manner, the most critical concern is the politi- country undertaking decentralization, but cal accountability of locally elected officials to should include a general vision and framework local residents. Political representation appears for reform, mechanisms for coordination and to be insufficient,however,reflecting capture by resolving conflicts, a prioritization of reforms elites and the weaknesses of local political and a plan for sequencing them, information to processes (Crook and Manor 1998; Conning monitor outcomes and adjust the reform pro- and Kevane 2001). Experience also shows that gram, and incentives to change central and local even if local elected officials are strongly com- behavior. I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 295 BOX 9.12 Hard Budget Constraints: The Challenge of Fiscal Decentralization in Argentina P rovinces in Argentina depend on federal trans- scores that central commitment problems and limita- fers for the bulk of their fiscal resources. Rev- tions in effective regulation can conspire to weaken enue sharing (coparticipaciones) was subnational hard budget constraints. Strong efforts introduced in the 1930s to compensate provinces for by the central government to regulate can exacerbate the introduction of national income and sales taxes, problems, as, for example, when subnational govern- and in 2000, 56 percent of provincial resources came ments circumvent the spirit of regulations through from this common pool. The majority of provinces recourse to affiliated state-owned enterprises, (roughly 60 percent) relied on their own resources for including regional banks. Argentina's experience less than 30 percent of their spending. shows that such problems have no quick fixes. They The challenge of ensuring hard budget constraints require the evolution of credible policies--and polit- under fiscal decentralization depends on a country's ical will--buttressed by effective institutional social, cultural, and institutional features. Experi- arrangements, whether predominantly market based ence in countries such as Argentina and Brazil under- or hierarchical and dependent on central oversight. Sources: Eaton 2003; Perry and Servén 2002; Rodden, Eskeland, and Litvack 2003; Saiegh and Tommasi 1999; Tom- masi 2002; Webb 2003. Information and Media Help Build The power of good research that feeds into Citizen Power the media should not be underestimated (box 9.13). Experience in Uganda illustrates how The media can be a powerful instrument for gal- access to information can galvanize civic action. vanizing citizen action.A free press raises popular The Ministry of Finance launched a public awareness of inappropriate actions by elected and expenditure tracking survey to monitor the flow appointed officials.When people are aware of of funds from the budget for per capita education corrupt officials or networks, their reaction may grants to local school districts.The survey showed include voting against incumbents, protests, and that, in 1994, the local districts received on aver- manifestations of unrest, or disengagement from age only 13 percent of the funds due them. the formal economy.36 Widespread press cover- Alarmed by the huge leakage, the ministry age was notable in the deposition of former launched a nationwide awareness program that Brazilian President Collor in the early 1990s, informed communities of the funds that were which brought citizens to the streets to protest due their respective districts, thus giving com- (Stapenhurst 2000). It was also instrumental in munities a benchmark for monitoring the flow building the people power movement in the of funds.As a result, the leakage has fallen to less Philippines to depose then-President Estrada in than 20 percent of the budgeted funds.37 2001,and widespread protests in Ecuador to oust Laws on the right to information empower then-President Bucharam in the late 1990s. NGOs, business organizations, and civil society Media pressure (more precisely,from progres- more generally. For example, a grassroots cam- sive elements),can create the impetus for reform. paign in Rajasthan, India, led by a local non- Just as brave reporting brought down the Nixon governmental organization, MKSS, used administration in the United States, so too it can information gleaned from government files to shake the foundations of corruption in develop- expose and then combat massive corruption at ing countries. the local level, showing how public officials 296 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s BOX 9.13 Investigative Journalism: Lifestyle Checks of Public Officials I n 2003, over an intense period of six months, a yet he earns less than P300,000 (approximately team of researchers from the Philippines Center $5,400) a year. Parked in his garage on the day PCIJ for Investigative Journalism (PCIJ) conducted a visited were a Ford Expedition, a Toyota Land lifestyle check on personnel from the Bureau of Inter- Cruiser, and a brand new BMW." nal Revenue (BIR) and published their findings in a The report also uncovered interesting schemes, three-part report, "BIR Officials Amass Unexplained such as BIR officials petitioning the Civil Service Wealth," which later was picked up by the daily news- Commission to change their birth records so that they papers. Throughout the years, survey after survey has could delay retirement and hang on to their lucrative indicated that the Bureau is one of the most corrupt postings. The PCIJ team discovered 24 such applica- government agencies. tions from the BIR between 1989 and 2001. The research produced a wealth of information on Since the publication of the report, one senior the lavish lifestyles of BIR officials and employees, official has resigned and several have been sus- ranging from grand houses in highly exclusive neigh- pended pending investigation by the ombudsman. It borhoods to expensive luxury vehicles, despite mod- remains to be seen whether any of the officials under est official incomes. The research covered 25 investigation will be indicted and convicted of cor- officials at various levels and found that many of ruption, as many challenges still confront the legal them could not explain how they acquired their and judicial system. But lifestyle checks have now assets, including shares in businesses and compa- been added to the arsenal of the anticorruption agen- nies. As the report stated, "One regional director, for cies, making it more difficult for public officials to example, lives in a big house in posh Ayala Alabang, enjoy illicit wealth. Sources: Bacalla 2003; Porcalla 2004; Nocum 2004. skimmed money off the wages of workers and results of these approaches were major changes paid friendly contractors for work never done.A in formal rules: new or amended constitutions, recently passed freedom of information law new legislation, ostensibly independent courts made it possible for MKSS to conduct this vig- and audit institutions, and so forth. Such orous and successful campaign.38 changes are not unimportant. But in practice they rarely shape behavior unless there is an equal commitment to better aligning informal 3. Strategy and Implementation: rules to improve the incentives that face politi- The Challenge for Gover- cians, bureaucrats, and citizens (Burki and Perry nance Reforms 1998).39 Ad hoc incrementalism has also been prob- Governance reform strategies in the 1990s typ- lematic.With few exceptions,the ad hoc reforms ically fell into two broad categories:"big bang" were often symbolic, intended to preserve the or ad hoc incrementalism. Big bang approaches old rules and informality while pretending to proved to be largely inconsistent with capacity reform. In some cases they represented well- constraints and political realities--in Hirsch- motivated attempts of individual or small groups man's words, countries with the wherewithal to of reformers that, for lack of support, were carry out a coordinated big push "would not be undermined by jealousy, intrigue, or fatigue. underdeveloped in the first place." The main More important, they tended to be unrelated to I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 297 a more coherent reform strategy and thus over time many lost their steam. FIGURE 9.4 What may be needed instead is highly Classification of States by Governance Profile focused, pragmatic interventions that may be Medium High 45 termed "strategic incrementalism."These inter- 40 ventions are opportunistic because they exploit MoldovaAzerbaijan High 35 the willingness to reform, but they are better Latvia Russian Fed. Ukraine grounded in political realities and consistent with 30 Bulgaria index Croatia Kyrgyz Rep. the capacity constraints of the country con- 25 Slovak Rep. Georgia Romania cerned. Knowing what is appropriate in which 20 capture country situation is often half the battle.Though 15 Albania Poland Czech Kazakhstan Medium providing a detailed road map to guide strategy is 10 Estonia State Rep. Lithuania Slovenia Armenia a task requiring fundamentally new research and 5 Hungary analysis, the following discussion suggests a pos- 0 0 1 2 3 4 5 6 sible approach to governance reform strategies in Administrative corruption index developing countries. A recent survey of firms conducted by the Source: World Bank Business Environment and Enterprise Performance Survey (BEEPS), 1999 and 2003. World Bank in Eastern Europe and Central Asia provided information that can be used to array the countries of that region along a two-dimen- and support that coalesces around the reform sional matrix, with an administrative corruption objective. index on one axis and a state capture index on Weak: Weak states lack many of the basic the other.40 Since administrative corruption structures needed to manage the public sector. reflects the quality of the compact and state cap- Many have only recently emerged from conflict ture affects the strength of political accountabil- or attained statehood. Bureaucratic capacity and ity,the quality or state of governance in a country accountability are weak, and administrative cor- can be broadly characterized by these two ruption is high. Often weak states have largely indexes.41 The matrix in figure 9.4 suggests a escaped capture by business interests,not because classification of countries into four possible accountability mechanisms are effective, but types: capable, weak, captured, and restrained. because the state is itself insufficiently developed Each type faces different challenges and different to be captured. In fact, as these basic structures opportunities for reform. are established, the risks of state capture quickly Capable: In capable states, administrative cor- increase. Examples of such states are Albania (in ruption tends to be low and state capture not 1999) and Armenia (in 1999). Nepal and Tanza- heavily entrenched. Examples are Korea, Chile, nia may also be examples, as may many low- Hungary, and the Czech Republic.To a lesser income countries under stress.42 In weak states, extent, Botswana and the Indian states of Andhra the primary challenge is to ensure that taxes are Pradesh and Karnataka may fall into this cate- collected, key services are delivered, and budget gory. In capable states, the challenge is usually to execution is sufficiently controlled. Given lim- increase the quality and efficiency of public ser- ited bureaucratic capacity, it is especially impor- vices,so as to best utilize limited public resources. tant that reform efforts be targeted and that Episodic scandals, reported by vigilant media or international support for these reforms be highly civil society organizations, usually result in pub- coordinated. lic dialogue and ultimately in a set of actions to Captured:These states have serious problems reduce opportunities for corruption. In these of administrative corruption and their environ- countries it is often possible to undertake diffi- ment makes them highly subject to capture. cult systemic reforms using a more or less tech- Many have an urgent need to build capacity in nocratic approach, providing there is leadership the public sector, but investments in capacity are 298 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s unlikely to produce sustainable improvements, in support of reform, and sufficient capacity to because political corruption (grounded in rents) carry the reform process forward. Clearly, these permeates the system at all levels. Examples of conditions exist in only a minority of developing such states include Azerbaijan (in 1999), the countries and rarely in those countries in most Russian Federation (in 1999), and the Kyrgyz urgent need of governance reform.This is the Republic. States that are weak but resource-rich crux of the challenge for the decade ahead. or dominated by a few valuable industries can fall Meeting the challenge requires a good under- easily into this category.The challenge in these standing of the political dimensions of reform, states is to break the stranglehold of special inter- and, in particular, of how reform can be used to ests--for example, by breaking up powerful identify and build constituencies that are capable monopolies if capture is by private interests, or of sustaining the reform momentum. This by reducing military expenditures if capture is by requires fundamental changes in current meth- the military. Not surprisingly, these types of ods of analysis. In this context, a focus on "driv- reforms are the least likely to be adopted while ers of change" is promising (Duncan 2003). vested interests remain strong. While the particular drivers will naturally vary Restrained: The bureaucracy in these states from country to country, the common thread of tends to have sufficient capacity and accountabil- this approach is a focus on solving the specific, ity so that administrative corruption is relatively highly salient problems facing individual com- mild.Political accountability is likely the weakest munities--for example,in health care,sanitation, link in the chain, which results in a high level of or business regulation. These are problems state capture. Examples of such states are Croatia around which constituencies for reform both (in 1999), the Slovak Republic (in 1999), Serbia inside and outside government may be easier to and Montenegro (in 1999), Latvia, and possibly build and maintain than, say, upstream reforms in Argentina, the Philippines, and some states in civil service reform or financial management. India. Reform options are limited in such states Almost all successful reform efforts have been while the existing leadership is well entrenched. shepherded through by dynamic leaders (World When a genuine change in leadership occurs, as Bank, World Development Report 1997, 154). In in Croatia, Latvia, Serbia and Montenegro, and many countries, the drivers of change may be a the Slovak Republic, and where civil society is group of young, perhaps inexperienced, leaders relatively robust and can play an important role in need of training and support. in stimulating demand for change, reforms can The challenge is creating and nurturing an emerge fairly fast and can potentially be sus- environment that encourages dynamic, forward- tained. looking individuals to push much needed reforms. Table 9.2 highlights the potential entry points In some cases,this may be achieved through polit- for strategic interventions in each of the four ical decentralization and economic liberalization, types of states. "Breaking through" a captured where the former gives local public officials state may be the most difficult strategic challenge autonomy over their localities and the latter cre- in governance reform. ates pressures for competition among localities. In other cases, the pressure for reform may come from outside.Civil society groups,media,business 4. Conclusion associations,and/or religious organizations are fer- tile sources of change.While these groups are often Improvements in governance are critical to plagued by collective action problems,experience ensuring sustainable development. Perhaps the shows that they are fully capable of overcoming most important lesson of the 1990s is that tech- these problems.The challenge they face is one of nocratic responses to improve governance work sustainability, a problem that donors and private only in very auspicious settings--where there is foundations can address through technical assis- committed leadership, a broadly based coalition tance and funding. Because better governance is a I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 299 TABLE 9.2 Types of States and Entry Points for Strategic Interventions: A Governance Typology Type of state Governance profile Main implications for the triad Key challenge Possible entry point Capable Low to modest Some gaps in bureaucratic Increasing efficiency of Technocratic reforms in state administrative capacity and internal public service delivery public administration, corruption accountability as needed; compre- Mild state capture hensive reform strategy may be feasible Weak High levels of Very weak bureaucratic Ensuring delivery of basic Highly targeted reforms state administrative capacity and/or internal public goods in key sectors only, corruption accountability supplemented by Mild state capture limited reforms in bud- get execution to ensure financial accountability Captured High levels Very weak political and Breaking hold of vested Build demand for reforms; state of administrative internal accountability; interests on the process possibly explore corruption possibly weak bureaucratic of policy and institutional opportunities at High state capture capacity reform subnational level Restrained Low to modest Very weak political Increasing "voice" Build demand for reform; state administrative accountability await change in leadership corruption caused by crisis related to High state capture corruption, after which technocratic reforms to increase political account- ability may be possible Source: Author's elaboration. public good, groups working on governance of Transparency International's Corruption Percep- reforms will generally find it more difficult to tion Index made governments and their constituent solicit contributions from the general public. publics more aware of problems of corruption and helped trigger the development of cross-country Whether this focus on problem solving and empirical studies and survey-based diagnostic work results-oriented drivers of change will help on corruption (see, for instance, Kaufmann, Kraay, countries to navigate the difficult terrain of gov- and Zoido-Lobaton 1999; Reinikka and Svensson ernance reform in the next decade remains to be 2003). seen.What is certain is that governance reform 4. This section builds on an analytical framework from will retain a high place on the reform agenda. the World Bank's World Development Report (WDR) 2004. 5. Formal organizations such as the central bank are also often referred to as institutions. In the abstract, Notes an organization is a collectivity functioning within a predetermined set of formal rules.That is, it is an 1. See, for instance, Rose-Ackerman (2004) and Klit- agglomeration of rules that affects the behavior of a gaard (1988). Much of the literature on rent seeking given set of individuals. from the mid-1970s to the 1980s,for example,Bhag- 6. WDR 2004 characterizes the bilateral relationship as wati (1978), essentially tackled conceptual issues. one of accountability (of agents to principals). But 2. The causality can work both ways, with growth also the relationship can also be affected by other fac- perhaps inducing better governance. Some recent tors--in particular, capacity issues. If the agent lacks studies address this (Rodrik and Subramanian 2003; the ability or the resources to perform his or her task, Subramanian and Roy 2001). then no matter how strong the accountability link, 3. The 1996 "Cancer of Corruption" speech given by he or she will be unable to deliver on his or her man- World Bank President James Wolfensohn is consid- date.If,for instance,bureaucrats have a poor financial ered a watershed in the Bank in the fight against cor- management information system to work with, they ruption and the push for reforms of governance cannot produce adequate information upon which institutions in developing countries.The publication to judge their performance. 300 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s 7. See also WDR 2004, chapter 5. Inter-American Development Bank had made 18 8. More specifically, the possibility of obtaining rents loans totaling $418 million by June 2001,and entered drives influential groups and/or individuals to bribe into 65 technical cooperation agreements compris- politicians or high-ranking civil servants.In turn,the ing another $43 million to support reform of judicial latter introduce and maintain bad laws, policies, or institutions (Biebesheimer and Payne 2001, 12). regulations in order to perpetrate their illicit earn- 16. For example, in the early 1990s the Ukrainian legis- ings. In this context, corruption causes bad gover- lature was asked to enact a verbatim translation of the nance. Uniform Commercial Code--a lengthy American 9. Several empirical studies have shown that deficien- statute governing sales, leasing, and related transac- cies in civil service processes and procedures tend to tions--and advisors to the Mongolian government increase administrative corruption (Van Rijkeghem suggested that to create a stock market Mongolia and Weder 2001; Evans and Rauch 2000; World simply copy U.S. securities laws. Bank,World Development Report 1997) and have con- 17. This Anglo-Saxon concept--and related ones in use firmed the findings of numerous case studies (for in Europe, Latin America, Asia, and Africa--can instance Rose-Ackerman 1978; Wade 1985; Klit- mean many different things. But in the parlance of gaard 1988, 1990).The relation between public sec- the development community the rule of law became tor pay and corruption is controversial: it is synonymous with a state where laws effectively commonly presumed that the lower is public sector restrained rulers from opportunistically seizing pri- compensation, the greater is administrative corrup- vate property and simultaneously promoted welfare- tion, but various empirical studies suggest that other enhancing cooperation among the citizenry. factors confound this relationship. See http:// 18. Though there are differing interpretations of the www1.worldbank.org/publicsector/civilservice/ composition of New Public Management reforms, ineffectivemon.htm#4. there is general agreement that the key components 10. Citizens can use shame tactics to pressure local-level include deregulation of line management, conver- civil servants who live within the same community. sion of civil service departments into free-standing 11. "Client power"also includes the ability of citizens to agencies; performance-based accountability, particu- choose among different providers of the same ser- larly through contracts;and competitive mechanisms vice.This does not require collective action on their such as contracting-out and internal markets.Various part; rather it requires policy decisions to promote authors also include privatization and downsizing as some form of competition among service providers. a part of the package (Polidano 1999). Voice (and thus organization) can play a role in get- 19. See Poterba (1996) for a review of the literature circa ting politicians to promulgate such policies. 1996. 12. Political accountability during elections can be 20. Alesina and Perotti (1996) discuss institutions in enhanced by independent electoral commissions. terms of the degree of centralization of authority in However, such institutions may themselves also be the budget process and the degree of transparency. subject to capture. Von Hagen (1992) provides summary information 13. Work has been done on reforming political institu- on the budget process in European Community tions, including attempts at reform in political cam- nations.Von Hagen and Harden (1996) suggest that paign finance. But such efforts have been far fewer tighter budget rules are associated with smaller than those devoted to legal, judicial, and legislative budget deficits and lower levels of government bor- reforms. rowing. 14. The heightened concern with improving legal and 21. An MTEF has five key features: (1) a top-down judicial systems in the 1990s was predated by similar process for establishing hard budget constraints at the concerns and reform efforts in the 1960s--some of aggregate and sectoral level; (2) a bottom-up process which are only now beginning to bear fruit. in which line ministries prepare forward estimates of 15. The World Bank's first significant effort was a 1990 expenditures over a three- or four year period; (3) a adjustment loan to Bangladesh, followed two years system for reconciling the forward estimates and the later by support to Tanzania for training judges, hard budget constraints,which includes processes for upgrading legal libraries, and publishing court deci- making intersectoral reallocations; (4) a transparent sions. In 1992 the Bank also extended a $30 million system for incorporating changes to the forward esti- loan to Venezuela solely for judicial reform.This was mates during rollovers to the following year; and (5) the first of 11 investment loans totaling close to $200 a tractable system for undertaking program evalua- million that the Bank extended principally for judi- tion (World Bank 1998b). cial reform during the 1990s.Data from other organ- 22. World Bank/OECD Survey on Budget Practices izations show a similar increase in judicial reform and Procedures (2003). Available online at activities. Starting from scratch in the 1990s, the http://ecde.dyndns.org. I M P ROV I N G P U B L I C S E C TO R G OV E R N A N C E : T H E G R A N D C H A L L E N G E ? 301 23. Their finding is consistent with related work by 33. An integrity pact embodies a transparent process of Evans and Rauch (2000), and with research at the procurement that all participating bidders agree to sector level. Gunnarsson et al. (2004) find strong bind themselves to.A civil society organization, such evidence indicating that the granting of autonomy as a local chapter of Transparency International, to local district school principals (akin to delega- monitors the process step by step to assure all bidders tion under NPM) improves student performance that each has kept to the joint commitment. only if the principal and the school staff have ade- 34. For examples see John Pomfret,"SARS Reported in quate capacity to take advantage of the increased Rural China,"available online at http://stacks.msnbc autonomy. .com/news/904928.asp?cp1=1, and John Pomfret, 24. In related reforms, more than 40 developing and "China to Open Field in Local Elections," Washing- transition countries have set up special large taxpayer ton Post, June 12, 2003. In Cuba, while dictatorial units to improve the tax compliance of the largest tactics have certainly kept the Castro regime in taxpayers and to pilot new organizational structures, power for almost four decades, the high quality of systems, and procedures. health services--which are among the best in the 25. For a particularly strong critique, see Fjelstad (2002); developing world--is a factor that has contained cit- and for a more general country survey and balanced izen dissent. assessment, seeTaliercio (2003). 35. Based on IMF Government Financial Statistics. 26. The same political support might perhaps have 36. An excellent study that analyzes the effect of type of achieved the same results without the move to an electoral system on policy outcome is Myerson autonomous revenue agency, but in some instances (1999).Persson,Tabellini,andTrebbi (2000) assess the introducing such an agency has energized and crys- impact of corruption and voting in different elec- tallized support for more wide-ranging administra- toral systems. tive and civil service reform. 37. Reinikka and Svensson (2003, 2004). 27. Harding (2003) analyzes different approaches to 38. For more on lessons on freedom of information laws, health care provision: direct provision by the public seeWorld Bank (2004a). sector, through nonprofit providers, through for- 39. There is"often a vicious circle whereby the failure of profit providers,and through informal providers.The the state breeds more corrective rules which both lessons suggest possibilities for focused strategic reformer and opportunist applaud--the reformer interventions in the health sector. under mistaken formalistic notions about how to 28. E-government applications have normally evolved reform and the opportunist in the knowledge that through a four-stage process.The first stage includes reform will be frustrated and his opportunism can the publication of information on aWeb site for cit- continue"(Burki and Perry 1998,128).This is not to izens to seek knowledge about procedures governing imply that in developed-country settings there are the delivery of different services; the second stage is no informal rules.These never disappear. But they interactivity online, allowing clients to download tend to be better aligned with formal institutions applications for receiving services; the third stage rather than being inconsistent with them (North involves electronic delivery of documents; and the 2002). fourth results in electronic delivery of services,which 40. This is the Business Environment and Enterprise may involve more than one department in process- Performance Survey (BEEPS) conducted first in ing a service request or service. 1999 and most recently in 2003.The findings of the 29. E-government can also enhance transparency. For 1999 survey were analyzed and published in World instance,the use of electronic bulletin boards in gov- Bank (2000a). ernment procurement has made information on 41. Ideally, a third dimension measuring client power government contracts much more widely available would be desirable.This dimension could be con- (Bhatnagar and Deane 2003). structed using available data on citizen voice. 30. Bhatnagar and Deane (2003) estimate that roughly 42. World Bank data suggest that 25 to 30 countries 40 percent of an e-government project's cost must be qualify as low-income countries under stress.In these allocated to managing the change process. countries, securing law and order and ensuring the 31. http://www1.worldbank.org/publicsector/egov/ delivery of very basic public services remain the pri- seoulcs.htm. mary governance tasks. Beyond this, the evidence 32. WDR 2004 provides examples and indicates the suggests that comprehensive and/or more techni- extensiveness of citizen-based mechanisms for rais- cally demanding governance reforms are unlikely to ing the demand for better governance. be feasible. Country Note H Natural Resources: When Blessings Become Curses S ince at least the time of Adam where natural resources account for 26 percent of Smith and David Ricardo there GDP (Collier and Hoeffler 1998, 2001). In far has been a belief that natural too many countries, including Iraq, Nigeria, resources are a blessing: that countries richly Sierra Leone,Venezuela, former Zaire, Zambia, endowed with natural resources have an advan- and many others,enormous oil or mineral wealth tage over countries that are not. For centuries, has not translated into economic and social well- people moved to where natural resources were being for the majority of the population. abundant: to the Americas, to Australia, to oil- The natural resources that depress countries' rich countries in the Middle East. Natural long-run growth are those whose rents are tech- resource endowments have helped many coun- nically easy to appropriate: so-called point- tries, including Australia, Canada, Finland, and source natural resources such as diamonds, gold, Norway, to grow and diversify, in part by provid- oil,and minerals.Other resources,such as land or ing a basis for developing associated technologies human resources, have more diffuse rents and do and capital goods industries (World Bank not seem to have such an effect. 2001b). Two lines of explanation have emerged to Since the end of World War II, however, and explain the "natural resource curse." The first particularly since the 1960s, evidence has accu- focuses on how natural resources affect the econ- mulated that natural resources are less often a omy, and the second on how they affect institu- blessing than a curse1 (see figure H.1).This find- tions (Eifert, Gelb, andTallroth 2003). ing is statistically robust, invariant to changes in specification, variable definitions, or inclusion of Economic Effects additional explanatory variables--including those commonly used in empirical growth stud- The so-called Dutch disease is perhaps the most ies, such as geography and climate.After control- well-known effect of natural resource rents on ling for all possible influences and interactions, the real economy. High exports of natural the evidence is that countries rich in natural resources cause an appreciation of a country's real resources grow more slowly. exchange rate, which moves its productive Not only economic growth is affected nega- resources away from tradables such as manufac- tively (Gelb 1988). Controlling for country tured goods. If manufacturing produces signifi- income level, countries that are rich in natural cant positive externalities that are crucial for resources have more unequal income distribution long-term development, such as learning-by- and a larger share of their population in poverty; doing, the country's economic growth rate will they exhibit greater corruption, have more suffer (Sachs andWarner 1997). authoritarian regimes, spend more on the mili- Another well-known problem of resource- tary, and face a higher probability of an armed rich economies is volatility, with cycles of boom conflict (Palley 2003).The probability of a civil and bust. For example, the high resource prices conflict is 0.5 percent in a country with limited of the 1970s led resource-rich countries to bor- natural resources, but 23 percent in a country row heavily,and the collapse of prices that ensued 304 NATURAL RESOURCES: WHEN BLESSINGS BECOME CURSES 305 FIGURE H.1 Natural Resources and Growth, 1970­89 8 Singapore Korea,, Rep. of Taiwan (China) 6 Hong Kong (China) Malta Botswana Indonesia Mauritius 89­ 4 Cyprus Iceland Malaysia 1970 3 Algeria Fiji capita, Nigeria Gabon Gambia Mauritania per 0 Bahrain Venezuela Saudi Arabia Côte d'Ivoire Chad Niger growth ­2 Nicaragua Iran, Islamic Rep. of Guyana Zambia Liberia GDP United Arab Emirates ­4 Real ­6 Kuwait ­8 0 10 20 30 40 50 60 70 80 Exports of natural resources, in percentage of GDP Source: Sachs and Warner 1997. in the early 1980s left them with large debts and nonresource tradables. In turn, the less the econ- little capacity to service them (Manzano and omy produces nonresource tradables,". . . the Rigobon 2001).The countries affected ranged greater the volatility of relative prices, the higher from Bolivia to República Bolivariana de the interest rate the sector faces, causing it to Venezuela to Côte d'Ivoire and Nigeria; many shrink even further, until it disappears" (Haus- resource-rich countries have not yet recovered. mann and Rigobon 2002). Equally serious, unless a resource-rich econ- Saving rates in oil-exporting countries are omy has a large non-resource-based tradable sec- much higher than in other developing countries. tor to begin with,the uncertainty associated with But even so,these relationships mean that without cycles of boom and bust can reinforce a down- corrective policies, volatility causes oil-exporting ward cycle.The smaller the nonresource tradable economies to specialize inefficiently in the pro- sector, the fewer opportunities workers have to duction of nontradables,retarding their long-term find new jobs when resource prices decline; as a growth.Thus, it has been argued that República result, a price decline can cause the whole econ- Bolivariana de Venezuela's growth implosion in omy to contract. Interest rates will reflect the the early 1980s was the result of the high real risks associated with this volatility:the greater the interest rates facing the nonresource tradable sec- volatility, the higher is the interest rate, and, in tor, and of uncompetitive and volatile exchange turn, the smaller are investments in nonresource rates, which caused the country to specialize tradables.These two effects combine to cause the almost exclusively in resource extraction and non- economy to specialize away from production of tradables (Hausmann and Rigobon 2002). 306 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s Institutional Effects While natural resources were a curse for Nigeria, the discovery of diamonds became a These economic explanations do not answer blessing for development in Botswana, as did oil questions about the differences in performance in Indonesia. across resource-rich countries, such as why dia- Botswana has effectively maintained law and monds have been a curse for Sierra Leone but a order, limited state predation, and enforced hard blessing for Botswana,or why oil has been a bless- budget constraints on its parastatal organizations ing for Indonesia. Neither can they explain why (Acemoglu, Johnson, and Robinson 2003).The point-source natural resources affect growth dif- government has invested heavily in the expansion ferently from natural resources with more diffuse of infrastructure and efficient delivery of educa- rents.Such differences in performance have given tion and health services.While AIDS has reversed rise to a large literature offering political and insti- some of the health gains, and led to a sharp tutional explanations of the resource curse. decline in life expectancy, Botswana's other social Large and concentrated rents, easier to appro- indicators are among the better in Africa and in priate than the more diffuse rents associated with the developing world.The bureaucracy is largely land or human resources,make societies less entre- meritocratic, relatively noncorrupt, and efficient. preneurial by increasing the private returns to Fiscal revenues from resource rents have been unproductive rent-seeking. Several studies focus used to smooth revenue over commodity price on the "voracity effect," or common pool prob- cycles,rather than financing consumption booms; lems, that move an economy into a low-growth indeed,Botswana was one of the first countries to equilibrium because of political fights over establish a stabilization fund, which it managed resource rents (Lane and Tornell 1999; Mehlum, well. One of the reasons behind Botswana's suc- Moene,andTorvik 2003).In Nigeria,for example, cess is believed to be its benign neglect by colo- governance institutions were weak and large oil nial powers: Botswana was on the periphery of resources were wasted. Large windfall oil profits the British empire, not known to have valuable corrupted Nigeria's institutions and changed its resources,and hence of little interest.Thus,unlike politics,which came to be shaped by the incessant in most other African countries, colonialism had fight over resource revenues. Public spending a negligible effect on traditional social and politi- turned into outright patronage, crippling the civil cal institutions. Botswana's pastoral traditions tra- service. Starting with the Biafra war in the late ditionally encouraged broad-based participation 1960s, successive military dictators plundered and constraints on political leaders;rural interests, Nigeria's oil wealth,wasting resources on an enor- chiefs, and cattle owners retained their political mous scale; the country's total factor productivity power throughout the colonial period.They have has declined by 1.2 percent a year over the last few been the source of checks on the executive, and decades (Sala-i-Martin and Subramanian 2003). explain why diamond rents have been exception- Numerous econometric studies confirm that ally well managed. countries rich in natural resources have weaker Indonesia did not have democratic and par- institutions,measured in terms of checks and bal- ticipatory institutions, at least until very recently. ances on the executive, rule of law, and corrup- But although corruption and governance prob- tion (Sala-i-Martin and Subramanian 2003; lems were widespread,Soeharto's regime focused Isham et al. 2003; Sala-i-Martin, Doppelhofer, on economic and social development. On the and Miller 2003; Robinson,Torvik, and Verdier one hand, it provided checks on state and indi- 2002; Mehlum, Moene, and Torvik 2003).These vidual predation and, on the other, it provided studies also show that,controlling for institutions, predictability and consistency in policy making natural resources have no effect on long-term (Temple 2003). Internal accountability mecha- growth.That is, institutions are not just the prin- nisms enabled the bureaucracy to deliver a wide cipal but the only channel through which array of social and infrastructure services, and resources influence the course of the economy. antipoverty programs. Growth in Indonesia was NATURAL RESOURCES: WHEN BLESSINGS BECOME CURSES 307 not only rapid but quite widely shared, through mented with oil funds or stabilization pro- programs such as the Instruction of the President grams--with disappointing results. Successful (INPRES), a rural development program that management of a natural resource curse calls for was started at the time of the first rise in oil prices a combination of policies and institutions. On in 1973 and subsequently expanded. INPRES the economic policy front, countercyclical stabi- included village support grants, rural infrastruc- lization policies have a critical role to play, as do ture, and a massive expansion of schooling policies that maintain the competitiveness of the (World Bank 1993). One reason for Indonesia's real exchange rate for the nonresource tradable success was that the Soeharto regime shielded sector, and financial policies that encourage technocrats from political pressures: the group of investments in that sector. On the institutional high-level technocrats responsible for policy front, institutions such as transparency, and making (the "Berkeley mafia") was empowered checks and balances on the use of rents, that to make economic policy decisions with long- increase the costs of nonproductive activities can term growth and development objectives in help countries to move away from rent-seeking mind.Thus, the response to a fall in oil prices in equilibria to more dynamic, diversified, and the early 1980s was a textbook adjustment that growing economies.EastTimor's oil stabilization triggered comprehensive microeconomic fund illustrates this approach.While it is too early reforms--from competition policy to exchange to determine how the fund will work in practice, rate adjustments and trade liberalization (see also the intent of the fund is to rely on institutional Country Note B,"Lessons from Countries That improvements that ensure resource rents are Have SustainedTheir Growth"). effectively used for long-term development. It Economic and political explanations are diffi- emphasizes transparency and public awareness of cult to disentangle. In the course of development, the issues that concern the good use of oil rev- economic institutions are shaped by economic enues, thus developing constituencies in support incentives and opportunities,and political dynam- of prudent policies. ics respond to underlying economic forces (Engerman and Sokoloff 2002).Political and other Note institutions may be the main explanatory forces, but economic forces also play a role in explaining 1. Sachs andWarner (1995b,1997,2001);Lane andTor- nell (1999); Auty and Mikesell (1998); Gylfason why the institutions are the way they are. (2001); Leite and Weidmann (1999); Dalmazzo and de Blasio (2001).While there are many ways to define natural resource abundance--for example as the share Conclusion of natural resources in the gross domestic product (GDP) or exports (as in figure H.1), with further Simply copying or adopting policies that have breakdown for fuel,ores,and metals;or oil-producing been effective elsewhere rarely succeeds. Many versus other developing countries--they all suggest resource-rich developing countries have experi- that countries rich in resources grow more slowly. Chapter 10 Does Democracy Help? A STRIKING PHENOMENON OF seeking,and venality--in some countries than in the 1990s was the rise in the others? Why are commitments by some govern- number of countries selecting ments more credible than others? their leaders through competitive elections.1The To answer these questions,this chapter focuses number rose from 60 countries in 1989 to 100 in on two propositions. First, elected governments 2000. Among poorer countries (those with less are most likely to make policies favoring narrow than the median country's per capita income), segments of the population at the expense of the the number nearly tripled,from 11 in 1989 to 32 majority when citizens are ill informed,or cannot in 2000; 15 percent of the poorer countries trust promises made prior to elections, or are elected their governments in 1989 and 42 per- deeply polarized. Second, elected governments cent in 2000. are most credible and most likely to respect pri- Unfortunately, democratization does not vate property rights when they confront checks ensure economic development.The simple fact and balances on their decision making.2These are of competitive elections did not enable Haiti's not the only explanations for democratic per- government to contain predation by the power- formance. For example, outside forces direct the ful or to establish minimal law and order. Nor policies of some countries, and the consequences did it prevent Kenya's government from exerting of a country's history and culture are surely its authority to benefit a small, privileged elite. important.3 But the arguments in this chapter Certainly, most poor democratic countries con- suggest that it is through their effects on political trol predation better and treat citizens more gen- credibility (party development), clientelism, citi- erously than in these examples. But typically zen information,and social polarization that these contractual and property rights, widely recog- other forces probably operate. nized as fundamental for investment and eco- Section 1 looks at the relationship between nomic growth, are less well enforced in poorer democracy and development, finding that com- democracies than in richer ones. Similarly, while petitive elections have only a modest effect on the rent seeking and corruption are higher in poorer quality of government; elected governments do democracies, public services such as education, not exhibit a systematic advantage in achieving critical for both growth and poverty alleviation, economic development. Section 2 examines rea- are less well provided. sons why political decision makers do not always Accelerating economic development in adopt policies in the broad public interest. Imper- developing countries with elected leaders stands fections in electoral markets--lack of voter infor- as one of the important challenges of the 2000s. mation, the inability of political competitors to Why are democratic institutions less account- make credible promises, and social polarization-- able--more vulnerable to narrow interests, rent are important to understanding policy formula- 309 310 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s tion and explaining differences in economic per- cies of newly democratized countries are better formance between rich and poor democracies. explained by increases in income per capita. Section 3 looks at reasons for the lack of credibil- Among all countries that held competitive elec- ity of government commitments.It finds evidence tions in the 1990s, purchasing power parity that imperfections in political markets have a sig- (PPP)­adjusted incomes rose by a third during nificant impact on economic growth and hence the decade.4 Using a six-point scale to compare need to be taken into account in designing strate- the quality of government in these countries, gies to speed growth and development. Section 4 three points separated the lowest-scoring 25 per- discusses reform strategies for remedying some of cent of countries from the highest-scoring.And the fundamental distortions that can plague dem- in half of these countries, the rule of law was no ocratic decision making.Adjustments in the way better than it was in the median country lacking that governments and their development partners competitive elections.Among countries that had approach the more traditional development competitively elected governments in 1995,gross agenda, from service delivery improvements to secondary school enrollment varied more than broader public sector reforms, can go a long way 140 percentage points from the minimum to the toward mitigating the shortcomings in informa- maximum,and 60 percentage points separated the tion and credibility that otherwise undermine top and bottom quartiles.After accounting for the government accountability and performance.Sec- effect of income per capita, 40 percent of the tion 5 concludes the chapter. countries lacking competitive elections exhibited higher gross secondary school enrollment than 40 percent of the countries that held them. 1. Elections Have an Uneven Consistent with these findings, a large litera- Impact on Development ture finds no consistent,significant effect of elec- tions on economic growth. For example, Intuitively, one might expect that in countries Przeworski et al. (2000) find no difference in where most of the public cannot hold the gov- growth rates between countries that have com- ernment accountable, government decision petitive elections and those that do not. making will tend to disregard the public interest. Another factor that blurs the distinction Moreover,the richest countries in the world (the between democracies and nondemocracies is the countries with the longest record of sustained heterogeneity of the latter group.Some autocrats growth) have experienced relatively long periods find that they can extract more rents and stay in of uninterrupted elections. Thus the political office longer if they encourage investment and upheaval and democratization in the 1990s promote long-run economic growth; indeed, in offered reasons for optimism regarding economic countries with nonelected leaders, property development. rights become more secure the longer the lead- Some policy progress could be seen among ers have been in office (Clague et al. 1996).5 But the democratizing countries. For example, many autocrats are unable to trigger this virtuous among countries that lacked competitively circle:investors are deterred by the fear that prof- elected governments in 1988 but had them by its will be expropriated,the rents obtained by the 1998, secondary school enrollment rose by autocrat fall, economic performance drags, and about 14 percentage points. Similarly, a measure threats to the autocrat's tenure grow. It appears of the rule of law--capturing the extent to that in nondemocracies in which the likely rates which government acts arbitrarily--improved of return to investment are low (for example, by three-quarters of a point on a six-point scale. countries with uneducated workforces and no Cross-country analysis shows, however, that easy access to foreign markets), or in which the there is little association between competitive rates of return to natural resource exploitation elections and the quality of government. The are high, leaders are less likely to curb their own modest improvements that took place in the poli- authority to attract greater investment. D O E S D E M O C R AC Y H E L P ? 311 Beyond geographic explanations,nondemoc- public interest in poor democracies but not rich racies that emerge from broad social movements ones: lack of voter information, the inability of appear to place more controls on their leaders.In political competitors to make credible promises Mexico, for example, during the period that it and be trusted, and social polarization. Each of dominated politics, the Partido Revolucionario these is important to understanding policy for- Institucional (PRI) provided checks on the behav- mulation and reform. ior of presidents. Nondemocracies also differ in the extent of their institutionalized sources of Special Interests authority (even within a single party) that might counterbalance the authority of the top leader. Before the 1990s, attempts to explain govern- Even nondemocracies with an unelected legisla- ment policy failures centered on the role of spe- ture have significantly less corruption and greater cial interests.The logic--"the logic of collective rule of law than countries without an unelected action," as Mancur Olson coined it in 1965--is legislature.6 Gandhi (2003) finds that nondemoc- clear. Small, homogeneous groups with much at racies with unelected legislatures grow faster than stake confront relatively low costs to acting col- those without such legislatures. lectively in their common interest. In compet- ing for benefits from government, this gives them advantages over large groups whose inter- 2. Characteristics of Democra- ests are heterogeneous. Unfortunately, narrow cies That Influence Policy interests rarely benefit from public goods, such Success and Failure as the provision of universal education or an improved court system,as much as they do from In all settings where people come together to act diverting some fraction of societal resources to collectively, complaints of high-handed behavior themselves. Hence, to the extent that govern- by leaders and of its converse, endless consensus- ment incentives encourage targeting of benefits building, are endemic.Whether in town coun- to special interest groups, policy failure--the cils, sports clubs, or Musikvereinen, issues of underprovision of public goods and the over- fairness and equity, efficiency, and consistency provision of regulations and laws that benefit regularly arise. Special interests curry favor in special interests at the expense of the whole every country in the world, and individuals society--is more likely. everywhere succumb to the temptations of Ample evidence points to the importance of venality and rent seeking. collective action considerations in the making of Why, though, are rent seeking, special interest public policy in all countries. Bates (1981, 1983) influence, and venality--the effects of govern- and Frieden (1991) make compelling cases for ment inefficiency--worse in poor countries than the role of special interests,indigenous or foreign, in rich?As shown in the following discussion,the in shaping policies in Africa and Latin America, activities of special interest groups explain policy for example. outcomes generally,but they do not explain why There is no strong evidence,however,that the policy outcomes in developing countries differ logic of collective action can explain differences from those in developed ones. Similarly, differ- in development outcomes. For example, there is ences in political and electoral institutions no indication that the intrinsic characteristics of explain variations in policy outcomes across special interests differ in developing countries, countries,but they do not explain the divergence better enabling special interests in these countries between policies in developing and developed to extract rents from the unorganized majority. countries. Nor is there evidence that special interests in Instead, there are three other explanations, developing countries are more unified than in all related to imperfections in electoral markets, developed countries, and therefore less likely to for why policies are more likely to neglect the "cancel out" their respective influences. 312 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s Formal Rules FIGURE 10.1 Great attention has focused on the role of insti- Electoral Rules in Richer and tutions in development.The formal rules deter- Poorer Democracies mining how politicians attain office and make 100 decisions decisively influence policy outcomes. However, it is less clear that institutional differ- 75 ences can explain the differences in development performance among democracies. 50 Substantial research in the 1990s focused pre- Percent cisely on the effects of political and electoral 25 institutions on the magnitude of government spending, broad public goods, and rent seeking.7 0 Researchers found that under some conditions 1990 2000 1990 2000 Richer Poorer majoritarian rules (first-past-the-post electoral Proportional Plurality systems with small electoral districts) lead politi- cians to focus on pivotal narrow constituencies, Sources: Database of Political Institutions (Beck et al. 2001) and World Bank, World Development Indicators. biasing spending downwards and away from broadly based public goods.8 Comparisons of Note: Countries represented are those that held competitive elec- tions for executive and legislative elections (the Legislative and presidential and parliamentary forms of govern- Executive Indexes of Competitive Elections from the Database of ment yielded similar predictions: parliamentary Political Institutions were both equal to seven). systems, under some conditions, promote greater allocations to broad public goods than do presi- proportional electoral systems than in presidential dential systems.9 These findings are potentially democracies with majoritarian systems.11 How- important for our understanding of develop- ever, the key public good is education itself, not ment,to the extent that public goods such as uni- education spending, which turns out to have little versal education or law and order are essential to effect on gross secondary school enrollment.12 economic growth. Political and electoral institutions are insignificant Do poorer democracies, with less robust pro- determinants of secondary school enrollment pre- vision of public goods and greater rent seeking, cisely because they have less of an influence in exhibit the electoral and political institutions that poorer countries.13Among the richer democracies are thought to promote these outcomes? in 1997, school enrollment was about 38 percent- In 2000, of the countries with competitively age points greater in parliamentary than in presi- elected governments, plurality or first-past-the- dential systems. In the poorer democracies, it was post rules dominated among the electoral sys- essentially the same regardless of political systems.14 tems of the poorer countries but not among Corruption is another indicator of the extent those of the richer (figure 10.1).10 Similarly with to which government decisions on spending or respect to political systems, presidential systems policy are likely to translate into improved social were much more common among the poorer welfare.There is little evidence that political and democracies than among the richer (figure 10.2). electoral institutions can explain the greater Though these associations might suggest an prevalence of corruption in developing countries institutional explanation for the differences in the than in developed ones. Arguments formulated policy experiences of the two sets of countries,the in the 1990s (for example, by Persson and evidence is weak that these institutions are responsi- Tabellini 2000) suggested that under some con- ble for the differences. As researchers have pre- ditions, presidential systems would reduce cor- dicted, spending on education is about 2 ruption. However, there is little evidence of this percentage points of gross domestic product in either rich or poor countries (Adserà, Boix, (GDP) greater in parliamentary democracies with and Payne 2003). D O E S D E M O C R AC Y H E L P ? 313 the arguments that these elements should matter FIGURE 10.2 are persuasive and seem to have great validity in Political Systems in Richer and richer countries. Their relative weakness in Poorer Democracies explaining outcomes in poorer countries suggests 100 that the underlying conditions of political com- petition in these countries differ from those in 75 richer countries.We explore these conditions and their effects next. 50 Percent Imperfections in Electoral Markets 25 Differences in economic performance across 0 democracies can be explained with respect to 1990 2000 1990 2000 imperfections in electoral markets. Numerous Richer Poorer imperfections in electoral markets make it difficult Presidential Parliamentary for citizens to hold politicians accountable for Source: Database of Political Institutions (Beck et al. 2001) and policies.The discussion below focuses on three World Development Indicators. imperfections--uninformed voters, noncredible Note: Countries represented are those that held competitive elec- political competitors,and social polarization--that tions for executive and legislative elections (the Legislative and offer powerful insights into the underperformance Executive Indexes of Competitive Elections from the Database of Political Institutions were both equal to seven). of many democracies. Uninformed Voters There is more evidence that electoral systems In political markets, the information that voters affect corruption, but the effect is subtle.Adserà, have about the characteristics of political com- Boix,and Payne (2003) find no effect.But Persson, petitors and government performance is crucial. Tabellini, and Trebbi (2003) break down electoral Without information about the attributes of institutions into their component parts--district political competitors, about what politicians are magnitude (measured by the number of seats up doing, and how their doings affect citizens' well- for election in the district) and voting rule--and being, citizens cannot easily reward high-per- argue that the larger the district, the lower are the forming politicians. This encourages poor barriers to entry faced by competing parties and performance. Politicians confronting unin- the more likely it is that voters can drive out cor- formed voters can invest resources to persuade rupt parties.When voters can express a preference them of their accomplishments, through adver- for individual candidates, as in plurality systems, tising or meetings, for example. But financing they are better able to remove corrupt legislators. these efforts, whether from their own pockets or Persson,Tabellini, andTrebbi (2003) find evidence those of special interests, or from government that both effects are at work. In practice, however, funds, carries a social cost: special interests the two effects cancel each other out, since coun- demand policies that diverge from the social tries with proportional electoral rules generally interest in exchange for campaign financing, require voters to choose parties rather than candi- while government funding diverts resources dates. Hence, electoral rules cannot explain the away from the provision of goods and services to greater levels of corruption in poor countries. the electorate. There is, then, no strong evidence that either No data directly measure how well informed special interest group organization or formal dif- citizens are about the contributions that their rep- ferences in political and electoral institutions resentatives make to their welfare. One commonly account for the different policy choices of devel- used proxy for citizen information is newspaper oped- and developing-country democracies.Still, circulation per 1,000 as a proportion of population. 314 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s In 1995, among countries that hold competitive elections,newspaper circulation was,not unexpect- FIGURE 10.4 edly, considerably higher in richer than in poorer Newspaper Circulation and Corruption countries (figure 10.3). Controlling for income and other factors, 2.16705 higher newspaper circulation is associated with lower corruption (Adserà, Boix, and Payne 2003), and with greater rule of law, better Corruption bureaucratic quality, and greater secondary (adjusted) school enrollment (Keefer 2003a). As discussed later in this chapter, newspaper circulation and access to radios increase the probability of receiv- ­3.32613 ing government transfers (Besley and Burgess ­206.656 343.875 2002 and Strömberg 2002, respectively). Newspaper circulation per 1,000 people (adjusted) Figure 10.4 illustrates these effects, showing how newspaper circulation, controlling for other Source: Database of Political Institutions (Beck et al. 2001) and influences, suppresses corruption.15 World Development Indicators. Note: The figure depicts the effect of the component of newspa- Credibility of Politicians per circulation that is uncorrelated with the other explanatory variables on the component of corruption that is uncorrelated When challengers cannot make credible policy with the other explanatory variables (the orthogonal component commitments to citizens, citizens have no reason of each), based on the regression below. The sample is of coun- to prefer them over incumbents. Even if incum- tries that exhibit competitive elections (LIEC = EIEC = 7 from the Database of Political Institutions), 1990­2000; economic vari- bents do badly, citizens have no reason to believe ables are from World Development Indicators; t-statistics are in parentheses; ordinary least squares regression controls for clus- ters of observations from the same country that artificially inflate statistical significance. FIGURE 10.3 Corruption = 3.85 + 0.00002PPP-adjusted income/capita + Indicators of Political Market Imperfections in (2.25e­08) land area­5.33) (0.66)(0.65) (9.3e­10) population ­ Countries Holding Competitive Elections, 1995 0.44 electoral system ­ 3.87percent population young + 1.2 per- cent rural (­1.14) (­2.49) (­1.98) ­ 0.01 political system + 0.02 50 continuous years competitive elections + 0.002 newspaper circula- 45 42.8 tion per 1,000 (2.62) (­0.07) (3.94) (2.88). 35.4 25 22 21.7 that challengers will do better.16 This insulates incumbents from pressure to perform. 11.8 In practice,politicians never entirely lack cred- ibility. Some political competitors are credible on 0 only one or a few issues unrelated to economic Newspaper Party age Continuous years of development,such as a country's struggle for inde- circulation/1,000 people (yrs) competitive elections (yrs) pendence,or issues of religious importance,but in Richer Poorer such cases, the votes they attract do not provide a motivation for better economic policy perform- Source: Newspaper circulation from World Development Indicators; party age (the aver- age age of parties under their current name); and continuous years of competitive elec- ance. tions, are from the Database of Political Institutions. Credibility may also be partial in the sense Note: Countries are all those with Legislative and Executive Indexes of Electoral Com- that politicians can make credible promises to petition (LIEC and EIEC) equal to the highest score of seven (see Figure 10.1). The some voters only.Credibility resides in individual income per capita threshold between richer and poorer democratic countries, dividing politicians or in "patrons" rather than in political them into roughly equally sized groups, is US$6,193. parties.The problem of credibility is therefore D O E S D E M O C R AC Y H E L P ? 315 closely related to the phenomenon of clien- A dysfunctional public sector limits the abil- telism, which is widely argued to characterize ity of politicians to make credible promises.This political relationships in poorer countries, and is the problem of capability that was discussed in involves patrons and clients who are bound chapter 9. If an education ministry is deeply dys- together by reciprocal, long-lasting patterns of functional and is likely to take years to reform, exchange.These exchanges form the foundation and if citizens cannot observe changes in the of reputations that allow patrons to deliver votes ministry until these are reflected in schools, even at election time. Unfortunately, narrowly based favorably inclined politicians are unlikely to credibility gives politicians incentives to under- make promises about education. For example, provide public goods and to extract large rents when Alberto Fujimori became President of (box 10.1). Peru in 1990,he privatized enterprises,revamped BOX 10.1 Clientelism, Credibility, and Politics O nly since the late 1990s have scholars rather than collectively because he wishes to create a begun to understand why clientelism is a personal obligation of clientship." He cites the work of more dominant characteristic of public pol- Nash on the 1960 elections in Burma: "When a local icy in some countries than in others. One explanation patron was approached to join U Nu's faction of the derives from the struggle to make credible promises to AFPFL on the promise of later patronage, he was able to citizens. Clientelism in public policy prevails when get thirty-nine others--his relatives and those who average citizens cannot believe the promises of politi- owed him money or for whom he had done favors, i.e., cal competitors with whom they have no personal con- his clients--to join as well." The rents to patrons were nection. Such a connection emerges most strongly in potentially high, since parties often had to give a local the context of patron-client relations. Scholars have patron significant authority over local administrative long noted that these relations have two important and development decisions in exchange for vote deliv- characteristics: patrons and clients interact over a long ery (Scott 1972, 110). period and they exchange goods and favors. Bista Patron-client relations drive politicians to focus on (1991, 91­92) describes the key role of reciprocity in targeted favors and goods over broad public goods and the operation of clientelism in Nepal (where it is called public policy: to the extent that only clients believe chakari): "The gift donor in chakari has certain rights. patron promises (given the absence of well-developed There is an obligation on the part of the recipient to political parties, for example), political competition respond to the chakariwal when the chakariwal so concerns primarily targeted transfers to clients rather determines. . . . Ultimately, there has to be a balance than public policy issues more generally. Wilder (1999) in exchange relations." quotes former members of the Pakistani National Scholars of clientelism from Africa to Southern Assembly from the state of Punjab as saying, "People Europe to East Asia confirm this pattern (Lemarchand now think that the job of an MNA and MPA is to fix 1972; Powell 1970; and Scott 1972). Extended com- their gutters, get their children enrolled in school, pliance with reciprocal obligations forms a basis for arrange for job transfers. . . . [These tasks] consume credible commitment, which patrons can use if they your whole day. . . ." (p. 196). "Look, we get elected decide to become politically active. because we are ba asr log [effective people] in our area. In fact, Scott (1972) quotes Wurfel as pointing out People vote for me because they perceive me as some- that "the Filipino politician . . . does favors individually one who can help them"(p. 204). Source: Keefer 2002. 316 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s the tax administration, removed price controls, reformed the customs system,and built up a large FIGURE 10.5 and successful social fund, but he explicitly Continuous Years of Competitive Elections and Corruption ignored education, which was at least as troubled as the sectors that he did address. The credibility of preelectoral promises is diffi- 2.13072 cult to measure empirically. However, it is likely to be associated to some extent both with the number of years that countries have experienced continu- Corruption ous elections and with the age of their political par- (adjusted) ties.The passage of time allows (though it does not require) political competitors and parties to build up a reputation for their stances on policy issues. ­3.87134 Among countries that hold competitive elections, ­40.7955 43.446 as figure 10.3 illustrates for 1995,both of these fac- Continuous years of competitive elections (adjusted) tors are considerably higher in richer countries than Source: Database of Political Institutions (Beck et al. 2001) and in poorer ones. World Development Indicators. Similarly, where political reputations are stur- Note: See note to Figure 10.4 for specification, data sources, and dier, the effects of clientelism should be reduced, estimation methodology. Figure drawn from regression below: public good provision should be greater, and rent Corruption = 3.85 + 0.00002PPP income/capita + (2.25e-08)land seeking lower,since political competition encour- area ­ (9.3e-10)total population (5.33) (0.66) (0.65) (­1.14) ­ ages politicians to provide high-quality public ser- 0.44 electoral rule ­ 3.87 percent pop. young + 1.2 percent pop. vices. Keefer (2003a) finds that this is the case in rural ­ 0.01 political system (­2.49) (­1.98) (2.62) (-0.07) + 0.02 continuous years competitive elections + 0.002 newspaper cir- practice:the longer a country's unbroken series of culation per 1,000 (3.94) (2.88). elections, the greater are secondary school enroll- ment,the rule of law,and bureaucratic quality,and the less are corruption and public investment as a Taken together, then, the evidence suggests fraction of GDP (public investment having the that the divergent performances of rich and poor greatest political payoffs to targeted constituen- democracies can be traced to differences in their cies).These effects are often large.The number of exposure to electoral market imperfections. continuous years of elections has a greater impact on corruption than do any of the other usual Social Polarization determinants, from newspaper circulation to for- Social polarization undermines the accountabil- mal constitutional rules to demographics (figure ity of government to citizens. One type of social 10.5). It has a greater impact on secondary school polarization emerges when substantial groups of enrollment than do education spending and pri- citizens have deeply opposing interests on most mary school enrollment (figure 10.6).17 salient political issues.These divisions can run so Figure 10.6 also indicates that education deep that one group of citizens cannot contem- spending,which has little effect on gross second- plate electing a representative from the other. ary school enrollment in general, has a strong Elected representatives from one group then have conditional effect: once one controls for contin- no incentive to satisfy the concerns of citizens in uous years of competitive elections, education the other. Moreover, they may have little incen- spending has a significant positive effect on tive even to satisfy the concerns of citizens from enrollments. This is a clear indication of an their own group; this can happen if groups increasingly well-identified phenomenon: that choose their candidates in a distorted manner without appropriate political incentives,financial (for example if backroom deals determine who resources do little to improve government per- will be the candidate from each group for the formance.18 general election). D O E S D E M O C R AC Y H E L P ? 317 abuse of minorities has been documented in a FIGURE 10.6 wide range of countries. Gross Secondary School Enrollment and Both the nature and consequences of social Continuous Years of Competitive Elections polarization depend on the political environment. 41.9462 In the first case above, of "classic" polarization, there may be third groups that are indifferent to the ideological divide between the other two groups and whose support is needed to win elec- Secondary school tions. In the second, political institutions and cir- enrollment (adjusted) cumstances can mitigate or exacerbate the effects of discrimination against minorities. For example, scheduled tribes and castes in India received greater benefits once they were guaranteed seats on local ­42.2341 legislative bodies. Wilkinson (2000) finds that ­40.3976 41.4514 Continuous years of competitive elections (adjusted) Hindu-Muslim violence was less common and elicited a more aggressive government response in Source: Database of Political Institutions (Beck et al. 2001) and World Development Indicators. those Indian states in which Muslims were pivotal voters.Rodrik (1999b) argues that ethnically frag- Note: See Figure 10.4 for interpretation. Specification and esti- mation as in Figure 10.4 and Figure 10.5 , with the addition of mented countries had the greatest difficulty reach- primary gross school enrollment and education spending. ing the agreement necessary to emerge from crisis, Gross secondary school enrollment = 82.23 + 0.0009PPP though countries with a better governance envi- income/capita ­ (3.68e­07) land area (6.96) (2.21) (­0.59) ­ ronment were able to offset this effect. (1.87e­08) total population ­ 176.7 percent pop. young ­ 16.78 The consequences of social polarization can percent pop. rural (­1.23) (­8.53) (­1.79) + 0.99 electoral rule + be worsened by all the factors that undermine 0.89 political system + 0.16 continuous years of competitive elec- tions (0.39) (0.52) (1.56) + 0.001 newspaper circulation per 1,000 voters' ability to hold politicians accountable. If + 0.34 gross primary enrollment + 77.7 total education spend- voters are better informed about actions taken by ing/GDP (0.11) (3.7) (2.29). members of their own social groups, or if they are more likely to trust promises by members of Majority disdain for the interests of identifi- their own social groups, relative to members of able minorities is another manifestation of social other groups,the effects of social polarization are polarization.The more pronounced the disdain, likely to be worsened,and the rewards from con- the greater the distortion in the provision of trolling the government are more likely to flow public goods,and the more likely that minorities to the groups whose representatives control gov- will be excluded from government services. ernment.By contrast,where the information and There is substantial evidence for these effects, credibility gap between own-group and other- and not only in developing countries. In the group representatives is smaller, social polariza- United States,Alesina, Baqir, and Easterly (1999) tion is less likely to have damaging effects. find that the more ethnically fragmented is a Controlling for numerous other factors, community (the smaller is the white majority), including income per capita,analysis shows that a the more limited is the provision of public common subjective measure of ethnic tensions is goods. In such cases, the failure of political significantly worse, the lower are the average age accountability does not show up as an excessive of political parties (one measure of the credibility willingness to serve special interests at the of political competitors) and newspaper circula- expense of average citizens, but as the opposite: tion.20 The extent to which there are multiple an exceptional unwillingness to protect the large ethnic or linguistic groups in a country, rights of minorities. Subjective measures of eth- though, has no effect on ethnic tensions. nic tensions are dramatically higher in poorer Another key element of social polarization is democracies than in richer ones.19 Outright the ability of competing groups to make credible 318 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s commitments to each other. Previously quies- tutional weaknesses in countries.The inability of cent intergroup relationships can suddenly countries to secure property and contractual rights explode into conflict when the foundation for is a core element of these weaknesses. credible commitment crumbles. Bates, de Controversy emerged in the 1990s regarding Figuereido, and Weingast (1998) point to key two issues surrounding the growth­property events--such as the election of Slobodan Milo- rights debate:whether trade or other factors mat- sevic in the former Yugoslavia--that disrupt ter more to growth than does the security of arrangements that all groups believe have pro- property rights, and whether the security of tected them from aggression by other groups. property rights is rooted in countries' more fun- The interaction of political and ethnic effects damental geographical features.The first debate explains why social identity (ethnic, tribal, reli- is not resolved and may not be, bound up as it is gious, geographic) is often not politically salient in intractable problems (Rodrik, Subramanian, and has no discernible effect on policy. It also and Trebbi 2002; Dollar and Kraay 2003).21 The explains why, as Posner (forthcoming) shows for implications of the second debate loom larger. If Africa, the "identities" that matter for politics geography determines the security of property often shift within the same population. rights--that is, if geography is fate--the range of options for accelerating development is more limited.The role of geography is discussed below, 3. Government Credibility as a in the context of institutional and other determi- Prerequisite for Development nants of government credibility. All of the foregoing relates to the reluctance or . . . and Undermines Policy inability of political decision makers to adopt policies in the broad public interest. A related A vast array of government policies need to be problem for development emerges when poli- credible to be effective.A key problem in mone- cies, once enacted, are not credible. tary policy,for example,is the threat that the gov- ernment will enact a surprise increase in the money supply at the expense of economic agents Lack of Government Credibility that have signed long-term contracts.Anticipat- Undermines Growth . . . ing this, economic agents factor extra inflation The most notable effect of credibility is on invest- into their contracts, raising the long-term rate of ment and growth. Investors rely on government inflation (Barro and Gordon 1983). promises to respect investors' rights to their assets. Lack of government credibility also dampens When those promises are not credible,investments incentives to invest in public infrastructure or to slow down or take inefficient forms: power plants make other social investments. The payoffs to are set up on barges rather than on land; older these investments depend on the willingness of machinery is used at the expense of greater effi- economic actors to make complementary invest- ciency;innovation falls,in part because production ments that take advantage of them.Where expro- techniques are not at the cutting edge and in part priation is more likely,private investors are slower because the fruits of innovation are themselves to respond to improved public infrastructure,and vulnerable to expropriation.The growth effects governments correspondingly reduce their allo- are immediate: annual growth in income per cations to these investments. Keefer and Knack capita in poor countries with the most secure (2002) show that in countries with insecure property rights is between 2 and 4 percentage property rights, measured public investment is points faster than in poor countries with the least largely rent seeking.When property rights are secure property rights (Keefer and Knack 1997). insecure, an additional percentage point of pub- Earlier chapters in this report attribute the weak lic investment as a fraction of GDP significantly effects of policy reform on growth partly to insti- reduces the rate of growth of income per capita; D O E S D E M O C R AC Y H E L P ? 319 when property rights are secure, it adds 0.3 per- ments with short horizons. As noted earlier, centage point to the growth rate. Clague et al.(1996) show that among nonelected The security of property rights has a similar leaders, the longer is their horizon, the more effect on other long-term investments. In the likely they are to respect property rights.These 1990s, improved security of property rights results do not imply that governments should be increased gross secondary school enrollment by an immune to threats of removal.They do imply amount as large as did a similar increase in govern- that in countries where accountability mecha- ment expenditure.22 Confirming the link between nisms are flawed, extending the horizons of gov- rent seeking and public sector performance,Rajku- ernments by making them more secure in office mar and Swaroop (2002) find that child mortality may be the only means to create sufficient incen- rates and primary school attainment improve in tives to maintain secure property rights. response to increased public health and education spending only in countries with low corruption Sources of Low Government Credibility: and high bureaucratic quality. Political Institutions Sources of Low Government Credibility: Multiple institutional arrangements have been proposed to solve the problem of government Lack of Reputation and ShortTime credibility, but in the end, only political institu- Horizons tions-- particularly institutional checks and bal- What makes government policies credible? Cer- ances--have demonstrated a consistent effect on tainly the elements of political competition that the credibility of government decision making. allow political competitors to make credible pre- For example, Keefer and Stasavage (2003) find electoral promises help to ensure the credibility of that only in countries that exhibit political the policies they implement after they take office. checks and balances does the legal independence There is some evidence of this,if one accepts that of central banks suppress inflation. Moreover, the number of years that countries with compet- even in countries without legally independent itively elected governments have continuously central banks, political checks and balances can had such governments corresponds to the oppor- inhibit governments from reneging on monetary tunities that political competitors have had to commitments. Inflation is lower in countries build reputations.The greater the number of years with political checks and balances than in coun- of uninterrupted competitive elections, the more tries without them. established is the rule of law.23 This finding is consistent with a large body of However, many policies are not the subject of research on institutions and government credibil- preelectoral debate.Even when they are,the gains ity. North and Weingast (1989) argued that the from reneging on policies,once implemented,are introduction of checks on the English monarchy often greater than the gains from reneging on after the Glorious Revolution reduced the pre-electoral promises to implement them in the monarchy's ability to renege on sovereign debt first place.24 Finally, investors are always con- obligations to foreign creditors, and eventually cerned about a change in political control, from brought down interest rates.25 Acemoglu, John- politicians who have promised to support a par- son, and Robinson (2002) point, though less ticular policy to those who have not. explicitly,to the role of the right to vote and polit- The horizons of political actors--how long ical checks and balances as the key link between they expect to be in power or to be competing the economic drivers of political power (relative for power--can mitigate these additional threats prices and natural resource endowments) and the to credibility. Governments that expect to be in ultimate security of property rights.26 office many years have more to lose from current Engerman and Sokoloff (2002) make similar policies that upset future growth, such as invest- arguments.These authors highlight the role of ment-deterring expropriation, than do govern- natural endowments and other exogenous fac- 320 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s tors in determining the types of institutions that is evidence that the extent of imperfections in countries exhibit.The essential logic is that in political markets has a significant impact on eco- countries where ownership of economic activ- nomic growth. ity is concentrated in the hands of a few, and Easterly and Levine (1997) look at the effects assuming that political power follows control of on growth of one measure of social polariza- the country's most valuable assets, there is no tion: ethnolinguistic fractionalization.Although incentive for the politically and economically they do not explicitly consider the role of elec- powerful to enfranchise the powerless.The gains tions, they find that their polarization variable that the elite could reap--in the form of has a significant impact on economic growth. increased productivity of the masses, whose Keefer (2003b) provides evidence that newspa- greater political rights would protect them from per circulation, checks and balances, and the expropriation--would not offset the losses that number of continuous years of elections have a the elite would face in being forced to share the significant impact on economic growth. Ger- high returns from their plantations, mines, or oil ring, Brandt, and Bond (2003) similarly find wells. Hence, these countries grow slowly, and that--controlling for whether a country is the price of slow growth is borne by the disen- democratic or not (which has no impact on franchised. growth)--the total number of years that a If geography were the main determinant of country had elected governments through most institutional development, there would be little of the 20th century had a significant impact on purpose in institutional reform,and development growth. possibilities would be limited. But in practice Taken together, these results make a com- many countries have made institutional changes pelling case for reformers and development that seem to represent an escape from geograph- activists to take political market imperfections into ically determined destinies. One example is the account in designing strategies to speed growth wave of democratization in the 1990s.Another is and development. the spread of democracy in Latin America in the 1980s and 1990s, precisely where conditions for 4. Lessons: Making Politics democracy are supposed to have been the least Work for Policy When propitious. In all of these cases,the question remains why Governments Are Not the introduction of formal institutions is not suf- Credible and Electoral ficient to ensure sustained development across Markets Are Imperfect countries. Our earlier analysis suggests that the reason may be rooted in the underlying condi- How should we formulate strategies of policy tions of political competition. Improvement in reform, given imperfections in the market for these conditions, therefore, is likely to be an political office and limitations on the credibility important complement to institutional of government commitments?And what reforms reform.27 might mitigate these political and institutional problems directly? The traditional answer to the first question is Growth and Accountability to buy off the opposition to reform.This formula Boiled down to its essence, the foregoing argues requires political leadership: buying off the losers that elections alone are insufficient to ensure who are in a position to block reform, and accountability of governments to citizens.If this is exploiting windows of opportunity such as crisis true,one might expect to find a stronger relation- or a change in government.However,nothing in ship between growth and democracy if one takes the traditional formula hints at the fact that account of the different accountability mecha- reform may be systematically more difficult in nisms used in different democracies. In fact, there some countries and policy areas than in others. D O E S D E M O C R AC Y H E L P ? 321 Political Market Imperfections Explain block subsequent efforts to reduce the generous Why Buying off Reform Losers Usually pension. Reform has undermined the political Fails power that would allow them to enforce the agreement.28 Realizing that they would then be That compensation strategies have rarely suc- vulnerable to government efforts to recapture the ceeded is not surprising. First, the compensation pension from the now-disorganized workers, needed to persuade reform losers to support workers therefore reject the proposal. reform can be prohibitively high;for example,the Dramatic increases in prices are relatively easy benefits to the fertilizer industry of fertilizer subsi- to attribute to political failure (Keefer and Khe- dies in India amount to 0.7 percent of GDP each mani,forthcoming).But some other types of pol- year (Panagariya 2003).Even large payouts may be icy reforms--banking and social service delivery, feasible if the gains are correspondingly large. But for example--are more vulnerable to political when the imperfections in the market for political market imperfections and institutions that fail to office loom as large as they do in many countries, solidify government credibility. or when political institutions provide few checks Banking crises emerge after years of regula- on opportunistic behavior by politicians,adequate tory neglect and imprudent lending practices. It compensation may be impossible. If politicians is difficult to assign political responsibility for cannot make credible promises to voters,they can- them because these practices may occur under not make credible promises of compensation to multiple governments, and because politicians losers from reform. And if citizens are poorly can easily blame regulators for shirking and informed about what politicians do in office, los- bankers for criminal behavior. Such claims are ers may be unable to observe whether govern- difficult for voters to evaluate in every country, ments have actually delivered the promised so it is not surprising that the fiscal costs of bank- compensation. Hence remedying the underlying ing crises are almost exactly the same in poorer imperfections in electoral markets is a prerequisite and richer countries.29 for successful reform. Social services are also vulnerable to electoral Second, institutional deficiencies can also market failure.The goal of universal education is undermine compensation strategies. If reform exactly contrary to clientelist political motiva- losers control government decision making, they tions.30 It is quite difficult for citizens to assign cannot credibly promise to refrain from intro- blame to politicians for health and education ducing inefficient policies that benefit them at failures, which could be due to idiosyncrasies of the expense of others: once they have received individual health status, or to shirking by service compensation, nothing prevents them from providers, or to the fact that the country lacks reverting to policies that run counter to the pub- resources. lic interest. In this case, institutional reform is an The essential lesson is that for policies and essential prerequisite of policy reform. countries in which electoral market failures loom Third,a reform itself can undermine the bases large, reform efforts should focus on mitigating for making credible agreements. Consider an these failures rather than on paying off losers or effort to downsize a ministry or to close a encouraging leadership or awaiting the opening money-losing state-owned enterprise. If they are of windows of opportunity.Where market failures well organized, threatened workers can oppose are too large, the first may be too expensive, the these efforts by demonstrating, targeting contri- second unrealistic, and the third may never occur. butions to politicians who will help them, and purchasing advertisements to sway public opin- Mitigating Electoral Market Failures ion.To offset this opposition, reform proponents could offer the workers a large pension. Should What measures might alleviate imperfections in this proposal be accepted, the workers will be the markets for political office?31 Mitigating sent home. Once scattered, they cannot easily electoral market failures essentially means reduc- 322 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s ing politicians' incentives to engage in clientelist the provision of public goods (gross secondary behavior. How to shift political competition school enrollment).This result, consistent with away from clientelism is a key challenge of insti- those of other studies,suggests that the availability tutional reform that is not yet well understood. of information has its greatest impact on the pro- Some steps are probably key to reform,however: vision of transfers to voters, who can easily mon- increasing public information, and increasing the itor such transfers with the assistance of a thriving credibility of political promises. media industry. For example, even in societies with educated societies and unrestricted media, Increasing Public Information voters tend to be relatively uninformed about the An important step is to encourage, or remove specifics of government performance.While gov- impediments to, nongovernmental sources of ernment-controlled media are more likely to information on reform needs and direction.If their limit citizens to information favorable to the gov- credibility is established, such sources can validate ernment,private media can be controlled by spe- reform strategies outlined by interested political cial interests that have their own biases.This is less actors.The report cards undertaken in India by the problematic if there are low barriers to entry into Public Affairs Centre of Bangalore, mentioned in the news business,but even if barriers to entry are chapter 9,are one way in which nongovernmental low, it might be the case that other types of news organizations (NGOs) can credibly collect infor- are more profitable to report than information on mation about the performance of public officials government performance (Strömberg 2002). For and use it to stimulate reform. example, the media might prefer to report The media also appear to be key for increas- extreme outcomes that are not typical of the ing government responsiveness. Research on government's performance and that bias voter India and the United States during the Great perceptions. Depression highlights how information can Indirect evidence of this emerges from work improve access to government assistance. on campaign finance reporting in the United Between 1933 and 1935 in the United States, States.The research suggests that newspapers sys- federal assistance to low-income households was tematically bias the information that citizens significantly greater in those counties where receive about campaign finance.Specifically,news- more households had radios and were thus more paper reporting conveys the impression that politi- likely to be informed about government policies cians receive more contributions overall and a and programs.The spread of the radio particu- higher fraction from corporations than they in fact larly improved information access for rural vot- do. In turn, college-educated Americans--those ers, who had previously been disadvantaged who are most likely to read newspapers--believe relative to urban voters, with the latter's ready campaign financing flows are approximately what access to other information sources such as newspapers report, while less educated Americans newspapers. It accounted for as much as 20 per- believe they are considerably less (Ansolabehere, cent greater allocation of social assistance funds Snowberg, and Snyder 2004). to a rural county as compared to an identical The media may solve the following coordina- urban county (Strömberg 2001). Besley and tion problem:voters unhappy with a government, Burgess (2002) find that state governments in for whatever reason, may be reluctant to oppose India are significantly more responsive to the government if they think their own experi- declines in food production and crop flood dam- ence is isolated. By conveying a general impres- age via public food distribution and calamity sion of government performance to which all relief expenditure where newspaper circulation, voters are exposed, individual voters who share particularly in local languages, is greater. that impression can be more confident that others These findings raise some unresolved issues. share it as well.This reduces their reluctance to The studies suggest that information (newspaper support or oppose performing or nonperforming circulation) seems to have only a limited effect on governments. D O E S D E M O C R AC Y H E L P ? 323 Information reforms must also grapple with more difficult to address. In principle, political the conditions under which politicians respond credibility should provide political competitors to the revelation of information about their per- with a competitive advantage. Clientelism (the formance.Scandalous information frequently has default option for political competition when no political impact: even public knowledge of politicians cannot make credible promises) is criminal behavior by politicians is not a sufficient expensive.The resources needed to give 50 voters condition for politicians to leave office, in either jobs could finance broad public goods, such as developed or developing countries. Newspaper improved education, offering equivalent benefits circulation can reduce corruption, as seen earlier to hundreds of voters.That is,politicians who can in this chapter, but appears to have no effect in offer credible policy or public goods to a large countries that lack competitive elections.Among number of voters can defeat politicians who can countries with competitive elections, the influ- only operate in a patronage mode. ence of newspaper circulation on corruption However, moving out of clientelism is risky depends,though less robustly,on the existence of for politicians. Shifting resources to public goods political checks and balances.32 Competing may leave clients sufficiently dissatisfied to desert political forces inside government, each with the their patron, while public good benefits may right to influence government decisions, have materialize too slowly to attract new bases of both the incentive and ability to use evidence of political support before the next round of politi- each other's mal- or misfeasance for their own cal competition. In any case, the beneficiaries of political advantage. improved public services may not credit the The efficacy of other information reforms incumbent politician for the improvements. also depends on the political environment.A key How can politicians build the credibility of characteristic of government in many developed their promises to improve the quality of public countries is the transparency with which new goods? Leaders can build credibility by being regulations emerge from the executive branch of vocal, emphatic, and specific about their reform government.These range from the issuance of goals. Specificity makes it easier for citizens to white and green papers in the United Kingdom judge when leaders have failed.Emphasis makes it to open meeting requirements and freedom of clear that leaders expect to be judged on their per- information laws.These transparency require- formance regarding these goals, rather than on ments are almost always imposed by politicians other issues, and independent of shocks or diffi- on themselves, and are potentially but not always culties that might emerge.Together these improve enforceable by courts.This means, however, that credibility. As is often the case, there may be a the requirements have less effect to the extent tradeoff between reform success and building up that there is little political cost to politicians who credibility. Publicizing reform may incite resist- decide not to abide by them and to the extent ance that stifles reform, while successful reform that the courts are reluctant to require adherence undertaken unpublicized has fewer political bene- to them. Unfortunately, the political costs of fits and may be less sustainable. ignoring transparency laws are likely to be lowest Public sector reform can help too.A political precisely where government performance in competitor is unlikely to promise improved pro- general is likely to be poor: where there are few vision of public goods if the organization needed political checks and balances, and where political to supply those goods is dysfunctional, since citi- competition is organized around clientelist favors zens cannot easily distinguish whether reform rather than overall government performance. failure is caused by bureaucrats or politicians.33 Unfortunately, public sector reform is itself an Increasing the Credibility of Political arduous process that requires political commit- Promises ment. In systems where politicians have a strong The other major electoral market imperfection-- political interest in satisfying clientelist demands, the lack of credibility of political competitors--is their incentives to improve the functioning of 324 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s the civil service are weak. Cox (1987) demon- construct patronage machines or vehicles for strates that the professionalization of the justly personal advancement rather than rely on the acclaimed British Civil Service followed, rather institutionalization of their party's stance on pol- than preceded, the shift in the basis of political icy issues (box 10.2). competition from clientelism to partisan or pol- icy differences. Mitigating Political Market Failures: Donors can help developing-country gov- Institutional and Legal Reforms ernments with this dilemma by coordinating their assistance for public sector reforms with Even though institutional factors do not system- their assistance for improving the provision of atically explain the underperformance of some public goods, while being sensitive to the politi- democracies relative to others, institutional cal timetable according to which citizens express reforms can promote policy reform. Such their judgments about these reforms.A successful reforms include changing electoral rules, rein- reform strategy is one that devises and links "on- forcing checks and balances, introducing laws the-ground" outcomes to intermediate stages of that regulate campaign contributions,and decen- public sector reform,such that politicians can get tralization. credit for reform in a timely fashion.(Again,gov- ernment leaders need to be vocal in promising Electoral Reforms Can Spur Sustainable results, or the credibility effects will be dimin- Policy Reform ished and voters will have little reason to change Reform of electoral laws can both spur reform their judgments about incumbents based on the and serve as a vehicle for mitigating electoral reform experience.) market imperfections.One indication of the pol- Donor strategies for project implementation icy effects of such reform emerged in the 1990s are relevant here.A donor focus on specific proj- in Japan. Prior to its 1994 reform, the electoral ects touching a fraction of the population, rather system in Japan was a mix of plurality voting and than on broad policy goals and public good multimember districts that essentially compelled improvements that affect most of the population, candidates from the same political party to com- may accelerate project implementation, but it pete with one another. Because they could not reinforces patterns of political accountability in use party labels to distinguish themselves from which voters expect only targeted or clientelist competitors, candidates spent considerable sums benefits from their leaders. Donors also often of money distinguishing themselves in other agree with governments to set up enclaves of ways, thereby building up personal constituen- bureaucratic excellence to carry out particular cies.These constituencies had clientelist attrib- tasks. As discussed in chapter 9, while enclaving utes. Politicians, for example, would appear at can assure governments that promises will be car- weddings and funerals, making cash contribu- ried out, enhancing the sustainability of reform, tions to the newly married or bereaved.Their this potential benefit is rarely realized, since the need for financial resources led incumbent politi- end of a task often means the dissolution of the cians to be especially generous toward special enclaved agency. interests, including the banking industry.The lax The development community is also doing regulatory standards to which banks were held considerable work to "institutionalize" political contributed to soaring nonperforming loans. parties, improving their ability to communicate These were exposed when rapid economic with voters or to organize at the grassroots level. growth ground to a halt in 1990. This is potentially important for achieving the The electoral reform of 1994 introduced sin- ultimate goal of improving policy credibility and gle-member districts and changed rules in multi- voter information about the policy stances of member districts to proportional rather than political competitors.However,it has no guaran- plurality electoral rules.These changes raised the tee of success, since party leaders may prefer to electoral value of partisan affiliation and reduced D O E S D E M O C R AC Y H E L P ? 325 BOX 10.2 Political Parties and Reform I n many countries political parties are suppressed ers can more easily assign blame and credit to the or limits are placed on the extent to which they parties in control, relieving them of the need to iden- can make ethnic appeals. Candidates in some tify specific individuals to hold responsible. elections in Pakistan and Uganda have been prohib- Unfortunately, history is replete with parties ited from running under a party affiliation. In Bul- hijacked by personal interests or dedicated to patron- garia, ethnically based parties have been excluded. age politics or serving as a locus for ethnic rivalry or While parties are far from a sufficient condition religious conflict. Parties often fail to offer voters a for eliminating electoral market imperfections, they credible choice in terms of economic policies. may be necessary. Mature political parties with well- At the same time, policy-based political parties defined positions on economic and social issues help can emerge from or succeed in a clientelist milieu, as solve problems of both information and credibility may be indicated by the fall of the PRI in Mexico from that otherwise plague competition for political dominance and the persistence of the Partido dos Tra- office. Mature parties convey information to voters balhadores (Workers Party) in Brazil. Nor is policy on the policy stances of party members, particularly reform impossible in clientelist environments--the relative to members of other parties. Unlike individ- most effective means for politicians to capture the ual candidates, they are more likely to have policy vast majority of disaffected voters who do not bene- reputations that allow them to make credible prom- fit from clientelist payoffs is to develop a reputation ises to voters. When parties are credible entities, vot- for policy performance that benefits the majority. the need for money in campaigns. Soon after, in authority over government decision making; and 1996,the ruling Liberal Democratic Party forced they require that those individuals or branches banks to bail out their mortgage-lending sub- enjoy independent sources of political authority. sidiaries and absorb huge losses rather than The first is relatively easy to accomplish, through socialize the losses with taxpayer-financed statutory or constitutional amendment.The sec- bailouts.34 Policy reforms that had been urged ond is difficult. on Japan for years finally occurred, but only after When formal institutions of checks and bal- the adoption of institutional reforms that ances are present (formal legislative, judicial, and changed political incentives. Whether such executive branches of government, for example), reforms would have an equal effect in countries steps can be taken to reinforce checks and bal- without well-established political parties and ances even when the branches do not enjoy inde- informed voters is less clear. pendent sources of political authority, as is often the case.Their effects are likely to be small, how- Checks and Balances: Difficult to Introduce, ever, until political authority is more equally Easy to Undermine shared among the branches. Political checks and balances have a significant For example, public sector financial manage- effect on government credibility and, as a conse- ment reforms increase the information available quence, on the effects of policies in areas ranging to legislators inside government. Often, though, from taxation to public investment and monetary legislators have no incentive to act on this infor- policy.35 It is difficult to introduce political checks mation: their prospects for reelection depend on and balances where none exist, however.They maintaining good relations with the executive require both formal institutions that endow mul- branch, such that the executive branch will fund tiple branches or individuals of government with projects in their constituencies, and this weakens 326 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s their incentives to supervise the executive's over- reforms. In the United States, caps on one form all performance.This dependency is less impor- of contribution have led to dramatic increases in tant in countries (such as the United Kingdom) other forms.Even when caps are comprehensive, where strong parties provide an offsetting check as in France and Germany, reports on campaign on political excess. But without strong political finance scandals suggest that the flows continued parties, budget rules that deny legislator influ- nonetheless. Evasion and nonenforcement are ence over spending undermine political checks more likely in countries in which politics is and balances. clientelist and large policy issues are not germane Moreover, formal institutions are often to political competition. incomplete in these circumstances: budget- Even where caps are binding, some observers making procedures deny them the policy-mak- argue that they actually increase the returns to lob- ing leverage they need to act on the bying.Drazen,Limão,and Stratmann (2004) argue information. Legal and constitutional changes that moderate caps on political contributions can that endow legislatures with very limited induce more lobbyists to enter the political mar- authority over spending prevent them from ket, offsetting the reduction in contributions by imposing budgetary sanctions on government existing groups.They find some evidence for this, ministries that diverge from agreed allocations moreover, across U.S. states, which exhibit sharply and amounts. different campaign finance regulations. Where political checks and balances are weak, implementation of reforms--or of donor-sup- Decentralization: Finding More Perfect ported projects--is more likely to be undermined. Political Markets Closer donor supervision is the most effective Decentralization embraces a range of institutional short-run response to avoid this.At the same time, reforms that have the possibility both of upsetting political checks and balances are not a substitute clientelist political patterns and of reinforcing for solving electoral market failures.Among coun- them.To the extent that political competition and tries that exhibit political checks and balances,the decision making are less subject to political market rule of law and corruption are still strongly imperfections--information, credibility, or social affected by variables that capture the effects of polarization--in subnational than in national gov- some of these failures. However, their absence ernments, policy outcomes are likely to be more undermines prospects for sustainable reform and conducive to development. Similarly, by splitting their development is therefore important. up issues between national and subnational gov- ernments, decentralization facilitates voter efforts Campaign Finance Reform: Attacking the to hold politicians accountable for specific policy Symptom, If Not the Disease areas, and also assists political efforts to develop Other institutional reforms can reduce both policy reputations that go beyond clientelism.36 electoral market failures and the lack of credibil- However, these preconditions for successful ity--although they can potentially exacerbate decentralization are frequently absent,and in their them as well. One is campaign finance reform. absence decentralization can exacerbate the policy Popular in both developed and developing coun- distortions of clientelism. tries, the general notion is that to prevent special interests from using money to distort political outcomes, one must place caps on campaign 5. Conclusion finance or increase public financing of elections. The evidence is not in on the efficacy of either The arguments in this chapter paint a broad pic- solution, though the latter is likely to be more ture of the role of political economy in develop- effective than the former. ment and highlight a few characteristics of Evasion and enforcement have everywhere political systems that help explain some develop- been a serious problem with campaign finance ment outcomes: D O E S D E M O C R AC Y H E L P ? 327 · Can voters observe the decisions of government 4. These results persist even when countries in Eastern officials and the effects of these decisions? Can Europe and Central Asia are excluded. 5. The effect is nearly as large as that of a standard and even informed observers attribute political always powerful control, income per capita: an responsibility for policy failure? They cannot if increase of one standard deviation in the years a non- political parties are amorphous and individual elected leader is in office reduces the risk of expro- participation in political decision making is priation almost as much as does one standard opaque. deviation in a country's income. 6. Controlling for total population,population living in · Are policy differences at all relevant to politi- rural areas,land area,population under the age of 16, cal competition? Do party platforms exist and purchasing power parity­adjusted income per and, if so, do they diverge? Can the average capita, and looking only at countries that did not exhibit fully competitive elections in 1995, the citizen recognize and rely on policy differ- absence of a legislature of any kind, elected or not, ences among the parties? If not,political com- was associated with a one standard deviation worsen- petition is sure to focus on the allocation of ing of the rule of law and corruption measures. narrowly targeted benefits--projects, jobs, 7. Persson andTabellini (2000). exemptions from onerous regulations--and 8. Kontopolous and Perotti (1999) and Persson, promises of broadly based reform are unlikely Roland, andTabellini (2003) argue that proportional representation systems encourage small parties, to be credible. which increases the prevalence of minority govern- · Are checks and balances present and operative? ments or multiparty coalition governments,which in turn increases taxes and spending. Majoritarian sys- tems, as argued by Milesi-Ferretti, Perotti, and Ros- These questions are important in seeking to tagno (2002) and others, should lead to greater understand societies' collective decision-making attention to pivotal voters, and therefore more tar- process. geted spending,rather than spending on broad-based public goods or redistributive programs. 9. Persson and Tabellini (2000) argue that vote of con- Notes fidence procedures in parliamentary democracies bind legislative majorities together,allowing them to 1. Countries are regarded as having competitively make credible agreements that taxes raised will serve elected governments if they are reported in Beck et the interests of the majority.This encourages them to al. (2001), Database of Political Institutions, as having establish higher taxes and spending. In presidential the highest score (seven) on the Executive Index of systems,legislative minorities (for example,chairper- Electoral Competition (EIEC) and on the Legislative sons of legislative committees) are more powerful, Index of Electoral Competition (LIEC), where seven but they are assumed to be unable to make credible implies that there are multiple parties competing and agreements with each other. Spending is targeted to no party gets more than 75 percent of the vote.The the constituencies of these legislative minorities, but rule of law measure is from Political Risk Services' because they cannot credibly agree with one another International Country Risk Guide. that higher tax revenues will be targeted to their con- 2. These same sources of heterogeneity, especially infor- stituencies, overall spending is lower. mation and social polarization, may also matter in 10. Elections in which there were multiple competing nondemocracies. This possibility is not explored candidates or parties, more than one party contest- below--there is little evidence bearing on the ques- ing,and no candidate or party winning more than 75 tion--but future work needs to explore the overarch- percent of the vote, taken from the Legislative and ing determinants of good government performance Executive Indexes of Electoral Competition,in Beck that might be common to both democracies and non- et al. (2001). democracies. 11. This result uses institutional data from Beck et al. 3. Banerjee and Iyer (2002) show that British colonial (2001) and economic and social data from World practices affect land tenure relationships and land Development Indicators. productivity in parts of India to this day.Acemoglu, 12. Education spending, controlling for primary school Johnson, and Robinson (2002) argue that a complex enrollment, has a small effect on secondary school interaction of relative price changes, natural endow- enrollment. One estimate suggests that a full per- ments,and institutional choice has consequences that centage point increase in education spending as a last for generations. fraction of GDP (where the average country spends 328 E C O N O M I C G ROW T H I N T H E 1 9 9 0 s approximately 3.3 percent of GDP) increases gross cal development. It does not explain either corrup- secondary school enrollment by fewer than 5 per- tion or secondary school enrollment but is strongly centage points (where gross secondary school enroll- related to years of continuous competitive elections. ment in the average country is approximately 65 19. One and one-half standard deviations higher,using a percent). measure of ethnic tensions from the International 13. These results are from an ordinary least squares Country Risk Guide. regression of gross secondary school enrollment on 20. The regression controls for income per capita, per- PPP-adjusted income per capita; the percentage of centage of the population that is young or rural,land the population that is young;land area;gross primary area, and total population of a country, for all years school enrollment; whether a system is parliamen- since 1989. tary, presidential, or semipresidential; the voting rule 21. Rodrik, Subramanian, and Trebbi (2002) argue that used to elect the majority of representatives in the the security of institutions (measured as the security lower chamber of the legislature;and the average dis- of property rights, the rule of law, and so on) matters trict magnitude of the chamber.The data are yearly, more for economic development than geography from 1990 to 2000.Significance tests based on robust and trade. Dollar and Kraay (2003) show that the standard errors assuming country observations from instrumental variables used to control for the endo- different years are not independent.The economic geneity of both trade and measures of governance or variables are from World Development Indicators; the the security of property rights yield a high correla- political variables from the Database of Political Insti- tion between the two, making their independent tutions (Beck et al. 2001). effects difficult to assess. 14. In neither group are electoral institutions a significant 22. Results from regressing yearly data on gross second- determinant of gross secondary school enrollment. ary school enrollment from 1990 to 1997 on gross Specification is as described in footnote 15. Standard primary school enrollment, PPP-adjusted income errors are White-corrected (robust).The economic per capita, land area, the fraction of the population variables are from World Development Indicators; the that is young,total population,education spending as political variables from the Database of Political Insti- a fraction of GDP, and expropriation risk. tutions (Beck et al. 2001). 23. Controlling for PPP-adjusted income per capita, 15. The sample includes only country-years in which total population, the fraction of the population that countries had competitive elections, since for coun- is young, land area, whether the political system is tries for which this is not the case, there is no obvi- presidential or parliamentary, whether the electoral ous reason that would impel a government to allow system is proportional or plurality, and the average a free press that reports on how well the government district magnitude (yearly data only for countries that is performing. have leaders chosen by competitive elections). 16. A common refrain of voters in many countries is,"All 24. For example,the gains from reversing tax cuts meant politicians are the same, and none is interested in the to encourage investment are potentially substantial people." Uninformed voters would naturally express after fixed investments have been made in response this opinion. So too would voters confronting politi- to the tax cut. cians who cannot make credible promises. 25. Stasavage (2003) revisits this episode and concludes 17. From cross-sectional regressions using democratic that parliament only restrained opportunistic behav- episodes (1975­2000) as the units of observation, ior by the government when the minority of parlia- controlling for land area,total population,percentage mentary members who favored honoring sovereign of the population that is young, and political and obligations were able to make a deal involving reli- electoral systems. gious freedom with those who were less favorable. 18. See, for example, Rajkumar and Swaroop (2002). It That is, he shows that not only did institutions mat- is possible that the relationships between continuous ter, but so did politics. years of competitive elections and corruption or 26. This point is explicit in the theoretical work of Ace- education, for example, are due to omitted effects moglu and Robinson (2001),who argue that only by that in turn influence both of these.One can control sharing power can the disenfranchised be persuaded for this possibility by identifying instrumental vari- that the enfranchised will not expropriate them. ables that explain competitive elections but not edu- 27. Acemoglu and Robinson (2001) explain the poor cation or corruption. Results are robust to performance of some democracies by arguing that, instrumental variable estimation. using the share of in countries exhibiting high inequality, as in Latin nonmanufacturing activity in total industrial activity America, democratization would give rise to signifi- in a country in 1965 and/or 1975.These capture cant redistribution and lay the groundwork for reliance on natural resources (especially mining), democracy's collapse as the elites aimed to take back which in turn is often thought to discourage politi- power. Certainly, Latin American democracy D O E S D E M O C R AC Y H E L P ? 329 throughout the 20th century has been notably unstable. held, the effect of an increase in newspaper circula- However, simple averages of government expenditure tion on corruption approximately doubles, moving as a fraction of GDP and of education spending specif- from a parliamentary political system in which the ically, as a fraction of GDP, show little difference party of the prime minister party controls the legis- between democratic and nondemocratic periods since lature to one in which a four-party coalition govern- 1975. If anything, government spending was slightly ment is in power. higher in the nondemocratic country­years than in the 33. See Shepherd (2003) for a thorough review of the democratic; education spending was almost identical. argument that meritocratic and well-performing 28. See, for example,Acemoglu and Robinson (2000). civil servants improve government credibility. 29. The total cost is 12 percent of GDP. Data from Hono- 34. Rosenbluth andThies (2001). han and Klingebiel (2000), data on real incomes from 35. Development assistance can have the unfortunate Aten, Heston, and Summers (2001). side effect of undermining political checks and bal- 30. Social funds, in contrast, which are intended to distrib- ances where they do exist. Chapter 9 described a ute resources to particular groups, are potentially useful number of reforms, particularly medium-term to clientelist politicians. Schady (2000), for example, expenditure frameworks, that are meant to ensure shows that the Peruvian social fund, FONCODES, was that all public spending is subjected to the scrutiny of well targeted to the poor,conditional on the poor resid- multiple actors in the political system. It also identi- ing in areas where President Fujimori thought political fies the hazards of funneling outside resources transfers would be most useful.The poor in opposition directly to line ministries,outside the normal budget strongholds were particularly unlikely to receive funds. processes. 31. In advocating efforts to organize the demand for 36. Besley and Coate (2001) argue in the context of cit- reforms, chapter 9 points out that reform winners are izens' initiatives that a key problem in politics is that often disorganized and confront significant barriers to governments make decisions on numerous policy the collective action that would make them effective dimensions, but voters can only cast votes for a sin- supporters of reform. Lowering these barriers is one gle politician or party.They are confronted, there- recommendation of chapter 9. The political market fore, with an "all or nothing" offer: politicians can imperfections and institutional deficiencies discussed shirk on some margins, but still be reelected if they here are additional obstacles to successful reform, and are sufficiently forthcoming on the "salient" dimen- also undermine the efforts of reform proponents to shift sions of policy. Decentralization eases this problem government policy. by splitting issues between multiple levels of govern- 32. Controlling for income per capita,population variables, ment, allowing more policies to become politically and the years that elections have been continuously salient than would otherwise be the case. Bibliography Aalt,J.,and D.Lassen.2003."FiscalTransparency and Fis- North African Countries, 1980­2000." World Devel- cal Policy Outcomes in OECD Countries." Eco- opment 31(12): 2027­48. nomic Policy Research UnitWorking Paper 2003-02. Adelman, Irma, and Cynthia Taft Morris. 1965. A Factor Abiad,Abdul, and Ashoka Mody. 2002."Status Quo Bias Analysis of the Interrelationship between Social and Politi- in Financial Reform." IMF Working Paper (Octo- cal Variables and per Capita Gross National Product. ber).Washington,D.C.:International Monetary Fund. Evanston, Ill.: Northwestern University Press. ------. 2003."Financial Reform:What Shakes It? What Adserà,Alícia, Carlos Boix, and Mark Payne. 2003."Are Shapes It?" IMF Working Paper WP/03/70 (April). You Being Served? Political Accountability and Gov- Online at http://www.imf.org/external/pubs/ft/wp/ ernmental Performance."Journal of Law,Economics,and 2003/wp0370.pdf. Organization 19(2): 445­90. Abramovitz, Moses. 1956."Resource and Output trends Aghion, Philippe. 1998. "Inequality and Economic in the United States since 1870." American Economic Growth." In Philippe Aghion and Jeffrey G. Review 46(2): 5­23. Williamson, Growth, Inequality, and Globalization:The- Acemoglu,Daron,and James Robinson.2000."Why Did ory, History, and Policy (5­102). Raffaele Mattioli Lec- the West Extend the Franchise? Democracy, Inequal- tures, Cambridge. New York and Melbourne: ity, and Growth in Historical Perspective." Quarterly Cambridge University Press. Journal of Economics 115(4): 1167­99. Aghion, Philippe, and Peter Howitt. 1998. Endogenous ------. 2001."A Theory of Political Transitions." Ameri- GrowthTheory.Cambridge,Mass.:Massachusetts Insti- can Economic Review 91(4): 938­63. tute ofTechnology Press. Acemoglu, Daron, Simon Johnson, and James A. Robin- Aghion, Philippe, Eve Caroli, and Cecilia Garcia son. 2001. "The Colonial Origins of Comparative Penalosa. 1999."Inequality and Economic Growth: Development:An Empirical Investigation." American The Perspective of the New Growth Theories." Jour- Economic Review 91(5): 1369­1401. nal of Economic Literature 37(4): 1615­60. ------. 2002. "Reversal of Fortune: Geography and Ahluwalia,M.S.2003."IMF Operations and Democratic Institutions in the Making of the Modern World Governance: Some Issues." Online at http://www. Income Distribution." Quarterly Journal of Economics imf.org/external/np/ieo/2003/sp/110103.htm. 117(4): 1231­94. Ahmed,Belal.2001."The Impact of Globalization on the ------. 2003."An African Success Story: Botswana." In Caribbean Sugar and Banana Industries." Society for Dani Rodrik, ed., In Search of Prosperity:Analytic Nar- Caribbean Studies Annual Conference Papers,Vol. 2. ratives on Economic Growth. Princeton, N.J.: Princeton Aitken, Brian J., and Ann E. Harrison. 1999."Do Domes- University Press. tic Firms Benefit from Direct Foreign Investment? Evi- Adams, D., and R.Vogel. 1986."Rural Financial Markets dence fromVenezuela."American Economic Review 89. in Developing Countries: Recent Controversies and Aitken, Brian J., Gordon Hanson, and Anne E. Harrison. Lessons." World Development 14: 477­87. 1997. "Spillovers, Foreign Investment, and Export Adams, D., D. Graham, and J. D.Von Pischke, eds. 1984. Behavior." Journal of International Economics 43: Undermining Rural Development with Cheap Credit. 103­32. Boulder, Colo.:Westview Press. Aizenman, Joshua, and Brian Pinto. 2004. "Managing Adams, Richard, Jr., and John Page. 2003. "Poverty, Volatility and Crises." Draft chapter (March) for Man- Inequality, and Growth in Selected Middle East and agingVolatility and Crises:A Practitioner's Guide.Online at 331 332 BIBLIOGRAPHY http://www-wbweb.worldbank.org/prem/prmep/ Appiah-Kubi, K. 2001."State-Owned Enterprises and economicpolicy/documents/mv/pgchapter00.pdf. Privatization in Ghana."Journal of ModernAfrican Stud- Alba, P., L. Hernandez, and D. Klingebiel. 1999. Financial ies 39(2). Liberalization and the Capital Account: Thailand Arioshi,A., K. Habermeier, B. Laurens, I. Otker-Robe, J. 1988­1997. World Bank Policy Research Paper No. Canales-Krijenko, and A. Kirilenko. 2000."Capital 2188. Controls: Country Experiences with Their Use and Alcala, Francisco, and Antonio Ciccone. 2004."Trade and Liberalization." Occasional Paper No. 190. Interna- Productivity." Quarterly Journal of Economics 119(2): tional Monetary Fund,Washington, D.C. 613­46. Ashraf, Nava, Margaret McMillan, and Alix Peterson- Alesina, A.,V. Grilli, and G. Milesi-Ferreti. 1994."The Zwane. Forthcoming."My Policies or Yours: Have Political Economy of Capital Controls." In L. Lieder- OECD Agricultural Policies Affected Incomes in man and A.Razin,eds.,Capital Mobility:The Impact on Developing Countries?" In Harrison (forthcoming). Consumption, Investment, and Growth. Cambridge, Asian Development Bank. 2000. Asian Development Out- U.K.: Cambridge University Press. look. Manila, Phillippines:Asian Development Bank. Alesina, Alberto, and Roberto Perotti. 1996. "Budget Aten, Bettina,Alan Heston, and Robert Summers. 2001. Deficits and Budget Institutions." NBER Working PennWorldTables 6.0.Philadelphia:University of Penn- Paper No. 5556. National Bureau of Economic sylvania, Center for International and Interarea Com- Research, Cambridge, Mass. parisons. Alesina,Alberto, and Lawrence Summers. 1993."Central Auty, R., and R. Mikesell. 1998. Sustainable Development Bank Independence and Macroeconomic Perfor- in Mineral Economies. Oxford: Clarendon Press. mance: Some Comparative Evidence." Journal of Bacalla,Tess. 2003."BIR Officials Amass Unexplained Money, Credit and Banking (May 25): 151­62. Wealth." Philippine Center for Investigative Journal- Alesina,Alberto, Reza Baqir, and William Easterly. 1999. ism. Reprinted in Malaya, Manila Times, Manila Stan- "Public Goods and Ethnic Divisions." Quarterly Jour- dard, Cebu Daily News and Sun-Star Daily Cebu, May nal of Economics 114(4): 1243­84. 12­14; in Abante and BusinessWorld, May 13­14. Alesina,Alberto,Sule Ozler,Nouriel Roubini,and Philip Online at http://www.pcij.org/stories/2003/bir Swagel. 1996. "Political Instability and Economic .html. Growth." Journal of Economic Growth 1: 189­211. Bahl, R. 2000."How to Design a Fiscal Decentralization Alfaro, Laura, Areendam Chanda, Sebnem Kalemli- Program."In S.Yusuf,W.Wu,and S.Evenett,eds.,Local Ozcan, and Selin Sayek. 2000."FDI and Economic Dynamics in an Era of Globalization.NewYork:Oxford Growth:The Role of Local Financial Markets."Work- University Press. ing Paper 01-083, Harvard Business School. Bahl, R., and Paul Smoke. 2003. Restructuring Local Gov- Allen, Donald S., and Michelle T. Gyles. 1995."Trade, ernment Finance in Developing Countries; Lessons from Growth,and Capital:A Case Study of Jamaica."Work- South Africa. Cheltenham, U.K.: Edward Elgar. ing Paper 1995-012B, Federal Reserve Bank of St. Baier, Scott L., and Jeffrey H. Bergstrand. 2001. "The Louis. Growth of World Trade:Tariffs,Transport Cost, and Allen, Richard, Salvatore Schiavo-Campo, and Thomas Income Similarity." Journal of International Economics Columkill Garrity. 2004. Assessing and Reforming Pub- 53. lic Financial Management:A NewApproach.Washington, Balassa, Bela. 1971. The Structure of Protection in Develop- D.C.:World Bank. ing Countries. Baltimore: Johns Hopkins University Anderson,J.,and D.Marcouiller.1999."Trade,Insecurity, Press. and Home Bias." NBER Working Paper 7000. ------. 1978."Exports and Economic Growth: Further National Bureau of Economic Research,Cambridge, Evidence."Journal of Development Economics 5:181­89. Mass. Balasubramanyam,V. N., M. Salisu, and D. Dapsoford. Anderson, Jim, Gary Reid, and Randi Ryterman. 2003. 1996."Foreign Direct Investment and Growth in EP "Understanding Public Sector Performance in Tran- and IS Countries." Economic Journal 106. sition Countries:An Empirical Contribution."World Balat, Jorge, and Guido Porto. 2004."Globalization and Bank,Washington, D.C. Complementary Policies. Poverty Impacts in Rural Andreasson, Bo. 1998. "Privatization in Sub-Saharan Zambia." In Harrison (forthcoming). Africa: Has It Worked and What Lessons Can Be Baldwin, E. Robert. 2003. "Openness and Growth: Learnt?" Swedish Development Advisers, Gothen- What's the Empirical Relationship?"NBERWorking burg, Sweden. Online at www.swedevelop.com. Paper No. 9578. National Bureau of Economic Ansolabehere, Steven, Erik Snowberg, and James Snyder. Research, Cambridge, Mass. Published (2004) in 2004."Statistical Bias in Newspaper Reporting:The Robert E. Baldwin and L.Alan Winters, eds., Chal- Case of Campaign Finance." Paper presented at the lenges to Globalization: Analyzing the Economics. 2004 Meeting of the Public Choice Society,Baltimore. Chicago: University of Chicago Press. BIBLIOGRAPHY 333 Balino,T.,A. Bennett, E. Borensztein, and others. 1999. Baulch, Robert, and John Hoddinott. 2000. Economic "Monetary Policy in Dollarized Economies." IMF Mobility and Poverty Dynamics in Developing Occasional Paper No. 171. International Monetary Countries. Journal of Development Studies 36: 1­24. Fund,Washington, D.C. Bayless,Charles E.1994."Less Is More:Why GasTurbines Bandiera, Oriana, Gerard Caprio, Patrick Honohan, and WillTransform Electric Utilities." Public Utilities Fort- Fabio Schiantarelli. 2000."Does Financial Reform nightly 132(22): 21­25. Raise or Reduce Saving?" Review of Economics and Beck, Thorsten, George Clarke, Alberto Groff, Philip Statistics 82(2) (May): 239­63. Keefer, and Patrick Walsh. 2001."New Tools in Com- Banerjee,Abhijit,and Esther Duflo.2003."Inequality and parative Political Economy:The Database of Political Growth:What Can the Data Say?" Journal of Economic Institutions."World Bank Economic Review 15(1):165­76. Growth 8: 267­99. Beckerman,P.1995."Central-Bank`Distress'and Hyper- Banerjee,Abhijit,and Lakshmi Iyer.2002."History,Insti- inflation in Argentina, 1989­90." Journal of Latin tutions, and Economic Performance:The Legacy of American Studies 27: 663­82. Colonial Land Tenure Systems in India." Working Bekaert, G., C. Harvey, and C. Lundblad. 2003."Equity Paper, Department of Economics, Massachusetts Market Liberalization in Emerging Markets." Federal Institute ofTechnology. Reserve Bank of St. Louis Review 85: 53­74. Banerjee,AbhijitV., Paul J. Gertler, and Maitreesh Ghatak. Bell, Stuart. 1995. FPD Note, "Privatization through 2002. "Empowerment and Efficiency: Tenancy Broad-Based Ownership Strategies." World Bank, Reform in West Bengal." Journal of Political Economy Washington, D.C. March. 110(2):239­80. Ben-David,Dan.1993."Equalizing Exchange:Trade Lib- Banerjee,Abhijit,Dilip Mookherjee,Kaivan Munshi,and eralization and Income Convergence." The Quarterly Debraj Ray. 2001."Inequality, Control Rights, and Journal of Economics 108(3): 653­79. Rent Seeking: Sugar Cooperatives in Maharashtra." Ben-David, Dan, and David H. Papell. 1994."The Great Journal of Political Economy 109(1). Wars, the Great Crash, and the Unit Root Hypothe- Bardhan, Pranab. 2002."Decentralization of Governance sis:Some New Evidence about an Old Stylized Fact." and Development." Journal of Economic Perspectives NBERWorking Paper No. 4752. National Bureau of 16(4): 85­205. Economic Research, Cambridge, Mass. Bardhan, Pranab, and Dilip Mookherjee. 2000."Capture Benston,G.,and G.Kaufmann.1997."FDICIA after Five and Governance at Local and National Level." Ameri- Years." Journal of Economic Perspectives 11: 138­58. can Economic Review 90(2): 135­39. Berezlin, Peter,Ali Salehizadeh, and Elcior Santana. 2002. Barr,Nicholas.2000."Reforming Pensions:Myths,Truths, "The Challenge of Diversification in the Caribbean." and Policy Choices."IMFWorking PaperWP/00/139. IMFWorking Paper No.02/196.World Bank,Washing- International Monetary Fund,Washington,D.C. ton,D.C. Barro, Robert J. 1991."Economic Growth in a Cross- Berg,Andrew, and Anne Krueger. 2003."Trade, Growth Section of Countries." Quarterly Journal of Economics and Poverty:A Selective Survey." IMFWorking Paper 106(2): 407­43. No. 03/30. International Monetary Fund,Washing- ------. 1996."Democracy and Growth." Journal of Eco- ton, D.C. nomic Growth 1: 1­27. Bernard, Andrew, and J. B. Jensen. 1999. "Exceptional Barro, Robert, and David Gordon. 1983. "A Positive Export Performance: Cause, Effect, or Both?" Journal Theory of Monetary Policy in a Natural Rate of International Economics 47(1): 1­25. Model." Journal of Political Economy 91: 589­610. Besley, Tim, and Robin Burgess. 2002. "The Political Barth, J., G. Caprio, and R. Levine. 2001. Regulation and Economy of Government Responsiveness:Theory Supervision of Banks around theWorld:A New Data Base. and Evidence from India." Quarterly Journal of Eco- World Bank Policy Research Paper No. 2588. nomics 117(4): 1415­52. Barzel,Yoram. 1989. Economic Analysis of Property Rights. Besley,Tim,and Stephen Coate.2001."Issue Unbundling Cambridge, U.K.: Cambridge University Press. via Citizens' Initiatives." Center for Economic Policy Bates,Robert H.1981.Markets and States inTropicalAfrica: Research Paper No. 2857 (June),Washington, D.C. The Political Basis of Agricultural Policies.Berkeley:Uni- Bhagwati, Jagdish. 1978. Foreign Trade Regimes and Eco- versity of California Press. nomic Development: Anatomy and Consequences of ------. 1983. Essays on the Political Economy of Rural Exchange Control Regimes. Cambridge, Mass.: Africa. Cambridge, U.K.: Cambridge University Ballinger. Press. ------. 1998. A Stream of Windows: Unsettling Reflections Bates, Robert H., Rui de Figuereido, and Barry Wein- onTrade,Immigration,and Democracy.Cambridge,Mass.: gast.1998."The Politics of Interpretation:Rationality, Massachusetts Institute ofTechnology. Culture, and Transition." Politics and Society 26: ------. 2002."Globalization and Appropriate Gover- 603­42. nance." 2000 WIDER Annual Lecture.World Insti- 334 BIBLIOGRAPHY tute for Development Economics Research, the Bloom, D., and J. Sachs. 1998."Geography, Demography, United Nations University, Helsinki. and Economic Growth in Africa." Brookings Papers on Bhagwati, Jagdish N., and T. N. Srinivasan. 2002."Trade Economic Activity, 2. Washington, D.C.: Brookings and Poverty in the Poor Countries." American Eco- Institution. nomic Review 92(2): 180­83. Bloom,David E.,Ajay Mahal,Damien King,Aldrie Henry- Bhalla, Surjit S. 2002. ImagineThere's No Country: Poverty, Lee, and Philip Castillo. 2001."Globalization, Liberal- Inequality,and Growth in the Era of Globalization.Wash- ization,and Sustainable Human Development:Progress ington, D.C.: Institute for International Economics. and Challenges in Jamaica."UNCTAD/UNDP Global Bhatnagar, Subash, and Arsala Deane. 2003. "Building Program Occasional Paper. Blocks of e-Government: Lessons from Developing Bloomberg database.Bloomberg Financial Services,New Countries.World Bank,Washington, D.C. Online at York. http://www1.worldbank.org/prem/PREMNotes/ Bokros,L.,A.Fleming,and C.Votava.2001.FinancialTran- premnote91.pdf. sition in Europe and Central Asia: Challenges of the New Bhattasali, Deepak, Li Shantong, and Will Martin. 2002. Decade. Washington, D.C.:World Bank. "Impacts and Policy Implications of WTO Accession Bolaky, Bineswaree, and Caroline Freund. 2004."Trade, for China." Draft,World Bank,Washington, D.C. Regulations, and Growth." Draft (March). University Biebesheimer, Christina, and J. Mark Payne. 2001."IDB of Maryland andWorld Bank. Experience in Justice Reform."Technical Paper, Sus- Bordo,M.,B.Eichengreen,D.Klingebiel,and M.S.Mar- tainable Development Department. Washington, tinez-Peria. 2001. "Is the Crisis Problem Growing D.C.: InterAmerican Development Bank. Online at More Severe?" Economic Policy:A European Forum 32 http://www.iadb.org/sds/doc/sgc-IDBExperiences- (April): 51-82. E.pdf. Borensztein, Eduardo, Jose De Gregorio, and Jong-Wha Birdsall, Nancy. 2002."From Social Policy to an Open- Lee. 1998. "How Does Foreign Direct Investment Economy Social Contract in Latin America." Center Affect Economic Growth?" Journal of International for Global Development Working Paper No. 21, Economics 45, June. December. Online at http://www.cgdev.org/docs/ Bosworth,Barry,and Susan M.Collins.1996."Economic cgd%20wp021.pdf. Growth in East Asia:Accumulation vs.Assimilation." Bista, Dor Bahadur. 1991. Fatalism and Development. Cal- Brookings Papers in Economic Activity, 2. Brookings cutta: Orient Longman. Institution,Washington, D.C. Black, B. 2001."The Legal and Institutional Precondi- ------. 1999."Capital Flows, Investment, and Growth." tions for Strong Securities Markets." UCLA Law Tokyo Club Papers,Vol. 12. Review 48: 781­858. ------. 2003."The Empirics of Growth: An Update." Black, B., R. Kraakmen, and A.Tarassova. 2000."Russian Unpublished paper (March 7). Brookings Institution, Privatization and Corporate Governance: WhatWent Washington, D.C. Wrong?" Stanford Law Review 52: 1781. Boubakri, Narjess, and Jean-Claude Cosset. 1998."The Blair, Harry, and Gary Hansen. 1994."Weighing in on Financial and Operating Performance of Newly Pri- the Scales of Justice:Strategic Approaches for Donor- vatized Firms:Evidence from Developing Countries." Supported Rule of Law Programs." Program and Journal of Finance 53(3) (June): 1081­1110. Operations Assessment Report 7. U.S. Agency for ------. 2002."Does Privatization Meet the Expecta- International Development,Washington, D.C. tions? Evidence from African Countries." Forthcom- Blanchard,O.,and F.Giavazzi.2003."Improving the Stabil- ing in Journal of African Economics. ity and Growth Pact through Proper Accounting of Bourguignon, François. 2004a."Global Distribution and Public Investment." Center for Economic and Policy Redistribution." Keynote Address, the 16th Annual Research Discussion Papers No.4220,Washington,D.C. Bank Conference on Development Economics Blank, Rebecca M. 2002. Evaluating Welfare Reform in (ABCDE),held atWorld Bank Headquarters inWash- the United States. National Bureau of Economic ington, D.C., May 3­4. Research, NBERWorking Papers: 8983. ------. 2004b."The Poverty Growth Inequality Trian- Blomstrom, Magnus. 1986."Foreign Direct Investment gle." Paper presented at the Indian Council for and Productive Efficiency:The Case of Mexico."Jour- Research on International Economic Relations,New nal of Industrial Economics 15. Delhi, February 4. Blomstrom, Magnus, Robert Lipsey, and Mario Zejan. Bourguignon, Francois, and Christian Morrison. 2002. 1994. "What Explains the Growth of Developing "Inequality among World Citizens, 1820­1992." Countries?" In William J. Baumol, Richard R. Nel- American Economic Review 92(4): 727­44. son, and Edward N.Wolff, eds., Convergence of Produc- Boycko, Maxim, Andrei Shleifer, and Robert Vishny. tivity: Cross-National Studies and Historical Evidence. 1996."Privatizing Russia." Cambridge, Mass.: MIT Oxford and NewYork: Oxford University Press. Press. BIBLIOGRAPHY 335 Brook, Penelope J., and Timothy C. Irwin. 2003."Infra- Calvo, G. 1998. "Capital Flows and Capital Market structure for Poor People."World Bank. Crises:The Simple Economics of Sudden Stops."Jour- Brown,David,and John Earle.2004."Economic Reform nal of Applied Economics 1: 35­54. and Productivity-Enhancing Reallocation in the Calvo,G.,and C.Reinhart.2000."When Capital Inflows Post-Soviet Transition." Upjohn Institute, Staff Come to a Sudden Stop: Consequences and Policy Working Paper No. 04-98. Options."In P.Kenen andA.Swoboda,eds.,Key Issues Brownbridge,M.,and C.Harvey.1998. Banking in Africa. in Reform of the International Monetary System. Wash- Trenton, N.J.:AfricaWorld Press. ington, D.C.: International Monetary Fund. Brueckner, Jan. 1999."Fiscal Decentralization in LDCs: Campillo, M., and J. Miron. 1997."Why Does Inflation The Effects of Local Corruption and Tax Evasion." Differ across Countries?"In C.Romer and D.Romer, Unpublished manuscript,Department of Economics, eds., Reducing Inflation: Motivation and Strategy. University of Illinois at Urbana-Champaign. Chicago: University of Chicago Press. Bruno,Michael,Martin Ravallion,and Lyn Squire.1995. Campos, J. Edgardo, ed. 2001. Corruption:The Boom and "Equity and Growth in Developing Countries: Old Bust of East Asia.Quezon City,Philippines:Ateneo de and New Perspectives on the Policy Issues."InV.Tanzi Manila University Press. and K.Chu,eds.,Income Distribution and High-Quality Caprio, Gerard. 1997. "Safe and Sound Banking in Growth. Cambridge, Mass.: MIT Press. Developing Countries: We're Not in Kansas Any- Buiter,W.1990.Principles of Budgetary and Financial Policy. more." Research in Financial Services: Private and Public Cambridge, Mass.: MIT Press. Policy 9: 79­97. Buiter, W., Akram Esanov, and Martin Raiser. 2001. Caprio, Gerard, and Patrick Honohan. 2001. Finance for "Nature's Blessing or Nature's Curse? The Political Growth: Policy Choices in a Volatile World. Oxford and Economy of Transition in Resource-Based NewYork: Oxford University Press. Economies."EBRDWorking Paper No.65.London: Caprio,Gerard,and D.Klingebiel.2002."Episodes of Sys- European Bank for Reconstruction and Develop- tematic and Borderline Banking Crises."In D.Klinge- ment. biel and L. Laeven, eds., Managing the Real and Fiscal Burkhart, Ross, and Michael Lewis-Beck. 1994."Com- Effects of Banking Crises.World Bank Discussion Paper parative Democracy:The Economic Development No. 428. Thesis." American Political Science Review 88: 903­10. ------. 2003."Bank Insolvencies: Cross-Country Expe- Burki, Shahid Javed, and Guillermo Perry. 1998. Beyond rience." Draft.World Bank,Washington, D.C. the Washington Consensus: Institutions Matter. Washing- Caprio, Gerard, and L. Summers. 1996. "Financial ton, D.C.:World Bank. Reform: Beyond Laissez Faire." In D. Papadimitriou, Burki, Shahid Javed, Guillermo E. Perry, andWilliam R. ed.,Stability of the Financial System.NewYork:McMil- Dillinger. 1999a. Beyond the Center: Decentralizing the lan Press (earlier version published asWorld Bank Pol- State.Washington, D.C.:TheWorld Bank. icy ResearchWorking Paper 1171,August 1993). ------. 1999b."Mas Alla Del Centro: La Descentraliza- Caprio, Gerard, P. Honohan, and J. Stiglitz, eds. 2001. cion Del Estado."World Bank,Washington, D.C. Financial Liberalization: How Far, How Fast? Cam- Burnside, Craig. 2004."Currency Crises and Contin- bridge, U.K.: Cambridge University Press. gent Liabilities," Journal of International Economics Caprio,Gerard,D.Klingebiel,L.Laeven and G.Noguera. 62(1): 25­52. 2003. Banking Crises database. Updated October Burnside, Craig, and David Dollar. 2000."Aid, Policies, 2003. Available online at http://www1.worldbank and Growth." American Economic Review 90(4): .org/fin ance/html/database_sfd.html. 847­868. Carkovic, Maria, and Ross Levine. 2002."Does Foreign Calderón,C.,R.Duncan,and K.Schmidt-Hebbel.2003. Direct Investment Accelerate Economic Growth?" "The Role of Credibility in the Cyclical Properties Draft (October). University of Minnesota, Depart- of Macroeconomic Policies in Emerging ment of Business Finance and Carlson School of Economies." Draft. Central Bank of Chile. Management. Calderón, C.,William Easterly, and Luis Servén. 2002. Cespedes, Luis Felipe, Roberto Chang, and Andres "How Did Latin America's Infrastructure Fare in the Velasco. 2002."IS-LM-BP in the Pampas."Working Era of Macro-Economic Crises?" Working Paper Paper, John F. Kennedy School of Government, Har- No.185. Banco Central de Chile. vard University. Online at http://ksghome.harvard Calomaris,C.,andA.Powell.2001."Can Emerging Mar- .edu/~.avelasco.academic.ksg/Files/Research/pam ket Regulators Establish Credible Market Disci- pasimf.pdf. pline?" In F. Mishkin, ed., Prudential Supervision:What Chang, Ha-Joon. 2001. "Institutional Development in Works and What Doesn't. Chicago: University of Developing Countries in a Historical Perspective-- Chicago for the National Bureau of Economic Lessons from Developed Countries in EarlierTimes." Research. Paper presented at the EuropeanAssociation of Evolu- 336 BIBLIOGRAPHY tionary Political Economy Annual Meeting, Novem- Clarke,G.,R.Cull,and M.Shirley.2003."Empirical Stud- ber 8­11. ies of Bank Privatization." Paper presented at the Chari,V., and P. Kehoe. 2003."Hot Money." Journal of World Bank Conference on Bank Privatization, Political Economy 111(6): 1262­92. November 21­23. Chege, Michael. 1998. Liberal Democracy and Its Malcon- Clarke, G., R. Cull, M. Martinez Peria, and S. Sanchez. tents: Contrasting Perspectives of a Democratic Capitalist 2004. "Bank Lending to Small Businesses in Latin World Order from Sub-SaharanAfrica in the 1990s. Cam- America: Does Bank Origin Matter?" Forthcoming, bridge,Mass.:HarvardAcademy for International and Journal of Money, Credit, and Banking. Area Studies,Weatherhead Center for International Clement, Jean A. P., Johannes Mueller, Stephane Cosse, Affairs. and Jean Le Dem. 1996. Aftermath of the CFA Franc Chen,Shaohua,and Martin Ravallion.2004."How Have Devaluation. IMF Occasional Paper No. 138 (May). the World's Poorest Fared since the Early 1980s?" Clerides, Sofronis K., Saul Lach, and James R.Tybout. World Bank Research Observer 19(2): 141­70. 1998."Is Learning by Exporting Important? Micro- Cho,Y. 2001."The Role of Poorly Phased Liberalization Dynamic Evidence from Colombia, Mexico, and in Korea's Financial Crisis."In G.Caprio,P.Honohan, Morocco." Quarterly Journal of Economics 113(3): and J. Stiglitz, eds., Financial Liberalization: How Far, 903­47. How Fast? Cambridge, U.K.: Cambridge University Coase, Ronald H. 1937."The Nature of the Firm." Eco- Press. nomica 4: 386­405. Chong, Alberto, and Cesar Calderón. 2000. "On the ------. 1960."The Problem of Social Cost," Journal of Causality and Feedback between Institutional Mea- Law and Economics 3: 1­44. sures and Economic Growth." Economics and Politics Cohen,Avi, and G. C. Harcourt. 2003."Whatever Hap- 12(1): 69­81. pened to the Cambridge Capital Theory Controver- Chuang,Yih-Chyi, and Chi-Mei Lin. 1999. "Foreign sies?" Journal of Economic Perspectives 17(1): 199­214. Direct Investment, R&D, and Spillover Efficiency: Cohen, D., and R. Portes. 2003."Dealing with Destabi- Evidence from Taiwan's Manufacturing Firms." Jour- lizing `Market Discipline.'" Center for Economic and nal of Development Studies 35(4). Policy Research Discussion Paper No. 4280. Claessens, Stijn, and K. Forbes. 2001. International Finan- Cole, David C., and Betty F. Slade. 1998. Building a Mod- cial Contagion. Boston: Kluwer Academic Press. ern Financial System:The Indonesian Experience. Cam- Claessens, Stijn, Simeon Djankov, and Daniela Klinge- bridge, U.K.: Cambridge University Press. biel. 2001."Stock Markets in Transition Economies." Collier, Paul, and J. Gunning. 1997."Explaining African In FinancialTransition in Europe and Central Asia.Wash- Economic Performance."Center for Study of African ington, D.C.:World Bank. Economies Discussion Paper, University of Oxford. Claessens,Stijn,Simeon Djankov,and L.Lang.1998.Cor- Collier,Paul,andAnke Hoeffler.1998."On the Economic porate Growth, Financing, and Risks in the Decade before Causes of CivilWar." Oxford Economic Papers 50. East Asia's Financial Crisis. World Bank Policy ------.2001."Greed and Grievance in CivilWar."World Research Paper No. 2017. Bank,Washington, D.C. Claessens, Stijn, D. Klingebiel, and S. Shmuckler. 2003. Collins, Susan, and Robert Z. Lawrence, eds. 2000. "Explaining the Migration of Stocks from Exchanges "Brookings Trade Forum 1999." Brookings Institu- in Emerging Economies to International Centers." tion,Washington, D.C. Working Paper No. 168, Center for Research on Condon,Timothy, and S. Ramachandran. 1993."Cash Economic Development and Policy Reform, Stan- Constraints and Credit Corsets: The Chimera of ford University, Palo Alto, Calif. Inter-Enterprise Credits." Clague, Christopher, Philip Keefer, Stephen Knack, and Conning, J., and M. Kevane. 2001."Community-Based Mancur Olson. 1996."Property and Contract Rights Targeting Mechanisms for Social Safety Nets."World under Democracy and Dictatorship." Journal of Eco- Bank,Washington, D.C. nomic Growth 1(2): 243­76. Cox,GaryW.1987.The Efficient Secret:The Cabinet and the Clark,Ximena,David Dollar,andAlejandro Micco.2002. Development of Political Parties inVictorian England.New "Maritime Transport Costs and Port Efficiency." York: Cambridge University Press. Working Paper No. 2781.World Bank,Washington, Crandall, R., and T. Hazlett. 2000."Telecommunications D.C. Policy Reform in the United States and Canada." Clarke, G., and R. Cull. 1999."Why Privatize?The Case Working Paper 00-9, American Enterprise Insti- of Argentina's Public Provincial Banks." World Devel- tute­Brookings Joint Center for Regulatory Studies, opment 27: 865­88. Washington, D.C. ------. 2002."Political and Economic Determinants of Crawford,Young.1994.TheAfrican Colonial State in Com- the Likelihood of Privatizing Argentine Banks." The parative Perspective. New Haven, Conn.:Yale Univer- Journal of Law and Economics 45: 165­89. sity Press. BIBLIOGRAPHY 337 Crook, R., and J. Manor. 1998. Democracy and Decentral- ------. 2000. The Mystery of Capital. New York: Basic ization in SouthAsia andWestAfrica."Cambridge,U.K.: Books. Cambridge University Press. ------. Forthcoming."Missing Ingredients of Globaliza- Cukierman,A., G. Miller, and B. Neyapti. 2001."Central tion." In Ernesto Zedillo, ed., The Future of Globaliza- Bank Reform, Liberalization, and Inflation in Transi- tion:Explorations in Light of RecentTurbulence. Princeton, tion Economies:An International Perspective." Cen- N.J.: Princeton University Press. ter for Economic and Policy Research Discussion Deaton,Angus.2003a."Health,Inequality,and Economic Paper No. 2808. Development." Journal of Economic Literature 41(1): Cukierman,A., S.Webb, and B. Neyapti. 1992."Measur- 113­58. ing the Independence of Central Banks and Its Effect ------. 2003b."Household Surveys, Consumption, and on Policy Outcomes." World Bank Economic Review the Measurement of Poverty." Economic Systems 6(3): 353­98. Research 15(2): 135­59. Dalmazzo, A., and G. de Blasio. 2001."Resources and ------. 2003c."Measuring Poverty in a Growing World Incentives to Reform:A Model and Some Evidence (or Measuring Growth in a Poor World)." NBER on Sub-Saharan African Countries." IMF Working Working Paper No. 9822. National Bureau of Eco- Paper No. 86. International Monetary Fund,Wash- nomic Research, Cambridge, Mass. ington, D.C. deLisle, Jacques. 1999. "Lex Americana? United States Damijan, Joze P., Mark Knell, Boris Majcen, and Matija Legal Assistance,American Legal Models, and Legal Rojec. 2003."Technology Transfer through FDI in Change in the Post-Communist World and Beyond." Top-10 Transition Countries: How Important Are Journal of International Economic Law (University of Direct Effects, Horizontal and Vertical Spillovers?" Pennsylvania) 20:179. William Davidson Institute Working Papers Series DeLong,J.B.2000."Cornucopia:The Pace of Economic 549.William Davidson Institute at the University of Growth in the 20th Century." NBERWorking Paper Michigan Business School. No. 7602. National Bureau of Economic Research, Davidson, Basil. 1992. The Black Man's Burden:Africa and Cambridge, Mass. the Curse of the Nation State. NewYork:Times Books. Del Valle, C., and P. Ugolini. 2003. "Development of De Ferranti, David, G. Perry, I. Gill, and L. Servén. 2000. Domestic Markets for Government Bonds." In R. Securing Our Future.Washington, D.C.:World Bank. Litan, M. Pomerleano, and V. Sundarajan, The De Ferranti, David, Guillermo E. Perry, Daniel Leder- Future of Domestic Capital Markets in Developing man, and William Maloney. 2001. From Natural Countries, 45-76. Washington, D.C.: Brookings Resources to the Knowledge Economy:Trade and Job Qual- Institution. ity. Washington, D.C.:World Bank. Demirgüç-Kunt, A., and E. Detragiache. 1998. "The De Haan,J.,andW.Koi.2000."Does Central Bank Inde- Determinants of Banking Crises: Evidence from pendence Really Matter? New Evidence for Devel- Industrial and Developing Countries." IMF Staff oping Countries Using a New Indicator." Journal of Papers 45: 81­109. Banking and Finance 24(4): 643­64. ------. 2001. "Financial Liberalization and Financial De Juan,A.2002."From Good Bankers to Bad Bankers." Fragility." In G. Caprio, P. Honohan, and J. Stiglitz, In G. Caprio, P. Honohan, and D.Vittas, eds., Financial eds.,Financial Liberalization:How Far,How Fast? Cam- Sector Policy for Developing Countries--A Reader. bridge, U.K.: Cambridge University Press. Oxford: Oxford University Press. ------.2002."Does Deposit Insurance Increase Banking De la Torre, A., and S. Schmukler. 2003."Coping with Stability? An Empirical Investigation." Journal of Mon- Risk through Mismatches: Domestic and Interna- etary Economics 49(7): 1373­1406. tional Financial Contracts for Emerging Economies." Demirgüç-Kunt,A.,and H.Huizinga.1999."Market Dis- Draft.World Bank,Washington, D.C. cipline and Financial Safety Net Design."World Bank De la Torre,A., E. Levy-Yeyati, and S. Schmukler. 2003. Policy Research Paper No. 2183. "Living and Dying with Hard Pegs:The Rise and Fall Demirgüç-Kunt,A., and E. Kane. 2002."Deposit Insur- ofArgentina's Currency Board."Economía: Journal of the ance around the Globe:Where Does It Work?" The Latin American and Caribbean Economic Association 3(2): Journal of Economic Perspectives 16(2): 175­96. 43­99. Demirgüç-Kunt,A., andT. Sobaci. 2001."Deposit Insur- De Long, J. Bradford, and Andrei Shleifer. 1993."Princes ance around theWorld:A Data Base."World Bank Eco- and Merchants: European City Growth before the nomic Review 15(3): 481­90. Industrial Revolution." Journal of Law and Economics Demsetz, Harold. 1964."The Exchange and Enforce- 36: 671­702. ment of Property Rights." Journal of Law and Econom- De Soto, Hernando. 1989. The Other Path:The Invisible ics 7: 11­26. Revolution in the Third World. New York: Harper and Diamond, Jared. 1997. Guns, Germs, and Steel:The Fates of Row. Human Societies. NewYork: Norton. 338 BIBLIOGRAPHY Diaz-Alejandro, C. 1985."Good-Bye Financial Repres- ------. 2002."How Did Heavily Indebted Poor Coun- sion, Hello Financial Crash." Journal of Development tries Become Heavily Indebted? Reviewing Two Economics 19(1­2): 1­24. Decades of Debt Relief." World Development 30(10): Djankov, Simeon, and Bernard Hoekman. 2000."Pro- 1677­96. ductivity Growth and Foreign Investment in Czech ------. 2003a."Can Foreign Aid Buy Growth?" Journal Enterprises." World Bank Economic Review 14: 49­64. of Economic Perspectives 17(3): 23­48. Djankov, Simeon, and Peter Murell. 2002."Enterprise ------. 2003b. "National Policies and Economic Restructuring in Transition:A Quantitative Survey." Growth: A Reappraisal." CGD Working Paper No. Journal of Economic Literature XL (September):739­92. 27, May. Center for Global Development,Washing- Dollar, David, and Aart Kraay. 2001."Trade, Growth, and ton, D.C. Poverty." Draft (June).World Bank,Washington, D.C. Easterly,William, and S. Fischer. 2001."Inflation and the ------. 2003."Institutions,Trade, and Growth: Revisit- Poor." Journal of Money, Credit, and Banking 33: ing the Evidence." World Bank Policy Research 160­79. Working Paper No. 3004 (March). Easterly, William, and Ross Levine. 1997. "Africa's Dooley, M. 1996."A Survey of Literature on Controls GrowthTragedy:Politics and Ethnic Divisions."Quar- over International Capital Transactions." International terly Journal of Economics 112(4) (November):1203­50. Monetary Fund Staff Papers 43(4): 639­87. ------.2003."Tropics,Germs,and Crops:How Endow- ------.2000."A Model of Crises in Emerging Markets." ments Influence Economic Development." Journal of Economic Journal 110: 256­73. Monetary Economics 50 (1): 3­39. ------. 2003."Financial Policies." In M. Feldstein, ed., Easterly,William,and L.Servén,eds.2003.The Limits of Sta- Economic and Financial Crises in Emerging Market bilization:Infrastructure,Public Deficits,and Growth in Latin Economies. Chicago: University of Chicago Press for America.Palo Alto,Calif.:Stanford University Press. the National Bureau of Economic Research. Easterly,William,R.Islam,and J.Stiglitz.2001."Volatility Dooley,M.,W.Helkie,R.Tyron,and J.Underwood.1986. and Macroeconomic Paradigms for Rich and Poor." "An Analysis of External Debt Positions of Eight In J.Dréze,ed.,Advances in MacroeconomicTheory.New Countries through 1990." Journal of Development Eco- York: Palgrave. nomics 21(2): 283­318. Easterly,William, Michael Kremer, Lant Pritchett, and Dornbusch,Rudi.2001."A Primer on Emerging Market Lawrence Summers. 1993. "Good Policy or Good Crises." NBER Working Paper 8326. National Luck? Country Growth Performance and Tempo- Bureau of Economic Research, Cambridge, Mass. rary Shocks." NBER Working Paper No. 4474. Dornbusch, Rudi, and S. Edwards, eds. 1991. The Macro- National Bureau of Economic Research, Cam- economics of Populism in LatinAmerica.Chicago:Univer- bridge, Mass. sity of Chicago for the National Bureau of Economic Eaton, K. 2003."Menem and the Governors: Intergov- Research. ernmental Relations in the 1990s." In S. Levitsky and Dorotinsky,William, and Robert Floyd. 2004."Expendi- M. N. Murillo, eds., Rethinking DualTransitions:Argen- ture Accountability in Africa: Progress, Lessons and tine Politics in the 1990s in Comparative Perspective. Challenges." In Brian Levy and Sahr Kpundeh, eds., Manuscript online at http://www.wws.princeton Building State Capacity inAfrica:NewApproaches,Emerg- .edu/keaton. ing Lessons. Washington, D.C.:World Bank Institute. Ebel, Robert D., and Serdar Yilmaz. 2002. On the Mea- Drazen, Allan, Nuno Limão, and Thomas Stratmann. surement and Impact of Fiscal Decentralization. Washing- 2004."Political Contribution Caps and Lobby For- ton, D.C.:World Bank Institute. mation:Theory and Evidence."Paper presented at the Economist. 2004."More or less equal?" March 11. 2004 Meeting of the Public Choice Society, Balti- Edison, H., R. Levine, L. Ricci, and T. Sloken. 2002. more. "International Financial Integration and Economic Duncan,Alex. 2003."Drivers of Change: Reflections on Growth." NBER Working Paper No. 9164. National Experience to Date." Discussion Note. U.K. Depart- Bureau of Economic Research, Cambridge, Mass. ment for International Development. Edwards, Sebastian. 1998."Openness, Productivity and Easterly,William. 1999."When Is Fiscal Adjustment an Growth:What Do We Really Know?" Economic Jour- Illusion?" Economic Policy 14(28): 55­76. nal 108(March): 383­98. ------. 2000."Small States, Small Problems? Income, ------. 1999."How Effective Are Capital Controls?" Growth,andVolatility in Small States."World Develop- Journal of Economic Perspectives 13(4): 65­84. ment 28(11): 2013­27. ------. 2003."Exchange Rate Regimes, Capital Flows, ------. 2001."The Lost Decade: Developing Countries' and Crisis Prevention." In Feldstein (2003). Stagnation in Spite of Policy Reform 1980­1990."Jour- Ehrlich,Isaac,and Gary Becker.1972."Market Insurance, nal of Economic Growth 6: 135­57. Online at http:// Self-Insurance and Self-Protection." Journal of Political www.cgdev.org/fellows/easterly_lostdecades.pdf. Economy 80: 623­48. BIBLIOGRAPHY 339 Eichengreen, Barry. 2001."Capital Account Liberaliza- Fischer, Stanley. 2001."Exchange Rate Regimes: Is the tion:What Do Cross-Country StudiesTell Us?"World BipolarView Correct?" Journal of Economic Perspectives Bank Economic Review 15(3): 341­66. 15(2): 3­24. ------.2002.Financial Crises andWhat to DoAboutThem. ------.2003."Globalization and Its Challenges.Richard Oxford: Oxford University Press. T.Ely Lecture."American Economic Review 93(2):1­30. Eifert,Benn,Alan Gelb,and Nils BorjeTallroth.2003."The Fjelstad, O-H. 2002."Fighting Fiscal Corruption:The Political Economy of Fiscal Policy and Economic Case of the Tanzania Revenue Authority."Working Management in Oil-Exporting Countries." In J. M. Paper 2002.3. Bergen, Norway: Christian Michelsen Davis, R. Ossowski, and A. Fedelino, eds., Fiscal Policy Institute. Formulation and Implementation in Oil-Producing Coun- Forbes,K.2003."One Cost of the Chilean Capital Con- tries.Washington,D.C.:International Monetary Fund. trols:Increased Financial Constraints for SmallerTrade Engberg-Pedersen,Poul,and Brian Levy.2004."Building Firms." NBER Working Paper No. 9777. National Governance Capacity: Learning from Performance Bureau of Economic Research, Cambridge, Mass. and Results." In Brian Levy and Sahr Kpundeh, eds., ------. 2004."Capital Controls: Mud in the Wheels of Building State Capacity inAfrica:NewApproaches,Emerg- Market Discipline." NBER Working Paper No. ing Lessons.Washington, D.C.:World Bank Institute. 10284. National Bureau of Economic Research, Engerman, Stanley, and Kenneth Sokoloff. 2002."Factor Cambridge, Mass. Endowments, Inequality, and Paths of Development Fox,L.,and E.Palmer.2001."New Approaches to Multi- among New World Economics." NBER Working Pillar Pension Systems:What in the World Is Going Paper 9259.National Bureau of Economic Research, On?" In R. Holzmann and J. Stiglitz, eds., New Ideas Cambridge, Mass. Published in Economia 3(1): 41­88. about Old Age Security.Washington,D.C.:World Bank. Enke, Stephen. 1963. Economics of Development. London: Frankel, Jeffrey A., and David Romer. 1999."DoesTrade Denis Dobson. Cause Growth?" American Economic Review 89(3) Enoch,C.,B.Baldwin,O.Frecaut,and A.Kovanen.2001. (June): 379­99. Indonesia:Anatomy of a Banking Crisis.TwoYears of Liv- Frankel,Jeffrey A.,and A.Rose.1996."Currency Crashes ing Dangerously, 1997­1999. International Monetary in Emerging Markets: Empirical Indicators." NBER FundWorking Paper No. 01/52. Working Paper No. 5437. National Bureau of Eco- Esfahani, Hadi Salehi. 2002. "Political Economy of nomic Research, Cambridge, Mass. Growth in Iran,1963­2002."Draft (September).Uni- Frankel, Jeffrey A., and Shang-Jin Wei. 2004."Managing versity of Illinois at Urbana. Macroeconomic Crises: Policy Lessons." NBER EBRD (European Bank for Reconstruction and Devel- Working Paper No. 10907. National Bureau of Eco- opment). 2003. Transition Report 2003. London. nomic Research, Cambridge, Mass. Evans, Peter. 1995. Embedded Autonomy: States and Industrial Frankel, Jeffrey A., E. Fajnzylber, S. Schmukler, and L. Transformation. Princeton, N.J.: Princeton University Servén. 2001."Verifying Exchange Rate Regimes." Press. Journal of Development Economics 66(2): 351­85. Evans, Peter, and James Rauch. 2000. "Bureaucratic Frieden, Jeffry A. 1991. Debt, Development, and Democracy. Structure and Bureaucratic Performance in Less Princeton, N.J.: Princeton University Press. Developed Countries." Journal of Public Economics 75: Friedman, Lawrence. 1985. A History of American Law. 49­71. NewYork: Simon and Schuster. Faguet, Jean-Paul. 1997. "Decentralization and Local Frydman, Roman, Cheryl Gray, Marek Hessel, and Government Performance."World Bank,Washington, Andrzej Rapaczynski. 1999."When Does Privatiza- D.C. tion Work? The Impact of Private Ownership on Feenstra, Robert C., and Hiau Looi Kee. 2003."On the Corporate Performance in the Transition Measurement of Product Variety in Trade." Draft. Economies." Quarterly Journal of Economics (Novem- University of California andWorld Bank. ber):1153­91. Feenstra,Robert C.,Dorsati Madani,Tzu-HanYang,and Funke,Michael,and Ralf Ruhwedel.2001."ExportVari- Chi-Yuan Liang. 1999."Testing Endogenous Growth ety and Export Performance: Empirical Evidence in South Korea and Taiwan." Journal of Development from EastAsia."Journal ofAsian Economics 12:493­505. Economics 60: 317­41. Furman, J., and J. Stiglitz. 1998."Economic Crises: Evi- Feldstein, M., ed. 2003. Economic and Financial Crises in dence and Insights from East Asia."Brookings Papers on Emerging Market Economies. Chicago: University of Economic Activity, 2: 1­135. Chicago Press for the National Bureau of Economic Furtado, Celso. 1963. "The Concept of External Research. Dependence in the Study of Underdevelopment." Fernandez-Arias, E., and P. Montiel. 2001."Reform and In Kenneth P. Jameson and Charles K.Wilber, eds., Growth in Latin America: All Pain and No Gain?" Political Economy of Development and Underdevelop- IMF Staff Papers 48(3) (December): 522­46. ment. NewYork: McGraw-Hill. 340 BIBLIOGRAPHY Galal, Ahmed, Leroy Jones, Pankaj Tandon, and Ingo Glenday, G. 2000. "Trade Liberalization and Customs Vogelsang. 1994. Welfare Consequences of Selling Public Revenues: Does Trade Liberalization Lead to Lower Enterprises:An Empirical Analysis. New York: Oxford Customs Revenues? The Case of Kenya." African University Press. Economic Policy Discussion Paper No. 44. John F. Galasso,Emanuela,and Martin Ravallion.2000."Distrib- Kennedy School of Government,Harvard University. utional Outcomes of a Decentralized Welfare Pro- Goh,Chor-Ching,and Beata Smarzynska Javorcik.2004. gram."World Bank Policy Research Working Paper "Trade Protection and Industry Wage Structure in No. 2316,World Bank,Washington, D.C. Poland." In Harrison (forthcoming). Gallup, John Luke, Jeffrey D. Sachs, and Andrew Goldberg, Penny, and Nina Pavcnik. 2004. "Trade, Mellinger. 1999."Geography and Economic Devel- Inequality, and Poverty:What Do We Know? Evi- opment."CIDWorking Paper No.1.Center for Inter- dence from Recent Trade Liberalization Episodes in national Development, Harvard University. Online at Developing Countries." NBER Working Paper No. http://ideas.repec. org/s/wop/cidhav.html. 10593. In Susan Collins and Carol Graham, eds., Gandhi,Jennifer.2003."Dictatorial Institutions andTheir BrookingsTrade Forum.2004.Washington,D.C.:Brook- Impact on Economic Performance." Draft, Depart- ings Institution Press. ment of Politics, New York University. Presented at ------.2005."The Effects of the ColombiaTrade Liber- the Meetings of the American Political Science Asso- alization on Urban Poverty." In Harrison (forthcom- ciation, Philadelphia, September. ing). Garrett, G. 1995. "Capital Mobility, Trade, and the Gonzalez-Vega, Claudio. 1984. "Credit Rationing Domestic Politics of Economic Policy." International Behavior of Agricultural Lenders:The Iron Law of Organization 49(4): 657­87. Interest Rate Restrictions." In D.Adams, D. Graham, ------. 2000."Capital Mobility, Exchange Rates, and and J.D.Von Pischke,eds.,Undermining Rural Develop- Fiscal Policy in the Global Economy." Review of Inter- ment with Cheap Credit. Boulder, Colo.:Westview. national Political Economy 7(1): 153­70. Goodpaster, Gary. 2003."Law Reform in Developing Gavin,M.,and Ricardo Hausmann.1996."The Roots of Countries." Transnational Law and Contemporary Prob- Banking Crises: The Macroeconomic Context." lems,A Journal of the University of Iowa College of Law 13 Working Paper 318, Office of the Chief Economist, (2): 659­98. InterAmerican Development Bank,Washington,D.C. Graham, Carol, and Stefano Pettinato. 2002. Happiness Gelb,Alan,ed.1988.OilWindfalls:Blessing or Curse? New and Hardship: Opportunity and Insecurity in New Market York: Oxford University Press. Economies. Washington, D.C.: Brookings Institution Gelb,Alan,Brian Ngo,and XiaoYe.2004."Implementing Press. Performance-Based Aid in Africa:The Country Pol- Graham, Carol, and Sandip Sukhtankar. 2004. "Does icy and Institutional Assessment." Africa Region Economic Crisis Reduce Support for Markets and Working Paper Series No. 77, November. Democracy in Latin America?" Lessons from Surveys of Gerring, John,WilliamT. Brandt, and Philip Bond. 2003. Public Opinion andWell Being 36 (May): 349­77. "Democracy and Economic Growth: A Historical Grossman, Gene, and Elhanan Helpman. 1992. Interest Perspective." Draft, Political Science Department, Groups and Trade Policy. Princeton, N.J.: Princeton Boston University, September. University Press. Gerschenkron,Alexander. 1962. Economic Backwardness in Gunnarsson,Victoria,Peter Orazem,Mario Sanchez,and Historical Perspective: A Book of Essays. Cambridge, AimeeVerdisco.2004."Does School Decentralization Mass.: Belknap Press of Harvard University. Raise Student Outcomes? Theory and Evidence on Ghirmay,Teame, Richard Grabowski, and Subhash C. the Roles of SchoolAutonomy and Community Par- Sharma. 2001. "Exports, Investment, Efficiency, and ticipation."Unpublished manuscript.World Bank and Economic Growth in LDCs:An Empirical Investiga- InterAmerican Development Bank,Washington,D.C. tion." Journal of Applied Economics 33: 689­700. Gupta, Poonam, Rachel Kleinfeld, and Gonzalo Salinas. Gill, Indermit S.,Truman G. Packard, and Juan Yermo. 2002. Legal and Judicial Reform in Europe and Central 2004. Keeping the Promise of Social Security in Latin Asia.Washington,D.C.:World Bank Operations Eval- America. Washington, D.C.:World Bank. uation Department. Glaeser, Edward, and Andrei Shleifer. 2001."Legal Ori- Gupta, Sanjeev, Hamid Davoodi, and Rosa Alonso- gins." NBER Working Paper No. 8272. National Terme. 2002. "Does Corruption Affect Income Bureau of Economic Research, Cambridge, Mass. Inequality and Growth?"Economics of Governance 3(1): ------. 2003."The Rise of the Regulatory State." Jour- 23­45. nal of Economic Literature 41(2): 401­25. Gutiérrez, E. 2003."Inflation Performance and Constitu- Glaeser, E., S. Johnson, and A Shleifer. 2001."Coase ver- tional Central Bank Independence: Evidence from sus the Coasians." Quarterly Journal of Economics 116: LatinAmerica and the Caribbean."IMFWorking Paper 853­99. 03/53.International Monetary Fund,Washington,D.C. BIBLIOGRAPHY 341 Gylfason,T.2001."Nature,Power,and Growth."CESinfo Harding,April. 2003."Service Delivery Reforms:A Key Working Paper No. 413. Ingredient for Accelerating Progress toward MDGs." Haber,S.,and S.Kantor.2003."Getting PrivatizationWrong: Draft.World Bank,Washington, D.C. The Mexican Banking System 1991­2003."World Bank Harris, Clive. 2003."Private Participation in Infrastruc- Conference on Bank Privatization,Nov.20­21.Online at ture in Developing Countries:Trends, Impacts and http://www.worldbank.org/research/projects/ bank_ Policy Lessons."Working Paper No.5.World Bank, privatization_conference.htm. Washington, D.C. Haddad,Mona,andAnn Harrison.1993."AreThere Pos- Harrison,Ann,ed.Forthcoming.Globalization and Poverty. itive Spillovers from Direct Foreign Investment?" Chicago: University of Chicago Press for National Journal of Development Economics 42. Bureau of Economic Research. Halac,M.,and S.Schmukler.2003."Distributional Effects Harrison, Ann, and Gordon H. Hanson. 1999. "Who of Crises: The Role of Financial Transfers." Policy Gains from Trade Reform: Some Remaining Puz- ResearchWorking Paper 3173.World Bank,Washing- zles." NBER Working Paper No. 6915. National ton, D.C. Bureau of Economic Research, Cambridge, Mass. Hall, Robert E., and Charles I. Jones. 1999."Why Do Harrison,Ann, and Jason Scorse. 2003."The Impact of Some Countries Produce So Much More Output per Globalization on Compliance with Labor Standards: Worker Than Others?" Quarterly Journal of Economics A Plant-Level Study." Brookings Trade Forum, 114: 83­116. Brookings Institution,Washington, D.C. Hall,Thomas. 1999."Economic Reform and Financial Harrison, Lawrence, and Samuel Huntington. 2000. Cul- Sector Regulation in Emerging Markets." Ph.D. dis- ture Matters: How Values Shape Human Progress. New sertation, University of Southern California, Los York: Basic Books. Angeles. Hausmann, Ricardo. 2003. "Good Credit Ratios, Bad Hanson,Gordon.2004."Globalization,Labor Income,and Credit Ratings:The Role of Debt Structure."Working Poverty in Mexico."In Harrison (forthcoming). Paper, John F. Kennedy School of Government, Har- Hanson,J.1994."An Open CapitalAccount:A Brief Sur- vard University. Online at http://ksghome.harvard vey of the Issues and Results." In G. Caprio, I.Atiyas, .edu/~.rhausma.cid.ksg/publication.htm. and J. Hanson, eds., Financial Reform:Theory and Expe- Hausmann, Ricardo, and Barry Eichengreen. 2002. rience. Cambridge, U.K.: Cambridge University Press. "Original Sin:The Pain,the Mystery,and the Road to ------. 2001. "Indonesia and India: Contrasting Redemption." Working Paper, John F. Kennedy Approaches to Repression and Liberalization." In G. School of Government,Harvard University.Online at Caprio, P. Honohan, and J. Stiglitz, eds., Financial Lib- http://ksghome.harvard.edu/~.rhausma.cid.ksg/pub eralization: How Far, How Fast? Cambridge, U.K.: lication.htm. Cambridge University Press. Hausmann, Ricardo, and Roberto Rigobon. 2002."An ------. 2002."Dollarization, Private and Official: Issues, Alternative Interpretation of the `Resource Curse': Benefits,and Costs."In G.Caprio,P.Honohan,and D. Theory and Policy Implications." NBER Working Vittas, Financial Sector Policy for Developing Countries-- Paper No. 9424. National Bureau of Economic A Reader. Oxford: Oxford University Press. Research, Cambridge, Mass. ------. 2003a."Indian Banking: Market Liberalization Hausmann,Ricardo,and Dani Rodrik.2002."Economic and the Pressures for Institutional and Market Frame- Development as Self-Discovery." NBER Working work Reform." In A. Krueger and S. Chinoy, eds., Paper No. 8952. National Bureau of Economic Reforming India's External, Financial, and Fiscal Policies. Research,Cambridge,Mass.Published (2003) in Jour- Stanford, Calif.: Stanford University Press. nal of Development Economics 72(2): 603­33. ______. 2003b. Banking in Developing Countries in the Hausmann, Ricardo, D. Mathieson, and J. Roldos. 2003. 1990s.World Bank Policy Research Paper No. 3168. "Trends in Developing Country Capital Markets Hanson,J.,and Sanjay Kathuria,eds.1999.India:A Finan- Around the World." In R. Litan, M. Pomerleano, and cial Sector for the Twenty-First Century. Washington, V.Sundarajan.eds.,The Future of Domestic Capital Mar- D.C.:World Bank and NewYork and Oxford:Oxford kets in Developing Countries. Washington, D.C.:The University Press. Brookings Institution. Hanson, J., P.Honohan, and G.Majnoni, eds. 2003. Glob- Hausmann, Ricardo, Ugo Panizza, and Ernesto Stein. alization and National Financial Systems. NewYork and 2001."Original Sin, PassThrough, and Fear of Float- Oxford: Oxford University Press for theWorld Bank. ing."Working Paper,John F.Kennedy School of Gov- Harberger,Arnold C. 1971."Three Basic Postulates for ernment, Harvard University. Online at http:// Applied Welfare Economics:An Interpretive Essay." ksghome.harvard.edu/~rhausma/paper/OSpasstrhro Journal of Economic Literature 9(3): 785­97. ughfearoffloating.pdf. ------. 2004. "Exploring the Process of Economic Hausmann, Ricardo, Dani Rodrik, and Lant Pritchett. Growth." Draft. 2004."GrowthAccelerations."NBERWorking Paper 342 BIBLIOGRAPHY 10566. National Bureau of Economic Research, Financial Systems.NewYork and Oxford:Oxford Uni- Cambridge, Mass. Online at www.nber.org/papers/ versity Press for theWorld Bank. w10566. Huang,Yasheng. 2003."One Country,Two Systems: For- Hausmann, Ricardo, Dani Rodrik, and Andres Velasco. eign-Invested Enterprises and Domestic Firms in 2004."Growth Diagnostics." John F. Kennedy School China." China Economic Review 14(4): 404­16. of Government,Harvard University.Online at http:// Huang,Yasheng, and Tarun Khanna. 2003."Can India ksghome.harvard.edu/~drodrik/barcelonasep20.pdf. Overtake China?" Foreign Policy (July). Heilbrunn, John. 2002."Anti-Corruption: Lessons from Hutchison, M., and I. Noy. 2002."Sudden Stops and the the 1990s." Draft, PREM Public Sector.World Bank, Mexican Wave." Working Paper Series No. 02-12. Washington, D.C. Economic Policy Research Unit, Copenhagen. Helliwell, John. 1994. "Empirical Linkages between Huther, Jeff, and Anwar Shah. n.d. "Anti-Corruption Democracy and Economic Growth." British Journal of Policies and Programs--A Framework for Evalua- Political Science 24: 225­48. tion." Draft, 13 pp.World Bank,Washington, D.C. Hellman, Joel, G. Jones, and Daniel Kaufmann. 2000. Inter-American Development Bank (IDB). 1995. Over- "Seize the State, Seize the Day: State Capture, Cor- coming Volatility in Latin America. Washington, D.C.: ruption, and Influence in Transition." World Bank Inter-American Development Bank. InstituteWorking Paper.Washington, D.C. ------. 1997. "Latin America after a Decade of Herbst,Jeffrey.2000.States and Power inAfrica:Comparative Reforms." In Economic and Social Progress in Latin Lessons in Authority and Control. Princeton, N.J.: America.Washington, D.C.: IDB. Princeton University Press. International Monetary Fund (IMF).2001."Jamaica:Staff Heron,Tony. 2002."Export Processing Zones and Policy Report for the 2001 Article IV Consultation." Inter- Competition for Foreign Direct Investment:The Off- national Monetary Fund,Washington, D.C. shore Caribbean Development Model." July. ------. 2002a."Jamaica Staff Report for the 2002 Arti- Heybey, Berta, and Peter Murell. 1999."The Relation- cle IV Consultation." International Monetary Fund, ship between Economic Growth and Speed of Liber- Washington, D.C. alization duringTransition." Journal of Policy Reform 3: ------. 2002b. "Financial Stability in Dollarized 121­37. Economies." Draft. Hirschman, Albert O. 1958. The Strategy of Economic ------. 2003a. Fiscal Adjustment in IMF-Supported Pro- Development. New Haven, Conn.: Yale University grams.Washington, D.C.: IMF. Press. ------.2003b.The IMF and Recent CapitalAccount Crises. Hnatkovska, V., and N. Loayza. 2004. "Volatility and Washington, D.C.: IMF. Growth." Policy Research Working Paper 3184. ------.2003c."Redrafting the ReformAgenda."Finance World Bank,Washington, D.C. and Development (September).Washington, D.C.: IMF. Hoekman, Bernard, Constantine Michalopoulos, Mau- ------. 2003d. World Economic Outlook. Washington, rice Schiff, and David Tarr. 2001. "Trade Policy D.C.: IMF. Reform and Poverty Alleviation." Developed as part ------. 2003e."Assessing and Promoting Fiscal Trans- of the World Bank's Poverty Reduction Strategy parency:A Report on Progress."FiscalAffairs Depart- Sourcebook (December).Washington, D.C. ment,International Monetary Fund,March 5 (online Hoff,Karla.2003."Paths of Institutional Development:A at http://www.imf.org/external/np/pdr/sac/2003/ View from Economic History." World Bank Research 030503s2.htm). Observer 18(2): 205­26. ------. 2004a. World Economic Outlook. Washington, Hoff, Karla, and J. E. Stiglitz. 2001."Modern Economic D.C.: IMF. Theory and Development."In Gerald Meier and J.E. ------. 2004b. Sovereign Debt Restructurings and the Stiglitz, eds., The Future of Development Economics in Domestic Economy: Experience in Four Recent Cases. Perspective. Oxford: Oxford University Press. Washington, D.C.: Policy Development and Review Holzmann, Robert, Robert Palacios, and Asta Zviniene. Department, IMF. 2001."Implicit Pension Debt: Issues, Measurement, International Monetary Fund.(IMF), World Bank, and Scope in International Perspective." Pension OECD, and EBRD. 1991. "A Study of the Soviet Reform Primer.World Bank,Washington, D.C. Economy." Honohan, Patrick, and Daniela Klingebiel. 2000."Con- Isham, Jonathan,Thomas Kelly, and Sunder Ramaswamy, trolling Fiscal Costs of Banking Crises." Policy eds.2002.Social Capital and Econmic Development:Well- Research Working Paper No. 2441, World Bank, Being in Developing Countries. Cheltenham, U.K. and Washington, D.C. Northampton, Mass.: Elgar. Honohan, Patrick, and A. Shi. 2003."Deposit Dollariza- Isham, Jonathan, Michael Woolcock, Lant Pritchett, and tion and the Financial Sector."In J.Hanson,P.Hono- Gwen Busby.2003."TheVarieties of Resource Expe- han, and G. Majnoni, eds., Globalization and National rience: How Natural Resource Export Structures BIBLIOGRAPHY 343 Affect the Political Economy of Economic Growth." Kamarck,Andrew. 1967. Economics of African Development. Middlebury College Working Paper Series 0308. NewYork: Praeger. Department of Economics. Kaminsky, G., and C. Reinhart. 1999."The Twin Crises: Islam, Roumeen. 2002."Institutional Reform and the The Causes of Banking and Balance of Payments Judiciary:WhichWay Forward?"World Bank Institute Problems." American Economic Review 89(3): 473­500. Working Paper No. 3134.World Bank,Washington, Kanbur, Ravi. 2000."Aid, Conditionality, and Debt in D.C. Africa."In FinnTarp,ed.,ForeignAid and Development: Ize, A., and E. Levy-Yeyati. 1998."Financial Dollariza- Lessons Learned and Directions for the Future. London: tion."IMFWorking Paper No.1998/28.International Routledge. Monetary Fund,Washington, D.C. ------. 2001. "Economic Policy, Distribution, and Jaffee, Steven, and Spencer Henson. 2004."Agro-Food Poverty:The Nature of Disagreements." World Devel- Exports from Developing Countries:The Challenges opment 29(6): 1083­94. Posed by Standards." Prepared as chapter for book ------. 2002."Conceptual Challenges in Poverty and edited by Ataman Aksoy, forthcoming,World Bank. Inequality: One Development Economist's Perspec- Jaffee, Steven, and Jeanette Sutherland. 2003."Integrated tive." Draft (April), Cornell University, Ithaca, N.Y. Framework Trade Diagnostic Studies: Findings and Karl,Terry Lynn. 1997. The Paradox of Plenty: Oil Booms Lessons from Agricultural Sector Analysis." Draft. and Petro-States. Berkeley: University of California World Bank,Washington, D.C. Press. Jalan, Joyotsna, and Martin Ravallion. 1998."Transient Kaufmann, Daniel. 2003. "Rethinking Governance: Poverty in Post-Reform China." Journal of Compara- Empirical Lessons,Challenge Orthodoxy."Discussion tive Economics 26: 338­57. draft, March 11.Washington, D.C.:World Bank. Jamasb,T.2002."Reform and Regulation of the Electric- Kaufmann, Daniel,Aart Kraay, and Massimo Mastruzzi. ity Sector in Developing Countries." University of 2003."Governance Matters III: Governance Indica- Cambridge, Department of Applied Economics, tors for 1996­2002."World Bank Institute Working Cambridge, U.K. Paper. Online at http://www.worldbank.org/wbi/ Jin, Hehui,Yingyi Qian, and Barry R.Weingast. 2001. governance/pdf/ govmatters3_wber.pdf. "Regional Decentralization and Fiscal Incentives: Kaufmann,Daniel,Aart Kraay,and Pablo Zoido-Lobaton. Federalism,Chinese Style."Draft.Stanford University. 1999."Governance Matters." Policy Research Work- Johnson, S., and T. Mitton. 2002."Cronyism and Capital ing Paper No. 2196.World Bank,Washington, D.C. Controls:Evidence from Malaysia."Journal of Financial ------. 2002."Governance Matters II: Updated Indica- Economics 67(2): 351­82. tors for 2000/01."Working Paper No. 2772.World Jones, Charles. 2004."Growth and Ideas." Forthcoming Bank,Washington, D.C. in Handbook of Economic Growth. Amsterdam: North Keefer, Philip. 2002. "Clientelism, Credibility and Holland. Democracy." Draft, Development Research Group, Jones,E.L.1981.The European Miracle.Cambridge,U.K.: World Bank. Cambridge University Press. ------. 2003a."Democratization and Clientelism:Why Jones, Leroy P.,Yahya Jammal, and Nilgun Gokgur. 1998. Are Young Democracies Badly Governed?" Draft, "Impact of Privatization in Cote d'Ivoire." Draft final Development Research Group,World Bank. report, Boston Institute for Developing Economies, ------. 2003b. "All Democracies Are Not the Same: Boston University. Identifying the Institutions That Matter for Growth Joseph, Richard A. 1998."Patrons, Clients, and Factions: and Convergence." Draft, Development Research New Dimensions of Conflict Analysis in Africa." In Group,World Bank. Peter Lewis, ed., Africa: Dilemmas of Development and Keefer, Philip, and Stuti Khemani. Forthcoming. Change. Boulder, Colo.:Westview Press. "Democracy, Public Expenditures, and the Poor." JP Morgan.Emerging Market Bond Index Plus (EMBI+) World Bank Research Observer. database. Keefer, Philip, and Stephen Knack. 1997."Why Don't Kahkonen, Satu, and Anthony Lanyi. 2001."Decentral- Poor Countries Catch Up?" Economic Inquiry 35 ization and Governance: Does Decentralization (July): 590­602. Improve Public Service Delivery?" PREM Note ------. "Boondoggles and Expropriation: When Are (June).World Bank,Washington, D.C. Property Rights Secure and Public Investment Kahn,Alfred. 2004."Lessons from Deregulation:Telecom- Growth-Promoting?"Policy ResearchWorking Paper munications and Airlines after the Crunch."American No. 2910,World Bank (October). Enterprise Institute and Brookings Institution,Wash- Keefer, Philip, and David Stasavage. 2003."The Limits of ington,D.C. Delegation:Veto Players,Central Bank Independence, Kaldor, Nicholas. 1955. An Expenditure Tax. London: and the Credibility of Monetary Policy." American George Allen and Unwin. Political Science Review 97(3): 407­23. 344 BIBLIOGRAPHY Kenward, L. 2002. From the Trenches--The First Year of Kokko,Ari, Ruben Tansini, and Mario C. Zejan. 1996. Indonesia's Crisis 1997/98As Seen from theWorld Bank's "Local Technological Capability and Productivity Office in Jakarta.Center for International and Strategic Spillovers from FDI in Uruguayan Manufacturing Studies, Jakarta. Sector." Journal of Development Studies 32. Keynes, John Maynard. 1931. Essays in Persuasion. Lon- Konings, Jozef. 2000. "The Effects of Foreign Direct don: Macmillan. Investment on Domestic Firms:Evidence from Firm- Khan, Mushtaq H. 2002."State Failure in Developing level Panel Data in Emerging Economies."William Countries and Strategies of Institutional Reform." Davidson Institute Working Papers Series 344. Draft paper for ABCDE conference, Oslo, June 24- William Davidson Institute at the University of 26. Michigan Business School.Also published as CEPR Kharas, H., and B. Pinto. 2001."An Analysis of Russia's Discussion Paper No. 2586, Centre for Economic 1998 Meltdown: Fundamentals and Market Signals." Policy Research, London. Brookings Papers on Economic Activity, 1: 1­50. Kontopolous,Yoannis, and Roberto Perotti. 1999."Gov- Kikeri, Sunita, John Nellis, and Mary Shirley. 1992."Pri- ernment Fragmentation and Fiscal Policy Outcomes: vatization:The Lessons of Experience."World Bank, Evidence from the OECD Countries." In James Washington, D.C. Poterba and Juergen von Hagen,eds.,Fiscal Institutions Kindleberger, Charles P. 1984. A Financial History ofWest- and Fiscal Performance.Chicago:University of Chicago ern Europe. London:Allen and Unwin. Press. ------. 2000. Manias, Panics, and Crashes: A History of Kornai, Janos. 1990. The Road to a Free Economy: Shifting Financial Crises. NewYork:Wiley. from a Socialist System--The Example of Hungary. New King, R., and R. Levine. 1993."Finance and Growth: York: Norton. Schumpeter Might Be Right."Quarterly Journal of Eco- ------.1992."The PostsocialistTransition and the State: nomics 108(3): 717­37. Reflections in the Light of Hungarian Fiscal Prob- Klenow, Peter J., and Andres Rodriguez-Clare. 1997. lems." In Robert E. Goodin and Deborah Mitchell, "The Neoclassical Revival in Growth Economics: eds., The Foundations of the Welfare State, Volume 2. Has It Gone Too Far?" In Ben S. Bernanke and Julio Cheltenham, U.K and Northampton, Mass.: Edward J. Rotemberg, eds., NBER Macroeconomics Annual Elgar. 1997. Cambridge and London: MIT Press. ------. 2000a."Making the Transition to Private Own- Klingen, Christoph, Beatrice Weder, and Jeromin ership." Finance and Development 37(3): 12­13. Zettelmeyer. 2004."How Private Creditors Fared in ------. 2000b. "What the Change of System from Emerging Debt Markets, 1970­2000." IMF Working Socialism to Capitalism Does and Does Not Mean." Paper WP/04/13. International Monetary Fund, Journal of Economic Perspectives 14(1): 27­42. Washington, D.C. ------. 2000c. "Ten Years after `The Road to a Free Klitgaard, Robert. 1988. Tropical Gangsters. New York: Economy':The Author's Self-Evaluation." Economic Basic Books. Systems 24(4): 353­59. ------. 1990."Cleaning Up and Invigorating the Civil Kose, A., E. Prasad, and M.Terrones. 2003. "Financial Service." Public Administration and Development 17(5): Integration and MacroeconomicVolatility." IMF Staff 487­509. Online at http://wbln0018.worldbank Paper No. 50: 119­142. International Monetary .org/Network/PREM/PREMDocLib.nsf/58292AB Fund,Washington, D.C. 451257BB9852566B4006EA0C8/256D2F1ADB2A Kossick, Roberto M. 2003."Modernizing Mexico's Tax E2F085256713000468E0. Administration:The Development and Implementa- Klitgaard, Robert, Ronald Maclean-Abaroa, and H. tion of the E-Sat Mexico."Case study,E-Government Lindsey Parris.2000.Corrupt Cities:A Practical Guide to Thematic Group.World Bank,Washington, D.C. Cure and Prevention. Oakland, Calif.: Institute for Kovacic,William. 1995."Procurement Reform and the Contemporary Studies and Washington, D.C.:World Choice of Forum in Bid Protest Disputes." Adminis- Bank Institute. trative Law Journal (American University) 9: 461. Knack, Stephen, and Philip Keefer. 1995."Institutions Kraay, Aart. 2004.When Is Growth Pro-Poor? Cross- and Economic Performance: Cross-Country Tests Country Evidence." IMF Working Paper 04/47. Using Alternative Institutional Measures." Economics International Monetary Fund,Washington, D.C. and Politics 7: 207­27. Kraay,Aart, andV. Nehru. 2003."When Is External Debt Kogot, Bruce, and Andrew Spicer. 2002."Capital Market Sustainable?" Policy Research Working Paper 3200, Development and Mass Privatization Are Logical February.World Bank,Washington, D.C. Contradictions: Lessons from Russia and the Czech Kritzer, Barbara E. 2001. "Social Security Reform in Republic."Industrial and Corporate Change 11(1):1­37. Central and Eastern Europe:Variations on a Latin Kokko,Ari. 1994."Technology, Market Characteristics, AmericanTheme."Social Security Bulletin 64(4).Wash- and Spillovers." Journal of Development Economics 43. ington, D.C: U.S. Social Security Administration. BIBLIOGRAPHY 345 Krueger, Anne O. 1978. Foreign Trade Regimes and Eco- Lau,Lawrence J.,Yingyi Qian,and Gerard Roland.2000. nomic Development: Liberalization Attempts and Conse- "Reform without Losers: An Interpretation of quences. Lexington, Mass.: Ballinger. China's DualTrack Approach toTransition." Journal of ------. 1983. Trade and Employment in Developing Coun- Political Economy 108(1). tries. Chicago: University of Chicago Press. Laursen,T.,and S.Mahajan.2004."Volatility,Income Dis- ------. 1997. "Trade Policy and Economic Develop- tribution and Poverty." In J.Aizenmann and B. Pinto, ment: How We Learn." American Economic Review eds., Managing Volatility and Crises: A Practitioner's 87(1): 1­22. Guide.Washington, D.C.:World Bank. ------. 2004."Meant Well,Tried Little, Failed Much: Le Houerou, Philippe, and Robert Taliercio. 2002. Policy Reforms in Emerging Market Economies." "Medium-Term Expenditure Frameworks: From Roundtable Lecture (March 23), Economic Honors Concept to Practice. Preliminary Lessons from Society, NewYork University. Africa."Africa RegionWorking Paper No.28 (Febru- Krugman, Paul. 1979."A Model of Balance of Payments ary).World Bank,Washington, D.C. Crises." Journal of Money, Credit, and Banking 11(3): Lee,HaYan,Luca Anotnio Ricci,and Roberto Rigobon. 311­25. 2004."Once Again, Is Openness Good for Growth?" ------. 1998."Firesale FDI."Working Paper, Massachu- Journal of Development Economics 75(2) (December): setts Institute ofTechnology. 451­72. ------. 1999."Balance Sheets, theTransfer Problem and Leite, C., and J.Weidmann. 1999."Does Mother Nature Financial Crises." In P. Isard,A. Razin, and A. Rose, Corrupt? Natural Resources, Corruption, and Eco- eds., International Finance and Financial Crises. Wash- nomic Growth."IMFWorking Paper No.85.Interna- ington, D.C.: International Monetary Fund. Partly tional Monetary Fund,Washington, D.C. reprinted from InternationalTax and Public Finance 6(4). Lemarchand,Rene.1972."Political Clientelism and Eth- Boston: Dordrecht and London: Kluwer Academic. nicity in Tropical Africa: Competing Solidarities in Kuznets, S. 1955. "Economic Growth and Income Nation-Building." American Political Science Review Inequality." American Economic Review 65: 1­28 66(1): 68­90. Laeven, L. 2002a."Pricing of Deposit Insurance." Work- Levine, David. 1995. Reinventing theWorkplace: How Busi- ing Paper No. 2871.World Bank,Washington, D.C. ness and Employees Can Both Win.Washington, D.C.: ------. 2002b. "Bank Risk and Deposit Insurance." Brookings Institution. World Bank Economic Review 16: 109­37. Levine, R. 1998."Financial Development and Economic ------. 2002c."International Evidence on the Cost of Growth:Views and Agenda." Journal of Economic Liter- Deposit Insurance." Quarterly Review of Economics and ature XXXV(2): 688­726. Finance 42(4): 721­73. ------. 2003. "More on Finance and Growth: More Lall, S. 1999. The Technological Response to Import Liberal- Finance, More Growth?" Federal Reserve Bank of St. ization in Sub-Saharan Africa. London: Macmillan. Louis Review 85: 31­46. Landes, David. 1998. The Wealth and Poverty of Nations: Levine, R., and S. Zervos. 1998."Stock Markets, Banks, Why Some Are So Rich and Some So Poor? NewYork: and Economic Growth." American Economic Review Norton. 88(3): 537­88. Lane, Philip R. 2003."The Cyclical Behavior of Fiscal Levine,R.,N.Loayza,andT.Beck.2000."Financial Inter- Policy: Evidence from the OECD." Journal of Public mediation and Growth:Causality and Causes."Journal Economics 87: 2661­75. of Monetary Economics 46(1): 31­77. Lane, Phillip R., and Aaron Tornell. 1999."Voracity and Levinsohn, James. 1999. "Employment Responses to Growth in Discrete Time." Economic Letters 62(1): International Liberalization in Chile." Journal of Inter- 139­45. national Economics 47: 321­44. LaPorta, Rafael, and López de Silanes. 1999."The Bene- Levinsohn, James, and Margaret McMillan. 2004."Does fits of Privatization:Evidence from México."Quarterly Food Aid Harm the Poor: Household Evidence from Journal of Economics 114(4): 1193­1242. Ethiopia." In Harrison (forthcoming). ------.2003."Related Lending."Quarterly Journal of Eco- Levy, Brian, and Sahr Kpundeh, eds. 2004. Building State nomics CXVII: 231­68. Capacity in Africa: New Approaches, Emerging Lessons. LaPorta, Rafael, F. López de Silanes, and A. Shleifer. Washington, D.C.:World Bank Institute. 2002a."Government Ownership of Banks." Journal of Lewis, P., and H. Stein. 2002."The Political Economy of Finance 57(1): 265­301. Financial Liberalization in Nigeria." In H. Stein, O. ------.2002b."WhatWorks in Securities Laws? Lessons Ajakaiye, and P. Lewis, eds., Deregulation and the Bank- for LatinAmerica."Draft,Harvard University andYale ing Crisis in Nigeria. London: Palgrave. University. Lieberman, Ira, and John Nellis. 1994."Russia: Creating Lardy, Nicholas R. 2002. Integrating China into the Global Private Enterprises and Efficient Markets." World Economy.Washington,D.C.:Brookings Institution Press. Bank,Washington, D.C. 346 BIBLIOGRAPHY Lim, Ewe-Ghee. 2001."Determinants of, and the Rela- for the Russian Federation."World Bank,Washington, tion between, Foreign Direct Investment and D.C. Growth: A Summary of Recent Literature." IMF Manning, Nicholas, and Richard Stapenhurst. 2002. Working Paper 01/175. International Monetary "Strengthening Oversight by Legislatures." PREM Fund,Washington, D.C. Note No.74 (October).World Bank,Washington,D.C. Lindemann, David, Michal Rutkowski, and Oleksiy Manzano, Osmel, and Roberto Rigobon. 2001. Sluchynskyy. 2000."The Evolution of Pension Sys- "Resource Curse or Debt Overhang?"NBERWork- tems in Eastern Europe and Central Asia:Opportuni- ing Paper No. 8390. National Bureau of Economic ties, Constraints, Dilemmas, and Emerging Practices." Research, Cambridge, Mass. World Bank,Washington, D.C. Marques, Gustavo, and Carmen Pagés. 1998."Trade and Lippman, Hal, and Jan Emmert. 1997."Assisting Legisla- Employment: Evidence from Latin America and the tures in Developing Countries:A Framework for Plan- Caribbean." IDB Working Paper No. 366. Inter- ning, Programming, and Implementation." Center for American Development Bank,Washington, D.C. Development Information and Evaluation, USAID, Martinez-Peria, M. S., and S. Shmuckler. 2001. "Do Washington,D.C. Depositors Punish Bank for Bad Behavior? Market Little,Ian,Tibor Scitovsky,and Maurice Scott.1970.Indus- Discipline, Deposit Insurance, and Banking Crises." try andTrade in Some Developing Countries:A Compara- Journal of Finance 56(3): 1029­51. tive Study. London and NewYork: Oxford University Martinez-Vazquez,Jorge,and Jameson Boex.2001."Rus- Press for the Development Centre of the Organisation sia's Transition to a New Federalism."WBI Learning for Economic Co-operation and Development. Resources No. 97.World Bank Institute,Washington, Litvack, J., J. Ahmad, and R. Bird. 1998. "Rethinking D.C. Decentralization at the World Bank."World Bank, Masson, P. 2001."Exchange Rate Regime Transitions." Washington, D.C. Journal of Development Economics 64(2): 571­86. Loayza N., P. Fajnzylber, and C. Calderón. 2002."Eco- Mauro, Paolo. 1995."Corruption and Growth." Quarterly nomic Growth in Latin America and the Caribbean. Journal of Economics 110: 681­712. Stylized Facts, Explanations, and Forecasts." Draft, ------.1997."WhyWorry about Corruption?" Economic World Bank (June). Online at http://wbln0018 Issues No. 6.Washington, D.C.: International Mone- .worldbank.org/lac/lacinfoclient.nsf/8d6661f6799ea tary Fund. 8a48525673900537f95/92400b44aeb50a0185256cad ------. 1998a."Corruption: Causes, Consequences, and 00708a4a/$FILE/N.Loayza.pdf. Agenda for Further Research." Finance and Develop- Lora, Eduardo. 2001a. "Structural Reforms in Latin ment 35(1) (March). America: What Has Been Reformed and How to ------. 1998b. "Corruption and the Composition of Measure It." Inter-American Development Bank, Public Expenditures." Journal of Public Economics 29: Washington, D.C. (December). 263­79. ------. 2001b."El crecimiento económico en América McKinnon, R. I. 1973. Money and Capital in Economic Latina después de una década de reformas estruc- Development. Washington, D.C.: Brookings Institu- turales." Draft. InterAmerican Development Bank, tion. Research Department,Washington, D.C. McMillan, John, and Christopher Woodruff. 2002."The Lucas, Robert E., Jr. 1987. Studies in Business-Cycle The- Central Role of Entrepreneurs in Transition ory. Cambridge, Mass. and London: MIT Press. Economies." Journal of Economic Perspectives 16(3): Luttmer, Erzo F. P. 2001."Measuring Poverty Dynamics 153­70. and Inequality in Transition Economies: Disentan- McMillan, Margaret, Dani Rodrik, and Karen Horn gling Real Events from Noisy Data."Working Paper Welch. 2002. "When Economic Reform Goes No. 2549.World Bank,Washington, D. C. Wrong: Cashews in Mozambique." Draft (July). Maddison, Angus. 1995. Monitoring the World Economy: Meagher, Patrick. 2002."Anti-Corruption Agencies: A 1820­1992. Paris: Development Centre of the Review of Experience." Institutional Reform and Organisation for Economic Co-operation and Devel- Informal Sector Center, University of Maryland opment. (August 2). ------.2002.TheWorld Economy:A Millennial Perspective. Megginson, William, and Jeffrey Netter. 2001. "From Paris: Development Centre of the Organisation for State to Market: A Survey of Empirical Studies on Economic Co-operation and Development. Privatization." Journal of Economic Literature (June): Mamdani, Mahmood. 1996. Citizen and Subject: Contem- 321­89. porary Africa and the Legacy of Late Colonialism. Prince- Megginson,William L., Robert C. Nash, and Matthias ton, N.J.: Princeton University Press, 1996. van Randenborgh. 1998. "The Privatization Divi- Manning, Nicholas, and Neil Parison. 2003."Interna- dend." Journal of Finance 49(2). Also summarized in tional Public Administration Reform: Implications FPD Note No. 68,World Bank. BIBLIOGRAPHY 347 Mehlum, Halvor, Karl-Ove Moene, and Ragnar Torvik. and Composition of Capital Flows? Evidence from 2003."Institutions and the Resource Curse." Draft, the 1990s." Journal of International Money and Finance 29/2002. Oslo University, Department of Economics 18(4) (August): 619­35. and EconomicsWorking PaperArchive atWashington Moog, Robert. 1997. Whose Interests Are Supreme? Orga- University in St. Louis. nizational Politics in the Civil Courts in India. Ann Meier, Gerald, and Dudley Seers, eds. 1984. Pioneers in Arbor, Mich.:Association for Asian Studies. Development. NewYork and Oxford: Oxford Univer- Moran,T. 1998."Foreign Direct Investment and Devel- sity Press opment: The New Policy Agenda for Developing Mellinger, Andrew D., Jeffrey D. Sachs, and John L. Countries and Economies inTransition." Institute for Gallup. 1999."Climate,Water Navigability, and Eco- International Economics,Washington, D.C. nomic Development."Working Paper No. 24 (Sep- Moreira, Mauricio Mesquita, and Sheila Najberg. 2000. tember). Center for International Development, "Trade Liberalization in Brazil:Creating or Exporting Washington, D.C. Jobs?" Journal of Development Studies (36)3: 78. Mendoza, E. 2001."Credit, Prices, and Crashes: Business Mukand, Sharun, and Dani Rodrik. 2002."In Search of Cycles with a Sudden Stop." NBER Working Paper the Holy Grail: Policy Convergence, Experimenta- No. 8338. National Bureau of Economic Research, tion, and Economic Performance." Draft, Harvard Cambridge, Mass. University. Messick,Richard.1999."Judicial Reform and Economic Murphy,Kevin,Andrei Shleifer,and RobertVishny.1989. Development: A Survey of the Issues." World Bank "Industrialization and the Big Push." Journal of Politi- Research Observer 14(1): 117­36. cal Economy 97(5): 1003­26. Michaely, Michael, Demetris Papageorgiou, and Murrell,Peter.1992."Evolution in Economics and in the Armeane Choksi. 1991. Liberalizing Foreign Trade: Economic Reform of Centrally Planned Lessons of Experience in the Developing World. Cam- Economies." In Christopher C. Clague and Gordon bridge, Mass.: Blackwell. Rausser, eds., The Emergence of Market Economies in Michalopoulos, Constantine. 1999. "Trade Policy and Eastern Europe. Cambridge, Mass.: Blackwell. Market Access Issues for Developing Countries: Mussa, M. 2002. Argentina and the Fund: From Triumph to Implications for the New Millennium Round." Pol- Tragedy. Washington, D.C.: Institute for International icy ResearchWorking Paper 2214.World Bank,Wash- Economics. ington, D.C. Myerson,Roger B.1999."Theoretical Comparisons of Elec- Milanovic, Branko. 2004. "Why We All Do Care about toral Systems."European Economic Review 43:671­97. Inequality (but Are Loath to Admit It)." Working Myrdal, Gunnar. 1972. Asian Drama:An Inquiry into the Paper, Carnegie Endowment for International Peace, Poverty of Nations. NewYork: Random House. Washington, D.C. Naim, Moises. 1995. "Latin America: The Morning Milesi-Ferretti, Gian-Maria, Roberto Perotti, and Marco After." Foreign Affairs (July/August). Rostagno. 2002."Electoral Systems and the Compo- ------. 1999."Fads and Fashion in Economic Reforms: sition of Government Spending." Quarterly Journal of Washington Consensus or Washington Confusion?" Economics 117: 609­57. Working Draft of a paper prepared for the IMF Con- Minsky, H. 1992."The Financial Instability Hypothesis, ference on Second Generation Reforms,Washington, Capitalistic Processes and the Behavior of the Econ- D.C. (October 26). Online at http://www.imf.org/ omy." In C. Kindleberger and J. Laffargue, eds., Finan- external/pubs/ft/seminar/1999/reforms/Naim.HTM. cial Crises:Theory, History and Policy. Cambridge, U.K.: Narayan, Deepa, Robert Chambers, Meera K. Shah, and Cambridge University Press. Patti Petesch. 2000. Voices of the Poor (2 vols.). New Mishkin, F., ed. 2001. Prudential Supervision:What Works York and Oxford: Oxford University Press. andWhat Doesn't. Chicago: University of Chicago for Nash, John, and Wendy Takacs, eds. 1998. Trade Policy the National Bureau of Economic Research. Reform: Lessons and Implications. Washington, D.C.: Mishra, Banamber, and Matiur Rahman. 1998."Exports, World Bank. Foreign Aid, Exchange Rate, and Economic Growth: Nash,Manning.1963."Party Building in Upper Burma." Empirical Evidence from 28 Low-Income LDCs."Scan- Asian Survey 3 (April): 196­202. Reprinted in Scott dinavian Journal of DevelopmentAlternatives 17:21­31. (1972). Moggridge, D. E. 1976. John Maynard Keynes. NewYork: Nayak, P. Jayendra. 1999. "Regulation and Market Penguin. Microstructure." In Hanson and Kathuria (1999): Montiel, P. 2002. "Ecuador: Una Estrategia de Crec- 266­304. imiento para una Economía Dolarizada." Cuestiones Ndegwa,Stephen N.,and Brian Levy.2003."The Politics Económicas 18(3): 133­225. of Decentralization in Africa:A Comparative Analy- Montiel,P.,and C.Reinhart.1999."Do Capital Controls sis."In Building State Capacity inAfrica:NewApproaches, and Macroeconomics Policies Influence theVolume Emerging Lessons.Washington, D.C.:World Bank. 348 BIBLIOGRAPHY Nellis, John. 2002."The World Bank, Privatization, and Social Protection Discussion Paper No. 9. World Enterprise Reform in Transition Economies:A Ret- Bank,Washington, D.C. rospective Analysis." Operations Evaluation Depart- Palley,Thomas I. 2003."Lifting the Natural Resource ment Background Paper.World Bank,Washington, Curse." Foreign Service Journal 80(12): 54­61. D.C. Panagariya, Arvind. 2003."Miracles and Debacles: Do ------. 2003a."Privatization in Africa:What Has Hap- Free-Trade Skeptics Have a Case?"EconomicsWork- pened? What Is to Be Done?" CGD Working Paper ing PaperArchive,Washington University in St.Louis, No. 25 (February). Center for Global Development, InternationalTrade Series No. 0308013. Washington, D.C. Paul,Samuel.1995.Strengthening PublicAccountability:New ------. 2003b."Privatization in Latin America." CGD Approaches and Mechanisms. Bangalore, India: Public Working Paper No. 31 (August). Center for Global Affairs Centre. Development,Washington, D.C. ------.2003."The Report Card Survey:ATenYear Ret- Nellis, John, and Mary Shirley. 1991."Public Enterprise rospective." Seminar presentation, World Bank, Reform: The Lessons of Experience." Economic November. Development Institute Development Study.World Perry, G. 2003."Can Fiscal Rules Help Reduce Macro- Bank,Washington, D.C. economicVolatility?"Policy ResearchWorking Paper Nocum,Armand. 2004."Lifestyle Checks to Resume." Series No. 3080.World Bank,Washington, D.C. Philippines Daily Inquirer, May 13. Perry, G., and L. Servén. 2002."The Anatomy of a Mul- North, Douglass C. 1990. Institutions, Institutional Change tiple Crisis:WhyWasArgentina Special andWhat Can and Economic Performance.NewYork:Cambridge Uni- We Learn from It?"Draft.World Bank,Chief Econo- versity Press. mist's Office, Latin America and the Caribbean ------. 1994."Economic Performance through Time." Region,Washington, D.C. American Economic Review 84(3): 359­68. ------. 2003."The Anatomy of a Multiple Crisis:Why ------. 2002."Institutions and Credible Commitment." Was Argentina Special andWhat CanWe Learn from St. Louis, Mo.:Washington University. It?"Working Paper No. 3081.World Bank,Washing- North, Douglass, and Barry Weingast. 1989."Constitu- ton, D.C. tions and Commitment:The Evolution of Institutions Perry, Guillermo, Daniel Lederman,William Maloney, Governing Public Choice in Seventeenth-Century and Luis Servén. 2003. "Lessons from NAFTA for England." Journal of Economic History 49(4): 803­32. LatinAmerica and the Caribbean."LatinAmerica and Obstfeld, M. 1994. "The Logic of Currency Crises." the Caribbean Region, World Bank, Washington, Cahiers Economiques et Monetaires 43:189­213. D.C.,April. Ocampo, Jose Antonio. 2004."Latin America's Growth Persson,Torsten, and Guido Tabellini. 2000. Political Eco- and Equity Frustrations during Structural Reforms." nomics: Explaining Public Policy. Cambridge, Mass.: Journal of Economic Perspectives 18(2): 67­88. MIT Press. Oliva, Maria-Angels. 2000."Estimation of Trade Protec- ------. 2004."Constitutional Rules and Fiscal Policy tion in Middle East and North African Countries." Outcomes." American Economic Review 94(1): 25­45. IMF Working Paper No. 00/27, February 1. Interna- Persson,Torsten, Gerard Roland, and Guido Tabellini. tional Monetary Fund,Washington, D.C. Online at 2003."How Do Electoral Rules Shape Party Struc- http://www.imf.org/external/pubs/ft/wp/2000/wp tures, Government Coalitions, and Economic Poli- 0027.pdf. cies?" NBER Working Paper No. 10176. National Olson, Mancur. 1965. The Logic of Collective Action: Public Bureau of Economic Research, Cambridge, Mass. Goods and the Theory of Groups. Cambridge, Mass.: Persson Torsten, Guido Tabellini, and Francesco Trebbi. Harvard University Press. 2000. "Electoral Rules and Corruption." CEPR ------. 1982. The Rise and Decline of Nations. New Working Paper. Centre for Economic Policy Haven, Conn.:Yale University Press. Research, London. ------.2000.Power and Prosperity:Outgrowing Communist ------. 2003."Electoral Rules and Corruption." Journal and Capitalist Dictatorships. NewYork: Basic Books. of the European Economic Association 1(4): 958­89. Organisation for Economic Co-operation and Develop- Philp, Mark. 2003."Why Systems Produce Corruption." ment (OECD). 1996­2002. Indicators of Tariff and Draft. University of Oxford. Non-TariffTrade Barriers. Pistor, Katharina, Martin Raiser, and Stanislav Gelfer. ------. 1998. The Role of the Legislature. Paris: OECD. 2000."Law and Finance in Transition Economies." Osborne,David,andTed Gaebler.1992. Reinventing Gov- EBRD Working Paper 48. European Bank for ernment: How the Entrepreneurial Spirit Is Transforming Reconstruction and Development, London. the Public Sector. Reading, Mass.:Addison-Wesley. Platteau, Jean-Philippe. 2004."Collective Action and the Palacios, Robert J., and Montserrat Pallares-Miralles. Commons:The Role of Inequality." In Baland, J. M., 2000."International Patterns of Pension Provision." P. Bardhan, and S. Bowles, eds., Inequality, Cooperation, BIBLIOGRAPHY 349 and Environmental Sustainability. Princeton, N.J.: ------. 2003b."Who Is Not Poor: Proposing A Higher Princeton University Press. International Standard for Poverty." CGD Working Pohl, Gerhard, Robert Anderson, and Simeon Djankov. Paper 33, November. Center for Global Develop- 1997."Privatization and Restructuring in Eastern and ment,Washington, D.C. Central Europe--Evidence and Policy Options."Tech- Pritchett, Lant, and Michael Woolcock. 2004."Solutions nical Paper No.368.World Bank,Washington,D.C. When the Solution Is the Problem:Arraying the Dis- Polidano, Charles. 1999."The New Public Management array in Development." World Development 32(2): in Developing Countries." Public Policy and Man- 191­212. agementWorking Paper No.13 (November).Institute Przeworski, Adam, Michael E. Alvarez, Jose Antonio for Development Policy and Management,University Cheibub,and Fernando Limongi.2000.Democracy and of Manchester, U.K. Development: Political Institutions and Well-Being in the Pollitt, Michael G. 2003."Electricity Reform in Chile World, 1950­1990. NewYork: Cambridge University and Argentina: Lessons for Developing Countries." Press. Paper presented at the Cambridge-MIT Institute Pyramid Research. 2001. Communications Markets in Mex- Electric Power Autumn Research Seminar. Cam- ico--Blurring Barriers and Escalating Competition. Cam- bridge,Mass. November 7. bridge,Mass.:Pyramid Research. Porcalla, Delon. 2004."BIR Exec in Life Style Check Qian,Yingyi.2000."The Process of China's MarketTran- Suspended." Philippines Star, May 20. sition 1978­98: The Evolutionary, Historical, and Porto,Guido.2003."Using Survey Data toAssess the Dis- Comparative Perspective." Stanford University, Calif., tributional Effects ofTrade Policy."Draft.World Bank, April. Washington, D.C. ------.2002."How ReformWorked in China."Depart- Posner,Daniel.Forthcoming."Measuring Ethnic Fraction- ment of Economics, University of California, Berke- alization in Africa."American Journal of Political Science. ley,August. Poterba, James M. 1996."Budget Institutions and Fiscal Radelet, Steven, and Jeffrey Sachs. 1998."The Onset of Policy in the US States."NBERWorking Paper 5449. the East Asian Financial Crisis." Draft. Harvard Insti- National Bureau of Economic Research,Cambridge, tute for International Development (March 30). Mass. Radelet, Steven, Cristina Garcia, and Mumtaz Hussain. Powell, John Duncan. 1970."Peasant Society and Clien- 2000."Trade and Exports." Background paper exam- telist Politics." American Political Science Review 64(2): ining the state of the Andean Region for the Andean 411­25. Competitiveness Project,April. Pradhan, Menno,Asep Suryahadi, Sudarno Sumarto, and Rajkumar, Andrew Sunil, and Vinaya Swaroop. 2002. Lant Pritchett.2001."Eating LikeWhich`Joneses?'An "Public Spending and Outcomes: Does Governance Iterative Solution to the Choice of a Poverty Line Matter?" Policy Research Working Paper No. 2840 `Reference Group.'"Review of Income andWealth 47(4) (May).World Bank,Washington, D.C. (December): 473­488. Ramachandran, S. 1997. "The Veil of Vouchers." FPD Prado,C.1972.História e Desenvolvimento. São Paulo: Edi- Note No.108.World Bank,Washington, D.C. tora Brasiliense. Ramey, G., and V. Ramey. 1995."Cross-Country Evi- Prasad, E., K. Rogoff, S.Wei, and M. Kose. 2003."Effect dence on the Link between Volatility and Growth." of Financial Globalization on Developing Countries: American Economic Review 85(5): 1138­51. Some Empirical Evidence." Draft. International Ravallion, Martin. 1994."Measuring SocialWelfare with Monetary Fund,Washington, D.C. and without Poverty Lines."American Economic Review Prayas Energy Group. 2003. "Performance of Private (American Economic Association) 84(2) (May): Electricity Distribution Utilities in India: Need for 359­64. In-Depth Review and Benchmarking." Occasional ------. 2001."Growth, Inequality and Poverty: Looking Report. Prune, India. beyond Averages." World Development 29(11): Pritchett, Lant. 1996."Measuring Outward Orientation 1803­15. in LDCs: Can It Be Done?" Journal of Development ------. 2003a."Inequality Convergence." Economics Let- Economics 49. ters 80(3): 351­56. ------.1997."Divergence,BigTime."Journal of Economic ------. 2003b."The Debate on Globalization, Poverty, Perspectives 11(3) (Summer): 3­17. and Inequality:Why Measurement Matters." Devel- ------. 2000. "The Tyranny of Concepts: CUDIE opment Research GroupWorking Paper 3038.World (Cumulated, Depreciated, Investment Effort) Is Not Bank,Washington, D.C. Capital." Journal of Economic Growth 5(4): 361­84. ------. 2004."Pessimistic on Poverty?" The Economist, ------. 2003a."Reform Is Like a Box of Chocolates: April 4. Understanding Growth Disappointments and Sur- Ravallion, Martin, and Michael Lokshin. 2004."Gainers prises." Draft (September), Harvard University. and Losers from Trade Reform in Morocco." Policy 350 BIBLIOGRAPHY Research Working Paper No. 3368. World Bank, Rodrik,Dani.1992."The Limits ofTrade Policy Reform in Washington, D.C. DevelopingCountries."JournalofEconomicPerspectives(6)1. Razin, Assaf, Efraim Sadka, and Chi-wa Yuen. 1999. ------. 1997a."TFPG Controversies, Institutions, and "Excessive FDI under Asymmetric Information." Economic Performance in East Asia." NBER Work- NBERWorking Paper No. 7400. National Bureau of ing Paper No. 5914. National Bureau of Economic Economic Research, Cambridge, Mass. Research, Cambridge, Mass. Redding, S., and Venables, A. 2003. "Geography and ------.1997b."Trade Policy and Economic Performance Export Performance: External Market Access and in Sub-Saharan Africa. Draft, Harvard University. Internal Supply Capacity." CEPR Discussion Paper ------. 1998."Who Needs Capital-Account Convert- No. 3807. Centre for Economic Policy Research, ibility?"In Stanley Fischer and others,Should the IMF London. Pursue Capital-Account Convertibility? Essays in Inter- Reinhart,Carmen,and Graciela L.Kaminsky.1999."The national Finance No. 207. Princeton, N.J.: Interna- Twin Crises:The Causes of Banking and Balance-of- tional Finance Section, Princeton University. Payments Problems." American Economic Review 89(3). ------. 1999a."The New Global Economy and Devel- Reinhart, Carmen, K. Rogoff, and M. Savastano. 2003. oping Countries: Making Openness Work." Policy "Addicted to Dollars." NBER Working Paper 10015. Essay No.24. Overseas Development Council,Wash- National Bureau of Economic Research, Cambridge, ington, D.C. Mass. ------. 1999b."Where Did All the Growth Go? Exter- Reinikka, Ritva, and Jakob Svensson. 2003."The Power nal Shocks, Social Conflict and Growth Collapses." of Information: Evidence from a Newspaper Cam- Journal of Economic Growth 4(4): 385­412. paign to Reduce Capture." Policy ResearchWorking ------. 2001a. "Development Strategies for the 21st Paper 3239. Development Research Group,World Century."In Boris Pleskovic and Nicholas Stern,eds., Bank,Washington, D.C. AnnualWorld Bank Conference on Development Econom- ------. 2004."Local Capture: Evidence from a Central ics 2000, 85­108.Washington, D.C.:World Bank. Government Transfer Program in Uganda." Quarterly ------.2001b."Why IsThere So Much Economic Inse- Journal of Economics 119(2). curity in Latin America? " CEPAL Review 73: 7­30. Reserve Bank of India. 1998. Report on Currency and ------. 2002a. "Trade Policy Reform as Institutional Finance 1998/99. Bombay: Reserve Bank of India. Reform." In B. M. Hoekman, P. English, and A. Mat- Reynolds, Lloyd. 1983."The Spread of Modern Eco- too. Development,Trade, and the WTO: A Handbook. nomic Growth, 1850­1950." Journal of Economic Liter- Washington, D.C.:World Bank. ature 21: 940­77. ------. 2002b. "Feasible Globalizations." Discussion Rigobon,Roberto,and Dani Rodrik.2004."Rule of Law, Paper No. 3524. Centre for Economic Policy Democracy, Openness, and Income: Estimating the Research, London. Interrelationships." NBERWorking Paper No. 10750. ------. 2003a."Growth Strategies." CEPR Discussion National Bureau of Economic Research, Cambridge, Paper No. 4100. Centre for Economic Policy Mass. Research, London. Robinson, James A. 2002."States and Power in Africa by ------, ed. 2003b. In Search of Prosperity:Analytic Narra- Jeffrey Herbst: A Review Essay." Journal of Economic tives on Economic Growth. Princeton, N.J.: Princeton Literature XL (June): 510­19. University Press. Robinson, James A., Ragnar Torvik, and ThierryVerdier. ------. 2003c. "Economic Reform without Rules of 2002."Political Foundations of the Resource Curse." Thumb."Draft,Harvard University,October.Available CEPR Discussion Papers 3422, June. online at http://ksghome.harvard.edu/~drodrik/ Robinson, M. 2002. The Microfinance Revolution, Vol. 2: Stiglitzconferencenotes.pdf. Lessons from Indonesia.Washington, D.C.:The World Rodrik, Dani, and Ricardo Hausmann. 2003."Discover- Bank. ing El Salvador's Production Potential." Draft (Sep- Rodden, J., Gunnar Eskeland, and Jennie Litvack, eds. tember). 2003. Decentralization and Hard Budget Constraints. Rodrik, Dani, and Arvind Subramanian. 2003."The Pri- Boston: MIT Press. macy of Institutions." Finance and Development 40(2): Rodríguez, Francisco, and Dani Rodrik. 2000."Trade 31­34. Policy and Economic Growth:A Skeptic's Guide to ------. 2004. "From `Hindu Growth' to Productivity the Cross-National Evidence." Electronic Working Surge:The Mystery of the Indian GrowthTransition." Papers No. 9901, University of Maryland, Depart- IMF Working Paper 04/77. International Monetary ment of Economics. Published in Macroeconomics Fund,Washington, D.C. Annual 2000. Cambridge, Mass: MIT Press. Online at Rodrik, D., and A.Velasco. 1999."Short-Term Capital http://ksghome.harvard.edu/~.drodrik.academic.ksg Flows." NBER Working Paper No. 7364. National /papers.html. Bureau of Economic Research, Cambridge, Mass. BIBLIOGRAPHY 351 Rodrik,Dani,Arvind Subramanian,and FrancescoTrebbi. Sala-i-Martin, Xavier. 2003."The Disturbing `Rise' of 2002."Institutions Rule:The Primacy of Institutions Global Income Inequality." NBER Working Paper over Integration and Geography in Economic Devel- No. 8904. National Bureau of Economic Research, opment." IMF Working Paper 02/189. National Cambridge, Mass. Bureau of Economic Research, Cambridge, Mass. Sala-i-Martin, Xavier, and Arvind Subramanian. 2003. Rofman,Rafael.2002."The Pension System inArgentina: "Addressing the Natural Resource Curse:An Illustra- Learning the Lessons."Background Paper for Regional tion from Nigeria." IMF Working Paper 03/139. Study on Social Security Reform, Office of the Chief International Monetary Fund,Washington, D.C. Economist, Latin America and the Caribbean Sala-i-Martin, Xavier, Gernot Doppelhofer, and Ronald Regional Office.World Bank,Washington,D.C. Miller. 2003."Determinants of Long-Term Growth: Rogoff, Kenneth. 1985."The Optimal Degree of Com- A Bayesian Averaging of Classical Estimates (BACE) mitment to an Intermediate MonetaryTarget." Quar- Approach."University of Cambridge Faculty of Eco- terly Journal of Economics 100: 1169­89. nomics and Politics, Columbia University Depart- Romer,Paul M.1983."Dynamic Competitive Equilibria ment of Economics, and Columbia University with Externalities, Increasing Returns and Department of Economics. Unbounded Growth." Journal of Monetary Economics Sanguinetti, P., and M.Tommassi. 2003."Intergovern- 26: 47­75. mentalTransfers and Fiscal Behavior:Insurance versus ------. 1986. "Increasing Returns and Long-Run Aggregate Discipline." Journal of International Econom- Growth." Journal of Political Economy 94: 1002­38. ics 62(1): 149­70. Rose-Ackerman, Susan. 1978. Corruption: A Study in Sarkar, S. 2000."On the Investment-Uncertainty Rela- Political Economy. NewYork:Academic Press. tionship in a Real Options Model."Journal of Economic ------. 2004."Poor Governance and Corruption." Pre- Dynamics and Control 24(2): 219­25. liminary draft, Copenhagen Consensus Challenge Savastano, M. 1992."The Pattern of Currency Substitu- Paper.Yale University. tion in LatinAmerica:An Overview."Revista deAnáli- Rosenbluth, Frances, and Michael F. Thies. 2001."The sis Económico 7(1): 29­72. Electoral Foundations of Japan's Banking Regula- ------. 1996. "Dollarization in Latin America." In P. tion." Policy Studies Journal 29(1): 23­37. Mizen and E.Pentacost,The Macroeconomics of Interna- Rosenstein-Rodan, Paul. 1943."Problems of Industrial- tional Currencies:Theory, Policy, and Evidence. Chel- ization of Eastern and South-Eastern Europe." Eco- tenham, U.K.: Edward Elgar. nomic Journal. Schady, Norbert R. 2000."Political Economy of Expen- ------. 1984."Natura Facit Saltum:Analysis of the Dis- ditures by the Peruvian Social Fund (FONCODES), equilibrium Growth Process." In G. Meier and D. 1991­1995." American Political Science Review 94(2): Seers, eds., Pioneers in Development. NewYork: Oxford 289­304. University Press. Schiantarelli, F., I. Atiyas, G. Caprio, J. Hariss, and A. Rostow, Walt Whitman. 1952. The Process of Economic Weiss. 1994."Credit Where It Is Due? A Review of Growth. NewYork: Norton. the Macro and Micro Evidence on the Real Effects ------.1960.The Stages of Economic Growth:A Non-Com- of Financial Reform." In G. Caprio, I.Atiyas, and J. munist Manifesto. New York: Cambridge University A. Hanson, eds., Financial Reform:Theory and Experi- Press. ence. NewYork and London: Cambridge University ------. 1962. The Process of Economic Growth, 2nd ed. Press. NewYork: Norton. Schick,Allen. 1998."Why Most Developing Countries Sachs, Jeffrey D., and Andrew M.Warner. 1995a."Eco- Should Not Try New Zealand's Reforms." World nomic Reform and the Process of Global Integra- Bank Research Observer 13: 123­31. tion." Brookings Papers on Economic Activity, 1 (25th ------. 2003."The Performing State: Reflection on an Anniversary Issue). Idea Whose Time Has Come but Whose Implemen- ------. 1995b."Natural Resource Abundance and Eco- tation Has Not." Paper presented at the OECD nomic Growth. Paper No. 517a, Harvard Institute for Global Forum on Governance, London School of International Development. Economics, December 2­3. ------. 1997. "Sources of Slow Growth in African Schmukler, S., and L. Servén. 2002."Pricing Currency Economies." Journal of African Economies 6(3): 335­76. Risk under Currency Boards." Journal of Development ------. 2001."The Curse of Natural Resources." Euro- Economics 69(2): 367­91. pean Economic Review 45(4): 827­38. Schroeder, Larry. 2002."Mechanisms for Strengthening Saiegh, Sebastian, and Mariano Tommasi. 1999."Why Is Local Accountability." Draft, World Bank, World Argentina's Fiscal Federalism So Inefficient? Entering Bank, Washington, D.C. Online at http://www1 the Labyrinth." Journal of Applied Economics II(1): .worldbank.org/publicsector/decentralization/Feb20 169­209. 04Course/Background%20materials/Schroeder.doc. 352 BIBLIOGRAPHY Scott,James C.1972."Patron-Client Politics and Political Smoke. 2001. Fiscal Decentralization in Developing Coun- Change in Southeast Asia." American Political Science tries: A Review of Current Concepts and Practices. Review 66(1): 91­113. Geneva: United Nations Research Institute for Social Seers, Dudley. 1962."A Model of Comparative Rates of Development. Growth in the World Economy." Economic Journal, Solow, Robert M. 1956."A Contribution to the Theory March. of Economic Growth." Quarterly Journal of Economics Selowsky, Marcelo, and Ricardo Martin. 1998."Policy 70: 65­94. Performance and Output Growth in the Transition ------.1957."Technical Change and theAggregate Pro- Economies." American Economic Review 87: 350­53. duction Function," Review of Economics and Statistics Sen,Amartya. 1999. Development as Freedom. NewYork: 39(3): 312­20. Knopf. ------. 1971."Growth Theory: An Exposition." New Senhadji,Abdelhak S. 2000."How Significant Are Depar- York and Oxford: Oxford University Press. tures from Certainty Equivalence? SomeAnalytical and Standard and Poor's. 2003. Emerging Markets Database. Empirical Results." Review of Economic Dynamics 3(3): NewYork: Standard and Poor's. 597­617. Stapenhurst,Rick.2000.The Media's Role in Curbing Cor- Servén, L. 2003."Real Exchange Rate Uncertainty and ruption.Working Paper,World Bank Institute,Wash- Private Investment in LDCs." Review of Economics and ington, D.C. Statistics 85(1): 212­18. Stasavage D.2003.Public Debt and the Birth of the Democra- Servén, Luis, Rui Albuquerque, and Norman Loayza. tic State: France and Great Britain 1688­1789. New 2003."World Market Integration through the Lens of York: Cambridge University Press. Foreign Direct Investors." Draft (February).World Stein, E., E.Talvi, and A. Gristani. 1998."Institutional Bank,Washington, D.C. Arrangements and Fiscal Performance: The Latin Shah,A., and S.Thomas. 1999."Developing the Indian American Experience." NBER Working Paper No. Capital Market."In J.A.Hanson and S.Kathuria,eds., 6358.National Bureau of Economic Research,Cam- India: A Financial Sector for the Twenty-First Century. bridge, Mass. NewYork: Oxford University Press. Stigler, George J. 1971."TheTheory of Economic Reg- ------. 2003."Securities Market Efficiency." In J. Han- ulation." Bell Journal of Economics and Management Sci- son, P. Honohan, and G. Majnoni, eds., Globalization ence. Reprinted in Stigler (1975). and National Financial Systems. NewYork and Oxford: ------. 1975. The Citizen and the State: Essays on Regula- Oxford University Press for theWorld Bank. tion. Chicago: University of Chicago Press. Shatz,Howard,and AnthonyVenables.2000."The Geog- Stigler,George J.,and Claire Friedland.1962."What Can raphy of International Investment." In Gordon L. Regulators Regulate?The Case of Electricity."Journal Clark,Meric S.Gertier,and Maryann P.Feldman,eds., of Law and Economics 5: 1­16. The Oxford Handbook of Economic Geography. Oxford: Stiglitz, Joseph. 2001."From Miracle to Crisis to Recov- Oxford University Press. ery:Lessons from Four Decades of East Asian Experi- Shaw, E. 1973. Financial Deepening in Economic Develop- ence." In Joseph E. Stiglitz and Shahid Yusuf, ment. NewYork: Oxford University Press. Rethinking the East Asian Miracle. Washington, D.C.: Shepherd, Geoffrey. 2003."Understanding Public Orga- World Bank. nizations:An Aid to Reform in Developing Coun- Stiglitz,Joseph,and Karla Hoff.1990."Rural Credit Mar- tries." Draft,World Bank, PREM Network, Public kets:Puzzles and Policy Perspectives."World Bank Eco- Sector Governance. nomic Review 5: 235­50. Sherif, K. 2003. The Fiscal Cost of State-Owned Banks in Stiglitz, Joseph, andThe Initiative for Policy Dialogue on Selected Economies of Central and Eastern Europe. Draft. behalf of The Commonwealth Secretariat. 2004."An World Bank,Washington, D.C. Agenda for the Development Round ofTrade Nego- Sherif, K., M. Borish, and A. Gross. 2003. State-Owned tiations in the Aftermath of Cancun." Banks in theTransition.Europe and CentralAsia Private Strauss, John, Kathleen Beegle, Agus Dwiyanto,Yulia Sector and Finance Sector Development Unit,World Herawati,Daan Pattinasarany,Elan Satriawan,Bondan Bank,Washington, D.C. Sikoki, Sukamadi, and Firman Witoelar. 2004. Shirley, M., and C. Menard. 2002. "Cities Awash: "Indonesian Living Standards before and after the Reforming Urban Water Systems in Developing Financial Crisis: Evidence from the Indonesia Family Countries." In M. Shirley, ed., Thirsting for Efficiency: Life Survey." Online at www.rand.org/nsrd/capp/ The Economics and Politics of UrbanWater System Reform. pubs/indo.html. Oxford: Elsevier. Strömberg, David. 2001."Radio's Impact on New Deal Shleifer,Andrei, and Daniel Treisman. 2004."A Normal Spending."Draft,Princeton University,Princeton,N.J. Country."NBERWorking Paper No.10057.National ------.2002."Mass Media Competition,Political Com- Bureau of Economic Research, Cambridge, Mass. petition, and Public Policy." Review of Economic Stud- BIBLIOGRAPHY 353 ies. Online at http://www.iies.su.se/~stromber/ Temple, Jonathan. 1999."The New Growth Evidence." MediaComp.pdf. Journal of Economic Literature 37(1): 112­56. Sturm, J, and J. de Haan. 2001."Central Bank Indepen- ------. 2000."Growth Regressions and What the Text- dence and Inflation in Developing Countries."Work- books Don't Tell You." Bulletin of Economic Research ing Paper No. 511. Center for Economic Studies and 52(3): 181­205. the Ifo Institute for Economic Research. ------. 2003. "Growing into Trouble: Indonesia after Subramanian,Arvind,and Devesh Roy.2001."Who Can 1966."In Dani Rodrik,ed.,In Search of Prosperity:Ana- Explain the Mauritian Miracle:Meade,Romer,Sachs, lytic Narratives on Economic Growth. Princeton, N.J.: or Rodrik?" IMF Working Paper 01-116. Interna- Princeton University Press. tional Monetary Fund,Washington, D.C. Online at Temu,Andrew, and Jean M. Due. 1998."The Success of http://www.imf.org/external/pubs/cat/longres.cfm Newly Privatized Companies: New Evidence from ?sk=15215.0. Tanzania." Canadian Journal of Development Studies Summers,Lawrence H.1999."Distinguished Lecture on 19(2): 315­41. Economics in Government: Reflections on Manag- Tendler, Judith. 1997. Good Government in theTropics. Bal- ing Global Integration." Journal of Economic Perspec- timore: Johns Hopkins University Press. tives 13(2). Thomas, Duncan. 2004."Financial Crises and Poverty: ------. 2000."International Financial Crises: Causes, The Case of Indonesia." In Harrison (forthcoming). Prevention, and Cures." In Papers and Proceedings of Thomas,Vinod, and John Nash. 1991a. Best Practices in the One Hundred Twelfth Annual Meeting of the Trade Policy Reform. Oxford, NewYork,Toronto, and American Economic Association, The American Eco- Melbourne: Oxford University Press for the World nomic Review 90(2): 1­16. Bank. Sundararajan,V., and T. Balino, eds. 1991. Banking Crises: ------. 1991b."Reform of Trade Policy: Recent Evi- Cases and Issues. Washington, D.C.: International dence fromTheory and Practice."World Bank Research Monetary Fund. Observer 6(2): 219­40. Suryahadi, Asep, Sudarno Sumarto, and Lant Pritchett. ------. 1992."Trade Policy Reform: Recent Evidence 2003. "Evolution of Poverty during the Crisis in from Theory and Practice." In R.Adhikari, C. Kirk- Indonesia." Asian Economic Journal 17(3): 221­41. patrick, and J.Weiss, eds., Industrial and Trade Policy Suryahadi,Asep,AnnaWetterberg,Sudarno Sumarto,and Reform in Developing Countries. Contemporary Issues Lant Pritchett. 1999. "A National Snapshot of the in Development Studies. Manchester, U.K.: Man- Social Impact of Indonesia's Crisis."Bulletin of Indone- chester University Press and New York: St. Martin's sian Economic Studies 35(3): 145­52. Press. Swann,T.W.1956."Economic Growth and CapitalAccu- Timmer, Peter, and Ashley Timmer. 2004."Reflections mulation." Economic Record 32: 334­61. on LaunchingThree Books about Poverty,Inequality, Tabibian, Mohammad. 2003. "Manufacturing Sector's and Economic Growth." WIDER Angle, No.1/2004. Long-Term Strategy and Development in Iran."Insti- World Institute for Development Economics tute for Research on Planning and Development in Research, Helsinki. Iran (February). Tirole, Jean. 2002. Financial Crises, and the International Taliercio, Robert, Jr. 2003."Designing Performance:The Monetary System. Princeton, N.J.: Princeton Univer- Semi-Autonomous Revenue Authority Model in sity Press. Africa and Latin America." Draft (August). World Tommasi, Mariano. 2002."Federalism in Argentina and Bank,Washington, D.C. the Reforms of the 1990s."Stanford University Cen- ------. 2004. "Administrative Reform as Credible ter for Research on Economic Development and Pol- Commitment:The Impact ofAutonomy on Revenue icy Reform, Palo Alto, Calif. Authority Performance in Latin America." World Topalova, Petia. 2005."Trade Liberalization, Poverty and Development 32(2): 213­32. Inequality:Evidence from Indian Districts."In Harri- Talvi, E. 1997."Exchange Rate­Based Stabilization with son (forthcoming). Endogenous Fiscal Response." Journal of Development Transparency International. 2004. The Global Corruption Economics 54(1): 59­75. Report. Berlin:Transparency International. Talvi, E., and C.Vegh. 2000."Tax Base Variability and Tsikata,Yvonne. 2003."Integrating into the Global Econ- Pro-Cyclical Fiscal Policy." NBER Working Paper omy:Challenges,Experiences,and Policy Lessons from No. 7499. National Bureau of Economic Research, the Integrated Framework." Draft,World Bank,Wash- Cambridge, Mass. ington,D.C. Tanzi,Vito, and Hamid Davoodi. 1998. "Corruption, United Nations Conference on Trade and Development Public Investment, and Growth." In H. Shibata andT. (UNCTAD).1987.Handbook ofTrade Control Measures Ihori, eds., The Welfare State, Public Investment, and of Developing Countries­Supplement. Growth.Tokyo: Springer-Verlag. ------. 1994. Directory of Import Regimes. 354 BIBLIOGRAPHY United Nations­WIDER. World Income Inequality Wagstaff,Adam. 2001."Inequalities in Health in Devel- Database. Online at http://www.wider.unu.edu/ opingTide?" Draft,World Bank,Washington, D.C. wiid/wiid.htm. Walmsley, Terri Louise, and Alan L. Winters. 2003. U.S. Census Bureau, International Programs Center. "Relaxing the Restrictions on theTemporary Move- http://www.census.gov/ipc/www/pas.html. ment of Natural Persons: A Simulation Analysis." U.S. General Accounting Office (GAO). 1993. Foreign CEPR Discussion Paper No. 3719. Centre for Eco- Assistance:Promoting Judicial Reform to Strengthen Democ- nomic Policy Research, London. racies.GAO/NSAID-93-149.Washington,D.C.:GAO. Webb, Steven B. 2003. "Argentina: Hardening the van deWalle,Nicolas.2001.African Economies and the Pol- Provincial Budget Constraint." In Jonathan Rodden, itics of Permanent Crisis, 1979­99. Cambridge, U.K.: Gunnar Eskeland, and Jennie Litvack, eds., Decentral- Cambridge University Press. ization and Hard Budget Constraints. Boston: MIT Van Rijkeghem, Caroline, and Beatrice Weder. 2001. Press. "Bureaucratic Corruption and the Rate of Tempta- Wei, Shang-Jin. 1996. "Intra-National versus Interna- tion: Do Wages in the Civil Service Affect Corrup- tional Trade: How Stubborn Are Nations in Global tion and by How Much?" Journal of Development Integration?" NBER Working Paper 5531. National Economics 65: 307­31. Bureau of Economic Research, Cambridge, Mass. Velasco,Andres, and Alejandro Neut. 2003."Tough Poli- ------. 2000."How Taxing Is Corruption on Interna- cies, Incredible Policies?" NBERWorking Paper No. tional Investors?" Review of Economics and Statistics 9932.National Bureau of Economic Research,Cam- 82(1) (February): 1­11. bridge, Mass. Weingast, Barry. 1995."The Economic Role of Political Virmani,Arvind. 2004."India's Economic Growth: From Institutions: Market-Preserving Federalism and Eco- Socialist Rate of Growth to Bharatiya Rate of Growth." nomic Development." The Journal of Law, Economics ICRIER Working Paper No. 122. Indian Council for and Organization 11(1). Research on International Economic Relations, New Weiss,Andy,and G.Nitkin.1997."Performance of Czech Delhi (January). Online at http://www.icrier.res.in/ Companies by Ownership Structure." Unpublished wp122.pdf. World Bank seminar paper. Von Hagen, Juergen. 1992."Budgeting Procedures and Wellenius,Bjorn.1997."Telecommunications Reform-- Fiscal Performance in the European Communities." How to Succeed." Public Policy for the Private Sec- European Economy--Economic Papers 96. Com- tor Note 130.World Bank,Washington, D.C. mission of the EC,Directorate-General for Economic Weyland, Kurt. 2003."Good Governance and Develop- and Financial Affairs (DG ECFIN). ment--A SkepticalView." Draft, University of Texas, Von Hagen, Juergen, and Ian Harden. 1996. "Budget Austin. Processes and Commitment to Fiscal Discipline."IMF Wilder,Andrew. 1999. The PakistaniVoter: Electoral Politics Working Paper 96/78.International Monetary Fund, andVoting Behavior in the Punjab.Karachi:Oxford Uni- Washington, D.C. versity Press. von Hirschhausen, C., and B. Meinhart. 2001. "Infra- Wilkinson, Steven. 2000."Consociational Theory and structure Policies and Liberalization in the East Euro- EthnicViolence." Asian Survey XL(5): 767­91. pean Transition Countries--Would Less Have Been Williamson, John. 1990."What Washington Means by More?" Internet Journal of the Centre for Energy, Petro- Policy Reform." Chapter 2 in JohnWilliamson, Latin leum and Mineral Law Policy--CEPMLP,Vol. 9. American Adjustment: How Much Has Happened?Wash- von Hirschhausen C., and P. Opitz. 2001."Power Utility ington, D.C.: Institute for International Economics. Re-Regulation in East European and CIS Transfor- Updated November 2000. mation Countries: An Institutional Interpretation." ------. 1995."The Management of Capital Inflows." Presented at the 2001 International Society for New Pensamiento Ibero Americano, January-June. Institutional Economics (ISNIE) Conference, Sep- ------. 2000."What Should the Bank Think about the tember 13­15, Berkeley, Calif. Washington Consensus?" Paper prepared as back- Wacziarg,Romain,and Karen HornWelch.2003."Trade ground to theWorld Bank's World Development Report Liberalization and Growth: New Evidence." NBER 2000, July. Online at http://www.iie.com/publi Working Paper No. 10152. National Bureau of Eco- cations/papers/ williamson0799.htm. nomic Research, Cambridge, Mass. ------.2003."From ReformAgenda to Damaged Brand Wade,Robert.1985."The Market for Public Office:Why Name:A Short History of theWashington Consensus the Indian State Is Not Better at Development."World and Suggestions for What to Do Next." Finance and Development 13 (April): 467­97. Development 40(3): 10­13. Wade, Robert, and Ian Hardin. 1995."National Budget ------. 2004a."The Washington Consensus as Policy Processes and Fiscal Performance."European Economy: Prescription for Development." Paper prepared for Reports and Studies 3: 311­418. lecture at theWorld Bank, January. BIBLIOGRAPHY 355 ------. 2004b."TheYears of Emerging Market Crisis:A ------. 2000d. Progress toward the Unification of Europe. Review of Feldstein." Journal of Economic Literature Washington, D.C.:World Bank. XLII (September). ------. 2000e."Securing Our Future in a Global Econ- Williamson,John,and Roberto Zagha.2002."From Slow omy." Latin America and the Caribbean Region. Growth to Slow Reform." Institute for International World Bank,Washington, D.C. Economics,Washington, D.C. (July). ------. 2000f."Trade Blocs." Policy Research Report. Williamson,Oliver E.1975.Markets and Hierarchies:Analy- World Bank,Washington, D.C. sis and Antitrust Implications. NewYork: Free Press. ------. 2000g. Can Africa Claim the 21st Century?Wash- ------.1985.The Economic Institutions of Capitalism.New ington, D.C.:World Bank. York: Free Press. ------.2000h."Sri Lanka: Recapturing Missed Oppor- Winters,Alan L., Neil McCulloch, and Andrew McKay. tunities." Poverty Reduction and Economic Manage- 2004. "Trade Liberalization and Poverty:The Evi- ment, South Asia Region.World Bank,Washington, dence So Far." Journal of Economic Literature XLII D.C. (March): 72­115. ------. 2001­04, various issues. Global Development Wood,Adrian. 1988."Global Trends in Real Exchange Finance.Washington, D.C.:World Bank. Rates: 1960­84." Discussion Paper No. 35. World ------. 2001a. Finance for Growth: Policy Choices in a Bank,Washington, D.C. VolatileWorld. Washington, D.C.:World Bank. Woodhouse, Andrea Fitri. 2002. "The Dynamics of ------. 2001b."From Natural Resources to Knowledge Rural Power in Indonesia: Fighting Corruption in a Economy: Trade and Job Quality." LAC Research World Bank Community Development Project." Report.World Bank,Washington, D.C. Draft (June).World Bank,Washington, D.C. ------. 2001c. Aid and Reform in Africa. Washington, World Bank. 1988­2004, various issues. World Develop- D.C.:World Bank. ment Report.Washington, D.C.:World Bank. ------. 2002a."Anticorruption in Transition:A Contri- ------. 1993­2004, various issues. Global Economic bution to the Policy Debate."World Bank,Washing- Prospects. Washington, D.C.:World Bank. ton, D.C. ------.1993.The EastAsian Miracle:Economic Growth and ------. 2002b. "Globalization, Growth and Poverty: Public Policy. NewYork and Oxford: Oxford Univer- Building an Inclusive World Economy." Policy sity Press for theWorld Bank. Research Report.World Bank,Washington, D.C. ------. 1994a. Averting the Old Age Crisis.Washington, ------. 2002c."Transition:The First TenYears:Analysis D.C.:World Bank. and Lessons for Eastern Europe and the former Soviet ------. 1994b. India: Issues in Trade Reform. Washington, Union."World Bank,Washington, D.C. D.C.:World Bank. ------. 2003a."Administrative Charges in Second Pillar ------.1996."Indonesia,Dimensions of Growth."Report Pensions in ECA: A Case Study Approach."World No.15383-IND.World Bank,Washington,D.C. Bank,Washington, D.C. ------. 1997a."China Engaged: Integration with the ------. 2003b. Better Governance for Development in the Global Economy." China 2020 Series, September. Middle East and North Africa: Enhancing Inclusiveness World Bank,Washington, D.C. and Accountability.Washington, D.C.:World Bank. ------.1997b."Indonesia:Sustaining High Growth with ------. 2003c."Combating Corruption in Indonesia-- Equity." Report No. 16433-IND.World Bank,Wash- EnhancingAccountability for Development." Report ington, D.C. No. 27246-IND, Poverty Reduction and Economic ------. 1998a. "Argentina: Financial Sector Review." Management Unit, East Asia and Pacific Region Report No.17864-AR.World Bank,Washington,D.C. (November 12).World Bank,Washington, D.C. ------. 1998b. Public Expenditure Management Handbook. ------. 2003d. "Doing Business"Web site. Online at Washington, D.C.:World Bank. http://rru.worldbank.org/DoingBusiness/. ------.1999."West Bank and Gaza:Strengthening Pub- ------. 2003e. East Asia Integrates. Washington, D.C.: lic Sector Management."World Bank,Washington, World Bank. D.C. ------. 2003f. Gender and Development in the Middle East ------. 2000a."Anti-Corruption in Transition,A Con- and North Africa:Women in the Public Sphere. Washing- tribution to the Policy Debate."World Bank,Wash- ton, D.C.:World Bank. ington, D.C. ------. 2003g. Jamaica Country Economic Memorandum. ------.2000b."Bolivia:From Patronage to a Professional Washington, D.C.:World Bank. State: Institutional Governance Review." Poverty ------. 2003h. Malawi Country Economic Memorandum: Reduction and Economic Management Network. Policies forAccelerating Growth.Washington,D.C.:World World Bank,Washington, D.C. Bank. ------. 2000c. East Asia, Recovery and Beyond. Washing- ------. 2003i. Malawi Integrated Framework Diagnostic ton, D.C.:World Bank. Trade Integration Study.Washington,D.C.:World Bank. 356 BIBLIOGRAPHY ------. 2003j. Trade, Investment, and Development in the ------. 2004f. World Development Indicators database. Middle East and North Africa--Engaging with theWorld. April.Washington, D.C.:World Bank. Washington, D.C.:World Bank. ------. 2005a. At the Frontlines of Development. Washing- ------. 2003k. "Trade Policies in South Asia: An ton, D.C.:World Bank. Overview,Volumes I and II" (Discussion Draft, May). ------. 2005b. Development Challenges in the 1990s. World Bank,Washington, D.C. Washington, D.C.:World Bank. ------. 2003l. Unlocking the Employment Potential in the ------. 2005c, forthcoming. Public Debt Trends and Middle East and NorthAfrica:Towards a New Social Con- Decompositions in Emerging Market Economies between tract.Washington, D.C.:World Bank. 1990 and 2003.Washington, D.C.:World Bank. ------. 2003m. Zambia Country Economic Memorandum: World CurrencyYearbook (for 1985,1990­93).International Policies for Growth,Diversification,and Poverty Reduction. Currency Analysis, Brooklyn, N.Y. Washington, D.C.:World Bank. World Trade Organization (WTO). 1994­2003, various ------. 2003n. Zambia:The Challenge of Competitiveness issues. Trade Policy Review: Country Reports. and Diversification.Washington, D.C.:World Bank. WorldTrade Organization (WTO). 1998."JamaicaTrade ------. 2004a."Legislation on Freedom of Information: Policy Review." Geneva:WTO. Trends and Standards." PREM Note 93.World Bank, Wurfel,David.1963."The Philippines."Comparative Stud- Washington, D.C. ies in Political Finance:A Symposium, Journal of Politics ------.2004b.Corruption in Enterprise:State Interactions in 25 (November): 757­73. Reprinted in Scott (1972). Europe and CentralAsia 1999-2002.Washington,D.C.: Wyplosz,C.2002."Fiscal Discipline in Emerging Market World Bank. Economies: How to Go about It?" Draft. Graduate ------. 2004c. Doing Business 2004: Understanding Regu- Institute for International Studies, Geneva, and Cen- lation.Washington, D.C.:World Bank. tre for Economic and Policy Research, London. ------. 2004d."LeveragingTrade for Development:The Yaron, J., M. Benjamín, and G. Piprek. 1997. Rural World Bank ResearchAgenda."Progress report to the Finance: Issues, Design, and Best Practices. Washington, Board of Directors.World Bank,Washington, D.C. D.C.:World Bank. ------. 2004e. Reforming Infrastructure: Privatization, Reg- Yusuf, Shahid, and Simon J. Evenett. 2002. Can East Asia ulation, and Competition. Washington, D.C.: World Compete? Innovation for Global Markets. NewYork and Bank. London:World Bank and Oxford University Press. Index Note: n and nn indicate note and notes. accountability, 24, 277, 279­282, 290, 300n.12, 316, 320 banks and banking activist role, industrialization, 83 assets, 208, 209 adjustment loan, 300n.15 crises, 69, 101, 102, 109, 242 administrative reforms, 292 debt markets, 226 Africa, 7, 9­10 function, 230 annual growth, 274 independence, 105, 107, 116nn.32, 33 financial liberalization, 205 licensing, 222­223 institutional perspective, 270­274 privatization, 220, 222 per capita GDP, 271 regulation and supervision, 213­215 privatization, 165, 171­172 state ownership, 212 rebound, 273 basic needs, 120 see also Middle East and North Africa behavior, 26 African Financial Community franc, devaluation, 7, 106 benefit sharing, 12, 13­14, 20 agencies, roles, 25, 26n.7 bilateral relationship, 299n.6 aggregate growth, trade reform and, 146­149 binding constraints, 15­16 agriculture, privatization, 197 Bolivia, 7 analysis, 24­25 pension plans, 191 Argentina, 7, 9, 10 bond index, emerging markets, 103 banking system, 247n.3 bond markets, 209­210 currency convertibility, collapse, 29, 34, 52n.5 regulation, 215 decentralization, 295 borrowing, 69, 211 electricity distribution, 178 Botswana, natural resources, 306 interest and exchange rates, 241 Brazil, 7, 9, 13 pension system, 192 interest and exchange rates, 241 privatization, 171 privatization, 171, 179 stability, 8­9, 116n.22 public service, 287 Asia bright line rules, 280 demography, 188 Brussels, recession vs distance from, 32 financial crisis, 33­34 budget and budgeting financial liberalization, 205 constraints, 295 lessons, 81­89 institutions, 110­111 social security, 190 participatory, 291­292 assets processes, 17­18 external, 233n.19 bureaucracy, 277­278 sales, 103 bureaucrats, citizenry and, 278­179 authority, 300n.20 autonomy, 301n.23 campaign finance reform, 326 capability, 265 Baltic states, privatization, 199 capable states, 297 Bangladesh capacity, 141, 280­281 cellular phones, 177 constraints, 283­285 loans, 300n.15 political imperatives, 285 357 358 INDEX capital access, 22, 226 accumulation, 2, 12, 13, 26n.5, 81­83, 158 allocation, 222 weak banks, 234n.38 domestic, 233n.23 capital account, 113­114, 117nn.51­56 private, 226­227, 233n.20 openness, 17 small borrowers, 230 policies, 242­244 crisis management, 260­261 restrictions, 74 currency capital controls, 232n.6 crises, 115n.15, 242 capital flows, 40, 98-99, 115nn.17, 18, 239 foreign, 232n.12, 233n.19 controls, 243­244 risk, 233n.24 international, 65­70, 75n.2 current account, 96 median developing countries, 70 Czech Republic, privatization, 169 private, 65 capitalism, 7, 8 debt, 84, 102, 103, 116nn.23, 26, 216, 235n.50, 245 capital markets, 225­226, 227 crisis economies, 246 framework, 223 foreign, 245­246 captured states, 297­298 government, 101, 233n.17 Caribbean, growth reform, 35 private sector, 240, 244 CFA franc, devaluation, 7, 106 public sector, 233n.15, 234n.30 checks and balances, 24, 325­326 decentralization, 111, 294, 295, 326, 329n.36 Chile default, 115n.16 electricity distribution, 178 deficits, 17, 241, 266n.7 social security, 186, 194n.27 democracies, richer and poorer, 312, 313 taxation, 22 democracy, 309­329 trade reform, 144, 145 characteristics, 311­318 China, 7, 9, 13 government credibility, 318­320 growth, 37­39 lessons, 320­326 poverty reduction, 39 demography, 115n.5, 188 inequality, 127­128 deposit, 235n.44 privatization, 167, 168 GDP and, 204 trade liberalization, 141­142, 152n.15 growth, 206, 208 CIS countries guarantees, 222 electricity prices, 186 insurance, 214­215 institution building, 197 depositors, 234n.39 citizen power, 295­296 deregulation, 19­21, 69, 163­194 citizens, 300n.10 devaluation, 245 bureaucrats and, 278­279 CFA franc, 7, 106 politicians and, 276­277 developed countries, growth, 59­61, 75n.5 civil service process and procedures, 300n.9 regional, 61 civil society, 293 developing countries, growth, 2, 3­5, 59­61 clientelism, 315, 323 issues, 6­9 client power, 291­293, 300n.11, 301n.41 results, 9­10 commodity prices, non-energy, 67 development competition, 263 assistance, 329n.35 conflict countries, privatization, 198 rapid, 266n.12 consumption, 115n.8 Diaz, Francisco Gil, 10 continuity, 265 dirigiste framework, 7 contraction, 31 disappointments, 29­37 converging, 5, 26n.2 discretion, 14, 252­260 corruption, 259, 288­291, 299n.3, 312, 328n.18 abuse, 256 anticorruption agency, 290 disenfranchised, 328n.26 elections and, 316 distribution, 158 newspaper circulation and, 314 diversification, 66 costs, 329n.29 higher-value products, 63­64 trade reform, 142­143 divestiture, winners and losers, 171 country specificities, 25 dollarization, 106, 107, 116­117n.34 credibility, 265, 315, 316 donors, 324 government, 318­320 Dutch disease, 304 political promises, 323­324 politicians, 314­316 East Asia. See Asia trade reform, 143 Easterly, 71 credit, 208­209, 233n.16 Eastern Europe. See Europe INDEX 359 economic growth, 2­6 GDP and tariffs, 137 (1960­2002), 3 parts and components, 69 (1990s), 7­9 external environment, 97­99 history, 1 external restraints, 289­291 issues, 6­10 lessons, 10­26 fairness, 50­52, 53n.24, 126, 127 operational implications, 24­26 finance policy and institutional reforms, 16­24 access to, 229­231 regional, 4 improving, 226­231 education. See school enrollment financial crises, 8, 14­15, 29, 215­222, 234nn.29, 32, 34 efficiency, 6, 10­11, 266n.4 East Asia, 33­34 E-government, 301nn.28, 29 lessons and controversies, 238­247 Egypt, 160­161 policy responses, 220­222 elections, competitive, 309, 310­311, 327n.10, 328n.15 financial intermediaries, 12, 211­213 corruption and, 316 financial liberalization, 10, 21, 22, 71­73, 75nn.4­6, defined, 327n.1 203­235, 232n.1, 235nn.46, 47 non-elected leader, 327n.5 effect, 75 school enrollment and, 317 evolution, 205 electoral markets future, 225­231 failures, 321­322 globalization, 205 imperfections, 313­318 high costs, 204­205 electoral reform, 324­325 lessons, 222­225 electricity outcomes, 206­222 CIS countries, 186 poor results, 204 cost-coverage ratios, 186 process, problems, 217 labor productivity, 178 financial markets, 233n.23, 234n.33 power markets, potential problems, 178­179, 180 financial resources, 224 privatization, 176­178 financial services, 231 regulatory capacity, 185 financial system emerging industries domestic, 112­113 protection, 143­144 efficient, 108 self-discovery, 143 fiscal balance, 96, 97, 103­104 emerging markets, 103 fiscal deficit, 11­12 enclaving, 285­286 fiscal flexibility, 111 Uganda, 286 fiscal management, 282­283 entrepreneurship, 11, 264 fiscal policy, 110­112, 117n.50 environment, external, 70­71 destabilizing, 104, 116nn.28, 29, 31 equity markets, 209­210, 235n.51 fiscal solvency, 101­103, 116nn.20, 21 regulation, 215 sustainable, 111­112 Eurasia, 55­57 focus, corruption, 290 Europe foreign currency, 232n.12, 233n.19 banks, 233­234n.26 foreign direct investment (FDI), 12, 67­68, 138 demography, 188 foreign entry, 222 economic miracle, 57­58 foreign exchange, 244 social security, 190 FSU, transition, 30­32 Europe, Eastern, 8, 30­32 functions, achieving, 12­14 demography, 188 funding, defined, 188 institution building, 196­202 funds, availability, 69 lessons, 200­201 social security, 190 Gaidar, Yegor, 13 events, 1990s, 29­41 Gaza, service delivery, 293 exchange rates, 64, 96­97, 98, 107­109, 112, 117n.39, 139, GDP per capita, 31, 52n.3, 61, 115n.5 216, 240­241 generating plants, U.S., 177 policies, 244­246 GEO ARM, privatization, 198 volatility, 84 geography, 56­57 expectations, 260­262, 263, 265, 273 Africa, 272 expenditures, timing, 103 Germany, East, 167 export processing zones (EPZs), 142­143 Ghana, access to services, 174 exports, 66, 134 global economy, 40 developing countries and, 64 trends, 61­71 GDP and, 64 globalization, poverty and, 148­149 GDP and goods and services, 65 governance, 23­24, 26n.6, 258 360 INDEX governance, public sector, 275­301 electric power purchases, 180 conundrum, 276­279 financial liberalization, 218­219 entry points, 299 GDP, 31 reforms, 279­296 natural resources, 306­307 strategy and implementation, 296­298 industrial countries, 266­267n.13 government, 2 industrial performance, 287 credibility, 318­320 industrialization, activist role, 83 debt, 101, 233n.17 inequality, 4, 5­6, 42, 50­52, 53nn.26, 27, 328­329n.27 elected, 24 causes and consequences, 125­128 industrial countries, 266­267n.13 global, 126 pension systems, 190­192 poverty and, 120­127, 128nn.2, 3 policies and, 260 trade liberalization and, 149 programs, 88­89 infant mortality rates, 123 role, 235n.42 inflation, 7, 16­17, 93, 95, 114n.4, 232nn.7, 13, 235n.49 social indicators, 89 stabilization, 238­239 technology and, 86 volatility, 84 government discretion, 14, 252­260 inflows, 15, 242­243 abuse, 256 informal rules, 301n.39 limiting, 256 information, 226, 291, 295­296, 327n.2 growth reform, 322­323 decomposition, 46­47 information and communication technologies (ICT), 287­288 determinants, 46 efficiency gains, 289 industrialized countries, 41­42 infrastructure, 258 long-term perspective, 56­58 cost, 174­175 new growth theory, 43, 45­46, 53n.13 privatization, 176 payoff, 109­110 trade-related, 139­141 per capita, forecast vs actual, 30 transaction and investment, 175 process, 41­43 instability, Africa, 273 promotion, 10­11 institutions, 2­3, 5, 14, 21­23, 24, 48­50, 53n.23, 57, 249, rapid, 7­8, 45 258, 326 rates, 42­43, 44, 53n.15, 82, 83 Africa, 270­273 reform weaknesses, 30, 34­37, 52nn.5­8 anticorruption, 290 regression, 36, 47, 47­48, 53nn.21, 22 capability, 259­260 shared, 87­89, 90n.3 complementary, 281, 290 successful, 78­81 constraints, 249, 253­260 theory, 41­48 defined, 253 worldwide, real GDP per capita, 1 Eastern Europe, 196­202 finance, 223 health care, 301nn.27, 34 investment and, 264­265 health indicators, poverty and, 123 macroeconomic policy formulation, 110­112 Hirschman, Albert, 11 money and power and, 51 horizons, short, 319 natural resources, 306­307 human capital, 53n.17, 193n.9, 283­288 policies and, 13, 253, 259 Hungary, privatization, 199 political, 319­320 quality, 39, 49­50 import substitution, 49 reform, 16­24, 324­326 incentives, 280­281 role, 10 income trade-related, 139­141 distribution, 146­149 variables, 47 inequality, 125, 129n.5 integration per capita, 328nn.20, 23 global, 62, 64, 85 volatility, 148 trade and, 63, 64 incrementalism, ad hoc, 296­297 integrity pact, 301n.33 India, 12, 13 interest rates, 16, 33, 68, 98, 102, 108, 115n.16, 116n.26, administrative reforms, 292 234n.32, 240­241 financial liberalization, 207 financial liberalization, 206, 232­233n.14 growth, 37­39 payments, 70 poverty, 39, 52n.10 intermediaries, 225, 229 regulatory capacity, and electricity, 185 international issues trade liberalization, 142, 152n.17 agenda, 126 Indonesia, 13 capital flows, 65­70, 75n.2 banking, 234n.28 financial system, 72­73 INDEX 361 terms of trade, 64­65 interest and exchange rates, 240 trade, 57, 63­65 malfeasance, 235n.52 investigative journalism, 296 market investment, 158­160, 264 access, 149­151 climate, 264­265 discipline, 227­228 credibility and, 318 economies, privatization, 164­165, 170­173 share of GDP, 81 market-friendly conditions, 252 investors, 69, 230 mass privatization, 169­170, 194n.15 Mauritius Jamaica growth episodes, 46 justice sector reforms, 281 trade reform, 142 trade-related institutions and macroeconomic stability, trade-related institutions and macroeconomic stability, 140 140 Japan, 13 media, 295­296, 322 jobs, trade reform, 147 medium-term expenditure framework (MTEF) , 283, 300n.21 judicial reforms, 279­282, 300n.14 Mexico judiciary framework, 231 access to services, 174 bank privatization, 220 Korea, 12, 13 interest and exchange rates, 240 interest and exchange rates, 240 privatization, 172 taxation, 288 labor toll-roads, 182 integration, 132, 134 Middle East and North Africa, 156­162 migration, 134 minorities, 317, 327n.8, 328n.25 mobility, 134 monetary policies, 18, 112, 246­247 productivity, water, 184 monetary stability, 96 land tenure. See property rights monopoly, antimonopoly laws, 167, 193n.12 Latin America, 8 Morocco, 160­161 banks, 234n.27 Mozambique, trade reform, 143 demography, 188 deregulation, 165 national economies, 53n.13 financial liberalization, 205 natural resources, 159, 304­307, 307n.1 growth reform weaknesses, 30, 34­37, 52nn.5­8 economic effects, 304­305 pension plans, 191 institutional effects, 306­307 policy reform, 258 New Growth Theory, 2­3, 250­252 privatization, 165, 170­171 New Public Management (NPM), 284, 300n.18, 301n.23 school enrollment, 40 newspaper circulation, 329n.32 social security, 188, 190 corruption and, 314 telecommunications, 176, 185 NGO advocacy, 292 legal framework, 231 Nigeria, banks, 234n.27 legal reforms, 279­282, 300nn.14, 17, 324­326 nominal stability, 104­107 legislation, 266n.8 nonblank financial intermediaries (NBFIs), 213 reforms, 281­282 nontariff barriers (NTBs), 72 , 142 legislature, 327n.6 lending, weak, 218­222, 228 OECD countries, 59, 60, 70, 75n.3 liberalization, 69 recessions, 31 external, 9 social security, 187­188 index, 202n.7 offshore borrowing, 227 loans, 235n.53 oil prices, 67 low-income countries under stress, 301n.42 oil shock, 82 openness, 263 macroeconomic management, 14­15, 21­23 operations, 26 macroeconomic progress, 100­109 opulence, natural progress of, 1 macroeconomic reform, 112­114 organization, 299n.5 macroeconomic stability, 16­18, 93­113, 114n.2, defined, 253, 254 115­116n.19, 139, 152n.9, 263, 160 output finance, 224­225 losses, 29, 30­32 lessons, 110­114 pre- and post­World War II, 31 outcomes, 93­95 pre-transition and, 32 trade-related institutions and, 140 ownership. See property rights Malaysia bank controls, 234n.40 parastatal firms, 267n.17 362 INDEX pensions, 21, 188, 194n.28, 215 multidimensionality, 123­124 defined, 187 sustained growth and, 121­123 government role, 190­192 trade liberalization and, 146­149 private accounts, 192 powerlessness, poverty and, 124 per capita income, 59, 88 powers, corruption, 290 perception, 258­259 prices performance stabilization, 17 external factors, 9 trade reform, 147­148 internal factors, 9 principal-agent problems, 266n.5, 277­278 personnel management, 291 principles, 262­263 Peru implementation, 11­12, 26n.4 access to services, 174 private sector, 12, 13 revenue authority, 286 CIS countries, 197 Poland, privatization, 169, 199 external finance for, 210­211 policies, 3, 11, 81­82, 86­87, 272­273 privatization, 193n.7 capital account, 242­244 privatization, 7, 10, 163­194 complementarities, 114 banks, 220, 221, 222 credibility and, 318­319 design, 19­21 defined, 253, 254 market economies, 164­165, 170­173 exchange rate, 244­246 mass, through vouchers, 169­170 financial, 223­224, 235n.49 process, 167, 169 functions, 12­14 results, 170­175 growth regressions and, 36 reversed, 175, 194n.23 incoherencies, 10 social security, 186­192 instruments, 24­25 speed, 165­170, 193n.7 interpretation, 11­12, 13 state-owned firms, 194n.24 monetary and fiscal, 246­247 trend, 164­165, 193n.4 political competition and, 327 utilities, 173­186 reform, 11 Procurement Watch Inc. (PWI), 293 rules-based, 257 production, incentives, 11 stability, 95­97, 100, 115n.14 productivity, 84­85 variables, 47 history, 1 policy makers, 15­16, 277­278 product standards, 141 institutions, 253 property rights, 13, 50­51, 13, 319, 327n.3, 328n.21 policy reform, 71­73, 249­267 public sector lessons, 249­262 consumption, 105, 116n.30 future, 262­265 governance, 23­24, 26n.6, 275­301 expectations, 260­262 officials, 296 impacts, 262 oversight, 291 lessons, 16­24 reform, 323­324 political economy, 24, 167, 194n.13, 315 solvency, 100 accountability, 300n.12 strengthening the compact, 282­292 buying off, 321 public service, 286­287 framework, 233n.21 calling and commitment, 287 imperatives, 285 institutions, 300n.13 railroads, 51 market failure, mitigating, 324­326 privatization, 181­182 market imperfections, 314, 321 rate-of-return rules, 194n.25 parties and reform, 325 real sector, privatization, 197, 200 support, 288, 301n.26 recessions, 31, 32 trade reform, 142­144 reform politicians, 277­278, 329n.36 agenda, incomplete, 108­109, 117nn.40­44 citizens and, 276­277 demand for, 329n.31 credibility, 314­316 growth-oriented, winners and losers, 126 Ponzi scheme, 266n.10 insufficient, 10 portfolio adjustment, 234n.31 nonstarters, privatization, 198 poverty reduction, 39, 87 rapidity, 7, 26n.3 access to services, 124­125 regional agreements, trade reform, 144­145 costs, 174­175 regulation, 51, 226 defined, 122, 128­129n.4 capacity, electricity, 185 India, 52n.10 reform, 20 inequality and, 120­129, 128nn.2, 3 revolution, U.S., 164 INDEX 363 utilities, 184­186 stopgap measures, 103­104, 111­112, 116n.27 remittances, 134 strategies, 25, 78, 89­90n.1, 137­141, 249, rents, 300n.8 263­264 reputation structural variables, 47 government, 319 Structuralists, 49 politicians, 316 Sub-Saharan Africa research, 25­26 bank restructuring and privatization, 221 resilience, 40­41, 52n.11 stagnation, 8, 30, 37, 52n.9 resources subsidies, 232n.4 allocation, efficiency, 12, 13, 83­84 sudden stops, 99, 115nn.13, 17, 18 corruption, 290 Summers, Lawrence, 40, 117n.51 use, 6 supervisors, 235n.52 restitution, 166 supply response, 261­262 restrained states, 298 surprises, 29, 30, 37­41, 52n.1 resurgence, 41­48 sustainability, 17, 79 revenue collection, 285 syndrome, symptom vs, 47, 48, 53nn.18, 19 risk, poverty and, 123­124 roads (toll), privatization, 182­183 Tanzania, 7 Romania, privatization, 199 loan, 300n.15 rules, formal, 312­313 tariffs, 150­151, 160, 162n.4, 266nn.2, 9 Russia, 7 changes, 136, 137 banks, 234n.27 dispersion, 73 crisis, 117n.44 reduction, 72 privatization, 193n.10 taxation, 22, 202n.4, 301n.24, 328n.24 Russian Federation efficiency, 288 interest and exchange rates, 241 public expenditures, 104 privatization, 198 reform, 73 VAT, 103 savings technology, growth and development, 12, 13, 84­87 oil-exporting countries, 305 telecommunications public sector, 240 Bangladesh, cellular phone operators, 177 school enrollment, 40, 327­328nn.12­14, 18, 22 Latin America, 185 poverty and, 123 privatization, 175­176 elections and, 317 U.S. market share, 165 seigniorage revenues, 97, 104 Thailand, interest and exchange rates, 240 services time, 9­10 access to, 124­125, 129n.6, 174 trade delivery, 293 integration, 12, 136 shocks, 82, 115n.9, 216, 243 international, 63­65, 75n.1 social funds, 329n.30 liberalization, 7, 19, 71, 131­153 social indicators, 39­40, 157 liberalization, partial, 141­142 social polarization, 316­317, 327n.2 openness, 18­19 social security policy, 133­134, 266n.6 defined, 187­188 political economy and, 142­144 privatization, 186­192 reform, 132­145, 151­152nn.1­5, 251­252 Soviet Union, 8, 12, 166 terms of, 66 special economic zones (SEZs), 141­142 trade and growth debate, 135 special interests, 311 total factor productivity (TFP), 46­47, 84­85, 90n.2 Sri Lanka, fairness, 127 successful countries, 86 stability, 11­12 transactions, corruption, 291 growth and, 100 transformational recession, 31 nominal, 104­107 transition countries, 126 state sector financial liberalization, 205 banks, 233n.25, 234n.37 key indicators, 198­199, 201n.1­2 enterprises, 10 policy reform, 258 ownership, 166 privatization, 173, 194n.22 privatization, 197 recession, 8 role, 6­7, 259­260 transparency, 69, 111, 283, 284, 291, 301n.29 stock transport, privatization, 180­183 companies, 233n.25 Treuhand, 167 exchanges, 193n.11 Tunisia, 160­161 markets, 232nn.9, 11, 233n.22 Turkey, interest and exchange rates, 241 364 INDEX Uganda, enclaving, 286 Africa, 273 Ukraine, 300n.16 flows, 227­229 privatization, 198 vote of confidence, 327n.9 United Kingdom voters, 327 electricity, 178 uninformed, 313­314, 328n.16 railroads, 182 vouchers, 169­170, 194n.16 United States vulnerability Great Depression, 31 poverty and, 123­124 growth rate, 75n.1 trade reform, 148 regulatory revolution, 164 unpredictability, 4­5 wages, trade reform, 147 utilities, privatization, 173­186 water labor productivity, 184 value added tax (VAT). See taxation privatization, 183­184 variability, 4 reform outcomes, 183­184 Vietnam supply systems, 183 growth, 37­39 West Bank, service delivery, 293 inequality, 128 world trade system, 19 poverty reduction, 39 voicelessness, poverty and, 124 Zambia, privatization, 171, 194nn.18­20 volatility, 82­83, 94, 115nn.6­12 "I welcome this departure from narrow orthodoxy to a "The central message of this book is that there are no broader appreciation of the realities of growth.The central simple panaceas for development. One can accept this message is that this experience has been highly varied, and still believe that it is often useful to allow markets to and that many different combinations and sequences of work,that trade is better than autarchy,and that crises are policy can all lead to the growth that we all desire for spawned by macroeconomic sloppiness as well as by poverty reduction. Country specificities matter in the dogmatism. Making a fetish of being unorthodox is as bad design of policy, and simplistic formulations of `best prac- as narrow-minded dedication to some ideological creed. tices' do not travel well across time and space. For policy Policies need to be tempered to both local objectives and analysts and aid agencies alike, this is a call for detailed economic realities,and in deciding how this would best be country-specific analysis before policy prescriptions are done. The experiences distilled in this book should be of recommended or required." enormous help." Ravi Kanbur John Williamson T.H. Lee Professor of World Affairs Senior Fellow, Institute for International Economics Cornell University This is an exceptionally informative book for economists "The central economic paradox of the last quarter cen- and other advisers on development policy. It provides a tury is that development successes took place where painfully honest and well-documented evaluation of what Washington economists had little influence (China, policy reforms worked, what didn't, and why, during a Vietnam, India) while the failures took place where they decade in which the fortunes of developing countries were practically running the show (Latin America, Africa). diverged in unexpected ways." This book is an honest, searching appraisal of this experi- Adrian Wood ence. It clears away many of the myths that surround Acting Director, Policy Division development policy, and lays the groundwork for a more Department for International Development, pragmatic approach better suited to local circumstances." United Kingdom Dani Rodrik Rafiq Hariri Professor of International Political Economy "The 1990s evolved in the shadow of the Washington John F. Kennedy School of Government, Consensus. By the end of the decade, many of the main Harvard University ideas in the consensus were believed insufficient for development.This book offers a wealth of insights into the nature of institutional and policy reform over this period. It distills the lessons in a way that will contribute to impor- tant debates going forward." Timothy Besley Professor of Economics and Political Science London School of Economics T his book is part of a larger effort undertaken by the World Bank to understand the development expe- rience of the 1990s, an extraordinarily eventful decade. Each of the project's three volumes serves a different purpose.Development Challenges in the 1990s:Leading Policymakers Speak from Experience offers insights on the practical concerns faced by policymakers, while At the Frontlines of Development: Reflections from the World Bank considers the operational implications of the decade for the World Bank. THE WORLD BANK ISBN 0-8213-6043-4