20924 September 2000 THE QtJALITY OF GROWTH _ Av.! Ido\} * . t .., D on_ " pV THE QJALITY OF GROWTH A Note about the Cover Future: Bright or Dark? by Maja Sasek, age 12, Skopje, FYR Macedonia. The future is a chance you should take. It is a creation and everyone has hi.s or her own vision. It will look the way we make it look. -Maja Sasek Over 5,000 children and students from Albania, Bosnia-Herzegovina, Bul- garia, FYR Macedonia, and Romania took part in art and essay contests on their vision of the future for themselves, their cities, and their countries and regions. A selection of their paintings was displayed in an exhibit en- titled "Children Look to the Horizon," which was shown at the World Bank, Washington, D.C., April 1999; in Vienna, Austria, September 1999; and in Florence, Italy, October 1999. THE QUJALITY OF GROWTH Vinod Thomas Mansoor Dailami Ashok Dhareshwar Daniel Kaufmann Nalin Kishor Ram6n Lopez Yan Wang Published for the World Bank Oxford University Press Oxford University Press OXFORD NEW YORK ATHENS AUCKLAND BANGKOK BOGOTA BUENOS AIRES CALCUTTA CAPE TOWN CHENNAI DAR ES SALAAM DELHI FLORENCE HONG KONG ISTANBUL KARACHI KUALA LUMPUR MADRID MELBOURNE MEXICO CITY MUMBAI NAIROBI PARIS SAO PAULO SINGAPORE TAIPEI TOKYO TORONTO WARSAW and associated companies in BERLIN IBADAN () 2000 The International Bank for Reconstruction and Development/The World Bank 1818 H Street, N.W., Washington, D.C. 20433, USA Published by Oxford University Press, Inc. 198 Madison Avenue, New York, N.Y. 10016 Oxford is a registered trademark of Oxford University Press. All rights reserved. No part of this publication may be reproduced, stored in a re- trieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford Uni- versity Press. Manufactured in the United States of America First printing September 2000 The findings, interpretations, and conclusions expressed in this study are entirely those of the authors and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Direc- tors or the countries they represent. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorse- ment or acceptance of such boundaries. Library of Congress Cataloging-in-Publication Data The quality of growth/Vinod Thomas ... [et al]. p. cm. Includes bibliographical references. ISBN 0-19-521593-1 1. Economic development. 2. Sustainable development. 1. Thomas, Vinod, 1949- HD75.Q35 2000 338.9-dc21 00-028324 CONTENTS Foreword ...................................... xiii Preface ...................................... xvii The Report Team ...................................... xxi Overview ...................................... xxiii Development Outcomes and Growth Processes .............................. xxiii Principles of Development ...................................... xxv Neglected Actions in the Growth Process ...................................... xxix Shifting Priorities ....................................... xxxii 1. A Mixed Development Record .......................................1 Assessing Development ......................................2 The Development Record .......................................5 Growth and Welfare ...................................... 14 External Factors Matter ...................................... 15 Domestic Policies Make a Crucial Difference ..................................... 17 Crucial Issues for Action ...................................... 21 2. Assets, Growth, and Welfare .................... .................. 27 A Framework ...................................... 28 Empirical Evidence ...................................... 36 Conclusions ...................................... 44 3. Improving the Distribution of Opportunities ...................... ..... 49 Potential Benefits of Education ............... ....................... 50 Quantity Is Not Enough-Quality Matters ...................................... 51 Achieving Equitable Education and Social Inclusion .................... ..... 56 Improving the Efficacy of Public Spending ....................................... 66 v CONTENTS Making Education More Productive .............................................. 73 Conclusions .............................................. 80 4. Sustaining Natural Capital .................................... .......... 83 Extensive Losses ............................................... 84 Significant Benefits of Environmental Action .................................... 88 The Growth-Natural Capital-Welfare Nexus .................................... 89 Incorporating Environmental Sustainability in Growth Policies ........ 98 Rethinking the State's Role ................................................ 102 Global Environmental Issues Must Be Confronted .................. ........ 107 Conclusions ................................................ 109 5. Dealing with Global Financial Risks ................................. ..... 113 Expansion of Capital Markets and Volatility of Capital Flows ......... 114 Causes and Consequences of Capital Flow Volatility ........................ 115 Past and Present Risk Management .............................................. 122 A Broad Framework of Risk Management ......................................... 124 Conclusions .............................................. 131 6. Govemance and Anticorruption ............................................ 135 Governance Affects the Quality of Growth ...................................... 136 Corruption Undermines Growth and Development ......................... 144 Causes of Corruption .............................................. 150 A Multifaceted Anticorruption Strategy ........................................... 152 Conclusions .............................................. 166 7. Seizing the Opportunities for Change ........................... ......... 169 The Framework and Themes .............................................. 169 Actions to Ensure Quality .............................................. 172 Where Have the Policies for Quality Worked-or Not? ................... 174 Political Economy of Quantity versus Quality .................................. 176 Going Forward .............................................. 177 Bibliography and References ............................. ................. 233 Annexes 1. Broad Objectives and Instruments ......................................... 179 Goals and Policy Measures ............................................... 179 Composite Indexes of Human and Sustainable Development .......... 183 2. Framework and Evidence .............................................. 185 A Welfare Function .............................................. 185 Private Sector Optimization .............................................. 186 Econometric Specification Used to Estimate Growth Functions ...... 196 VI CONTENTS 3. Distribution of Education, Openness, and Growth ................ 205 The Extended Production Function with Distribution of Education ........................................... 206 Empirical Analysis on Education and Investment Returns ............... 209 Selected Literature on Asset Distribution and Growth ..................... 212 4. Measuring Natural Capital ........................................... 215 5. Financial Openness .............. ............................. 217 Note on Country Vulnerability and Volatility Measures ..... ............. 220 Note on Gross Domestic Product Gaps ........................................... 220 Note on the Binomial Logit Model ........................................... 221 Summary Statistics for Variables Used in Chapter 5 ......................... 222 6. Governance and Corruption Indexes: Aggregation Methods, New Empirical Measures, and Econometric Challenges ........................................... 225 Defining and Unbundling Governance ........................................... 225 Measuring and Unbundling Corruption ........................................... 228 Boxes 1 Asset Accumulation, Growth, and Welfare .xxvi 2.1 Social Capital .33 2.2 Alternative Approaches to Sustain Growth: Brazil, Chile, and the Republic of Korea .37 3.1 Health Gaps between the Rich and the Poor Are also Large .61 3.2 Supporting Girls' Education in Bangladesh .64 3.3 Population and Development .71 4.1 Environmental Degradation in India .86 4.2 Worldwide Efforts for Environmental Action .90 4.3 Population, Poverty, and the Environment .94 4.4 Development and the Environment .103 4.5 Private Profit at Public Expense: Corruption in the Forest Sector .107 4.6 International Cooperation to Mitigate Global Climate Change .108 5.1 Openness to Intemational Capital Flows .118 5.2 Chile: Openness, Capital Controls, and Social Protection . 130 6.1 Governance and Survey Diagnostic Tools: The Power of Empirics .161 6.2 "Voice" as a Mechanism to Enforce Transparency and Accountability .162 Vil CONTENTS 6.3 Millions of "Auditors" Improve Transparency and Governance in Budgetary Accounts and Beyond ........................................... 165 7.1 Actions for Quality ................................................. 173 Figures 1 A Framework ....... xxvii 1.1 GDP Growth and Changes in Quality of Life, 1960s and 1990s ...4 1.2 Change in Human Development and Growth of Income, 1981-98 .6 1.3 Poverty Rates and Number of Poor, Selected Years . 7 1.4 Poverty Incidence in Selected Transition Economies, 1987-88 and 1993-95 .8 1.5 Environmental Changes Versus Growth of Income, 1981-98 .8 1.6 Total Suspended Particulates, Selected Cities, Early 1990s .9 1.7 Purchasing Power Parity-Adjusted GDP Per Capita, 1975-98 ... 10 1.8 Growth in GDP Per Capita, Selected Economies, 1975-98 . 11 1.9 Income Inequality within Countries: 1980s and 1990s .13 1.10 Volatility of GDP Growth Rates by Decade .14 1.11 Public Expenditures on Education by Region, Selected Years .18 1.12 Trade Barriers, Selected Regions, 1984-93 .20 1.13 Parallel-Market Premium, 1970s-1990s .21 3.1 Relationship between Public Spending Per Capita and Educational Attainment, Various Years .......................... 52 3.2 Secondary School Completion Rates for 20-25-Year-Olds by Household Income Level, Selected Latin American Countries and Years .56 3.3 Gini Coefficients of Education, Selected Countries, 1960-90 ....58 3.4 Education Lorenz Curves for India and Korea, 1990 .59 3.5 Education Gini Coefficients for 85 Countries, 1990 .60 3.6 Mortality of Children Two Years Old and Younger by Wealth, Brazil, 1996 ...... 61 3.7 Median Grade Attainment for 15-19-Year-Olds from Rich and Poor Households, Selected Countries and Years ................... 62 3.8 Years of Schooling for 25-Year-Olds from Rich and Poor Households in Latin America .63 3.9 Gender Gaps and the Inequality of Education, 1970 and 1990 ... 65 3.10 Trend Rates of Poverty Reduction and Nonfarm Output Economic Growth in India, 1960-94 .67 3.1 la Poverty and Landholding, Bangladesh, 1988-89 .74 3.1lb Income Share in the 1980s and Land Gini Coefficients in the 1960s ...... 74 Vill CONTENTS 3.12 Education, Openness, and Economic Rates of Return in 1,265 World Bank Projects ......................................... 79 4.1 Growth Paths and Environmental Quality .................................. 96 5.1 Global Financial Market Size and World Trade, 1980-96 ......... 115 5.2 The Rise and Fall of International Capital Flows, 1990-99 ...... 116 5.3 Relationship between Economic Growth Variability and Volatility in Private Foreign Capital Flows, 1975-96. 120 5.4 Relationships among Democracy, Financial Openness, Capital Mobility, and Government Social Expenditures. 128 5.5 Relationship between Financial Openness and Social Expenditure .128 5.6 Country Classification: Political Rights and Financial Openness .129 6.1 Quality of Rule of Law Indicator: The "Traffic Lights" Presentational Approach .140 6.2 Control of Corruption: The "Traffic Lights" Presentation Approach ....................................... 142 6.3 The Development Dividend of Good Governance .................... 143 6.4 Corruption Is Regressive: Results from Diagnostic Surveys ....... 148 6.5 Corruption and Absence of Meritocracy in Public Agencies Impair Access to Services to the Poor: Results from Public Officials' Diagnostic Surveys .149 6.6 "Petty Bribery" versus State Capture: Does Engaging in Corruption Benefit the Firm? ................... 150 6.7 Corruption and Civil Rights ................... 151 6.8 Meritocracy Can Reduce Corruption: Evidence for each Agency from Surveys of Public Officials in Three Countries .153 6.9 Multipronged Strategies for Combating Corruption and Improving Governance: Recognizing the Political Economy .154 6.10 High Variation in the Quality of Courts in Selected Economies .158 6.11 Legal and Judicial Capture by the Corporate Sector in Selected Transition Economies ......................................... 159 A1.1 Development Objectives and Policy Instruments ...................... 181 A2.1 Constant Returns to Scale and No Technological Spillovers .... 192 A2.2 Subsistence, Growth, and Poverty Traps among the Poor: The Case of Constant Returns to Scale and No Spillovers ....... 195 A 5.1 Correlations of Financial Openness with Political Rights and Civil Liberties ......................................... 224 [X CONTENTS Tables 1.1 Correlations of Measures of Development, 1981-98 .....................3 1.2 Development Outcomes by Growth Class, 1980s and 1990s ....... 15 1.3 External Factors Matter for Domestic Outcomes, Examples from 1997-99 ..................................... 16 1.4 Policy Performance, by Growth Class, 1980s and 1990s .............. 22 2.1 Selected Variables for Brazil, Chile, and Korea ............................ 38 2.2 Review of Development Indicators for 60 Reformers and Nonreformers, Selected Years .39 3.1 Primary School Repetition and Dropout Rates, Selected Years .54 3.2 Public Spending by Level of Education, Korea, Selected Years ....................................... 68 3.3 Public Expenditure Per Student by Level, 1960s to 1990s ........... 69 3.4 Public Current Expenditure Per Student, India and Korea, Selected Years ................................... 70 4.1 Annual Health Costs Associated with Air Pollution .................. 85 4.2 Annual Health Costs Associated with Waterborne Diseases and Pollution .85 4.3 Trade, Growth, Poverty, and Environmental Degradation, Selected Years .92 4.4 Classification of Selected Countries by Growth-Environment Trajectory .97 4.5 Environmental Services of Costa Rican Forests and Their Beneficiaries................... ......... 98 5.1 Growth of Derivatives Markets, 1991-97 ............................ 116 5.2 Country Grouping by Financial Openness ............................ 127 5.3 Estimation Results of the Binomial Logit Model on the Likelihood of Countries Belonging to the High Democracy and High Financial Openness Categories .............................. 127 6.1 A Synthesis Matrix: Corruption and Poverty ............................ 147 6.2 Impact of Civil Liberties on Project Socioeconomic Rates of Return ......................................... 164 A1.1 Relationships between Development Objectives and Policy Instruments, 1981-98 ......................................... 180 A2.1 The Growth Equation under Various Specifications .................. 199 A2.2 GDP Growth Rates Regressed on Stocks Per Worker, Using All Countries with Available Data from 1965 to 1990 .............. 200 A2.3 Elasticities for Stocks Per Worker on GDP per Capita Growth Rates ......................................... 201 x CONTENTS A2.4 Selected Empirical Studies on the Impact and Size of Capital Subsidies ....................................... 202 A3.1 Gini Coefficients of Education for Selected Countries, Selected Years ....................................... 207 A3.2 Production Function: Linear Estimation .................................... 208 A3.3 Production Function: Nonlinear Estimation .............................. 210 A3.4 Education, Openness, and Lending Project Performance .......... 212 A3.5 Selective Empirical Studies on Asset Distribution, Growth, and Poverty ....................................... 213 A5.1 International t'ransactions ....................................... 218 A5.2 Financial Openness Index, Selected Countries, 1997 ................ 219 A5.3 Developing Countries Classified by Degree of Volatility to Private Foreign Capital Flows ....................................... 221 A5.4 Summary Statistics for Selected Industrial and Developing Countries ....................................... 223 A5.5 The Relationship among Financial Openness, Democracy, and Social Expenditure ....................................... 224 A6.1 Percentage of Firms Affected by Different Forms of State Capture, and Overall State Capture Index, Selected Countries, 1999 ....................................... 231 xI FOREWORD W A r e have much to celebrate as we begin the new millennium. A child born today in the developing world can expect to live 25 years longer and be healthier, better-educated, and more productive than a child born 50 years ago. The spread of democracy has brought new freedoms and unprecedented opportunities for many people throughout the world. And the communications revolution holds the promise of universal access to knowledge. But, if we take a closer look, we see something else-something alarming. In the developing countries, excluding China, at least 100 million more people are living in poverty today than a decade ago. And the gap between rich and poor yawns wider. In many countries the scourge of AIDS has cru- elly cut life expectancy-in some African countries by more than 10 years. More than 1 billion people still lack access to safe water, and each year 2.4 million children die of waterborne diseases. As many as a billion people have entered the 21st century unable to read or write. Some 1.8 million people die every year of indoor air pollution in rural areas alone. Forests are being de- stroyed at the rate of an acre a second, with unimaginable loss of biodiversity. Those are some measures of our shortcomings: that despite prosperity for some, the quality of life has remained bleak for many. In the face of al- most two decades of rapid economic growth in some countries, others have not benefited from such progress. In many settings policies have favored vested interests of the elite, and thus not promoted adequate investments in human capital and natural capital, which are essential for broad-based growth. The quality of the factors contributing to growth requires funda- mental attention if poverty is to be reduced and a better quality of life at- tained by all. This is the central theme of this book. xlii F 0 R E D 0 R D A better quality of life for the poor calls for higher incomes. This re- quires sound economic policies and institutions conducive to sustained growth. Achieving higher incomes and a better quality of life also calls for much more-improved and more equitable opportunities for education and jobs, better health and nutrition, a cleaner and more sustainable natural environment, an impartial judicial and legal system, greater civil and politi- cal liberties, trustworthy and transparent institutions, and freedom of access to a rich and diverse cultural life. The World Bank's recent book Voices of the Poor: Can Anyone Hear Us? reinforces this message. Poor women and men from around the world noted emphatically the importance of dignity, respect, security, gender issues, a clean environment, health, and inclusion in addition to material well-being. As incomes per capita rise, several aspects of quality of life improve as well, but not all, not at the same rate, and not inevitably. In different coun- tries the same pace of economic growth has been associated with very dif- ferent degrees of improvements over time in education, health, civil liber- ties, citizens' participation in the decisions affecting their lives, freedom from corruption, and environmental quality and sustainability. This book demonstrates that how growth is generated and whether it is sustainable matter crucially for the quality of life of everyone. The World Bank's strategy is to design and evaluate its activities through the lens of poverty reduction, the vision that informs the Compre- hensive Development Framework we have embraced in the countries we work with. This framework encourages countries to pursue a balanced ap- proach to development, simultaneously trying to augment the human, so- cial, natural, and physical dimensions. Only then are the fruits of develop- ment widely shared and sustained. In aligning these complementary dimensions, this integrated framework also attempts to bring together the key development actors. It puts institu- tions, governance, corporate responsibility, and issues of inclusion, voice, liberties, and participation on par with conventional economic concerns and policymaking. In addressing these related issues simultaneously, the framework emphasizes the need for the country's leadership, as well as for partnership among govemment, the private sector, civil society, and the in- ternational community, in driving the development agenda. We are com- mitted to assisting this framework not only with financing, but increasingly also with state-of-the-art leaming and knowledge programs using new data, toolkits, and methodologies, and supported by the latest information and communication technologies. Indeed, travelling throughout all continents I am constantly reminded by the people-in rural villages and overcrowded urban centers alike-that XIV FOREWORD the quality of life for them transcends the contribution of just financing. That quality is about access for girls and boys to education, and to jobs when they graduate. It is about access for the rural poor to basic medicine when they go to their village clinic. It is about the cleanliness of the air and water and about protecting the precious biodiversity. It is about the dignity the poor might enjoy and the security of their lives. It is about the partici- pation of people along with reformers in the government in implementing an anticorruption program. It is about fighting the vested interests of an economic elite that unduly influences, or even purchases, the policies, regulations, and laws of the state. This volume makes a compelling case for focusing on these policy and in- stitutional dimensions, and for doing so with country ownership and partner- ship among participants in the development process. Investing in people, sus- taining natural resources, managing risks, and improving governance, it shows, are dimensions that make up quality growth. It is more of such growth that can best promote poverty reduction, environmentally and socially sus- tainable development, and an improved quality of life shared by all. James D. Wolfensohn President The World Bank Group xv PREFACE T he 1990s-at the end of a century and a millennium-were a period of stocktaking about development. Studies reexamined and assessed some of the central tenets of development. Sustained economic growth emerged unscathed as fundamental for reducing poverty. And the development record confirmed the efficacy of certain reforms for sustaining growth in both developing and industrial countries: investing more-and more efficiently-in education and health, reducing barriers to trade and investment, dismantling domestic price controls in agriculture and industry, and reducing fiscal deficits. Neither the economic ups nor the downs of the 1990s called these relationships into question. The assessments also uncovered some crucial gaps. Missing from coun- try policymaking as well as the advice, conditions, and financing of exter- nal entities has been adequate attention to the quality and sustainability of growth. Without that, the real potential of reforms cannot be realized. And the assessments highlighted some profound changes in develop- ment thinking over the past 50 years as our understanding of develop- ment processes has matured, shaped by experience. Not that interpreta- tions have been entirely uniform. For example, some have understood the "Washington consensus" solely as a policy prescription for liberalizing markets. Others have accepted the extended interpretation of the World Development Report 1991's market-friendly approach as a market-plus view, involving both liberalization and a strong, positive role for the state and other stakeholders. Differing interpretations aside, the assessments show an emerging con- sensus on some key lessons about complementarity and balance among xvI I' R F F A C I policies and institutions. Functioning markets and liberalization are crucial. But so is acknowledging the limits of the market and an essential role for governments and other stakeholders in the reform process. Sometimes expectations based on experience have been borne out, sometimes not. Early discussions predicted the success of such natural re- source-rich countries as Myanmar, the Philippines, and some in Africa and the failure of such natural resource-poor economies as the Republic of Ko- rea or Singapore. Expectations of rapid development through market liber- alization in the transition economies failed to materialize. And in the 1980s, the slowdown of productivity growth in North American and Euro- pean industrial economies contrasted with the marked success of Japan, prompting calls for changes in the growth paradigm. Sometimes reality fell short of expectations because changes in global and local circumstances blunted the impact of actions and forced govern- ments to revise priorities. "Heavy industries first" seemed the best way for- ward at the tum of the 19th century; information technology seems to be the key to success at the tum of the 20th. And as markets were liberalized in recent decades, sometimes disappointing outcomes exposed the impor- tance of institution building to make markets work. We present this book in this spirit of continuous inquiry and feedback in framing development thinking. It is addressed to policymakers, practitioners, and others in developing as well as industrial countries. It reaffirms the crucial contribution of market-friendly policies. It also highlights key missing ingredi- ents and fresh evidence. Not a complete review of development, the book ex- amines vital issues that are often overlooked as a basis for action: the distribu- tion of opportunities, especially education; environmental sustainability; management of risks; and govemance and anticorruption. It does not address such important factors as the political economy of change, the influence of so- cial instability, the consequences of communicable diseases such as HIV/ AIDS, or the impact of cross-border and global issues-population pressures, labor migration, global warming, information technology, and the global fi- nancial and business architecture. It concludes that growth is crucial, and so is the quality of growth. The work for this book was carried out by a team at the World Bank Institute. The work was supported by a research grant from the World Bank as a contribution to the learning material for a course on development as well as background for the World Development Report 1999/2000: Entering the 21st Century and the World Development Report 200012001: Attacking Poverty. It was inspired by a keynote paper and discussion by Stanley XVIII I R E FAC E Fischer at the 1998 Annual World Bank Conference on Development Eco- nomics. The team benefited from comments from many people outside and inside the World Bank. Special thanks are due to those who provided sug- gestions for the development of the book. These include Nancy Birdsall, Paul Collier, Eduardo Doryan, Ravi Kanbur, Mats Karlsson, Gautam Kaji, Rung Kaewdang, Vijay Kelkar, Mohsin Khan, Aart Kray, Nora Lustig, Rakesh Mohan, Mohamed Muhsin, Robert Picciotto, Jan Piercy, Jo Ritzen, Lyn Squire, T. N. Srinivasan, Nicholas Stern, Thomas Sterner, Joseph Stiglitz, Anand Swamy, Shahid Yusuf, Shengman Zhang, and the World Development Report 200012001 team. The team thanks the following for their inputs to the volume: Montek Ahluwahlia, Jane Armitage, Kaushik Basu, Surjit Bhalla, Jan Bojo, Deepak Bhattasali, Gerard Caprio, Shaohua Chen, Kevin Cleaver, Maureen Crop- per, Monica Dasgupta, Shanta Devarajan, Ishac Diwan, David Dollar, Will- iam Easterly, Gershon Feder, Andrew Feltenstein, Deon Filmer, Pablo Guerrero, Cielito Habito, Kirk Hamilton, Jeffrey Hammer, Joseph Ingram, Farrukh Iqbal, Ramachandra Jammi, Emmanuel Jimenez, Mary Judd, Philip E. Keefer, Homi Kharas, Elizabeth M. King, Kathie Krumm, Ashok Lahiri, Kyung Tae Lee, Andres Liebenthal, Magda Lovei, Muthukumara Mani, Michele de Nevers, David Nepomuceno, Jostein Nygard, Michael Pomerleano, Tanaporn Poshyananda, Lant Pritchett, Martin Ravallion, David Reed, Neil Roger, William Shaw, Mary Shirley, Ammar Siamwalla, Hadi Soesastro, T. G. Srinivasan, Tara Vishwanath, Christina Wood, Michael Woolcock, Roberto Zagha, and discussants and participants in seminars at the International Monetary Fund/World Bank Annual Meet- ings, the National Council of Applied Economic Research (India), Asian Development Forum (Singapore), Thailand Development Research Insti- tute, as well as a presentation at an International Monetary Fund confer- ence on reforms. Several units of the World Bank reviewed the manuscript. xix THE REPORT TEAM T his work was carried out by a team at the World Bank Institute led by Vinod Thomas and comprising Mansoor Dailami (chapter 5), Ashok Dhareshwar (chapter 1), Daniel Kaufmann (chapter 6), Nalin Kishor (chapter 4), Ram6n E. L6pez (chapter 2), and Yan Wang (chapter 3 and task manager). The team was assisted in its research by Cary Anne Cadman, Xibo Fan, and John Van Dyck. Taji Anderson, Alice Faria, and Jae Shin Yang provided support. Bruce Ross-Larson and Meta de Coquereaumont of Communications Development Incorporated and International Communications, Inc. (ICI) of Sterling, Virginia, edited the manuscript at different stages. ICI also pro- vided typesetting and proofreading. Product development, design, editing, production, and dissemination were directed and managed by the World Bank's Office of the Publisher. xxI OVERVIEW T he last decade of the 20th century saw great progress in parts of the world. But it also saw stagnation and setbacks, even in countries that had previously achieved the fastest rates of economic growth. These gaping differences and sharp reversals teach us much about what contributes to development. At the center is economic growth, not just its pace but-as important-also its quality. Both the sources and the patterns of growth shape development outcomes. Have those patterns been adequate for rapidly reducing poverty or im- proving the quality of people's lives? Why have so few countries sustained robust growth rates for prolonged periods? Why have some crucial dimen- sions-income equality, environmental protection-deteriorated in so many economies, both fast- and slow-growing? How does governance un- derpin the growth process? As answers, we offer three principles of develop- ment and a set of actions for enhancing the quality of growth processes. Development Outcomes and Growth Processes Development is about improving the quality of people's lives, expanding their ability to shape their own futures. This generally calls for higher per capita income, but it involves much more. It involves more equitable education and job opportunities. Greater gender equality. Better health and nutrition. A cleaner, more sustainable natural environment. A more impartial judicial and legal system. Broader civil and political freedoms. A richer cultural life. As per capita incomes rise, several of these aspects improve in varying degrees-but others do not. How can growth pro- cesses be influenced so that the qualitative dimensions of development XXIII O V E R V I E W outcomes also improve? This book explores these issues of faster and bet- ter growth. A recent study, Voices of the Poor: Can Anyone Hear Us? (Narayan and others 2000), indicates that raising incomes is one part of poverty reduc- tion. Greater security in life and a more sustainable environment are oth- ers. The experience of the past decades and the voices of the poor offer compelling reasons to emphasize these qualitative factors. Indeed, from Bolivia, Egypt, and Uganda to Romania, Sri Lanka, and Thailand, the development community is broadening the traditional defi- nition of poverty and welfare. Beyond an individual's or household's mea- sured income, welfare includes opportunity, as assessed by the functioning of markets and investments and improvements in health and education. It includes security, as reflected by reduced vulnerability to economic and physical shocks. It includes empowerment, as evaluated by social inclusion and the voice of individuals. And it includes sustainability, as represented by the protection of the environment, natural resources, and biodiversity. Economic growth has been associated positively with poverty reduc- tion. Early assessments projected a growth rate for the developing world for the 1990s of a little over 5 percent, or about 3.2 percent per capita. They projected a reduction in the number of poor people of some 300 million, or an annual rate of decline of nearly 4 percent. But actual growth during 1991-98 was about half that, at 1.6 percent per capita. If the countries of Eastern Europe and Central Asia are excluded from these estimates (as in the above-mentioned projections), actual per capita growth is closer to pro- jected growth, at 3.5 percent-with the number of poor people unchanged and the incidence of poverty down 2 percent a year (World Bank 2000a). The poverty reduction associated with growth has varied widely, as have social progress and welfare improvements, whether in education, health, voice, or participation (chapter 1). Where growth has stagnated or declined, social and welfare dimensions have deteriorated. The widely dif- fering measure in which growth contributes to welfare improvements means that there must be a direct concern for sustainable advances in wel- fare. It also means that the way growth is generated is very important. The quality of the growth process, not just its pace, affects development out- comes-much as the quality of people's diets, not just the quantity of food, influences their health and life expectancy. That is why it is essential to ex- plore the complex interactions of the factors shaping growth. The pace of growth has been more sustainable in developing and indus- trial countries that pay attention to the qualitative attributes of the growth process. Indeed, there is a two-way relationship between economic growth and improvements in social and environmental dimensions. Attention to the xxiv OVE RV I EW sustainability of the environment, for example, helps to deliver more sus- tained growth, especially where growth rates are highly variable and the negative impacts are particularly pronounced for the poor. This suggests a premium for steady growth rates over stop-and-go growth, even if the go in- cludes short periods of fast growth. As countries exhaust the possibilities for increasing growth through market reforms, the qualitative factors supporting long-term growth become much more important. These dimensions of the growth process often interact positively in a virtuous cycle. But there can also be some difficult tradeoffs between quan- tity and quality. Fast, temporary growth relying on such distorted policies as subsidies to capital, neglect of environmental externalities, and biased pub- lic expenditure allocations can actually diminish prospects for more sus- tained growth. Even more difficult to correct are situations in which growth conflicts with environmental and social sustainability, which both contrib- ute directly to development. Managing these qualitative aspects becomes essential for achieving sustainable improvements in welfare. What then is the quality of growth? Complementing the pace of growth, it refers to key aspects shaping the growth process. Country experi- ences bring out the importance of several such aspects: the distribution of opportunities, the sustainability of the environment, the management of global risks, and governance. These aspects not only contribute directly to development outcomes. They also add to the impact that growth has on these outcomes, and they address the conflicts that growth might pose to environmental or social sustainability. It is the mix of these policies and in- stitutions shaping the growth process that is the main focus of this study. Principles of Development Viewing the quantitative and qualitative sides of the growth process to- gether puts the spotlight on three key principles for developing and indus- trial countries: * Focus on all assets: physical, human, and natural capital * Attend to the distributive aspects over time * Emphasize the institutional framework for good governance. The Major Assets Broadly speaking, the assets that matter for development are physical capital, human capital, and natural capital. Technological progress affect- ing the use of these assets matters as well. For accelerating growth rates, xxv O V F R V I EW much attention has traditionally gone to the accumulation of physical capital. But other key assets also deserve attention-human (and social) capital as well as natural (and environmental) capital (box 1). These as- sets are also crucial for the poor, and their accumulation, technological progress, and productivity, along with that of physical capital, determine the long-term impact on poverty. By focusing predominantly on physical capital, industrial and developing countries can be tempted to implement policies that subsidize it at a cost (chapter 2). This can create a situation that benefits vested interests and is hard to reverse. Meanwhile, from the social viewpoint there is Bo .AstAcuuain Growth, and Welfare 0 00 0: 00 Fig 1 l oasimple sch of how human by regressively reallocating public expenditures (H), narural (RX and physical (K) capital contrib- t*Natural capital, by undermining tax, ute ro economic growth and welfare. Physical ties, and regulations that could sustain n capial cntriutesto elfae thougheconmic resources. growt. Humn (an socil) caital nd naural Distortions, market fai'lures, implicit govrmn als direct components o welfare. f guarantees, and inadequate regulation can Humana na tura c ao c to th Oerin estment or wasteful investment in accumulation iof ph c by ican e physical capital by increasing the profitability X turns.0 Physical capital increases the returnstohu- of certain physical assets through guarantees- man capital and natural capital and, if markers re- which influence risk-raking behavior by banrks, lect0 this;: their accumulation. Adding to all this, corporations, and investo sd by low investments in physical, human,andnatural, captal, 0 Xthe value of certain natural resources together with many policy reforms, ontribute to :0 *Underinvestment in human and natural re- technological progress and the growth of total factor sources by underpricing these assets and by re- productivity, thereby boosting growth (chapter 2). ducing resources devoted to them. But oliy dstotion, crrutio, migovrnace, The effects of these policy dsotoso h c on pah o ditotedor nbaancd sse acumua-cumulation of humnanan~dnatua aiaeaiet _on. This situation canholdincomegrowthan physical capital can reduce gr welfare improvementsbelow theirpotential. MoreConversely, if s ctroand gover nance is adequat,udsotdplce ol os speifialy, t cn ad o lwe toal hcor rodc-asset accumulaincotbungofserrwh tivity and underinvestments in(chapter 6). So,b eoigplc itrin,fs * Prductve pysial cpita, b redcingthe tering good govenne n drsigmre kl proftablit of nvetmet troug brbesand ure's and externalities, countriescnaheels red tape or by distoertingtheallocation of vdistortedo, more balanced asset investments A physical investm -a toward certain lu-l that can lead to more stable and sustaie gwth cive c and tobroadlybsd inees in welfare. xx HuaIaia,b rmoigsc aoe r OVE RV I EW I - - natural captal, oftenbecause of ts underprcaitgorwal) rpetrgt ssnsot, ng migvtaner amontd crutioan esiae KS70 bilo oU$0 oetir Rntedcn distlortions faorngs Kd oradClmi19) Sustaecning markerlatilues hurtingec H n f hysica (aphyical accmultio Welfar reu* Strenungditheniong. reulton cxml, asphyital)cptldepn,ss undt ermvemen ine oeduloiation offrssand helhcather 3)tuand overexploratins dofn natural capital oftnd because eniofmeitsl sundepriingbliy or weak, propert rdghts- (ichsapters 4).rA anbggegte l5 evel, (fgross) susdimestoc agricultue eGPnergy, road billionping thoerly. Co9r0s,e about tw-hirdspeto of iintenvirndustral couniesand ow oe-thrd gnin te delopetc aingons (dre Moory and Calamaie 1997) en, o GP aTtrac forlueignh capial ofurthermor, accelerasoeting growths though2 policiest ever genuine dmsi savings weregtie onya1stmtd1 percent, n husa Fdrtof, GDP. rates of 25 percent and negative 1.6 percent (World Bank 1l999d). XXVI I O V F k V I E W A less distorted (more neutral or balanced) approach to the accumula- tion of the three types of assets is preferable. Policies can contribute to the accumulation of these assets. Investments in education at all levels, while helping to generate growth, also contribute to the accumulation of human capital and welfare. Investing in natural capital is essential to human health and, for the many poor people who depend on natural resources for their live- lihoods, to economic security (chapter 4). As important as the accumulation of these assets is their efficient use. For that-and for greater total factor pro- ductivity of these assets-good governance, mitigating the undue influence of elite interests, and anticorruption actions are vital. Distributional Aspects This focus on quality brings to light the importance of distributional as- pects to the growth process. A more equitable distribution of human capi- tal, land, and other productive assets implies a more equitable distribu- tion of earning opportunities, enhancing people's capacity to take advantage of technologies and to generate incomes. That is why a given growth rate is likely to be associated with better poverty outcomes in set- tings where educational opportunities are distributed more equitably (chapter 3). Stability in growth outcomes over time is also likely to be important. The incomes of the poor can be very sensitive to cycles and crises, espe- cially because the poor lack assets-land, skills, and financial savings-to smooth their consumption in bad times. Living barely above the poverty line, millions of near-poor have been thrown back into poverty by external shocks. So for growth to reduce poverty, it usually needs to be relatively stable, and its benefits need to be widely spread. What, then, of the income gains expected from globalization in the 1990s? These have begun to materialize, but not everywhere. One reason is the inadequacy of regulatory and supervisory frameworks both at the global and country levels, and the overall lack of preparedness for partici- pating in the global economy. Another is the volatility sometimes related to moral hazard and the responses of external players. A third is that, by some estimates, incomes in the last decade or so have become more un- equal. Aims of development policy thus include reducing not only the in- equality of opportunities, but also the inequality and volatility of growth outcomes. In this it is important to enhance financial risk management and reduce the sensitivity of poor people to changing economic fortunes (chapter 5). XXVIII OVE RVI EW The Governance Framework The institutional structures of good governance underpin everything done to boost growth. The effective functioning of bureaucracies, regulatory frame- works, civil liberties, and transparent and accountable institutions for ensur- ing the rule of law and participation matters for growth and development. The effects of poor govemance, bureaucratic harassment, and corruption are regressive and harmful to sustained growth. The capture of state policies, laws, and resources by elite interests often biases incentives and public expen- diture toward less socially productive assets and, by eroding the benefits that would go to society, reduces the impact on welfare. Estimates of the "devel- opment dividend" in the form of higher incomes or better social outcomes are dramatic, in going from low levels of rule of law or high levels of corrup- tion to even middling levels. (A difference in corruption levels of even one standard deviation can be associated with enormous differences in the devel- opment impact.) Thus investing in the capacity for better governance is a top priority for better economic performance (chapter 6). A vibrant civil society-empowered by Internet computing tools, diag- nostic survey techniques, and the latest information on govemance-is in- dispensable in the fight against corruption and other forms of misgovernance. Civil liberties are not only linked positively to improved governance, reduced corruption, and increased productivity of public in- vestments, but they also contribute to welfare directly. Indeed, attention ought to go beyond getting the government side of the equation right. It also needs to go to enhancing civil rights and giving greater voice to di- verse groups, promoting competitive enterprises, and complementing top- down government policy reforms with bottom-up formulation and imple- mentation of development strategies. Neglected Actions in the Growth Process While we now see the development process more broadly, there is often in- adequate attention, especially in a crisis, to two of the three assets that the poor rely on: human and natural capital. This neglect in turn seems to have led to the neglect of some key actions: * Improving the distribution of opportunities * Sustaining natural capital * Dealing with global financial risks * Improving governance and controlling corruption. XXiX 0 \' 1 [\/ N' U 0 Addressing them contributes not only to asset accumulation, but also to technological progress and greater total factor productivity. Improving the Distribution of Opportunities The main asset of poor people is their human capital. Of 85 economies ex- amined, Poland and the United States (estimated to have the highest aver- age years of schooling) have the most equitable distribution of educational attainment among people in the labor force. And the Republic of Korea re- corded one of the largest improvements in educational equality over the past three decades. But inequality in education remains staggering, as in Algeria, India, Mali, Pakistan, and Tunisia. The relation of any given growth rate to poverty reduction depends on the investments in people. The more equitable the investments, the greater the impact of growth in lowering the incidence of poverty, as seen in a comparison of the effects of growth on poverty across Indian states (Ravallion and Datt 1999). If people's abilities are normally distributed across the population, the skewed distribution of education and health out- comes would seem to represent especially large welfare losses to society as a significant proportion of people are deprived of opportunities to use new technologies and to lift themselves out of poverty. A review of education spending for 35 countries finds it to be weakly related to educational attainment after controlling for incomes. In health, the United States, with the highest per capita health expenditure, ranks 37th among 191 countries in a measure of overall health system perfor- mance. France, with less than 60 percent of the U.S. per capita health ex- penditure, ranks first. Colombia, which is much further down in per capita health expenditure, ranks first in the category of fairness of finan- cial contribution (WHO 2000). So spending on education and health ser- vices is not enough. Also demanding attention are the breadth and depth of human capital-its quality and equity, as measured by girls' education, access for the poor, and school attainment. Governments need to reallocate public expenditure for basic educa- tion to ensure its quality and equitable distribution. Private-public part- nerships need to be encouraged through market-based policies to increase efforts in education at all levels, including higher education. Also needed are supportive labor market policies and social protection policies. In ad- dition, the human capital of the poor can be better applied by addressing the distribution of land and pursuing labor-intensive strategies in an open global environment (chapter 3). xxx o)V F R VI E W Sustaining Natural Capital Environmental degradation has worsened sharply. Contributing factors in- clude population growth; domestic and global pressures on scarce resources; economic policies, for example, subsidies that ignore environmental conse- quences; and the neglect of local and global commons. The costs of envi- ronmental pollution and resource overexploitation are enormous; the losses in many cases are irreversible. Indonesia's forest fires-the result of human, policy, and natural factors-produced some US$4 billion in direct losses in 1997 and again in 1998, with extensive damage in neighboring nations. And it is poor people, because of their reliance on such natural capital as land, forests, minerals, and biodiversity, who suffer disproportionately from environmental degradation. Few countries have adequately confronted the underlying causes of en- vironmental and resource degradation-the policy distortions, market fail- ures, and lack of knowledge about the full benefits of environmental pro- tection and resource conservation. Growth and higher incomes can create conditions for environmental improvement by increasing demand for better environmental quality and by making resources available to meet that de- mand. However, only a strong combination of domestic and global market- based incentives, investments, and institutions can make environmentally sustainable growth a reality-as in examples from China, Costa Rica, Indo- nesia and many European countries (chapter 4). Dealing with Global Financial Risks Global financial integration has large benefits, but it also makes countries more vulnerable to hidden risks and sudden swings in investor sentiment. Volatile private capital flows seem to be associated with volatile growth rates, which hurt especially the poor, who lack the assets to weather an economic storm. To deal better with such risks, countries need to maintain sound macroeconomic policies. They also need to deepen domestic finan- cial markets, strengthen domestic regulation and financial supervision, in- troduce corporate governance mechanisms, and provide social safety nets. For all this they need sound institutions and strong capabilities, which take time to cultivate. Developing them, while opening a country's capital markets, can help deal with the risks for the financial system and the economy. In the meantime, as governments open their capital accounts, they can consider a spectrum of actions as in Argentina, Chile, Mexico, and else- where. One is to shun special incentives for short-term flows. Another is to XXXI ( V E RV I E W set up reserve requirements and taxes for risky short-term flows. Yet another is to strengthen prudential regulation and supervision. International policy coordination and lender-of-last-resort activity can provide liquidity and emergency financial assistance (chapter 5). Improving Governance and Controlling Corruption Governance needs to move to center stage in institution-building strate- gies. That requires better analysis and measurement of the dimensions of governance and a clearer understanding of the vested interests of power- ful groups. Where the legal and judicial frameworks are weak and vested interests have taken over the state policymaking and resource allocation apparatus, the social cost can be enormous. In that case the institution building needed for effective development interventions may be exten- sive, warranting an active approach. Participation and voice would be vital for increasing transparency, providing the necessary checks and balances, and countering state cap- ture by the elite. The engagement of civil society in participatory and transparent processes with reformists in the executive, legislative, judi- ciary, and private sectors can make the difference between a well- governed and a misgoverned state, between a stagnant and a thriving so- ciety. A rigorous understanding of governance would need to be supported by new technologies (as in Albania, Bolivia, Georgia, Latvia, and several African countries). Creating a climate for successful development thus requires an inte- grated approach linking economic, institutional, legal, and participatory el- ements: building transparent and effective institutions for budgeting and public investment programs (as in Australia, New Zealand, and the United Kingdom), as complements to macroeconomic policies; establishing merit- based public administration (as in Malaysia, Singapore, and Thailand) and efficient and honest customs and procurement agencies; and promoting civil liberties and popular participation (chapter 6). Shifting Priorities Why focus on quality when the pace of growth is slow in many parts of the world? Growth has been modest in many countries-about 1.6 percent per capita for low- and middle-income countries since the 1980s and lower still when China and India are excluded. Some countries are also facing or com- ing out of financial crises. In these circumstances, the issue is not one of quality or quantity. Both are essential, and both are involved in a two-way relationship. xxxII OVE RV I EW A relative shift in priorities could boost the pace of long-term growth. In- vestments in human capital-education, health care, and population poli- cies-can directly improve the quality of life. They can also improve invest- ment incentives through the effect of a healthier, more educated work force on the productivity of capital. So shifting the emphasis more toward human capital could promote faster growth in the long run. The key point? A focus on the quality of outcomes could help to sustain more rapid growth. Addressing the quality dimensions that contribute to the pace of growth can in turn enhance welfare directly. For example, less air and water pollution or less degradation of natural resources, in addition to contribut- ing to growth, enhances welfare directly by improving health or providing greater opportunities for income and consumption. This book shows that some processes and policies in developing or indus- trial countries genetate economic growth with greater equality of human de- velopment, sustainability of the environment, and transparency of gover- nance structures. Others do not. Furthermore, sequencing actions is unlikely to be effective-whether liberalizing first and regulating later, privatizing first and ensuring competition later, growing first and cleaning up later, or grow- ing first and providing civil liberties later. To do the most for long-term growth, liberalization, for example, needs to go together with regulatory ac- tions, environmental management, and anticorruption measures. Defining the Shift The actions focusing on the quality of growth need to be a core part of the policy package, not add-ons to an already crowded agenda. That means that stakeholders will have to augment actions by governments, shifting the emphasis to * Asset accumulation and use, by reducing policy distortions, for ex- ample, those favoring or subsidizing physical capital, while comple- menting markets in valuing natural resources and investing ad- equately in human resources. The implication is to ensure broadly based, sustainable growth, not to slow growth. * Regulatory frameworks, by building regulatory frameworks for competi- tion and efficiency to accompany liberalization and privatization and giving legal and judicial reforms greater attention, while ensuring macroeconomic stability. The implication is to take supportive regu- latory actions along with liberalization, not to slow liberalization. * Good governance, by nurturing civil liberties, participatory processes, and accountability in public institutions; promoting anticorruption ef- forts; and actively involving the private sector to reduce the influence xxxII V E RV I F VW of vested interests, while building capacity for policy changes. The im- plication is to increase the attention to coalition building in civil soci- ety, not to detract from govemment policy and capacity building. Making the Shift-Now How can more and better investment in people and natural capital be fi- nanced? In several ways. First, improving govemance, reducing rent-seeking and corruption, and encouraging greater corporate responsibility can increase national savings. Second, increasing the charges for the use of natural re- sources and taxing such extemalities as pollution can make more resources available for development. Third, reducing distortions that favor physical capital can be beneficial-as with much of the experiences in removing such distortations. It can allow a reallocation of national savings in favor of hu- man development. And fourth, reducing subsidies within sectors for services that are regressive or damaging to the environment can reallocate public re- sources to benefit the poor or to promote sustainable development. In sum, this book supports broadening the focus of actions to encompass a comprehensive development framework, a qualitative and fuller agenda involv- ing structural, human, social, and environmental aspects of the growth process. This broader focus complements liberalization with a build-up in the assets and capabilities of the poor. It shifts attention from an exclusive reliance on gov- emment as the agent of change to the engagement of all parts of society. And it requires much more effective capacity building across the board. With all development partners complementing one another, a more in- tegrated framework can be implemented more effectively. First, the large inequalities in opportunities-especially in education-if addressed now, present the greatest promise for welfare gains to society. Second, the envi- ronmental damage and biodiversity losses from current growth patterns are frightening, but if they are addressed now, growth can achieve a better natural environment and reduce the number of poor. Third, globalization presents risks to the poor, but if those risks are addressed now, globalization could make possible the technological wherewithal for reducing poverty. Fourth, corruption, misgovernance, and a lack of civil liberties and voice threaten the gains from any action, but if those threats are addressed now, better governance presents great promise of improved welfare. The opportunities afforded by increased openness, knowledge, and technologies have never been more plentiful. Equally, the challenges of poverty, population growth, environmental degradation, financial distress, and misgovernance have never been greater. Needed is more growth with a focus on quality. This is not a luxury. It is crucial for countries to seize the opportunities for a better life for their present and future generations. XXXIV C H A P T E R 1 A MIXED DEVELOPMENT RECORD Economics is not only concerned with generating income, but also making good use of that income to enhance our living and our freedoms. -Amartya Sen, A Conversation with Sen n the 1990s, a group of economies in East Asia posted some of the fastest growth rates and sharpest declines, as well as recoveries, giving market liberalization policies both strong support and serious _ .L qualification. In many ways the 1990s concentrated the development experiences of the previous decades, offering approaches and cautions to guide action in the 21 st century. Looking at the previous decades of development, various studies from around the world in the 1990s spotlighted the successes in East Asia, the set- backs in Sub-Saharan Africa, and the modest gains elsewhere. The World Development Report 1991 (World Bank 1991) articulated an emerging con- sensus under the rubric of a market-friendly approach, calling for a reap- praisal of the roles of state and market. This and other reviews signaled the crucial roles of the state and markets in poverty reduction (World Bank 1990), in environmental protection (World Bank 1992), in infrastructure provision (World Bank 1994), and in the legal and governance frameworks and the financial system (World Bank 1997j). Here we examine how the lessons of development have played out over the past decade. We update earlier assessments of how countries are per- forming with regard to poverty reduction, sustainable development, and economic growth. And we examine the global factors and policy and insti- tutional changes underlying the countries' performance. T i I QU AL I T O F G R O W T H Evidence from the 1990s expands the development story, especially conceming institutional requirements for success, and provides a rich set of hypotheses to consider with regard to policy. First, investments in people need to be concerned with the quality and distribution of those invest- ments. Second, rapid growth, while it supports social development when broad based, can hurt environmental sustainability in the absence of appro- priate actions. Third, while market openness and competition continue to provide benefits, the financial risks must be managed with attention to country-specific factors. Fourth, good governance and institutional factors should be given priority and not postponed for later stages of reform. Assessing Development Development is about people and their well-being, which involves their ability to shape their lives. Accordingly, development must be inclusive of future generations and the earth they will inherit. It must engage people, for without their participation, no strategy can succeed for long. This no- tion of development as well-being means that measures of development must include not just rates of growth, but the dispersion, composition, and sustainability of that growth. Development practitioners have often used growth in gross domestic prod- uct (GDP) per capita as a proxy for development, partly because social progress is associated with GDP growth and partly because of expediency. However, re- liance on GDP as the sole measure of development is seriously limiting. GDP growth can be of high quality or of poor quality. Some processes and policies generate GDP growth along with the growth of human and natural assets that directly affect people's welfare beyond their productive roles. Others generate poor quality growth that is not associated with improvements of human and natural assets. To integrate the quality of growth in assessments of develop- ment, multidimensional indexes of welfare are needed. Economic theory distinguishes the concept of growth from the broader idea of development. How carefully this distinction has been made has var- ied over time.' The rapid growth of the 1950s and 1960s motivated an in- creased concem for broader development goals. Over the following decades, as stagnation set in, the emphasis shifted to economic growth. In the 1990s, the broader outlook reemerged, as exemplified in the United Nations Devel- opment Programme (UNDP) Human Development Report (produced annually since 1990) and the World Bank's A Proposal for a Comprehensive Develop- ment Framework (Wolfensohn 1999). In an ideal assessment of development, progress would be measured by human and environmental advances before considering intermediate indicators, such as GDP. However, we lack good quality data to construct 2 A MIXED DEVELOPMENT RECORD robust indicators of human and environmental progress and consequently rely heavily on GDP. We supplement the analysis with indexes of human development and environmental sustainability, keeping in mind serious data limitations on some variables. Lack of consistent data on the inci- dence of poverty, comparable internationally and over time, forced us to exclude a poverty reduction component in our human development index. However, we document, where possible, progress in poverty alleviation and the impact of growth and development policies on poverty (see also Dollar and Kraay 2000; Ravallion and Chen 1997; World Bank 2000i). Future work should improve the scope and empirical basis of these indexes and ex- pand the discussion to other dimensions, including cultural well-being. Table 1.1 shows correlations among the components of the three indi- cators of progress since 1981: human development, income growth, and en- vironmental sustainability. It shows that GDP growth is correlated A step toward better and broader measures of development Table 1.1. Correlations of Measures of Development, 1981-98 Income Huraan development growth Environmental sustainability Decrease Decrease Increase Decrease Increase Decrease in carbon Increase Decrease Measure of in in in infant in life in income Growth dioxide in forest in water development poverty literacy mortality expectancy inequality of GDP emission cover pollution Human development Decrease in poverty 1,00 -0.40 0.18 0.14 0.44 0.52 -0.45 -0.23 0.28 27 28 28 20 27 27 26 22 Increase in literacy 1.00 0.15 -0.19 -0.23 0.03 -0.14 0.15 -0.21 115 115 41 89 102 94 72 Decrease in infant 1.00 0.54 0.28 0.20 -0.20 -0.12 -0.13 mortality 146 43 104 121 107 81 Increase in life 1.00 0.54 0.17 -0.16 -0.15 -0.05 expectancy 43 104 121 107 81 Decrease in income 1.00 0.34 -0.33 -0.20 0.32 inequality 39 41 41 37 Income growvth Growth of GDP 1.00 -0.53 -0.06 0.33 100 81 65 Environmental sustainability Decrease in carbon 1.00 0.27 -0.38 dioxide emission 87 70 Increase in forest cover 1.00 -0.14 70 Decrease in water 1.00 pollution Note: The ns'o values in each cell are the correlation coefficient and number of countries. Entries in bold italics are significant at the 10 percent level or better. Sources: World Bank (2000c); authors' computations. 3 1- H E Q_J A l I T Y O F Ci R O V T II * Positively with reduction in poverty, income inequality, infant mortality, and increase in life expectancy, with considerable differ- ences in strength * Negatively with decline in carbon dioxide emissions and positively with decline in water pollution. Other associations between GDP growth and changes in the compo- nents of human development and environmental sustainability are not sta- tistically significant. These preliminary correlations suggest that GDP growth is a crucial, yet partial, indicator of development, as when it is im- perfectly associated with certain aspects of human development, and at times when it is associated with an increase in environmental damage. Easterly (1999a) applied several techniques to a large set of quality of life indicators, including tests of causal relationships. He found that fewer than 10 percent of the 81 examined indicators improved with growth. A similar fraction deteriorated with growth, and many showed no signifi- cant association with growth (figure 1.1). These findings strengthen the case for broadening the measures of development. It is very important to note that the relationships discussed above are be- tween growth of income and changes in human development and environmen- tal sustainability. Relationships in most cases are much stronger with levels of income and indicators, particularly for human development indicators (Dasgupta 1993; Fedderke and Klitgaard 1998; Kakwani 1993; Sen 1994; Figure 1.1. GDP Growth and Changes in Quality of Life, 1960s and 1990a Quality of life indicator Illustrative indictors Industrial countrie Infant survival rate S * Calorie and protein intake Some indicators improved during growth Some indicators were Secondary school enrollment insensitive to growth ~ * Gini coefficient of income Developing countisSoeiditr Somne indictors deteriorated during growth \ during groh * Carbon dioxide per capita 9 Sulfur dioxide per capita 1960s 1980s/90sl TLme Note: The schematic trends are applicable for countries with positive GDP growth. Source: Easterly (1999a). 4 A MIXED DEVELOPMENT RECORD World Bank 2000i). Easterly's study also looks at this discrepancy, hypothesiz- ing that cross-country analysis of income levels may capture long-term trends that are not discernible in analysis of shorter periods, and that growth may lead to improvements in human development with long and varying lags. Al- ternatively, country-specific factors, such as endowments, location, and social infrastructure, could be dominant determinants of levels of both income and human development indicators. In that case the cross-country correlations be- tween income and quality of life indicators would need to be qualified. The Development Record Progress in some areas of human development, especially in extending people's lives and increasing literacy, has been considerable over the 1960s through the 1990s, a period over which some data are available. However, many other qualitative aspects of life, including steady and sus- tained increase in incomes, poverty reduction, equality gains, and envi- ronmental quality, have lagged. Human Development Robust economic growth is accompanied by improvements in measures of human development, such as higher literacy and life expectancy. The broad association is seen in figure 1.2. Overall, the gains in human development over the past four decades have been enormous in some areas-partly reflecting technological im- provements-and modest in others. Infant mortality and adult illiteracy rates fell dramatically almost everywhere. Progress in raising incomes and reducing poverty has been mixed, based on the data and estimates available (figure 1.3). In the developing world, the poverty headcount index, defined as the proportion of people with an income of less than US$1 a day based on 1993 purchasing power parity (PPP) prices, decreased from 28.3 percent in 1987 to 24 percent in 1998. The East Asia and Pacific region showed the largest improvement, particu- larly China in the mid-1990s. Improvements were modest in the Middle East and North Africa and the South Asia regions. Poverty rates remained stubbornly high in Sub-Saharan Africa and Latin America and the Carib- bean and increased sharply in Europe and Central Asia. Overall the de- crease in the poverty rate could not keep pace with population growth, and the number of poor in the developing world outside China increased by about 106 million between 1987 and 1998 (World Bank 1999c). At the end of the 20th century, the incidence of poverty increased in many parts of the world. In particular, the East Asian countries directly 5 T 11 E Q_U A L I T Y 0 F G R 0 W 1 H Figure 1.2. Change in Human Development and Growth of Income, 1981-98 Change in human development (index) 100 - 80 - . 60 - 40 - * * 20 . . -4 -2 0 2 4 6 8 10 12 GDP growth (percent per year) Note: r = 0.22, p < 0.05, n = 89. The data are for 89 developing countries. Controlling for per capita income in 1981 gives a stronger pattern with a correlation coefficient of 0.33. Source: World Bank (2000c); authors' computations. affected by the 1997 financial crises and the consequent slowing of growth experienced reversals in poverty reduction achieved during their period of rapid growth (World Bank 2000f). Even greater is the poverty increase in the transition economies of Europe and Central Asia, where as recently as 1987 poverty and income inequality were both extremely low. Survey data show an enormous increase in the number of poor as a result of sustained declines in economic output and worsening income distributions (Milanovic 1997) (figure 1.4). Environmental Degradation The impact of economic growth on environmental conditions has been mixed and is a serious concern. In many instances GDP growth and higher incomes are associated with better sanitation and water quality, as well as investments in cleaner technologies. But growth is also related to increases in particulates and carbon dioxide emissions.2 With equal weights for the changes in indicators of water quality, air quality, and deforestation, be- tween 1981 and 1998 income growth was associated with environmental deterioration and depletion of natural resources as in figure 1.5. Between 1990 and 1995, the rate of forest clearing slowed in most de- veloping regions, but forest cover was still disappearing rapidly. Forest cover 6 A MIXED DEVELOPMENT RECORD Figure 1.3. Poverty Rates and Number of Poor, Selected Years Poverty Headcount Index Percentage (of population living on less than US$1 per day) 60 - 50 - 40 -1U18 50 U 1990 30 U 1993 Z2 1996 20 D 1998 10 Africa East South Europe Latin Middle East Developing Asia Asia and America and countries Central Asia North Africa Number of Poor Millions 1,400 1,200 _ _ 600 U 1987 U 1990 400 U 1993 171 1996 200 17 1998 Africa East South Europe Latin Middle East Developing Asia Asia and America and countries Central Asia North Africa Note: Based on 1993 purchasing parity rates. The 1998 values are estimates. Poverty is defined as income of less than US$1 a day. Source: World Bank (1999d). increased only in high-income countries and in developing Europe and Central Asia. It is unclear how much of the improvement in the latter is the result of concerted environmental action. Between 1980 and 1995, carbon dioxide emissions, total as well as per capita, increased across income groups and regions. Only Sub-Saharan 7 T I E QL A L I T Y o F G R O W T H Figure 1.4. Poverty Incidence in Selected Transition Economies, 1987-88 and 1993-95 Percent (population living on less than US$4 a day) 70 60- 50 - U 1987-88 40 - U 1993-95 30- 20- 10 Poland Baltics Moldova Russia Ukraine Central All Asia transition economies Note: The poverty line of US$4 a day is considerably higher than that used elsewhere. Source: Milanovic (1997). Figure 1.5. Environmental Changes Versus Growth of Income, 1981-98 Change in environmental quality (index) 60 .1 * 50 40 . 30 20 35 * _ * 10 * 0 - I I I -2 0 2 4 6 8 10 12 GDP growth (percent per year) Note: r = -0.27, p < 0.05,n = 56. The data are for 56 developing countries. Controlling for per capita income in 1981 gives a similar pattern and the same value for the correlation coefficient (-0.27). Source: World Bank (2000c); authors' computations. 8 A MIXED DEVELOP'MENTF RECORD Africa, probably because of general economic stagnation, did not experi- ence increased carbon dioxide output. East Asia had the fastest deforesta- tion and highest carbon dioxide emissions per capita, suggesting a conflict between growth and sustainable development (World Bank 2000c). In much of the developing world, the environmental quality is far worse than indicators portray. Air quality worsened as incomes rose.3 Exposure to high levels of air pollution, that is, total suspended particulates, sulfur diox- ide, and nitrogen dioxide, poses a major threat to human health. In Delhi, one of the world's most polluted cities, total suspended particulates were more than four times the level identified as safe by the World Health Orga- nization (WHO) (World Bank 1999d). For particulate levels in selected cities, see figure 1.6. The human cost of environmental deterioration is staggering. Poor water supply, inadequate sanitation, indoor air pollution, urban air pollution, malaria, and agro-industrial chemicals and waste account for an estimated one-fifth of the total burden of disease and premature death in the developing world, based on a standardized measure of health outcomes-disability adjusted life years, or DALYs. For Africa, poor water supply, inadequate sanitation, and indoor air pollution account for 29.5 percent of the disease burden, a share larger than that attributed to malnutrition, 26 percent (Lvovsky and others 1999). Air pollution is alarmingly high in many cities in developing countries Figure 1.6. Total Suspended Particulates, Selected Cities, Early 1990s Paris Tokyo X 90 World Health Organization Tokyo :~~~~~ guideline for air quality standards New York Nairobi Seoul Moscow Rio de Janeiro Manila Bangkok Tehran Jakarta Mexico City Beijing Delhi _~~ ~ ~ ~ ~ ~ ~~~~ I III 0 100 200 300 400 500 Micrograms per cubic meter Note: Most data are for 1995. The figure for New York is for 1990. Source: World Bank (1997i, 2000c). 9 T II F CQU LI A L I I Y O F G R O W T H Income Growth, Inequality, and Volatility Long-term progress in income growth in the world has been very un- even. Figure 1.7 show the trends in per capita incomes in the developing regions and the industrial countries since 1975. East Asia has improved liv- ing standards significantly, while Sub-Saharan Africa has seen the opposite trend. The large variation in growth rates at the level of individual econo- mies may be seen in figure 1.8. Of the 15 fastest growing economies, 8 are in East Asia. Many of those at the other end of the spectrum are countries affected by civil strife and other dislocations. judging by customary growth rates, weighted by incomes of countries, the 1980s were a lost decade for the developing world. The picture looks better when growth rates are weighted by population, because the decreases among middle-income countries, especially in Latin America, weigh less and the in- creases in the larger low-income countries, China and India, weigh more. In the 1990s, the difference between income-weighted and population-weighted aggregate growth rates for developing countries narrowed as growth improved in the middle-income countries in Latin America. Figure 1.7. Purchasing Power Parity-Adjusted GDP Per Capita, 1975-98 Constant 1995 US$ 50,000 Industrial countries 20,000- 10,000 - 5,000 . . Latin America Middle East and North Africa _ ~ _ __ East Asia ~ ~~~Developing countries 2,000 _ ___--- .z. ............ . -.-.-.-_=-___------------- 2,000 .~~~~~~~.. … - - - ~~~~~~~South Asia ..000. -_ --~~ ..... *-- Sub-Saharan Africa I ,000 -___ --_____ 500 I I I I 1975 1980 1985 1990 1995 Note: Europe and Central Asia is excluded for reasons of data availability. Source: World Bank data. 10 A MIXED DEVELOPMENT RECORD Figure 1.8. Growth in GDP Per Capita, Selected Economies, 1975-98 China Taiwan, China Korea, Rep. of Singapore Botswana Thailand Hong Kong, China Mauritius Chile Indonesia Malaysia Egypt, Arab Rep. of Lesotho Sri Lanka India Russian Federation Venezuela, RB Cote d'lvoire Iran, Islamic Rep. of Guinea-Bissau Central African Rep. Haiti Niger Gabon Madagascar Zambia Sierra Leone Nicaragua Georgia Congo, Dem. Rep. of -6 -4 -2 0 2 4 6 8 10 Percent change per year Source: World Bank (2000c); authors' computations. Income Inequality. Within this picture of overall income growth, it is also important to consider how the income was being shared by looking at changes in income inequality. Several dimensions of income inequality are relevant here: between countries, across households within countries, and within households. As noted in World Bank (2000i), the gap between the 11 ! I [: t' 1 \i I T F ( 1 _r ' Q, X) * I I I average income of the richest 20 countries and the average for the poorest 20 has doubled in the past 40 years-to more than 30 times. The data requirements for estimating the distribution of personal incomes in the world are onerous, and the available data have severe weaknesses. That said, Dikhanov and Ward (2000) estimated such distributions for 1988 and 1993 and found that the overall inequality of personal incomes in the world increased from a Gini coefficient of 0.63 to 0.67 (see also Comia 1999). Schultz (1998) looked at the trends in between-country inequality of income. The results differ considerably based on whether China is in- cluded in the analysis. Income inequality between countries increased from 1960 through 1968, stayed high through 1976, and gradually de- clined after that, ending up slightly higher in 1989 than in 1960. If China is excluded, the decline in inequality among countries from 1976 disap- pears. Extension of the analysis through 1994 for a slightly smaller set of countries confirmed these trends. Using comparable data on income Gini coefficients for 45 countries from the early 1960s to the early 1990s, Deininger and Squire (1996) found no general trend in within-country inequality, which stayed approximately the same in 29 countries, rose in 8, and fell in 8. For a different comparison, between the early 1980s and early 1990s inequality increased in 19 coun- tries and decreased in 24 (figure 1.9). Among the countries in which in- equality increased are those with large populations: Brazil, China, and In- dia. Population-weighted, average inequality for the 43 sample countries increased by 0.52 percent a year in the 1980s and the early 1990s.4 Growth Volatility. Economic fluctuations seem to affect the poor dis- proportionately, but the impact is likely to be particularly severe in coun- tries where social safety nets are typically less well developed (Furman and Stiglitz 1998). Declines in economic growth were directly associated with sharp increases in poverty in Eastern Europe, and more recently in East Asia. Economic downturns appear to have enduring adverse effects on the economy. Studies suggest that bigger fluctuations in growth rates are associated with slower average growth.' The volatility of growth would seem to matter. On average, volatility of growth by some measures is estimated to have declined in the 1980s for most country groups (except middle-in- come countries, mainly because of the debt crisis in Latin America), com- pared with the 1970s when the oil shocks occured. The picture is more mixed in the 1990s. Volatility is estimated to have declined for Latin America, the Middle East and North Africa, and South Asia, but in- creased slightly for industrial countries and for East Asia (figure 1.10). I12 A MIXED DEVELOPMENT RECORD No general trend in within-country inequality Figure 1.9. Income Inequality within Countries: 1980s and 1990s Sri Lanka Bangladesh Jamalca Mauritius Korea, Rep. of Mauritania Ghana Egypt, Arab Rep. of Turkey Guyana Tanzania C6te d'lvoire Peru Yugoslavia Tunisia Malaysia Indonesia Colombia Philippines Costa Rica Pakistan Honduras MexicoI Jordan Morocco Nigeria India Barbados Brazil Chile China Panama Dominican Republic Thailand USSR/FSU Guatemala Bulgaria Venezuela, RB Hungary Poland Zambia Romania Uganda _I I I l I -8 -6 -4 -2 0 2 4 Reduction in Gini coefficient (percent per year) FSU Former Soviet Union. Note: The quantity plotted is reduction in income Gini coefficients, in the early 1990s over the early 1980s, in percentage declines per year. Negative values indicate an increase in inequality. Source: Deininger and Squire (1996). 13 T H E QjU A 1 I T Y 0 F G R 0 W T H Figure 1.10. Volatility of GDP Growth Rates by Decade Percent (standard devation of growth rates) 12 10 S * 1960s U 1970s 6 U 1980s 4 El 1990s 2 I 0 Africa East SoLuth Europe Latin Middle East Industrial Asia Asia and America and couLntries Central Asia North Africa Note: Volatility in a decade was computed by taking the standard deviation of growth rates in the decade for each country and the unweighted average across the countries in the group. Sources: World Bank (2000c); authors' computations. Developing Europe and Central Asia had especially more volatile growth than other regions in the 1990s compared with the 1980s. Developing countries seem to have experienced higher volatility than industrial countries. Easterly, Islam, and Stiglitz (1999) explore the deter- minants of the increased volatility of countries, finding that * Openness to trade and the volatility of capital flows are associated with increased volatility of growth * Improvement in indicators of financial development are associated with lower volatility * Constraints on policy from institutional limitations and financial sector underdevelopment contribute to the variability of outcomes * Wage flexibility does not seem to be an important factor. Growth and Welfare Table 1.2 places developing countries in three groups, based on their per capita GDP growth rates: high-growth countries, those with moderate or improving growth rates, and low-growth countries.6 By the definition used here, 13 had rapid growth, 53 had moderate growth, and 39 had low 14 A MIXED DEVELOPMENT RECORD Table 1.2. Development Outcomes by Growth Class, 1980s and 1990s (unweighted means) Moderate or Change in indicator: High improved Low comparing 1 980s and 1 990s Unit Period growth growth growth Poverty Percentage with less 1990s 24.1 31.4 36.9 than US$1 a day 1980s 31.0 32.1 30.2 Infant mortality Per thousand 1990s 29.2 54.3 60.7 1980s 41.0 66.6 71.0 Illiteracy Percent 1990s 17.2 31.2 31.4 1980s 22.9 37.6 38.8 Life expectancy Years 1990s 70.0 62.9 59.8 1980s 66.8 60.6 58.4 Carbon dioxide emission Tons per capita 1990s 2.4 2.3 1.7 1980s 1.5 2.3 1.8 Deforestation Percent per year 1990-95 0.83 1.05 1.11 1980-90 1.08 0.65 1.15 Water pollution Kilograms per day 1990s 0.16 0.21 0.21 per worker 1980s 0.18 0.21 0.21 GDP growth Percent per year 1990s 5.3 4.2 0.3 1980s 6.5 2.3 2.1 Number of countries 13 53 39 Note: See text for details regarding country classification. Some variables are missing for some of the countries. In particular, poverty data are available for only a small number of countries. Sources: World Bank (2000c); authors' computations. growth. Also by this definition, the moderate-growth countries had the strongest improvement in growth. Several of the human development in- dicators generally improved for all three groups, with the high-growth countries showing the strongest improvements. The high-growth coun- tries had higher and increasing carbon dioxide emissions per capita. External Factors Matter In the 1990s, externally- and domestically-driven political and social up- heaval and wars continued to derail progress in numerous countries (Collier 1999; Collier and Hoeffler 1998) (table 1.3). Global and cross-border issues related to financial crises, population pressures, labor migration, and environmental distress continued to affect domestic outcomes. Despite 15 Table 1.3. External Factors Matter for Domestic Outcomes, Examples from 1997-99 Natural Human-caused Financial crises disasters Conflicts disasters Region or country East Asia Bangladesh Albania Indonesia in crisis Russia Central America Bosnia (forest fire) Brazil Congo, Dem. Rep. of Yugoslavia Rwanda Sierra Leone Impact Short-term Loss of human lives Destruction of Long-term increase increase in poverty and physical and human and social in poverty natural capital capital Source: Authors' compilation. progress in slowing population growth, population increases in many coun- tries could undermine efforts to achieve sustainable development. Global warming, environmental degradation, and the loss of biodiversity continue to worsen as an ever more crowded planet puts more pressure on limited glo- bal resources (World Bank 2000b and various editions of the World Bank's Global Economic Prospects and Global Development Finance). The quarter century or so after World War II marked a period of rapid and steady growth for both industrial and developing countries, and the eco- nomic environment wvas relatively free of major shocks. The international economic environment changed dramatically in 1973 with the oil price shock and the end of the Bretton Woods system of fixed exchange rates among major industrial countries. Subsequent decades saw sharp declines in the productivity growth of industrial countries, high inflation and interest rates, and cycles of large amplitude in commodity prices and exchange rates of major currencies. It has been argued that the low growth record of most developing coun- tries (with some exceptions, mostly in East Asia) after 1973 and into the 1990s was due primarily to the growth slowdown in industrial countries (East- erly 1999b). While that was a significant factor, the record of the developing countries that prospered during this period, such as those in East Asia, sug- gests that domestic policy, governance, and institutions also influence out- comes. The damage caused by shocks and conflicts depends on the institu- tions in place and their effectiveness in strengthening governance, civil rights, the rule of law, social programs, and safety nets (Collier 1999; Collier and Hoeffler 1998; Easterly, Islam, and Stiglitz 1999; Rodrik 1998, 1999). A Nl I X ED D EVE l O INA EN I RE CO RD The global economic environment experienced another significant change in the 1990s, becoming more conducive to development in some respects, less so in others (see various editions of the World Bank's Global Economic Prospects). Import demand by Organisation for Economic Co-op- eration and Development (OECD) nations was less volatile in the 1990s than in earlier decades, partly because the cycles of North America, Eu- rope, and Japan were no longer synchronized, and partly because of the in- creased weight of developing countries, especially East Asia, in world trade. Thanks to monetary restraint and progress in fiscal consolidation, real in- terest and inflation rates in the major OECD countries dropped in the 1990s, and volatility in the exchange rates of major currencies was consid- erably smaller relative to the pronounced dollar cycle of the 1980s. Particularly important was the relative steadying in developing country terms of trade with industrial countries, especially in prices of the nonenergy primary commodities. Non-oil exporters saw a serious deteriora- tion in their terms of trade from the mid-1970s to the early 1990s. How- ever, for much of the 1990s, non-oil commodity prices held firm, and the decline since 1997 has been less steep than in earlier price cycles. Although export prices are far more volatile for commodities than for manufactured products, commodity prices were less volatile in the 1990s than in the 1980s for 22 of 30 key commodities (World Bank 2000a; various issues of Commodities Quarterly). The robust growth in world trade sharply surpassed growth in world output through 1998. The international trading environment remained liberal on the whole, with increased multilateralism, notwithstanding the rise of such questionable practices as antidumping. And there was a phe- nomenal increase in private capital flows to developing countries, albeit only to a few. The East Asian financial crisis revealed that while opportunities have grown enormously, so have the demands on institutions and the costs of mistakes. Success in a highly globalized setting requires adequate mecha- nisms for managing risks, and successful policies for openness and competi- tion need effective regulatory and legal frameworks. Domestic Policies Make a Crucial Difference Underlying the varied development outcomes has been the effectiveness of policy, notably in the following four areas: the quality and distribution of education and health services, the stewardship of the environment, the man- agement of opportunities and the risks of globalization, and the effectiveness of govemance. These links are analyzed in the subsequent chapters. 17 T 1i F Q U A i I r Y O F G R O W T 11 Investing in People No country has achieved sustained development without investing sub- stantially and efficiently in the education and health of its people. Devel- oping countries have generally been spending more public resources on education, and many regions have expanded that spending in the 1990s (figure 1.11). Public spending in education has declined in East Asia and the Middle East and North Africa. There is evidence that in East Asia, the share of private expenditure is on the rise. Cross-country data on health expenditures are available only for the 1990s, so long-term trends are not known. What happens to the share of social expenditures in countries undergo- ing adjustment and fiscal austerity? Analyses have been divided on the question. The World Bank (1992) concluded that the shares remain un- changed, while Corbo, Fischer, and Webb (1992) found that education shares declined. A recent International Monetary Fund (IMF 1998) study of low-income countries undergoing adjustment found that the shares of education and health expenditures have generally been protected. Private spending is also important in social services funding, especially in East Asia, where its share rose with economic growth. But public spending does not always produce good outcomes. Those depend on the distribution and quality of the public spending and on the incentives for more private spending. These issues are explored in chapter 3. Figure 1.11. Public Expenditures on Education by Region, Selected Years Percentage of GNP (medians) 6 5 | 1965 * 1970 4 * 1980 3 E7 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~E1990 2 EJ 1996 0 Sub- East Asia South Europe Latin Middle Developing Industrial Saharan and the Asia and America East and countries countries Africa Pacific Central and the North Asia Caribbean Africa Source: World Bank (2000c). 18 A M I X ED D E V E LO P M1 E N 1- RECO RD Managing the Environment We know that government policies have neglected the environment, but we have no standard measures for evaluating a country's environmental policies. One recently developed indicator, genuine saving, measures the rate of saving after accounting for investments in human capital, depreciation of produced assets, and depletion and degradation of the environment (World Bank 1999f, pp. 175-77). Such measures are still experimental and reflect both policies and outcomes. We have looked at progress in agreements reached on environmental issues. From this, however, we get only a sense of the commitment of gov- ernments from having completed a country environmental profile, formu- lating conservation and biodiversity strategies, and participating in global treaties. These measures appear to be only weakly related to environmental outcomes. We need better ways to capture country policies for environ- mentally sustainable development. Creating Market-Friendly Policies Openness and Liberalization. Openness increased in developing countries in the 1990s. The ratio of trade to GDP rose in all developing regions. Lev- els of trade protection have declined in most regions, aided by successive rounds of multilateral trade negotiations. Average tariffs came down in the 1990s, sharply in many cases (figure 1.12). Nontariff barriers have also been reduced significantly in most regions, with the exception of Sub-Saharan Africa (Rodrik 1999; UNCTAD 1994). Openness to capital also increased dramatically in some regions. An in- dex of financial controls shows a sharp decline in the 1990s, following a sharp increase in the previous decade (chapter 5). Liberalization has also taken hold in domestic markets, as governments have become more willing to rely on markets and to increase incentives for private initiative by priva- tizing industries and have lifted other restraints on marketing and distribu- tion. Many primary commodity exporting countries in Africa are liberaliz- ing control boards, allowing for more pass-throughs of international commodity prices to producers (Akiyama 1995). Macroeconomic Stability. Two often used reflectors of economic man- agement are parallel-market exchange rate premiums and government deficits. Figure 1.13 shows that parallel-market premiums declined sharply in the 1990s in most countries. After sharp increases in the 1980s, govern- ment deficits declined in most regions, except Europe and Central Asia. Partly as a result, inflation was down in most developing countries. 19 T H E QU A L I T Y O F G R O W T H Trade barriers decline in most regions Figure 1.12. Trade Barriers, Selected Regions, 1984-93 Weighted Average Tariffs Incidence of Nontariff Measures Percent Percent 35 - 50 - 30- 40- 25- 20 -30- 1 5 20- 10 5 10 0 07 East Asia Latin Sub-Saharan East Asia Latin Sub-Saharan America Africa America Africa * 1984-87 * 1988-90 * 1991-93 * 1984-87 * 1988-90 * 1991-93 Source: Rodrik (1999). Growth Outcomes and Policy Performances. Table 1.4 shows policy profiles for three growth classes and for the 1980s and 1990s. They do not claim to establish the direction of the link between policies and outcomes, but the patterns and trends are noteworthy. A great deal of previous work has shown the impact of policies on growth (a literature survey summary is available on request.) Although some market-friendly policies remain con- tentious, many developing countries made significant efforts to adopt them in the 1990s. Average budget deficits were lower for all groups in the 1990s-sharply so for the high- and moderate-growth groups. All three groups had significantly lower tariffs and a higher trade-to-GDP ratio in the 1990s than in the 1980s. All three groups were more open for capital ac- count transactions in the 1990s-the high-growth group more cautiously so. Domestic financial systems were also generally less repressed in the 1990s, relative to the 1970s-again the high-growth group more cautiously so, relative to the moderate-growth group. The fast growers had greater fi- nancial depth, measured by the M2-to-GDP ratio, and more prudent mac- roeconomic policies, partly evident in increased reserves. And they had higher ratings on governance measures. There is an immense empirical lit- erature on some of these policy-outcome links. 20 A MIXED DEVELOPMENT RECORD Figure 1.13. Parallel-Market Premium, 1970s-1990s Medians (percent) 50 40 - * 1970s 30 - * 1980s L 1990s 20- 10- Sub- East Asia S-ith E.rope Latsin Middle Des-1pmng Saharan and the Asia and America East and countries Africa Pacific Central and the North Asia Caribbean Africa Note: The values plotted are (parallel market rate/official rate - I) as a percentage, for a unit of foreign currency in terms of local currency units.. Sources: Easterly and Yu (2000); World Bank (2000c). Crucial Issues for Action The developing world continued to move forward in the 1990s. Progress on policy was substantial: reducing fiscal deficits, investing more in edu- cation, lowering trade and investment barriers, and dismantling domestic price controls in agriculture and industry. The record was more mixed on development outcomes. But both the 1990s and the longer-term record confirm that these actions go with improved economic growth. They also corroborate the link between economic growth and poverty reduction. Thus, on the whole, the developing world recovered from the setbacks of the 1980s, but both the depth and the breadth of the recovery left much to be desired. The record also suggests that actions, by government and others, to af- fect the quality and sustainability of growth have lagged. Events in East Asia, Europe, Central Asia, and elsewhere underscore the fragility of advances in reducing poverty and achieving sustainable development. The numbers of poor people continue to increase, and today an estimated 1.2 billion live in absolute poverty, on less than US$1 a day. The incidence of poverty is highly sensitive to changes in income distribution and to popula- tion growth. So, policies affecting broad and equitable growth and popula- tion growth deserve considerable attention. 21 T H 1 1: ( U A t i r Y 0 F G R O \W i Table 1.4. Policy Performance, by Growth Class, 1980s and 1990s (unweighted means) Moderate or Change in indicator: High improved Low comparing 1 980s and 1 990s Unit Period growth growth growth Budget surplus Percentage of GDP 1990s -1.8 -1.4 -3.4 1980s -4.2 -2.9 -4.7 Effective tariff rate Percent 1990s 22.7 25.4 18.3 1980s 29.1 31.9 22.7 Trade/GDP Percent 1990s 92.1 77.0 70.2 1980s 82.0 71.0 59.9 Capital account openness Index 1996 2.4 3.0 3.1 1988 1.7 1.9 1.7 Financial repression Index 1996 3.6 3.2 4.0 1973 5.9 6.8 4.5 M2/GDP Percent 1990s 55.4 36.9 28.6 1980s 42.8 34.6 28.4 International reserves Months of imports 1990s 4.2 3.9 2.9 1980s 3.1 2.8 2.4 Rule of law Index 1997-98 0.2 -0.2 -0.7 Control of corruption Index 1997-98 -0.1 -0.2 -0.6 Education spending Percentage of GNP 1990s 3.7 4.4 4.3 1980s 3.6 4.2 4.4 Environmental action 0-1 Index International 0.89 0.95 0.88 0-1 Index Domestic 0.89 0.86 0.65 Number of countries 13 53 39 Note: See text for details regarding country classification. Some variables are missing for some of rhe countries. In particular, rhe following variables are available for only a small number of countries: effective tariff rate and financial repression. Variables are described in annex 1. Sources: World Bank (2000c); authors' computations. The relationships between development goals and policy instruments have been investigated in considerable detail in the development litera- ture. The annex to chapter 1 includes a set of correlation coefficients for goals and policies in the spirit of providing the basic data. As correla- tions, they say nothing about the direction of the causality or the mecha- nisms. Nevertheless, the combinations shown to be significant are worthy of further investigation as hypotheses. Equally important are the plausible combinations that are not significant with the expected sign. Many of the 22 A MIX ED D F V E LC) P N1 ENT RE CO RD hypothesized relations are taken up in chapters 3 through 6, and in chap- ter 2, which elaborates a basic framework. Here, we frame the relationships in the form of questions that motivate this study: * Are the observed improvements in human capital sufficient to sus- tain growth in countries that experienced accelerated growth in the 1990s? * Is the increased human capital in the slow-growth countries suffi- cient to propel faster and better growth in the near future? * Will the deterioration of natural capital reduce the potential for fu- ture growth, especially among poor countries? * Is the degradation of natural capital becoming a serious obstacle to improving the welfare of the population? * Can the risks of financial globalization be managed in a way that di- minishes the volatility of growth and improves its sustainability? * How serious is the manner of governing for growth processes and outcomes, and how can progress be made in controlling corruption? The chapters ahead provide insights into these questions, if not answers. The rest of this publication is organized as follows. Chapter 2 provides an analytical framework for interpreting the development experience out- lined in this chapter and derives lessons on the importance of undistorted growth of human, natural, and physical assets and the welfare significance of alternative growth patterns (see also the box in the Overview). Chapter 3 explores how investments in people-in quantity, quality, and distribu- tion-can augment welfare directly and also make growth processes more sustainable. Chapter 4 does the same for environmental and natural re- sources, where the conflict between growth and welfare is apparent and the tradeoff that much tougher. Chapter 5 revisits the issue of growth volatility and financial risks and considers the quality aspects of reforms that would render growth processes more sustainable in today's globalized setting. All three-human, natural, and financial sides-hinge on the quality of overall governance, which is fundamental to the quality and sustainability of the growth processes. Formal and informal institutions of governance are dis- cussed in chapter 6, with a special focus on fighting corruption. Chapter 7 considers an agenda for action. Why is there a failure to take up policies that have shown themselves sound? Policymakers' lack of understanding of these policies is unlikely to be the main reason. More likely is the political difficulty of implementing sound policies. Interest groups limit the range of feasible reforms and open a gulf between policy design and implementation. How best to counteract 23 T I F I O l A I I T Y O F G RPO W T FH these forces through enhanced participation and stronger government own- ership is an important topic not fully explored in the report (except for a brief discussion of selected aspects in chapter 6). Another crucial topic is the special circumstances facing the formerly centrally planned economies as they strive to make the transition to mar- ket economies. These economies are included in the analyses when data permit, and illustrations and cases drawing on their experience are used in some of the chapters, especially the one on governance and anticorrup- tion (chapter 6). However, a full discussion of the issues for transition economies is beyond the scope of this volume.7 Notes 1. Many studies have addressed the multidimensional view of development objec- tives (Dasgupta 1993; Hicks and Streeten 1979; Lewis 1955; Nordhaus and Tobin 1972; Sengupta and Fox 1969; Tinbergen and Theil in Hughes-Hallet 1989). Some studies used multivariate analysis of a large number of economic, social, and political variables (Adelman and Taft-Morris 1967; Baster 1972; Morris 1979; UNRISD 1970). Some constructed indexes of quality of life or hu- man development (Dasgupta 1990a; Diewert 1986; Drewnowski and Scott 1966; Griliches 1971; McGranahan 1972; Ram 1982a,b; Slottie 1991). The frameworks in some of these works are touched upon in chapter 2. 2. Carbon dioxide emissions do not have much local health impact directly, but are important in the context of greenhouse emissions and the associated prob- lem of global climate change. Also, carbon dioxide emissions are usually associ- ated with emission of other air pollutants that do have a local health impact, but about which data are less widely available. 3. Measures of air quality for many cities became available only in the mid-1990s. An inverted U-shaped relationship between pollution and per capita income, the environmental Kuznets curve, has been estimated for several types of pollut- ants (Grossman and Krueger 1995). This does not, however, invalidate the need for environmental intervention, because the per capita income turning points for improved environmental indicators are generally quite high. This is- sue is discussed in chapter 4. 4. Data on income distribution are scanty for the more recent years. A World Bank study of 29 countries estimated that 5 countries experienced a decline in inequality, while about 5 times as many countries (24) saw an increase (Buckley 1999). 5. For example, a regression of average rates of GDP growth for 112 countries against the volatility of GDP growth rates, measured by the standard deviation of growth rates, yielded a significant negative coefficient (see also Ramey and Ramey 1995). 24 A MIXED DEVELOPMENT RECORD 6. High-growth countries as defined here are those with per capita income growth of more than 2.3 percent per year in the 1980s and the 1990s, a rate that doubles incomes in 30 years. The second group-moderate or improved growth-includes countries that maintained a positive growth of per capita in- come in both decades, or improved growth in the 1990s by at least 2 percentage points. The rest are classified as low-growth countries. 7. See World Bank (1996b) for a full treatment of transition issues. See also recent papers by Aslund (1999); Commander, Dutz, and Stem (1999); Kornai (2000); Qian (1999); Stiglitz (1999); and Wyplosz (1999). 25 C H A P T E R 2 ASSETS, GROWTH, AND WE LFARE The difficulty lies, not in the new ideas, but in escaping from the old ones which ramify. . . into every corner of our minds. -John Maynard Keynes, The General Theory of Employment, Interest, and Money R_ apid economic growth has usually been considered the prime indicator of development. Yet, there has been dissatisfaction with using growth measured by national accounts as the yardstick (see, for example, Adelman 1975; Dasgupta 1993; Dreze and Sen 1995; Lewis 1955; Sen 1988). More meaningful is welfare, comprising consumption, human development, and environmental sustainability, and their quality, distribution, and stability. Often per capita income growth and welfare improvements go hand in hand. But sometimes they do not. Large divergences between growth and welfare improvements may arise when growth is volatile and unsustained. Can such divergences between growth and welfare change still arise when economic growth is sustained? That is, can countries maintain rapid growth for prolonged periods without commensurate increases in welfare? If not, the focus should be on policies that assure sustained growth-because those policies would also generally improve welfare. But if yes, the focus on growth has to be complemented with an examination of alternative patterns of (sustained) growth. The analysis here focuses on investment patterns in three key assets: physical and the closely associated financial capital,' human and the closely associated social capital, and natural and the closely associated environmen- tal capital. Technology affecting the use of these assets matters a great deal too. The central hypothesis, which is assessed empirically subsequently, is 27 T H1 E 0) A I I f (F C R O W r II that promoting adequate investment in all forms of capital is a way to in- duce more and better growth and improvements in welfare. But policies of- ten introduce distortions that encourage either over- or underinvestments in different forms of capital. Examples of these distortions are artificially low interest rates, underpricing of natural resources, or underemphasizing basic education in public policy. Focusing mostly on physical capital accu- mulation to the relative neglect of human and natural capital is no guaran- tee to sustain growth. Some recent evidence shows little correlation be- tween investment rates and growth rates in the short term (Easterly 1999c). Special efforts to encourage physical capital accumulation per se are likely to impose large costs. Some policy changes in the 1980s and the early 1990s would seem to have especially raised the rate of return to physical capital, reflected by investment booms in many countries. But these reforms by themselves have not automatically assured sustained growth, to the extent that there have not been complementary investments in human and natural assets. Moreover, some countries have not generated growth-partly because of wrong regulation (for example, licensing that reduces investment incen- tives), and insufficient regulation (for example, for financial markets and to deal with monopolies). Altematively, growth induced by relatively undistorted or a balanced ex- pansion of human, physical, and natural capital can be sustained for pro- longed periods.2 Balanced does not signify an equal expansion in the assets. Rather, it refers to assest accumulation in response to an undistorted policy framework. Such a pattern is more likely to reduce poverty and improve in- come distribution. That, in tum, creates the conditions for faster growth that improves welfare more rapidly. So, preventing underinvestment in human and natural capital is one way of promoting rapid and sustained growth. We begin with a framework that allows us to explore these hypotheses and their implications: pattems of asset accumulation, factor productivity, and social welfare. In particular, we look at the implications of distorted asset growth for the poor. The next section provides empirical evidence from a va- riety of sources. In addition to a historical review of 60 countries, we provide econometric evidence from two groups of countries on the determinants of growth. Finally, we tum to the empirical evidence on a variety of (gross) sub- sidies, followed by an evaluation of the impacts of capital subsidies. A Framework Improving the quality of national accounts by including human and natural capital at shadow prices (notwithstanding the complexities in computing 28 ,N > 9, 1 I s . , R, () \N I I1 IA N I) \ I I I A It 1. them) is one way of reconciling the divergence between growth and welfare improvements. But even the limited progress in valuing these assets has not yet been incorporated into national accounts, and there still are serious con- ceptual problems with incorporating (and weighting) them. For these rea- sons, a more practical (and more modest) approach is to identify measurable growth pattems and policy instruments likely to promote greater welfare. Three Patterns of Growth Consider these altematives: * Pattern 1. Unsustained growth, where the economy grows with some phases of fast growth, but at a declining rate, eventually leading to stagnation or near stagnation. * Pattern 2. Distorted growth bought at the expense of deteriorating natural resources, for example, from their underpricing; lagging in- vestments in human capital, for example, inadequate safeguards re- garding child labor; and subsidies to physical capital, such as tax ex- emptions, allowing tax arrears, giving financial grants to reward certain investments, and providing investment credit subsidies. * Pattern 3. Sustained growth through undistorted or balanced asset accumulation, with public support to developing primary and sec- ondary education, improving public health, and protecting natural capital. This prevents a decline in returns to private assets (espe- cially physical capital) and provides the minimum and increasing levels of human capital needed to facilitate technological innova- tion and the growth of total factor productivity (TFP). Pattem 1 usually is associated with slow and highly unstable or volatile growth. Slow and unstable growth prevents poverty reduction and leads to inadequate resources for investing in human capital and natural capital. That is, pattern 1 causes economic stagnation and welfare losses. Pattem 1 usually occurs in a context of poor govemance and corruption that brings about low investment and inefficient allocation of public expenditures. Compared to pattem 1, the stop-and-go growth of pattern 2 is better for welfare improvement and poverty reduction. But pattern 2 growth might depend on public support to physical capital, which is difficult to sustain. Pattern 3 is better for improving welfare and for reducing poverty. To sus- tain a reasonable rate of economic growth, therefore, the principal assets of the economy-physical and financial, human and social, natural and envi- ronmental-need to grow at undistorted or fairly balanced rates. The distri- bution of assets among the population, especially of human capital, is also 2c) i I E QOU A L I T Y 0 F G R 0 WI THF important. Stable, sustained growth is highly beneficial for the poor, who usually suffer the most in the reversals of stop-and-go growth. Externalities and Asset Accumulation All forms of capital may involve externalities. Components of human capital and natural capital often have a social value that goes beyond that accruing to the individuals using it. As (partly) public goods, they have positive spillovers that are not necessarily fully accounted for by the ac- tions of individuals or firms. That is why public policy and other mecha- nisms must prevent underinvesting in them. There has been some empha- sis on the positive production and technological externalities associated with physical capital accumulation (Barro and Sala-l-Martin 1995; Romer 1986). But the externalities associated with human and natural capital are much harder to account for, and are probably larger.3 Human and natural capital are important not only as factors of production, but as direct determinants of societal welfare. Governments can use market instruments to deal with these external effects. But the issue also involves the allocation of public expenditures. Government spending typically accounts for 25-30 percent of GDP, ex- erting a powerful direct effect (as opposed to the effect of policies and regulations) on resource allocation and income distribution. Few coun- tries have used market instruments successfully to account for the true so- cial value of natural and human capital. Governments responsible for the Amazon region, for example, have exacerbated the negative environmen- tal externalities. Public subsidies and tax incentives to large cattle pro- ducers and loggers were responsible for more than 50 percent of the defor- estation in the Amazon region in the 1970s and the 1980s (Binswanger 1991). Moreover, public investments in infrastructure into the frontier areas have magnified the externalities associated with the lack of well-de- fined property rights in such areas. Doing little to prevent underinvestment in human and natural assets is likely to lead to unbalanced asset accumulation, at least in the short term, by focusing on physical capital accumulation. Relying mostly on physical capital accumulation instead of balanced asset growth may increase the growth of GDP (using conventional national account methods). But wel- fare might not improve as fast-and it could even decline, if, for example, natural capital were to decline dramatically, or if the quality of public edu- cation and health care fell. The distributional consequences of distorted or unbalanced asset growth could also be severe, especially if the imbalance makes growth unstable, hurting the poor disproportionately. 30 ASSETS, GROWTH. AND WELFARE Fast GDP growth without some degree of balanced asset augmenta- tion may also be difficult to sustain. Unless there are very high techno- logical spillovers or scale economies, rapid physical capital accumulation with slow growth in human capital and a depletion of natural assets would lead to declining marginal productivity of capital-as capital stocks increase relative to other productive assets (see annex 2). Growth in Total Factor Productivity and Asset Accumulation So far most of the emphasis in this chapter has been on asset accumulation and asset structure as a source of growth. An important set of analyses ar- gues that the main source of growth is not asset accumulation-it is the growth of TFP (Easterly and Levine 2000; King and Rebelo 1993; Klenow and Rodriguez-Clare 1997a; Romer 1986, 1993). This conclusion, elabo- rated from theoretical models based on endogenous growth, is supported by earlier empirical studies showing that growth over time, especially in the United States and some other industrial countries, is indeed heavily explained by TFP. Analyses of East Asian countries, however, suggest that TFP growth may not be as important a source of growth for developing countries as it has been for the United States and some other industrial countries. East Asian coun- tries are practically the only developing countries that have experienced per- sistent, fast growth over long periods. Collins and Bosworth (1996), Kim and Lau (1994), Krugman (1996), and Young (1991, 1994, 1995) show that East Asia's rapid growth (before 1997) was based on strong asset accumulation. Two recent papers, however, point to factors that qualify these analyses. Klenow and Rodrfguez-Clare (1997b) and Nelson and Pack (1998) empha- size improvement in asset measurement and methodological refinements that could significantly alter the conclusions reached by the above authors. TFP in developing countries is potentially important for growth. It is also closely linked to asset accumulation for two reasons. First, a main vehicle of new technology is embodied in imported capital and new in- termediate goods. Second, to benefit from technological progress, the level of education needs to be continually increasing both in depth and breadth. Expanding general education is more crucial in the developing countries than in the industrial countries, where it is already broad- based. But in most developing countries, general education is still insuf- ficient to facilitate technological diffusion. So, TFP growth can be fast only if human capital rapidly broadens and deepens. That is why it is closely linked to asset accumulation and why it may be difficult to dis- entangle TFP and asset growth as sources of growth. 31 T H E QU A L I 1 Y O F G R O W T H Investments in Physical Capital Market reforms-trade and capital market liberalization, privatization, elimination of price controls, liberalization of labor and other markets- have been vital instruments in increasing the rewards to all forms of capi- tal. Given the greater responsiveness of private investments in physical relative to human and natural capital, they have helped some countries (especially those countries not severely affected by corruption) to enjoy an investment boom, accelerating growth. Several of these reforms, for example, trade liberalization or removing anti-agricultural biases, also raise the rewards to human capital. However, in the absence of comple- mentary investments in these assets (especially human capital), the ex- pansion of physical capital could bring about a declining return and even- tually a deceleration of growth (see annex 2). For some countries this tendency has been countered by deepening the reform process. Others have used increasing public resources to sustain distortions (thereby gen- erating pattern 2 growth). Moreover, as developing countries participate more in global markets, national (and subnational) governments can engage in competition to at- tract capital by artificially creating favorable conditions, as seen in recent evidence on subsidies to attract foreign investments in industrial and devel- oping countries. (For a review of countries such as Argentina, Brazil, Canada, China, India, Malaysia, Singapore, and the United States, as well as Western Europe, see Oman 2000; also the next section.) There is a vari- ety of evidence of incentives and subsidies relating to investments in indus- tries such as the automobile industry from various regions or in the overexploitation through the underpricing of natural resources as in mining or forestry. A mechanism to increase the attractiveness of domestic and for- eign investments is to "give away" human and natural resources at low costs, for example, by allowing child labor; not enforcing health and sanita- tion regulations in the workplace; not regulating banks and other financial institutions; not enforcing environmental regulations; and giving away mining, water, and logging rights.4 In some countries these capital subsidies and tax exemptions may offset the firms' costs associated with misgovernance and corruption that reduce their incentives to invest in productive activities (see chap- ter 6). This suggests that by reducing corruption and misgovernance, it may be possible for countries to save resources. In addition to gover- nance, another ingredient that can play a positive role in enhancing quality growth is the strength of informal institutions in a country, of- ten referred to as social capital (box 2.1). 32 A S S E -I T. t, R (V ID H A N D \ F L F A R L Box 2.1. Social Capital The notion of social capital has received a lot of at- formation? Evidence involving both aggregate tention from scholars and development professionals cross-country and within-county microdata is accu- of late. The phenomena grouped under the rubric of mulating to suggest the potential of social capital. social capital have included trust, cooperative norms, Knack and Keefer (1997) use data from the World voting, participation in referenda, and horizontal as- Values Survey for 29 market economies over 1980- sociational activities in diverse groups. 94 to probe the importance of trust and civic in- How does social capital affect economic volvement. After controlling for initial per capita performance? income, human capital, and capital goods prices, * Fewer resources have to be spent to protect they found that both social capital indexes show against fraud in economic transactions, which significant links to economic growth. They also woul almost be a corollary ofhigh-trust envi. found that trust is even more important for poorer would almost be a corollary of high-trust env- countries with weak legal systems and financial sec- ronments. * There is less need for entrepreneurs to moni- tors. A policy implication: establishing formal legal and credit institutions is especially important in tor suppliers and workers, freeing more re- low-trust societies- sources for innovative activity. The concept of social capital has generated dis- . Interpersonal trust can substitute for formal cussions and debates. Its proponents claim it to be property rights, as important as-or encompassing-physical, hu, i The greater confidence in goverment policy man, and natural capital. Others see this focus as is good for investment, excessive and inappropriate. Some of the work in * A higher degree of trust seems important for the area is also criticized for leaving out important human capital accumulation. Galor and Zeira social isions. templeand ohnsortant (1993) suggest that higher trust is associated social dimensions. Temple and Johnsori (1998) with higher enrollment in secondary education. sgtaers pectiveansimpl th at socie ty . Trust and civic participation are also associated vmatters. They anape the data on socioeconomic with bete pefrac of governmentinsti- variables compiled by Adelman and Taft-Morris witiobs, includin those of public , ed iatio. (1967) and show that several social variables have * Communiry or cooperarive action by local significant explanatory power for predicting long- ,, Community or cooperative actn bterm economic growth. These variables go beyond groups can alleviate "the tragedy of the com- the "trust variables" typically studied by research- tenance (Ostrom 1990). ers in social capital. Among these variables, the ones inportant in capturing differences in social * Greater links among individuals facilitate bet- important inccaptui e renc in social te ifrmto flow an fatr dison of in-. arrangements include the extent of mass commu- terinformation flowsl and faster diffun Fosr in- nication (newspapers and radios), the character of novation (Besley and Case 1994; Foster and basic social organization, the modernization of Rosenzweig 1995; Rogers 1983). outlook, the extent of social mobility, and the im- * Social capital may act as informal insurance, portance of the indigenous middle class. much the same as the diversification of a port- Some key readings are Dasgupta and Serageldin folio. Risk sharing by many households can act (1999), Narayan and Pritchert (1999), and as a social safety net and enable them to un- Woolcock (1998). See also two sets of articles that dertake higher-risk and higher-return activities (Narayan and Pritchett 1999), appeared in special sections of Worl1d Development (Evans 1996) and the Joural of Internaonal Devel- But can social capital be measured, and what is opment (Harriss 1997). Included in the latter is an its effectiveness in contributing to growth? And are article critical of the World Bank's use of the no- there policy interventions that can contribute to its tion of social capital (Fox 1997). 33 T- E (tU A L I T Y O F G R,O V T 11 Investments in Human and Natural Capital The other side of the coin of special incentives to physical and financial capital is the insufficient attention paid to human capital and the rapid de- struction of various forms of natural capital through overexploitation. Ef- forts to raise artificially the incentives to investment in physical and finan- cial capital could be linked to insufficient investment in human and natural capital. The private sector contributes to human capital accumulation-through training, private schools, and private health care. But private schooling and private health care go mostly to the better-off, who can afford to pay for their human capital up-front. Most people, particularly low- and middle-income people, depend on public support to accumulate human capital. Imperfec- tions in capital markets prevent them from borrowing against future earnings, making this dependence even more marked. Growth in physical capital may spill over into human capital through private investment in research and development and training in higher technologies-that is, in knowledge-driven growth. But to sustain this growth, a large (and growing) part of the work force must have enough general schooling to acquire skills and technology and participate in the expansion of research and development activities. So, publicly provided general schooling and privately generated knowledge are complementary. If the quality and coverage of general schooling do not increase fast enough, knowledge-driven growth may be stifled, particularly in poorer countries where most of the labor force does not have primary school edu- cation (chapter 3). Growth without complementary environmental policies may damage the environment as the accumulation of physical capital accelerates. This is especially likely in countries with comparative advantages in natural re- source-intensive industries that also require a lot of physical capital for their exploitation, such as mining, forestry, and fisheries. Preventing exces- sive environmental and natural resource degradation also depends on pub- lic policies and investments. Many environmental resources have social values-as inputs in production and as consumption-that are generally well above those that the private sector considers in its resource ailoca- tions. When natural resources are plentiful, degrading natural capital is not likely to have much effect on the productivity of physical capital. But after natural resources fall below certain thresholds, further degradation could re- duce the productivity of physical capital (chapter 4). While degrading natural capital is likely to reduce welfare, its impact on economic growth is subject to debate (see the exchange between Daly, 34 ASSETS, GROWTH, AN D WELFARE Solow, and Stiglitz in Daly 1997). That impact hinges on the substitution of other assets for natural capital (see annex 2). Some recent evidence im- plies that human capital, but not physical capital, can substitute for natural capital. So, economies that expand human capital can reduce the depen- dence of output growth on natural capital. The high levels of human capi- tal permit the economy to diversify into activities progressively less inten- sive in natural capital. For example, a country with a high level of human capital can specialize in knowledge-intensive activities, making the exploi- tation of natural capital less essential for sustaining income growth. But degrading natural capital is likely to be devastating for the poor, who generally have little human capital and continue to depend on natural capital (soils, natural water sources, fisheries) for their incomes, even in middle-income economies. Because the poor have few possibilities for sub- stituting other assets for natural resources, the degradation of those re- sources could lead to irreversible vicious circles of poverty and environmen- tal destruction (see Lopez 1997 for an analysis of the dynamic features of natural resource degradation and institutional change for the rural poor).5 Distorted Asset Growth and the Poor The poor, due to their lack of assets, would have more difficulties than the rich in smoothing their consumption in bad times. Close to the limits of subsistence, they usually work in activities hit most by economic cycles (ag- riculture, construction). So, unstable growth can have harsh effects for them, and an economic crisis can so degrade their human and natural assets that they may not be able to benefit from subsequent booms (see annex 2). The economy of the poor is often separate in many ways from the modern economy, but the demand for their products depends at least in part on the modern economy (exchange rates, for example, affect the prices of their export products). Instability in the modern economy thus affects the incentives for the poor, and a deterioration of these incentives hurts the poor. Even if incentives return to original levels, the poor may not be able to take advantage of them. This implies two possible alterna- tive equilibria: a sustained growth equilibrium and a stagnant subsistence equilibrium. During bust times, the poor lose the assets needed to main- tain consumption at subsistence levels and to respond to stronger incen- tives in the next boom. Some developing countries, for example, in Latin America, have a rela- tively high income inequality especially because of the skewed distribution of physical capital, education, and land. Expanding education could change that. Making education less concentrated through, say, reallocating public 35 T IL l0 U A 1 Y C) F GRG R W T H spending toward basic and secondary schooling, is probably the least con- troversial asset redistribution, and likely the most feasible. Asset inequality affects social welfare through two mechanisms. One is a direct effect: large segments of the population have few assets and consume little, while a minority has large amounts of assets and consumes a lot (see annex 2). The other is indirect: asset inequality has been shown to reduce the potential for economic growth and poverty reduction through a variety of channels (see, for example, Alesina and Rodrik 1994; Deininger and Squire 1998; Persson and Tabellini 1994; Ravallion and Sen 1994 on asset inequal- ity and growth and chapter 3 and annex table A3.5 for a literature review). Even small changes in income distribution can have large effects on the extent and depth of poverty in developing countries (Lundberg and Squire 1999). Several studies have tried to establish a relationship between in- come distribution and growth. However, as Lundberg and Squire argue, growth and inequality should be analyzed as joint endogenous variables. How asset inequality affects both growth and income distribution is closely related to the way that the level and composition of public spending on education and health affect the inequality of human capital. Skewed distributions of education are unlikely to produce the best growth outcomes.6 If human capital is relatively concentrated, any further concentration would slow growth, while efforts to improve its distribution would benefit growth (chapter 3). An economy with a small number of highly educated people and a large number that is illiterate may find it dif- ficult to sustain high rates of return to physical capital, because the poten- tial technological spillovers associated with capital accumulation may not materialize. Greater access to secondary and higher education would allow for more technological spillovers. Empirical Evidence In this section we provide four types of evidence as follows: * Experience in 60 developing countries. Growth experience has often fol- lowed patterns 1 and 2, relying mostly on a rise in investment in physical capital, while investment in human capital has lagged and investment in natural capital has been mostly negative (see box 2.2). * Econometric evidence. Growth based mostly on physical capital ex- pansion is unlikely to be sustainable. The possible positive spillovers of physical capital investment do not seem to be sufficient to main- tain a stable rate of growth in the absence of significant expansion in human capital and a sustainable use of natural capital. 3( A S f IS. G ROW XIH, AND WE LF AR E Box 2.2. Alternative Approaches to Sustain Growth: Brazil, Chile, and the Republic of Korea Two approaches to pursue sustained growth might terms. Also, the relatively small support to the so- be noted: cial sectors would seem to have contributed to so- * Approach 1. Increasingly large policy and ex- cial inequity. penditure distortions (incentives and subsi- Attention to Human Capital dies) in favor of capital (pattern 2 growth). * Approach 2. High levels of support to the Korea also seems to have subsidized investors growth of other assets as well, particularly hu- starting before the 1990s. Its subsidies were selec- man capital (pattern 3 growth). tive, focusing mostly on a few industries at a Approach 1 implies that maintaining a high time-aiming at developing a few industries into growth rate requires that the pro-capital bias has exporters within a reasonable time. Some favored to be increasing over time. Apart from being less industries have become leaders in causing growth effective than approach 2 in sustaining long-run spillovers to others. While this approach was growth, this approach means unstable growth in problematic in many ways, it implied relatively the short run and increasing concentrations of in- less of an explicit financial burden on the public come and wealth. The second approach is more sector. In addition, the allocation of public re- likely to sustain a reasonable growth rate in the sources to education has prioritized basic educa- long run, reducing short-run instability and pro- tion. This has allowed the pttblic sector to support moting equity. a fast buildup of human capital, along with a rapid decline of the education Gini coefficient Favoring Physical Capital (chapter 3). This has also balanced the incentives Most countries use a combination of the two ap- for the growth of physical and human assets, has proaches with different emphases. Brazil, like several permitted income inequality to remain at accept- others, at times seems to have used approach 1. Re, able levels, and has helped poverty to decline. views of several countries show examples of public There was sustained economic growth during the allocations to support the profitability of capital 1980s and the 1990s through 1997. Growth was through direct financial subsidies to domestic and relatively stable-possibly in part because the pub- foreign investors; efforts to build infrastructure and lic sector maintained its support to both human and services with public monies oriented to expand par- physcal capitat over the years ticular industries and develop environmentally sensi- Relative Neutralty tive areas; as well as credit, tax, and price policies in favor of capital. In many countries, the allocation of Since the early 1980s, Chile's public sector has gen- public resources to education has emphasized subsi- erally abstained from directly favoring physical dies to tertiary education and underinvested in pri- capital. Nor have the social sectors, particularly mary and secondary schools. education and health, received special support, ex- Over the past two decades, the standard devia- cept for the period 1997-2000. The public sector tion of annual growth rates has been more than has not taken on any significant role in especially the average growth rate (table 2.1). Such instabil- orienting growth strategy in these areas. However, ity could be due in part to the varying capacity of Chile has low taxation on using its natural re- the public sector to generate resources needed to sources, providing strong incentives for foreign in- continue to support physical capital in relative vestors to exploit mining, forestry, and fisheries. (box corninues on following page) 37 I 11 E o ! A !, I 1- h 0 r G, R\ ( W I I Box 2.2 continued There was a boom in 1987-95, which benefit- hind. The lack of dependence of capital on direct ted from a large acceleration of investment in public subsidies may have led to stable growth physical capital, with hufiman capital lagging be- rates in an eightyear expansion. Table2.1. SolectedV iabl" s OBrazil Chit,, and Korea Variable Brazil Chile Korea GDP growth (percent per year) Average level 2.8 5.9 7.6 Coefficient of variation' 1.4 0.9 0.4 Public expenditures on education and health (Percentage of GDP) Average level 2.9 5.6 3.4 Trend over time 0.1 -0.1 0.0 Gross domstic'ivesmnent (prcentge of GDP) Average level 20.5 19.7 32.6 Trend over rime -0.1 0.6 0.4 Memo itetms (latest available year) Poverty (percentage blow US$1 a day) 23.6 15.0 - Gini coefficient of iniom e 0.60 0.59 0.32 Gini coefficient of education 0.39 0.31 0o.2 Illiteracy (percent) 16.7 48 2.0 Infant mority 000) ;00 ; :00 ; 0: ; 0: 0; : :A \ ; X;34O 0: 00: L: ;1.134.0 11.0 : 9.0 --Not available. Note: T'he values are for 1978-97, except for expenditures on education and health, which are for 1980-97 (1980- 94 for Brazil), and specific years for some variables. a. Standard deviation of the growth rate divided by the growth rate. Source: Various issues of the World kWorldelopmentnica ardthe Inional M d's Government Finance Statiscs Ymearbook. * Evidence on subsidies. Industrial and developing countries have spent public resources on subsidies. In the case of capital, they involve a variety of mechanisms including tax concessions, credit subsidies, and grants. These subsidies absorb a sizable share of government revenues, which in developing countries seem comparable to what is spent on education, health, and social sectors. * Impact of subsidies. A finding in the literature is that capital subsi- dies have not contributed to increased productivity and have only modest effects on growth. Moreover, their effects on growth seem short-lived. 38 AS S >EF-, G ROWVTH, AND LV L LF ARE Reforms and Unbalanced Growth in 60 Countries A review of 60 countries in the late 1980s and 1990s shows that about 16 of the countries were considered serious reformers in implementing a set of policy changes (table 2.2). The other 44 countries did not implement such a set of reforms over the period. Reformers already had higher rates of physical capital accumulation in the 1980s than the nonreformers.7 Al- though a controlled experiment would better reveal counterfactuals, the contrast is suggestive. In the 1990s, the rates of physical capital accumu- lation increased by about 70 percent for reformers but declined for nonreformers. But the growth of human capital apparently has not in- creased much-for either reformers or nonreformers. Spending on educa- tion as a share of GDP was lower for reformers than for nonreformers, in- creasing modestly for both groups in the 1990s.5 Although deforestation rates, a rough proxy for natural resource degra- dation, were lower for reformers than for nonreformers in both periods, the deforestation by reformers almost doubled in the 1990s while that by nonreformers increased only slightly. Thus, reformers have significantly accelerated economic growth over the 1990s. This growth seems to be based on an increase in physical capital accumulation while, relatively speaking, investments in human and natural capital lagged. Table 2.2. Review of Development Indicators for 60 Reformers and Nonreformers, Selected Years Development indicator Years 16 reformers 44 nonreformers Per capita GDP growth rate (percent) 1984-89 2.8 -0.5 1990s 3.5 0.1 Physical capital stock (per worker) growth rate 1984-89 2.1 0.0 (percent) 1990s 3.5 -0.5 Deforestation rate (percent) 1984-89 0.7 1.2 1990s 1.1 1.4 Education spending as a percentage of GDP 1984-89 3.2 4.6 1990s 3.5 4.7 Note: Reformers in this table are defined based on the speed of integration index (World Bank 1996a). Countries that implemented significant economic reforms (reformers) in the late 1980s or early 1990s by this measure are Argentina, Bolivia, Chile, China, Ghana, Indonesia, Korea, Malaysia, Mauritius, Mexico, Morocco, Nepal, Peru, Philippines, Sri Lanka, and Thailand. Source: Author's calculations. 39 T 1i l Li A I I i X O r - ., P -: Are the increases in educational spending by reformers enough to sus- tain the new growth rates? Will the acceleration in degrading natural capi- tal seriously hurt the sustainability of growth for reformers and nonreformers alike? To answer these questions we need to know how spending improves human capital, how deepening physical and human capital affects growth, and how losing natural capital can affect growth. Econometric Evidence: 20 Middle-Income Countries Country econometric analysis of growth in 20 mostly middle-income coun- tries over 1970-92 shows the following (see annex table A2.1 and Lopez, Thomas, and Wang 1998):9 * The marginal productivity of capital, given other asset levels, de- clines with increases in physical capital. Economies of scale and technological spillovers from investment in physical capital appar- ently may not be enough to offset the declining marginal productiv- ity of physical capital. This suggests that growth based primarily on physical capital accumulation may not be sustained in the long run. * Human capital, represented here by formal education, would seem to have a powerful positive effect on economic growth in reform episodes, but not in the absence of reforms. This implies that educa- tion would not contribute much to the productivity of physical capital in overregulated economies with little space for markets. But it could do much to boost the marginal productivity of physical capital and economic growth in a market-friendly framework. This confirms our hypothesis presented earlier that human capital accu- mulation at sufficient speed can induce sustained growth. At the same time, this evidence suggests that key market reforms are a nec- essary condition to achieve long-run sustained growth. * In nonreforming economies and episodes, economic growth rates are not sustained, regardless of the additions to human capital, ac- cording to these results. Instead, they face stagnation after periods of moderate growth, triggered by favorable exogenous shocks that tem- porarily spur the returns to physical capital. * The good economic growth rates in reform episodes can be sus- tained if human capital grows fast enough to offset the declining marginal returns to capital caused by physical capital accumula- tion. Per capita growth of about 4 percent a year, according to these estimates, can be sustained if per capita human capital ex- pands at about 1.7-1.8 percent a year. 40 ASSErs, G ROWTH. AND WE LFARE So the pace of growth based mostly on physical capital accumula- tion-to the neglect of human capital-would not seem to be sustained. Market reforms can accelerate growth. But if the reforms are not accom- panied by investments in human capital, growth is likely to flag. Coun- tries that implement market reforms have a chance of sustained growth. Nonreformers do not. Econometric Evidence: 70 Developing Countries The previous study did not consider natural capital as a determinant of growth, but few of the 20 countries analyzed earlier show a heavy depen- dence on natural capital as a source of income. A related study of 70 developing countries that includes both middle-income and poor coun- tries, including several Sub-Saharan nations, considers natural, physical, and human capital as factors affecting growth (L6pez, Thomas, and Tho- mas 1998; see also note 8).10 Unlike most previous studies, this one uses a flexible functional form (translog for the growth equations) that allows for nonlinear effects of the ex- planatory variables and for interactive effects across these variables. The in- teractive effects are extremely important in elucidating interasset substitution or complementarity in the growth process (see annex tables A2.2 and A2.3). • According to these estimates, the rate of economic growth on aver- age declines with increases in the stocks of physical capital-for constant human and natural capital-but not for all countries. Countries that have very low physical-capital-to-labor ratios tend to have their growth rates increase. So, in capital-poor countries, capi- tal accumulation at first tends to speed growth even faster. But after reaching a certain capital intensity, further physical capital accumu- lation-for given human and natural capital-has a declining effect on economic growth. * Human capital on average would seem to boost the rate of eco- nomic growth, though this link is smaller than in the previous study. As human capital increases, the positive link to economic growth becomes larger. At low levels of human capital, its link to economic growth is negligible, but at higher levels of human capital it becomes larger, with the marginal effect of the stock of human capital on growth always increasing. * To sustain economic growth, human capital can to some extent substitute for natural capital, but physical capital may not. The growth rate of countries with high levels of human capital is much 41 T HE i QJ A L I T Y O F G R O W T I1 less sensitive to losses of natural capital. But that of human-capital- poor countries is highly sensitive to those losses. For them, natural capital is crucial for sustaining rapid economic growth. They there- fore need to invest in human capital to reduce their dependency on natural capital. These results suggest that growth especially based on physical capital ac- cumulation tends to be difficult to sustain. Economies of scale and techno- logical spillover arising from physical capital accumulation exist, but may not be sufficient to sustain growth. Physical capital accumulation needs to be accompanied by an expansion of human capital to permit sustained growth."' Disinvestment in natural capital hurts the sustainability of growth, especially in human-capital-poor countries. This result, that physical capital accumulation alone may not sustain growth, is consistent with recent em- pirical studies (Barro and Sala-l-Martin 1996; Jones 1995; Mankiw, Romer, and Weil 1992; Young 1994, 1995). Evidence on Subsidies The evidence accumulated over the last decade indicates that government subsidies to industries, agriculture, and infrastructure worldwide are large. Table A2.4 in annex 2 presents some examples that illustrate both the size and impact of such subsidies. The data are fragmented and partial, making it difficult to put in perspective the real magnitude of these subsidies relative to GDP and relative to government expenditures. In addition, the available data only include direct subsidies involving financial outlays (or foregone tax revenues) for the public sector. The evidence on indirect subsidies, such as the giveaway of public lands and natural resources, is mostly anecdotal. The available evidence, however, permits us to derive lower-bound esti- mates of the financial subsidies, at least for some countries. It is important to note that these remain gross estimates. They do not consider the net magnitude after accounting for taxes or other offsetting dis- tortions. These estimates also do not differentiate between cases where such subsidies might be justified on social grounds and where they might not be. Together with taxes, they influence the implicit tax rates, introducing elements of nontransparency, discrimination across different activities, and pressures on scarce resources-rendering them distortionary. During the early 1990s, industrial (OECD) countries spent an estimated US$490-$615 billion a year in subsidizing agriculture (US$335 billion), en- ergy (US$70-$80 billion), and road transport (US$85-$200 billion) (de Moor and Calamai 1997). The total is about 2.5-3.0 percent of the total 42 ASSEF S. GROWTH. AND WELFARE GDP of OECD countries and about 7.6-9.1 percent of the total govern- ment expenditures. Developing countries spent US$220-$270 billion per year in subsidizing energy, road transportation, agriculture, and water dur- ing the early 1990s. These amount to some 4.3-5.2 percent of GDP and 19-24 percent of total government expenditures. These subsidy estimates point to possible distortions, and do not necessarily suggest overinvestment in these sectors in the aggregate. On the one side, these are probably only a part of all subsidies, as subsi- dies to manufacturing are not included here. On the other, some of these subsidies (especially for energy) concern consumer demand and not corpo- rate production, which is our primary focus. However, a significant part of the energy subsidies seem to be captured by corporate entities, and the above estimates might still be close to representing corporate subsidies. From a different estimation, corporate subsidies in the United States in 1996 were US$170-$200 billion (Collins 1996), or 2.3-2.7 percent of GDP and 10-12 percent of total government expenditures. Government subsidies to Fortune 500 corporations, which in 1997 recorded profits of US$325 billion, were about US$75 billion-comprising government grants, cut-rate insurance, subsidized loans, and loan guarantees (Moore 1999). Apart from the energy and agriculture subsidies, countries provide subsi- dies directly to manufacturing industries. The evidence suggests that these in- dustrial subsidies may be larger than the energy and agriculture subsidies. Sub- sidies to foreign investors seem to be significant in a number of country cases. Preferential tax treatment for foreign firms sometimes costs the government in foregone tax revenues. Competition for foreign investments is, in some cases, a reason for these subsidies, which have gone to investors in mining and vari- ous industries ranging from automobile to steel (Aviation Week and Space Tech- nology 1999; Castaneda 1997; La Nacion 1997, June 10; Sieh Lee 1998; Oman 2000; also table A2.4 in annex 2). They are essentially discriminatory in na- ture and raise the question of the effectiveness of favoring some over others. These admittedly partial data suggest the significance of the corporate subsidies as a proportion of government expenditures-with implications for capital subsidies, although we have not been able to disentangle fully capital and corporate subsidies. In the previous sections we have empha- sized a less distorted or a more neutral asset growth pattern that includes the expansion of human and natural assets along with physical. These sub- sidies compete for scarce public resources with alternative uses. The ques- tion is whether or not they could be better spent from a social viewpoint on the sector in question or in other areas such as building human capital and in preventing a rapid deterioration of the natural capital. It is also pos- sible that the corporate subsidies contribute to a sustained expansion of 43 l l R:\ l ) l l ( E investment in physical capital, increasing economic efficiency and produc- tivity and generating positive social spillovers. If this is true, the case against subsidies would diminish. The Impact of Subsidies Recent studies based on industry or microfirm data have examined how corporate subsidies affect long-term economic growth and productivity. By and large they suggest that government subsidies to industries have a mod- est impact on firms' investment and growth in the first year, but over the medium run have little effect on growth. Capital subsidies also seem to in- duce a negative effect on total factor productivity of the industries that re- ceive subsidies. Beason and Weinstein (1996) for Japan; Bergstrom (1998) for Sweden; Bregman, Fuss, and Regev (1999) for Israel; Fakin (1995) for Poland; Fournier and Rasmussen (1986) for the United States; Harris (1991) for Ireland; and Lee (1996) for Korea conclude that corporate subsi- dies are inappropriate if increasing national income and productivity is the goal (see also table A2.4 in annex 2). The papers by Bregman, Fuss, and Regev (1999) and Bergstr6m (1998) are particularly important because they use detailed firm-level panel data. Bregman, Fuss, and Regev (1999) found that capital subsidization induced efficiency losses ranging from 5 to 15 percent. They also show that the sub- sidies were basically incorporated into profits or rents, as the subsidized firms earned higher rates of return than those that were not subsidized. Similarly, Bergstr6m (1998) found little evidence that subsidies affect pro- ductivity. Their effects on the growth rate of firms seemed temporary. This finding is consistent with the point in this chapter that capital subsidies could only offer a temporary relief to decreasing rates of economic growth associated with distorted asset growth. Conclusions This chapter presented a framework for the augmentation of three main as- sets: human, physical, and natural capital. Its main hypothesis: sustained growth and welfare improvement require the efficient expansion and use of all three assets. However, countries can be tempted to subsidize physical capital. The evidence is that such subsidies (tax exemptions, direct subsidies, easy access to natural resources, and so on) comprise large shares of government expenditures and of GDP. Such an approach is unlikely to produce sustained growth. It also neglects human and natural assets, which ASSETS. GROWTH. AND WELFARE directly contribute to welfare. So, such growth may provide only a small contribution to welfare. Investing a greater part of national savings in the expansion of human and social assets-and the sustainable use of natural assets-could contrib- ute to more growth and better growth in the long run. This sustained growth, attending to all three assets, is more likely to increase welfare. This could be because investments in human and natural capital contribute to welfare directly and because investment in such assets helps to improve the distribution of income and reduces the instability of growth. That is why a relatively undistorted or balanced approach to the accumulation of all assets is likely to be superior to a primary focus on physical and financial capital. Notes 1. Financial capital here does not refer to the development of financial institu- tions and the deepening of financial markets in a economy, which are desirable in supporting development (see chapter 5). 2. As discussed later, balanced asset growth does not imply that all assets should grow at the same rate. The focus of balanced growth, as the term is used in this chapter, is on the composition of assets, rather than on the sectoral composi- tion of output, which is the common convention (Hirschman 1958; Nurkse 1953). 3. The lack of balance in asset growth arises as a consequence of externalities and market failures. Physical capital is perhaps less subject to externalities than hu- man and natural capital. Imperfections in credit markets prevent the poor from investing in their education at desired levels even if they can obtain a high rate of retum. Extemalities affecting natural capital, including the environment, are extremely pervasive. Also, investments in human and natural capital require a long time to mature relative to most investments in physical capital. Capital market imperfections are likely to affect the financing of the former more nega- tively than the financing of the latter. Thus, the private market economy tends to concentrate more on the accumulation of physical capital than of the other two assets. Other reasons that could lead to unbalanced asset growth empha- sized in the literature are coordination failures. These are caused by agent inter- actions that are not fully mediated by market prices (see, for example, Stiglitz 1975 for an early model of multiple equilibria arising from imperfect informa- tion conceming ability and education and Murphy, Shleifer, and Vishny 1989 and Rodriguez 1993 for intersectoral coordination failures). 4. Other examples of subsidies to capital are abundant. Argentina and Mexico provided monopoly rights to privatized telephone companies for prolonged periods. Brazil gave subsidies and tax concessions to invest in automobiles (Fi- nancial Times, July 21, 1999). Chile has subsidized tree planting by a few large 45 1- H1 F Q l A L I T Y 0 F G R 0 W T H corporations to support the expansion of the private pulp and paper industry. Since the early 1980s, China has provided tax exemptions and tax reductions to foreign investors. In Central and Eastern Europe, direct government subsi- dies take the form of tax arrears that amount to 5-10 percent of GDP and in- crease by about 2 percent of GDP every year (Schaffer 1995). In Brazil, rubber producers received large subsidies from the government. Eight companies re- ceived R$5 billion (US$2 billion) (Gazeta Mercantil, May 21, 1999). In Ko- rea, two major steel producers received US$6 billion in 1993-99 in govern- ment subsidies, according to the U.S. complaints filed with the World Trade Organization (New Steel 1998). Herrera (1992) discussed in detail the regres- sive impact of the lack of regulation in the privatized telephone system in Ar- gentina. See table A2.4 in annex 2. 5. Because multiple equilibria and irreversible processes are likely outcomes, there is scope for public policy interventions aimed at avoiding vicious cycles of pov- erty and environmental degradation. 6. The distribution of education is measured by Gini coefficients and standard de- viations of education (see chapter 3 for details of these measures and Lopez, Thomas, and Wang 1998 for a statistical analysis). 7. The average growth rate in domestic investment among the most aggressive re- formers was much higher during the 1990s, after the reforms had been imple- mented, than in the 1970s and 1980s. In Argentina, Bolivia, Chile, and Peru, four of the most aggressive reformers in Latin America, the growth of gross in- vestment during 1990-97 was more than 9 percent a year, almost three times the historical rates (IDB 1998). 8. In table 2.2 we use expenditures on education as a percentage of GDP instead of per capita expenditures because the underlying stock of education is likely to be positively related to GDP. Thus, a change in the share of education expendi- tures in GDP is likely to be more closely related to the rate of growth in human capital than the level of expenditure per capita. 9. This study was based on an explicit growth theoretic behavioral model. This is important because the estimating empirical equations derived from such a model suggest a specification that is relatively free from the simultaneous equa- tion bias that has affected some previous studies. In particular, the empirical model consists of explaining annual growth rates by lagged stocks of assets rather than by rates of change of assets as is usually done. This considerably re- duces contemporaneous correlation with the error term that usually leads to se- rious difficulties in deriving causality from the results. Moreover, the fact that we use country fixed effects could decrease the possibility of bias due to omitted country-specific variables, another important source of difficulty in interpreting causal relationships. Controlling for omitted variable biases and simultaneous equation biases suggests that we are in large measure addressing causality prob- lems. Finally, the study used a detailed analysis of the policy reforms of the vari- ous countries over the two decades considered, so that the coefficients esti- mated were allowed to vary systematically across policy regimes. This allowed 46 ASSETS, GROWTH, AND WELFARE the study to show that the weak impacts of education on growth reported by other studies was right only under certain closed economy and distorted policy regimes, but not for market-friendly environments. For more details about the estimating procedure see annex 2. 10. This study used forest area as a proxy for natural capital. Loss of forest cover is usually associated with watershed deterioration, loss of commercial logging spe- cies, water depletion, and soil erosion, all crucial for production, and is likely to be a good proxy for the degradation of natural capital. 11. This finding is not necessarily inconsistent with the literature on growth con- vergence, which generally finds slow convergence across countries. In fact, we find that a stable growth rate can be maintained indefinitely if physical and hu- man capital grow at balanced (not equal) rates. The problem is only that the rate of economic growth declines as the stock of physical capital increases for a given level of human capital, or if human capital expands at a speed below a minimum required rate. 47 C H A P T E R 3 IMPROVING THE DISTRIBUTION OF OPPORTUNITIES Wealth to us is not mere material, but an opportunity for achievement. -Thucydides, 460-400 B.C. T he main asset of most poor people is their human capital. Investing in the human capital of the poor is a powerful way to augment their assets, redress asset inequality, and reduce poverty. This chapter examines the quality of education, associating the distribution of education with growth and poverty reduction. It then asks how to make education at all levels more productive. To be sure, access to good quality education is important in that it enhances people's capabilities to generate income. This is not enough, however. To be more productive, they need to be able to combine their human capital with other productive assets, such as land and equity capital, and with job opportunities in an open market. Chapter 2 discussed the importance of undistorted or balanced asset augmentation. This chapter focuses on assets that the poor possess, prima- rily human capital, and those on which they rely most heavily, such as land. For growth to have an impact on poverty, the assets of the poor, espe- cially their human capital, need to be augmented and distributed more eq- uitably.' Yet inequality in education and health outcomes is staggeringly high, reflecting market failures and underinvestment in the human capital of the poor. Asset distribution represents the distribution of opportunities and is a precondition for individual productivity and income. While redis- tributing existing assets and incomes is politically difficult, building new as- sets such as human capital is widely accepted. 49 T 1H E (QU A L I T Y 0 F C, R 0 W T H To be sustained, development must be equitable and inclusive. Ensuring adequate public spending in education and health care is important, but does not by itself guarantee progress. A multidimensional strategy to em- power people is needed. Actions to highlight include the following: * Augment assets of the poor by ensuring access to high-quality edu- cation and health services * Increase attention to the distributive effect of public investment and reduce subsidies to the types of education and health care that benefit the rich * Facilitate full use of human capital by empowering the poor with land, credit, training, and job opportunities * Complement all human capital investments with economic reforms and market openness, which increase the productivity of education. Potential Benefits of Education Education and good health improve people's ability to shape their lives- strengthening their functioning in society and contributing to their welfare directly. Educating women, for example, not only increases their income- earning capacity, but also improves their reproductive health, lowers infant and child mortality, and benefits both current and future generations. In- vesting in human capital is therefore crucial for economic growth, poverty reduction, and environmental protection. The benefits of investing in hu- man capital are well known, but some of the linkages with other dimen- sions of development-security, social justice, and sustainability-are bet- ter understood today than they were 10 years ago.2 Investing in people can protect workers and improve security-an im- portant aspect of quality of life. Education and good health increase the poor's ability to cope with changes in their environment. They allow them to switch jobs and provide some protection against economic downturns and financial crises (chapter 5). Social exclusion reduces an individual's incentive to attend school and to work (Bourguignon 1999; Loury 1999). Investment in human capital, if well distributed and targeted to the poor, can facilitate social inclusion. Better education and health services to vulnerable, often ex- cluded groups, such as those who are illiterate, disabled, elderly, chroni- cally ill, or separated by language barriers, can help them overcome social obstacles and increase their productivity. Investing in people may also help protect the environment. Better- educated women have healthier and, in many cases, fewer children, thereby reducing demographic pressure on natural resources and the environment. 50 I IMPROVING THE DISTRIBUTION OF OPPORTUNITIES With more education, people can assimilate more information and employ in- struments to protect the environment and better manage resources (chapter 4). Investing in people improves human rights and social justice, which provides direct satisfaction. Basic education enables the poor to learn about their civil and political rights; to exercise those rights by voting and run- ning for office; and to voice their concerns, seek legal redress, and exercise public oversight. That helps in building institutions, improving gover- nance, and fighting corruption (chapter 6). These benefits are far from automatic. Many studies show that addi- tional years of education per person increase real output or growth rates. However, a few researchers suggest that human capital accumulation has an insignificant or negative impact on economic and productivity growth (Benhabib and Spiegel 1994; Griliches 1997; Islam 1995; Pritchett 1996). More government spending on education, if misallocated, might contribute little to poverty reduction and instead increase inequality and rent seeking. As Murphy, Shleifer, and Vishny (1991, p. 503) point out: "A country's most talented people typically organize production by others... When they start new firms, they innovate and foster growth, but when they become rent seekers, they only redistribute wealth and reduce growth." Quantity Is Not Enough-Quality Matters Since 1980, developing countries have invested substantial amounts of public resources in education services (see figure 1.11). In the 1990s, more than three-quarters of school-age children in developing countries were enrolled in schools, up from less than half in the 1960s. Illiteracy rates dropped from 39 to 30 percent between 1985 and 1995 (World Bank 1999a). Progress has been uneven across regions. Enrollment rates fell in Sub- Saharan Africa: the proportion of 6-11 year olds enrolled in schools dropped from 59 percent in 1980 to 51 percent in 1992 (World Bank 1999a). Lack of access to basic education remains a major challenge in many countries. Increas- ing public spending is desirable, but not sufficient for the following reasons. Public Spending Is Only Weakly Related to Outcomes Cross-country analyses reveal a weak relationship between the generosity of education spending and education outcomes. Using cross-country data, Filmer and Pritchett (1999b) examined the correlationr between government education spending per student and the percentage of people aged 15 through 19 who had completed grade five. The correlation appeared positive and significant at first, but after controlling for per capita income, the corre- lation was found to be fairly weak (figure 3.1). A similarly weak correlation 51 I 1 I r Q F l A L I T ) 0 F (, R () W I -H Public spending on education is only weakly related to education outcomes Figure 3.1. Relationship between Public Spending Per Capita and Educational Attainment, Various Years Percentage of 15-19-year-olds who have Percentage of 15-19-year-olds who have completed grade 5 completed grade 5 (controlling for GNP per capita) 100 40 * 30- 80- 3 *3 20 - 60- 10- *E .U 40- 0 * -10 - a 20 - a -20 - * 0 - l l l -30 - lI I 0 100 200 300 -200 -100 0 100 200 Public spending on education per person Public spending on education per person, (U.S. dollars) controlling for GNP per capita (U.S. dollars) Nore: Expenditures refer to public spending on preprimary, primary, and secondary education only. Thirry-five developing countries were included in the study. Sources: Education outcome data are updated from Filmer and Pritchett (1999b) combined with expenditure data from the United Nations Educational, Social, and Cultural Organization (UNESCO) database. was found between govemment health spending and mortality rates for chil- dren under five years old (Filmer and Pritchett 1999c). Why is public spending only weakly related to outcomes? What makes the difference is the quality and distribution of education services and the produc- tivity of human capital. For developing countries that already allocate a sub- stantial share of public resources to social services, further spending may not improve education outcomes for the poor. Reallocating public spending and improving its efficacy often can improve outcomes, especially when public re- sources are subsidizing education for the wealthy. Economywide strategies and policies also matter: subsidies to attract foreign capital may, under certain cir- cumstances, bias the rate of return against human capital.3 Labor market dis- tortions create disincentives for investing in education. In addition, to be pro- ductive, people must have access to other productive assets, including land, credit, equity, and job opportunities in open and competitive markets. 52 IMP ROVING THE DISTFRIBUT ION OF OPPORTUN ITI ES Variability in School Quality Despite progress on access to education, the quality of schooling varies con- siderably across countries and regions. An extensive literature explored how best to define and measure the quality of schooling: whether inputs, pro- cesses, or student achievements should be used in assessments (see, for ex- ample, Behrman and Birdsall 1983; Card and Krueger 1992; Greaney and Kellaghan 1996; Lockheed and Verspoor 1991). We measured quality as a combination of indicators that reflect inputs, defined by expenditure per stu- dent and the number and quality of teachers; processes, that is, the length of school terms and the curriculum content; and outputs, measured by cognitive achievements, attitudes, test scores, and dropout rates. In high-income countries where these indicators are well developed, student achievement varies widely, even in countries with universal basic education. Functional literacy rates for young adults, 16-25 years old, in some industrial countries vary from 45 percent in the United States to 80 percent in Sweden, while the secondary net enrollment rates in these coun- tries are all above 85 percent (World Bank 1999a). In developing countries, where achievement indicators are scarce, less accurate indicators, such as repetition and dropout rates, have been used to assess education outcomes. Data generated by these imperfect measure- ments showed considerable variation in the quality of schools (table 3.1). Repetition and dropout rates for primary school are much lower and test scores higher in East Asia than in Latin American countries, where in- comes are higher. While public education spending rose in some Latin American countries in the 1990s, average primary dropout rates also in- creased.4 Other studies, based on the limited available data on internation- ally comparable test scores, also show that generous public spending did not guarantee high-quality education. What explains the large variations in quality? Education outcomes de- pend on both demand and supply factors, and thus on policies and incentive structures that affect the whole economy. Macroeconomic stability, repre- sented by international terms of trade and GDP volatility, for example, is found to be the most significant determinant of educational attainment in Latin America. Using data from 18 household surveys, Behrman, Duryea, and Szekely (1999) found that the debt crisis of the 1980s contributed to the slowdown in the accumulation of schooling in Latin American countries. Kaufmann and Wang (1995) found that macroeconomic policies affect social sector investment projects. As a country opens to intemational trade and in- vestment, the rate of return to education rises. People demand higher quality education and are willing to pay more for it. Stronger demand, higher private 53 T H E QU A L. I T Y O F G R O W T HI Repetition and dropout rates vary enormously across countries Table 3.1. Primary School Repetition and Dropout Rates, Selected Years (percent) Public spending on education Primary school Primary school (percentage repetition rate dropout rates of GNP) Country 1980s 1990s 1970 1980 1990 1970s 1980s 1990s Argentina - 6 36 34 34 1.65 1.79 3.07 Brazil 20 18 78 78 80 2.95 4.04 3.60 Chile - 6 23 24 23 4.60 4.52 2.84 Colombia 17 9 43 43 44 2.05 2.75 3.43 Mexico 10 8 11 12 28 2.90 4.06 4.45 Peru 17 15 34 30 30 3.30 3.09 3.40 Venezuela, RB 10 11 41 32 52 4.30 5.09 4.56 Average Latin America 15 10 38 36 42 3.11 3.62 3.62 China - 3 15 15 15 1.45 2.45 2.20 Indonesia 10 9 20 20 23 2.65 1.38 1.34 Korea, Republic of - - 5 6 1 2.80 3.89 3.92 Malaysia - - 1 1 4 5.10 6.61 5.37 Philippines 2 - 25 25 30 2.40 2.02 2.54 Thailand 8 - 57 23 13 3.35 3.58 3.88 Average East Asia 7 6 21 15 14 2.96 3.32 3.21 - Not available. Sources: World Bank data; UNESCO for expenditure data. investments, better-paid teachers, and more motivated students produce higher educational achievements, with differing time lags. The higher the de- mand for education, the higher its quality, and vice versa. If a country de- votes public resources to subsidize physical capital instead of basic education, it can bias the rates of return against unskilled labor and hurt the poor (see table A2.4 on capital subsidies). At the micro level, many studies have examined the links between schooling quality and student performance. Behrman and Knowles (1999) found a strong positive association between the quality of teaching staff, the quality of current inputs, and children's success at school. Hanushek and Kim (1995) found that conventional measures of school resources, that is, pupil-teacher ratios and educational spending, did not affect student test performance. In cross-country regressions, test scores were positively related to growth rates of real per capita GDP, indicating a potential feedback from growth to strong demand and good student performance. Lee and Barro 54 I M P ROV I N G TH E D I STRI BUTIO N O F OP PO RTU N IT I E S (1997) found that family background, strong communities, school inputs, and length of school terms are positively related to student performance; however, they cannot fully explain why East Asian countries experienced better education outcomes than did other developing countries. That sug- gests that other factors may be at play, including those associated with a more open and export-oriented economic environment. Consequences of Poor Quality Low-quality schooling disproportionately hurts the poor and limits their future earning opportunities. For example, Vietnamese students from high-income households enjoy greater access to high-quality education (Behrman and Knowles 1999). In Latin America, most students from low-income families attend public schools, which offer half the hours of instruction and cover only half the curriculum compared with the private schools. The higher the family's income, the greater the aversion to pub- lic schools (IDB 1998). Estimates based on household surveys from Latin America show that students from lower-income deciles received an inferior primary education. Quality, measured by students' labor market performance, was 35 percent lower for low-income students than for those at the next-higher income decile (IDB 1998, p. 54). Figure 3.2 shows the enormous gaps in secondary school completion rates for the rich and poor. Because private education is feasible only for the wealthy, the poor quality of public schooling severely reduces the income-generating potential of children from poor families. Quality and Quantity: A Tradeoff? Improvements in quality complement the expansion of access to education. If poor children can go only to low-quality schools, they have few opportu- nities to obtain high-paying jobs and parents are disinclined to send them to school. When education coverage is not universal, the best strategy is to focus on policy interventions that raise demand for both the quantity and the quality of education. For example, programs to reduce child labor and keep children in school-such as school lunches and cash stipends-would go well with teacher training to improve quality. However, with growing populations and tight budgets, the synergies of quantity and quality can turn into tradeoffs, especially if the quality mea- sures selected are not closely linked with student learning. What quality measure should be used for intervention? Should it be student incentives, or length of school terms, or the quality of teaching staff? Evidence shows that reduction of pupil-teacher ratios, which is expensive, has little impact on 55 I 1i E Q_U A L I T Y 0 F G R 0 W T fi Figure 3.2. Secondary School Completion Rates for 20-25-Year-Olds by Household Income Level, Selected Latin American Countries and Years Latin American countries Bolivia, urban 1995 - 40 Honduras, 1996 * X 48 Pertu 1996 __ 54 Uruguay, urban 1995 56 Venezuela, RB, 1995 59 Ecuador, 1995 _59 Costa Rica, 1995 60 Chile, 1994 = - 60 EL Salvador, 1995 t _ _ 61 Paraguay, 1995 62 Mexico, 1994 66 Beaz 1995 71 * Poorest 10 percent Ez Richest 10 percent Panama, 1995 i- -_73 Argentina, 1996 79 0 20 40 60 80 100 Percent Note: Numbers next to bars are gaps in completion rates (percent) The surveys for Argentina include only Greater Buenos Aires. Source: IDB (1998, p. 27). student learning (Mingat and Tan 1998).5 Despite the relatively high pupil- teacher ratio in the 1980s and 1990s, Korean students' average scores on in- ternational science and mathematics tests were among the highest. Spend- ing more to hire more teachers might imply a tradeoff against wider coverage and broader distribution of education, which would be inefficient and inequitable, particularly where many children still have no access to ba- sic education (Mingat and Tan 1998). Achieving Equitable Education and Social Inclusion Equal access to education and health services is among the basic human rights to which everyone is entitled. As with land and physical capital, an equitable distribution of human capital is important for broad-based growth s5 IMPROVING THE DISTRIBUTION OF OPPORTUJNITIES and poverty reduction. Moreover, equitable distribution of opportunities is preferable to the redistribution of existing assets, because investing in people creates new assets and improves social welfare.' Ensuring access by the poor by distributing education services more equally is a win-win policy that is gaining support in both industrial and developing countries. Why the focus on the distribution of education? This is because ensur- ing access to basic education by the poor is closely related to a better distri- bution of education. Given limited public resources for education, concen- trating public investment on education for the poor usually implies a reallocation of public spending away from subsidies to the types of educa- tion services that benefit the rich. Such policies are politically unpopular, and many countries have been unable to implement them. However, as shown in this section, there are compelling reasons why a government should pursue such policies. Measuring Dispersions in Education Outcomes Since the days of Adam Smith, education has been linked to equitable social and economic progress. There is a small but growing literature on schooling inequality or the distribution of education (see, for example, Lam and Levinson 1991; Londonio 1990; Maas and Criel 1982; Ram 1990). As data became available for measuring the distribution of education, the disparities became more apparent. Using standard deviation of schooling attainment, Birdsall and Londofio (1997) investigated the impact of initial asset distribu- tions on growth and poverty reduction and found a significant correlation be- tween initial educational inequality and reduced income growth. Later, researchers constructed education Gini coefficients, which are similar to the Gini coefficients widely used to measure distributions of in- come, wealth, and land. The Gini coefficient ranges from 0, which repre- sents perfect equality, to 1, which represents perfect inequality (see annex 3 for the two methods used to calculate the Gini coefficient). Education Gini coefficients can be calculated using enrollment, financing, or attain- ment data, recognizing that different cohorts in a population were edu- cated at different times. L6pez, Thomas, and Wang (1998) estimated Gini coefficients of educational attainment for 20 countries and found significant differences in the distribution of schooling. Korea had the fast- est expansion in education coverage and the fastest decline in the educa- tion Gini coefficient; it dropped from 0.51 to 0.22 in 20 years. India's education Gini coefficient declined moderately, from 0.80 in 1970 to 0.69 in 1990. Education Gini coefficients for Colombia, Costa Rica, Peru, and Venezuela have been increasing slowly since the 1980s, showing that inequality is on the rise (figure 3.3). 57 T HE QOU ALI TY OF G R O W T H Figure 3.3. Gini Coefficients of Education, Selected Countries, 1960-90 Education Gini Mali Less equal 0.8 - - -- - India 0.6 -. - 0. 6 - -~ - = = Venezuela, RB 0.4 - ~China More equal 0.2 -- Korea, Republic of 0.0 - I I I I I 1960 1965 1970 1975 1980 1985 1990 Source: L6pez, Thomas, and Wang (1998). An examination of education Lorenz curves for India and Korea in 1990 shows a great range among developing countries (figure 3.4). Despite progress in expanding primary and secondary enrollment in India, more than half of the population (age 15 and older) did not receive any educa- tion while 10 percent of the population received nearly 40 percent of total cumulated years of schooling. Providing universal access to basic education remains a huge challenge for the country. Korea expanded its basic education program more rapidly, with a far more equitable distribution in educational attainment, as indicated by a flatter Lorenz curve and a smaller Gini coefficient. Even in 1960, when Korea's per capita income was similar to that of India, Korea's education Gini coefficient was 0.55, much lower than that of India in 1990. Note that the distribution of education in Korea was more equitable than that of income, but the distribution of education in India was much more skewed than that of income between 1970 and 1990.7 A distribution of education as skewed as that of India implies a huge so- cial loss from the underutilization of potential human capital. Assuming that ability or talent is normally distributed across population groups, pro- duction increases to its optimum when the dispersion of education matches the distribution of human ability. When the distribution of education is too skewed to match the distribution of ability, there is a deadweight loss to the society of underdeveloped and underutilized talent. In this case, societies 58 IMPROVI NG TH E DISTRI BUTION OF OPPORTUNITI ES The distribution of education varies enormously, from highly skewed to more equal Figure 3.4. Education Lorenz Curves for India and Korea, 1990 Cumulative proportion of schooling, India Cumulative proportion of schooling, Korea (percent) (percent) 100 - t.o 100 100 - . - - _ 100 91 89.2 so ~~~~~ ~~~~~~83.8 8 8 80-. 8-' 38.... 80 - ..v;;8.3 60 - 63.3 60- 40 - X40 20 -2.8 20 249 200 0.0 0 20 40 60 80 100 0 20 40 60 80 100 Cumulative proportion of population (15 and over) (percent) Cumulative proportion of population (15 and over) (percent) Mean = 2.95 years Education Gini = 0.69 Mean = 10.04 years Education Gini = 0.22 Source: Thomas, Wang, and Fan (2000). would be better off to massively expand basic education, especially by im- proving access to education by the poor. Examining the cross-country pattern of the distribution of education, we found that education Gini coefficients decline as the average education and income levels increase, although there clearly are other possibilities. Does the education Gini have to get worse before it gets better? As sug- gested by Londofio (1990) and Ram (1990), there is a "Kuznetsian tale" with distribution of education. That is, as a country moves from the zero to maximum level of education, the variance first increases and then declines. However, country analysis suggests that this may not be the case if Gini co- efficients are used to measure inequality. In addition to the industrial coun- tries, Argentina, Chile, and Ireland had relatively low education Gini coef- ficients from the 1960s to the 1990s. The Gini coefficient for education in Korea and some other countries declined dramatically. Only a few coun- tries-Colombia, Costa Rica, Peru, and Venezuela-have seen a significant worsening of the education Gini coefficient. So worsening distribution of education is not inevitable (figure 3.5). Among 85 countries for which edu- cation Gini coefficients were calculated, Afghanistan and Mali had the 59 T f i E 0lLU A l I T Y 0 F C, R 0 W T H Education Ginis decline as average education level rises Figure 3.5. Education Gini Coefficients for 85 Countries, 1990 Education Gini coefficient 1.0 - Mai 0.8 - * <<]India 0.6 - < 0.4 - g Hong Kong, China 0.2 - Sri La i oe Poland United States 0.0 - I I I I I I 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Average yeaTs of education Source: Thomas, Wang, and Fan (2000). least equitable distributions in the 1990s at approximately 0.90, while most industrial countries were at the lower end, with the United States and Po- land having the most equitable distribution (Thomas, Wang, and Fan 2000). Similar to the large variations in the distribution of education, other studies found large variations in health outcomes across income groups (box 3.1 ). Causes of Inequality in Education Disparities in education is one of many aspects of poverty, but they are also associated with misallocation of public investment, war, wealth gaps, gen- der gaps, social exclusion, and economic crises. Numerous studies found that parents' education and household income, as well as wealth, affect children's education attainments. Wealth Gaps. Using data from the National Family Health Survey col- lected in Indian states in 1992 and 1993, Filmer and Pritchett (1999a) found that the wealth gap, defined as the difference between the top 20 percent of an asset index and the bottom 40 percent, accounted for a large proportion of differences in enrollment rates. Enrollment rates varied from 4.6 percent in Kerala to 42.6 percent in Bihar. 60 IMPROV ING r FE D IS frR I MU r ION OF OPPO RV1U N I IES Box 3.1. Health Gaps between the Rich and the Poor Are also Large Health gaps between the rich and the poor are as those among people who have higher incomes, Fig- large as education gaps, which reflects the difficul- ure 3.6 shows that in Brazil, child mortality rates ties of reaching the poorest people outside the were high among the poorest 10 percent of the mainstream of economic life. Many studies find population, and fell as wealth rose. It indicates that that the poorest of the poor are in the worst health the poorest of the poor are in worse health than (Behnran and Deolalikar 1988), and they are often others. They suffer from infectious diseases much hit hardest by wars, external shocks, and social and more than richer people do. Therefore, they are political upheavals. Child mortality rates among more dependent on good public policies than the the poorest of the poor are often much higher than rich (Bonilla-Chacin and Hammer 1999). The poor are sicker than other people Figure 3.6. Mortality of Children Two Years Old and Younger by Wealth, Brazil, 1996 Deaths per 1,000 12 - 10 - 6- 4- 2 0-I Poorest 2 3 4 5 6 7 8 9 Rkhest Source: Bonilla-Chacin and Hammer (1999). In some countries, the differences in educational outcomes between the rich and the poor are staggering. A study of youths aged 15-19 in 20 coun- tries showed that the poorest 40 percent of the population in five countries had a median of zero years of completed schooling; more than half of this group completed less than one year of school (figure 3.7). The education difference between the richest and poorest groups reached as high as 10 grades in India. Similar disparities in education attainment are found in Latin America (figure 3.8). One implication of this large wealth gap is that demand for education is not independent of other endowments. Providing access to education 61 F H E Q LJ A L I T Y O F G R O W f H Differences in grade attainment for rich and poor households are enormous in some countries Figure 3.7. Median Grade Attainment for 15-19-Year-Olds from Rich and Poor Households, Selected Countries and Years Country Tanzania, 1996 2 Zambia, 1996-97 l 2 Zimbabwe, 1994 . _l 2 Uganda, 1995 3 Indonesia, 1994 3 Egypt, 1995-96 3 Mali, 1995-96 4 Central African Republic, 1994-95 . l 4 Malawi, 1992 _ 4 Haiti, 1994-95 _ 4 Brazil, 1996 4 Cameroon, 1991 4 Colombia, 1995 4 Peru, 1996 4 Dominican Republic, 1996 4 Bangladesh, 1996-97 5 Cote d'lvoire, 1994 r 6 Guatemala, 1995 6 Morocco, 1992 8 Pakistan, 1990-91 9 India, 1992-93 10 0 1 2 3 4 5 6 7 8 9 10 11 Grade * Poorest 40 percent E Richest 40 percent Note: Numbers next to bars are the gaps (in grades) between rich and poor. Source: Filmer and Pritchett (1999b). 62 IMPROVING THE DISTRIBUTION OF OPPORTUNITIES Figure 3.8. Years of Schooling for 25-Year-Olds from Rich and Poor Households in Latin America Country Uruguay, urban 1995 6.2 Venezuela, RB, 1995 6.3 Argentina, 1996 6.5 Chile, 1994 16.6 Peru, 1996 6.9 Bolivia, urban 1995 1 7.2 Paraguay, 1995 _7.3 Costa Rica, 1996 7.4 Honduras, 1996 7.5 Ecuador, 1995 _ 8.4 Brazil, 1995 8.5 _ Poorest 10 percent El Salvador, 1995 8.6 El Richest 10 percent Panama, 1995 9.2 Mexico, 1994 10 0 4 8 12 16 Years of schooling Note: Numbers next to bars are the gaps (in years of schooling) between rich and poor. The surveys for Argentina include only Greater Buenos Aries. Source: IDB (1998, p.27). (supply) is not sufficient. Addressing many structural and social inequali- ties influencing demand, such as gender gaps and the distribution of other productive assets such as land (discussed later), is important as well. Social Exclusion. People who are excluded from mainstream society are less likely to be educated. Loury (1999) showed how social exclusion changes human behavior and reduces the demand for schooling in inner cities of the United States. One reason that students drop out of school is because their peers have dropped out. In Bolivia, the inability of parents to speak Spanish is associated with higher mortality rates for children under two years old. In India, members of scheduled castes have higher mortality rates than other groups (Bonilla-Chacin and Hammer 1999). 63 T H E QU A L I T Y G F G R O W T H1 Gender Gaps. In some countries gender gaps are an important cause of education inequality. Among many studies addressing gender gaps in education, Schultz (1998) found that some 65 percent of world inequal- ity is between countries, 30 percent is between households in a country, and 5 percent is between gender inequality. Bouis and others (1998) found a significant difference in human capital investments, such as in nutrition, health care, and educational attainment, between boys and girls in the rural Philippines. In Bangladesh, which has the largest gen- der gap out of the countries reviewed, women's attitudes toward their daughters' education have been slow to change (Amin and Pebley 1994). However, recent efforts have resulted in encouraging progress (box 3.2). Knight and Shi (1991) found that educational opportunities were still unevenly distributed in China despite considerable progress. The pattern of educational attainment is affected by gender as well as by other fac- tors, such as income of the provinces, rural-urban differences in income, and family background. Though on the decline, gender discrimination persists in China's rural areas (see Dubey and King 1996; King and Hill 1993; and World Bank 2000g for cross-country experiences). The correlation is strong between inequality in education and gender gaps in literacy. Using a sample of 85 countries for which education Gini coefficients are available, Thomas, Wang, and Fan (2000) found that 64 IMPROVING THE DISTRIBUTION OF OPPORTUNITIES correlation coefficients between gender differences in illiteracy and educa- tion Gini coefficients increased significantly from 0.53 in the 1970s to 0.69 in the 1990s. While educational inequality declined, gender inequality accounted for much of the remaining disparities in educational attainment (figure 3.9). Reducing gender gaps in education is crucial to addressing the inequality in education. Consequences of Large Dispersions in Education Outcomes A society cares about the unequal distribution of education because it di- rectly affects human welfare. Unequal distribution of education is both a source and a consequence of poverty and social exclusion. Poor children who drop out of school eventually form a core of disadvantaged citizens who will be left out of mainstream economic and social life. Unless such people can obtain training later in life to find a meaningful job, poverty re- duction and social inclusion will remain out of reach. A highly skewed distribution of education tends to be associated with reduced per capita income growth, even after controlling for labor and While education inequality has been declining, gender inequality accounts for much of what remains Figure 3.9. Gender Gaps and the Inequality of Education, 1970 and 1990 Education Gini, 1970 Education Gini, 1990 1.0 1.0 0.8 - 0.8 - 0.6 -0.6 - Im 0.4 - A 0.4 - * 0.2 - r1970 =0.53 0.2 - r1sso 0.69 0.0 - I I I I 0.0 . I I -40 -20 0 20 40 60 -40 -20 0 20 40 60 Difference in illiteracy rate (female-male, percent) Difference in illiteracy rate (female-male, percent) Note: The figures include data for 85 countries. Sources: Education Gini coefficients from Thomas, Wang, and Fan (2000); gender gap in illiteracy from World Bank (1999d). 65 T H E Q U A L I T Y O F G R O W T H physical capital (L6pez, Thomas, and Wang 1998). Unlike land and physi- cal capital, which are tradable across firms and individuals, education and skills are not perfectly tradable. As a consequence, both the distribution and level of education enter the production function and affect the level and growth of output. Using panel data from 20 developing countries, L6pez, Thomas, and Wang (1998) demonstrated the negative association between skewed distribution of education and economic growth. When a large part of the population is not educated, the low productivity of the la- bor force discourages investment in physical capital, and economic growth suffers (see regression analyses in table A2.1 and annex 3). The distribution of education also holds strong implications for the poverty-reducing impact of growth. Ravallion and Datt (1999), using data from 15 Indian states between 1960 and 1994, found that the poverty- reducing association of growth varied according to initial conditions: growth contributed less to poverty reduction in states with initially lower literacy rates, farm productivity, and rural standard of living relative to ur- ban areas. In Kerala, where basic education is well distributed and literacy rates are the highest, for males and females, a percentage point increase in the growth rate was more strongly associated with poverty reduction. In Assam and Bihar, which had similar nonfarm growth rates to that of Kerala, but low literacy rates and higher inequality in basic education, growth contributed little to poverty reduction (figure 3.10). For example, Bihar, with the lowest female literacy rate among the states studied, 29 per- cent, showed a 32 percent gender gap in literacy rates, and 6 million chil- dren ages 6-10 were not enrolled in school between 1992 and 1993. Other states, such as Maharashtra and Madhya Pradesh, had higher growth rates but lower poverty reduction rates than that of Kerala. More than fast growth, pro-poor growth is needed for poverty reduction. If all Indian states had an elasticity of poverty reduction like Kerala, poverty, as measured by the headcount index, would have fallen at a rate of 3.5 percent, instead of 1.3 percent, a year since 1960. Improving the Efficacy of Public Spending Markets alone cannot provide equitable access to basic education by the poor. As partly a public good, education provides positive spillovers that are not fully captured by individuals and firms. However, the market fails mainly at the lower end of the income distribution: without public invest- ment in the education of the poor, society's investment in education would be suboptimal. Yet, as we have seen, public spending is only weakly associ- ated with education outcomes, partly because of a bias toward the better- off. Increasing public spending is desirable, but not enough to deal with the 66 IMPROVING THE DISTRIBUTION OF OPPORTUNITIES Growth has a stronger poverty reducing impact in states with more equitable education such as Kerala Figure 3.10. Trend Rates of Poverty Reduction and Nonfarm Output Economic Growth in India, 1960-94 Trend rate of poverty reduction (headcount index; percent per year) 3 Kerala m Punjab & Haryana m Andhra Pradesh s 2 West Bengal a Orissa Gujarat Tamil Nadu 1 - Rajasthan * Kamataka x s Maharashtra * Jammu & Uttar Pradesh * Kashmir * Madhya Pradesh Bihar a O Assam U -1 - I III 1 2 3 4 5 Trend growth in nonfarm output per person (percent per year) Note: Trend rates of growth estimated by ordinary least squares regressions of the logarithms on time. Source: Ravallion and Datt (1999). inadequate human development outcome, therefore, we now turn to im- proving the allocation and efficacy of spending. Allocate More Publlc Spending to the Education of the Poor The composition of government expenditures on education and health in- fluences human development outcomes. Public spending needs to concentrate on areas where market failure is pervasive and where positive spillover is largest: in primary and secondary schooling, especially for the poor. Given limited public resources, the balance needs to shift more to- ward investments in primary and secondary education. Additionally, the private sector and public-private partnerships should be encouraged to pro- vide higher education where market failure is minimal. 67 T H E QU A L I T Y O F G R O W T H Korea showed how a strong emphasis on primary and secondary educa- tion could eliminate illiteracy and reduce educational inequality. Korea al- located two-thirds of its public education spending to primary schooling in the 1960s and early 1970s (table 3.2). Public spending on secondary educa- tion rose from 22 percent in 1965 to 33 percent in 1990. Yet, public expen- ditures on higher education rarely exceeded 12 percent of the total public spending between 1965 and 1990. Tertiary education was mainly financed by private investments. Before the 1990s, India spent a larger share on higher education than did Korea and a smaller, but increasing, share on pri- mary education. In the mid-1990s, India increased its spending on elemen- tary schools and adult literacy programs from 20 to 31 percent of its total public spending on education, which was still far below that of Korea. To provide broader access to education and reduce the inequality, more re- mains to be done to improve the allocation of public investment in India. Measured by public spending per student, public subsidies to higher education have been falling in many countries, but not fast enough to en- able reallocation of public funds to basic education (table 3.3). Resource allocation is still biased against primary and secondary education in most countries. In the United States, the allocation of public spending has been balanced for more than 30 years, with subsidies to primary schooling at more than 20 percent of gross national product (GNP) per capita, the highest in the world. In Korea, due to the large number of students in pri- mary schools, government support per student did not sufficiently empha- size primary education in the 1960s, even though more than 60 percent of total spending was allocated to primary education. This pattern was re- versed in the 1980s, when public spending per primary school student ex- ceeded that per college student. Associated with a strong emphasis on ba- sic education, Korea was able to reduce education inequality rapidly. The United States has maintained the lowest education Gini coefficient in the world since 1965. Venezuela, in contrast, has favored higher education over basic educa- tion for more than four decades. While total public spending on education Table 3.2. Public Spending by Level of Education, Korea, Selected Years (percentage of total expenditures on education) Levels 1965 1970 1975 1980 1985 1990 Primary 64.7 67.4 52.2 47.9 44.5 43.2 Secondary 21.8 20.9 37.1 33.8 37.7 33.1 Higher education 13.3 8.2 10.7 11.4 11.5 9.6 Source: UNESCO database. 68 IMPROVING THE DISTRIBUTION OF OPPORTUNITIES Table 3.3. Public Expenditure Per Student by Level, 1960s to 1990s Education Gini Public expenditure per student coefficient (national (percent of GNP per capita) average, all levels) Country Level 1 960s 1 970s 1 980s 1 990s 1980 1990 Argentina Primary - 3.06 6.49 8.32 0.29 0.27 Secondary 26.17 10.43 - - Tertiary 59.29 23.58 17.45 19.84 Chile Primary 6.92 6.08 12.53 9.20 0.32 0.31 Secondary - 12.01 12.58 8.80 Tertiary 151.71 67.46 79.69 23.36 Korea, Republic of Primary 6.21 7.86 12.79 14.86 0.34 0.22 Secondary 8.64 7.39 10.76 11.88 Tertiary 36.67 28.02 10.49 5.83 Mexico Primary 4.34 - 3.97 7.18 0.50 0.38 Secondary - - 8.61 13.93 Tertiary 70.72 - 32.43 35.66 United States Primary 22.05 28.45 26.28 19.83 0.12 0.15 Secondary - - 18.77 23.86 Tertiary 73.73 58.84 37.85 22.91 Venezuela, RB Primary 8.50 7.37 4.80 2.39 0.39 0.42 Secondary 21.26 17.60 18.34 7.07 Tertiary 121.76 100.00 65.74 37.38 - Not available. Sources: Public expenditure data are from the UNESCO database; education Gini coefficients are from Thomas, Wang, and Fan (2000). has increased from 4.3 percent of GNP in the 1970s to 5.1 percent in the 1980s and 4.6 percent in the 1990s, its allocation has worsened. In fact, the subsidies to primary and secondary education were reduced in the 1990s. This misallocation of public resources might partially explain the worsen- ing of the education Gini coefficient in the 1990s. The Interaction between Demographics and Education Public spending per primary-school-age student in Korea rose more than ten- fold between 1970 and 1995 as population growth rates slowed and the economy expanded (table 3.4). Public spending per secondary student also rose. Rapid economic growth, together with a stabilizing and even declining student base, meant that far more resources were being devoted to fewer chil- dren, allowing dramatic improvements in the quality of primary education. 69 I tI E ) L A 1. I T } C) F C; R ( \\' T :i Table 3.4. Public Current Expenditure Per Student, India and Korea, Selected Years Country Level 1965 1970 1975 1980 1985 1990 1995 Amount (1995 US$ per student) Korea, Republic of Primary 92 207 182 386 701 955 1,890 Secondary - 223 134 339 541 786 1295 Tertiary 545 757 622 589 546 460 599 India Primary 8 10 20 23 29 39 39 Secondary - 54 35 34 38 - 43 Tertiary - - - 189 227 299 260 Percentage of GDP per capita Korea, Republic of Primary 6.3 9.5 6.3 10.2 13.5 12.0 17.4 Secondary - 10.3 4.6 9.0 10.4 9.9 11.9 Tertiary 37.2 35.0 21.5 15.6 10.5 5.8 5.5 India Primary 4.3 4.8 9.2 9.7 10.6 11.8 9.9 Secondary - 24.9 15.8 14.8 13.9 - 11.0 Tertiary - - 81.8 84.0 90.3 66.4 - Not available. Note: Dollar amounts are not comparable across countries as they are not in PPP dollars, but are comparable over time. Sources: Calculated from UNESCO and World Bank data. In India, rapid population growth and constraints on public funding meant that a quantity-quality tradeoff was likely to occur. In 1995, India spent US$39 (in 1995 constant dollars) per pupil in primary schools, or 10 percent of its GDP per capita; Korea spent 17 percent (table 3.4). In Tamil Nadu, India, enrollment in primary and middle schools expanded 35 percent between 1977 and 1992, a major achievement, but the pupil- teacher ratio rose from 36 to 47 and school conditions worsened. Student achievement suffered as a result (Duraisamy and others 1998). These rela- tionships point to a need to consider the interaction between demograph- ics and education policy and a need for policies focusing on education of girls and women, education to improve reproductive health, and volun- tary family planning as part of an overall development strategy centered on people (see also box 3.3). Improve the Mix of Public and Private Spending Korea also achieved a good mix of public and private financing in educa- tion. Since the mid 1960s, private colleges and universities have accounted 70 I NT I'-OV IIN (f I H [)IS Ft, R IBUT ION OF O PPORT U NIT IE S Box 3.3. Population and Developmeint The link between population growth and economic contributed approximately 22 percent to changes in development is a subject of contentious debate. output growth between 1960 ,and 1992, a figure that The 1960s and 1970s were dominated by pessimis- corresponds to approximately 21 percent of the aver- tic, and sometimes alarmist, predictions that rapid age growth of per capita output, which was measured population growth would lead to famines, resource at 1.5 percent. exhaustion, deficiencies in saving, irreversible envi- Various components of demographic change ronmental, damage, and ecological collapse (Ehrlich have been successfully introduced into growth mod- 1968). The population optimists believed that rapid els. Bloom and Williamson (1998) showed that population growth would allow countries to capture rapid demographic tamnition in East Asia led to fast economies of scale and promote technological and growth in the working,age population between institutional innovation (Simon 1976). In the 1965 and 1990, expanded the per capita productive 1980s, the alarmist views were replaced by moder- capacity, and contributed to the East Asia eco- ated, time- and country-specific assessments of the nomic miracle. Other economic policies also facili- net negative impacts of rapid population growth, tated the East Asians to realize the growth potential which were considered to be small. Only weak or of the demographic transition. inconclusive links were found between demo' Less evidence was available on the lirnk between graphic changes and economic growth (Bloom and demographic change and poverty until recently. Freeman 1988; Kelley 1988). However, if rapid population growth has a negative More recent investigations revealed fairly large, effect on economic and wage growth, it would nega, negative effects of rapid population growth and re- tively affect poverty as well. Eastwood and Lipton lated demographic components on per capita eco- (1999) found that higher fertility increases poverty nomic growth. Kelley and Schmidt (1999) found both by retarding growth and by skewing the distri- that rapid population growth exerted a fairly strong, bution against the poor. In addition, evidence shows adverse impact on the pace of economic growth in that public sector programs targeted at the poor, such 89 countries between 1960 and 1995. The positive as basic education and health care programs, have impacts of density, size of population, and labor force contributed to reduced poverty. Rapid population entry were dominated by the costs, of rearing chil- growth will dilute the intersity of public investment, dren and maintaining an enlarged youth dependency and as a consequence make quality of service im- age structure. Declining mortality and fertility each provements more difficult to achieve. Sources: Bloom and Williamson (1998); Eastwood and Lipton (1999); Kelley (1998); Kelley and Schmidt (1999). for more than 70 percent of enrollments, private secondary institutions for more than 40 percent. Households assume a large share of educational costs, between 30 and 50 percent, depending on student education level. Tuition and related fees account for 40 percent of in-school expenditures for middle school, but rise sharply to 72 percent and more for high school and college students. The most effective public-private mix depends on the extent of mar- ket failures and a variety of other factors. Higher education is crucial for technological progress and productivity growth, but it can be considered a 71 T H E QU A L I T Y O F G R O W T H private good, because most of the returns can be internalized by individu- als and firms. Whereas primary and secondary education have large spillover effects that are not fully captured by individuals and firms. Thus while government has a direct role in primary and secondary education, it needs to encourage private investments and public-private partnerships in higher education. The United States, for example, provides valuable ex- periences in this regard. The policy environment, which can be defined by the degree of open- ness to trade and investment, for example, affects the demand for skilled workers and as a consequence people's willingness to pay for education. The quality of service provision for education, which is related to institu- tional capacities, also affects the willingness to pay. Similarly, the public- private mix in health care also depends on the nature of services and the degree of market failures in particular subsectors (Filmer, Hammer, and Pritchett forthcoming). One successful intervention is the Quetta Girls Fellowship Program in Pakistan. Launched in 1995, the pilot project aimed at determining whether establishing private schools in poor neighborhoods was a cost- effective way of expanding primary education for girls. The program en- couraged private schools controlled by communities, ensuring them gov- ernment support for three years. An evaluation analysis indicates that the program increased girls' enrollments by 33 percentage points, and boys' en- rollment rose as well. Such programs offer promise for increasing enroll- ment rates in poor urban areas (Kim, Alderman, and Orazem 1999). Decentralize Decisionmaking and Encourage Participation How decisions are made also affects the efficacy of public services. Where institutional capacity is low, public spending on centrally planned and or- ganized interventions is likely to be ineffective. Many countries are mov- ing to decentralized decisionmaking to better match expenditures to local needs. Empirical evidence on the benefits of decentralized school man- agement was rare until recently. A recent evaluation of El Salvador's EDUCO program (Community-Managed Schools Program) shows that enhanced community and parental involvement in EDUCO schools has improved students' language skills and diminished student absences, which may have long-term effects on achievement (Jimenez and Sawada 1999). Other studies have also shown that community-managed schools achieved better results in Indonesia and the Philippines (James, King, and Suryadi 1996; Jimenez and Paqueo 1996). Several counties have been experimenting with voucher programs, which transfer resources to parents to help pay private school tuition. Colombia 72 IMPROVING THE DISTRIBUTION OF OPPORTUNITIES used a national voucher program from 1991 to 1997 to decentralize manage- ment and expand enrollment. The program was meant to address the defi- ciencies in the public education system, especially the low transition rate from primary to secondary schools by the poor. Only the poor were qualified for vouchers, which avoided subsidizing the wealthy as in previous voucher programs. Participation was a problem, however; only 25 percent of Colombia's municipalities joined the program, limiting the benefits. A care- ful evaluation of the program found that demand for secondary education and availability of space in private schools were key determinants of munici- pal participation (King, Orazem, and Wohlgemuth 1999). Such voucher pro- grams are potentially beneficial to the poor. In countries with corrupt and predatory governments, however, decen- tralizing decisionmaking may not be the answer. Corrupt officials are likely to reallocate public resources from the poor to elite interest groups, subsi- dizing the types of social services that benefit the rich. Empowering people to influence policy through democratization and a greater role for civil soci- ety, and encouraging greater participation of the community and families are steps in the right direction (see chapter 6 on the role of participation and civil society in anticorruption and better governance). Making Education More Productive Improving the productivity of education for the poor takes more than in- vestments in their education. To be more productive, the poor must be able to combine their human capital with other productive assets such as land, equity capital, and job opportunities in open and competitive markets. Distribute Land more Equitably The poor are not just income poor; they also lack assets. In agrarian econo- mies, disadvantaged households are usually landless or land poor. In South Asia, southern Africa, and much of Latin America, poverty is highly corre- lated with landlessness (figure 3.1 la). Income inequality also seems to be associated with inequality in landholding (figure 3.1 Ib), although data on land ownership is weak. Land reform has many benefits for growth and poverty reduction, as suggested by empirical studies discussed later. In societies where a large seg- ment of the population does not have access to the productive resources of the economy, strong demand for redistribution gives rise to civil unrest. Studies suggest that inequality in land ownership and income are correlated with subsequent lower economic growth (Alesina and Rodrik 1994); a one 73 T H E OU A L I T Y O F G RPO W T H Figure 3.11a. Poverty and Landholding, Bangladesh, 1988-89 Headcount index (percent) 80 60 U 40 U 20 0- I 0 5 10 15 Median landholding (acres) Source: Ravallion and Sen (1994). Figure 3.11b. Income Share in the 1980s and Land Gini Coefficients in the 1960s Income share of the lowest quintile in the 1980s 0.12 - 0.10 - 0.08 -U. 0.06 - H * U 0.04 -* 0.02 - 0.00 - I 0.0 0.2 0.4 0.6 0.8 1.0 Land Gini coefficient in the 1960s Note: Data are country-specific decade averages. N = 27. r = -0.40. Source: Deininger and Squire (1996). 74 IMI'ROVING THE DISTRIBUTION OF OPPORTUNITIES standard deviation increase in equality is associated with increases in growth of one-half to 1 percentage point (Persson and Tabellini 1994). Other studies showed that the initial inequality of assets, measured by land distribution, is more significant than income inequality in affecting subse- quent growth (Deininger and Squire 1998; Li, Squire, and Zou 1998; Lundberg and Squire 1999). Still others have found initial land inequality, along with initial education inequality, to have strong negative links to economic growth and to the income growth of the poorest (Birdsall and Londonio 1998). In addition to being negatively correlated with growth, land inequality also appears to reduce the positive effect of human capital on growth through interaction effects (Deininger and Olinto 1999). Redistributive land reform gives land to more efficient producers and re- duces credit market imperfections, leading to improved investment decisions by the poor. Greater wealth, as measured by land ownership, also provides a safety net for the poor against external shocks and increases their ability to participate in the political process (Binswanger and Deininger 1997; Binswanger, Deininger, and Feder 1995). Ravallion and Sen (1994) noted that redistribution from land-rich to land-poor households would reduce ag- gregate poverty in rural Bangladesh. They also found that transfers from the budget would have the greatest impact on poverty if concentrated on landless and marginal farmers (see annex table A3.5 for a selective literature review). Widespread ownership of land improves not just equity, but also produc- tivity (Berry and Cline 1979) and efficiency (Banerjee 1999). Better land rights have facilitated investment in Ghana (Besley 1995), and possession of legal land ownership documents in Thailand has significantly impacted farm- ers' agricultural performances (Feder 1987, 1993). Many East Asian econo- mies have widespread landholdings, a result of traditional ownership or land reform. In Korea, confiscated land at the end of World War II was first dis- tributed to the tillers. Then in the 1950s the government distributed landlord properties, with nominal compensation, to 900,000 tenants, effectively elimi- nating tenancy. In Taiwan, China, the government obtained land from land- lords in the early 1950s, compensated owners with shares in state enterprises, and then sold the land to tillers on favorable terms. In China, the household responsibility system introduced in 1979 as- signed collectively owned land to households for up to 15 years. The system, which was renewed for another 30 years in 1998, tied rewards more closely to farming efforts. Together with price and other reforms, the initiative resulted in a 5.7 percent annual rise in average grain yields from 1978 to 1984 and 1.8 percent thereafter. Nearly half of the total output increase in the period can be attributed to the household responsibility system (Lin 1992). One study 75 T H E QU A L I T Y O F G R O W T H found that access to land can improve nutritional status in China, because it serves both as a means of generating income and as a source of cheap calories relative to the market (Burgess 2000). Another study found that in rural China, wealth, especially land, is distributed more equally (Gini coefficient of 0.31) than income (Gini coefficient of 0.34). The main source of rural in- come inequality is wage income rather than the returns from land, an atypi- cal pattern for a developing country (McKinley 1996). Land reform is contentious and politically difficult. Market-assisted land reform has emerged in recent years as an alternative to traditional land reform, and is being implemented by Brazil, Colombia, and South Africa. The basic idea is that the state gives qualified, landless people a grant or a subsidized loan to buy land. This market-assisted approach differs from fully compensated land reform in two ways: there are neither explicit targets for land distribution nor fixed time schedules. In addition, the reforms are demand driven; people who want the land most will come forward to buy it. Some researchers contend that market-assisted land reform has advantages, especially if combined with microcredit, extension programs, and complementary actions that facilitate agricultural cooperatives and contract farming (Banerjee 1999). The success of the programs can be enhanced if accompanied by efforts to make land mar- kets more transparent and fluid and to involve the private sector (Deininger 1999). While it is still too soon to reach definitive conclusions on the costs and benefits of these reforms, some other studies have found that this ap- proach benefits large landholders because land prices are likely to be bid up, requiring the poor to pay elevated prices (L6pez and Valdes 2000). Distribute Equity Capital and Foster Competition A case can also be made for better distribution of equity through employee ownership plans. In industrial countries, employee stock ownership plans have been positively associated with firms' performances. Firms in the United States have used employee ownership plans in restructuring. For ex- ample, United Airlines negotiated significant wage concessions in return for a majority equity stake for employees. By communicating the benefits of the restructuring plan to its investors and employees, the company reduced the up-front restructuring cost, enhanced the effects of the restructuring, and thereby created additional shareholder value. Both investors and em- ployees have benefited (Gilson 1995). In countries hit by the recent financial crises, the sale of equity shares to employees may provide a way to recapitalize companies in desperate need of capital, and can also redistribute wealth and risks. Where restruc- turing leads to retrenchments, laid-off workers may be given equity shares 76 IMPROVING THE DISTRIBUTION OF OPPORTUNITIES in lieu of severance pay, and so benefit from the companies' restructuring and recovery. Employee ownership plans can also help reduce workers' re- sistance to restructuring (Claessens, Djankov, and Klingebiel 1999). Pro- viding microfinancing to laid-off workers to establish small enterprises is another way to empower them to build physical and financial capital. Privatization offers additional opportunities for redistributing equity. Because public enterprises were built using tax revenues, a certain propor- tion of the equity shares can justifiably be distributed or sold at a discount to taxpayers during privatization. Properly designed privatization programs can reduce asset inequality and poverty. For example, using proceeds from the privatization of the six largest state enterprises, Bolivia established a pool of financial assets to fund a minimum flat pension for everyone in the country. While the amount provided is small, the program will reach the most vulnerable in society: the elderly poor who are unable to save for re- tirement. Hungary used its receipts from privatization to repay foreign debt, which raised its sovereign debt rating, reduced its interest payments, and benefited all citizens (Kornai 2000). Privatization entails efficiency gains as well as social losses, and soci- ety must maintain a balance between the efficiency gains and social losses (and compensate the losers), if the gains are to be sustainable. After privatization in Mexico, there was a 24 percentage point increase in the ratio of operating income to sales. Of those gains in profitability, 10 per- cent were due to higher product prices, 33 percent to a transfer from laid- off workers, and the remaining 57 percent to productivity gains (La Porta and Lopez-De-Silanes 1999). To compensate those who suffer losses as a result of privatization, equity shares in lieu of severance pay could be dis- tributed to laid-off workers, or other forms of income transfers could be financed by taxation. Competition and regulation are vital for a market economy. The effi- ciency of a market economy depends on both private property and com- petitive markets, but many developing and transition economies lack both. Before and during privatization, competition and a regulatory framework must be introduced (Stiglitz 1999). Evidence from the United Kingdom shows that when big public enterprises were privatized, antitrust regula- tions were crucial to ensure transparent, equitable, and efficient allocations of resources (see also Herrera 1992). Privatizing large public firms that have a natural monopoly without first setting up antitrust regulations, as was done in Russia, can worsen the distribution of wealth and income. And it could create powerful, entrenched interests that undermine the possibility of viable regulation and competition in the future, and block further broad-based reform measures (Kornai 2000). 77 H HF CLI A L I T Y O F GR OP W T Fi Combine Human Capital with Opportunities in Open Markets The urban poor must hire out their labor. Thus, the creation of job op- portunities is critically important to the productive use of their human capital and to poverty reduction. The World Development Report 1990 (World Bank 1990) proposed a strategy of broad-based, labor-intensive growth to generate income-earning opportunities for the poor. Some economies have pursued this strategy and more-they have combined investment in learning and education with openness, forming a virtuous circle. Examples include Japan in the 1950s and Hong Kong, China; Ko- rea; Singapore; and Taiwan, China, from the 1960s through the 1980s. The accumulation of knowledge influences a country's trade and com- petitiveness, and trade enhances the accumulation of knowledge, especially through imports. Lucas (1993) noted that to sustain knowledge accumula- tion, a nation must be outward oriented and a significant exporter. Young (1991) and Keller (1995) found that trade itself is not an engine of growth, but must operate through some mechanism, such as the formation of hu- man capital, to affect growth. Market openness facilitates technological progress and capacity building through various modes of leaming, such as the importation of capital and in- termediate goods, learning by doing, and on-the-job training. Foster and Rosenzweig (1995) found strong evidence of leaming-by-doing and leaming spillovers: farmers' own experiences and that of their neighbors with high- yield varieties significantly increased profitability. Farmers with experienced neighbors are significantly more profitable than others, and the spillover ef- fects associated with leaming from others are small, but not unimportant. The link between overall economic policies and the impact of educa- tion is clear. The World Development Report 1991 (World Bank 1991) found that among 60 developing countries from 1965 to 1987, economic growth rates were especially high for countries with high levels of educa- tion, macroeconomic stability, and market openness. The impact of trade openness on long-term growth thus depends on how well people can absorb and use the information and technology accompanying trade and foreign investment. Increases in the stock of human capital tend to accelerate growth during market reforms and under an outward-oriented economic structure, but in their absence, education has no significant impact on growth. The growth ef- fect of an interaction between openness and education was robust (L6pez, Thomas, and Wang, 1998; see also chapter 2 and annex 2). Similarly, for 1,265 World Bank projects, Thomas and Wang (1997) found that the rate of return was 3 percentage points higher in countries with both a more educated /8 I P ROVING -r TF FE DI sr R IhLI I I0 N O O r P 110 R-r TI N I r EIL S labor force and a more open economy than in countries that had only one or the other (figure 3.12 and annex table A3.4).8 Protect Workers against Shocks The urban poor usually lack adequate human capital for all but unskilled work. With increased openness and globalization, job opportunities for unskilled workers have become more scarce and incomes more volatile. Diwan (1999) found that labor shares in GDP have been falling for more than 20 years in most regions. Consistent with this evidence, unemploy- ment rates in Latin America have risen since the end of the 1980s. In 1989, only 5 or 6 of every 100 Latin Americans willing to work were un- employed; by 1996, nearly 8 of every 100 were not working. Unemployment rose in East Asian countries hit by the recent financial turmoil, from previously modest levels to 4.5 percent in Thailand, 5.5 percent in Indonesia, and 7.4 percent in urban Korea (World Bank 2000a, p. 59). Per- haps even worse was the fall in real wages, because the poor could not afford to remain unemployed. Real wages fell in 16 of 22 recessionary episodes in Education and openness interact and increase investment returns Figure 3.12. Education, Openness, and Economic Rates of Return in 1,265 World Bank Projects Economic rate of return (percent) 18.0 18 16 14 = Mo r e~~~~~~~~Mr open 12 Low Less open education High education Note: Economic rates of return are from the evaluation database of the World Bank's Operations Evaluation Department. Education is measured by the average level of schooling of the labor force, and openness by the logarithm of the foreign exchange parallel-market premium. Sources: Thomas and Wang (1997); annex 3. 79 THF QU A L I T Y O F G R O W T HF Latin America during the 1980s and 1990s. In 18 cases, after two years real wages remained lower than their precrisis levels (Lustig 1999). In East Asia, manufacturing real wages fell by 4.5 percent in Thailand, 10.6 percent in Ko- rea, and 44 percent in Indonesia between 1997 and 1998 (World Bank 2000a, p. 57). As a result of both a decline in real wages and in employment growth, labor shares in GDP fell sharply following the financial crises, perhaps because labor is less mobile than capital, and so is forced to bear a large share of the financial burden of crisis resolution (Diwan 1999). Urban unskilled workers are most vulnerable to external shocks, struc- tural adjustment, and economic downturns. Lacking adequate human capi- tal, they are often unable to adjust to changes in labor market demand. The problem is exacerbated by labor market distortions and weak labor market institutions that further hamper labor market adjustments. Labor market distortions need to be checked: the existence of child labor and distorted wage structures discourage demand for education. Governments need to help build labor market institutions and provide the labor market informa- tion that the poor need. There is also the need to train or retrain displaced workers and increase their mobility across sectors. Ghana trained more than 4,000 people in vo- cational schools or apprentice programs, which offered instruction in such skills as dressmaking, electrification, and carpentry. Participants received certificates and tools after completing the training, giving them the human and physical capital to begin work immediately as self-employed workers. Many labor exchange centers were established in China to retrain and re- deploy displaced state sector workers in the private sector. Some of the pro- ceeds from liquidating the assets of bankrupt state enterprises were used to redeploy unemployed workers. Such measures help to ease the rise in social tensions and inequality during transition periods.9 Conclusions For growth to have an impact on poverty reduction, the assets of the poor must be augmented. This can be achieved either by investing in new as- sets, specifically, human capital, or by redistributing existing assets. This chapter has focused on investing in new assets by examining the quality and distribution of education and the causes and consequences of, and remedies for, large dispersions in educational attainment. When the quality of schooling is low and educational inequality is high, the poor are hurt most because human capital is often their main asset. Inadequate in- vestment in the human capital of the poor exacerbates and perpetuates poverty and income inequality. 80 IMPROVING T1 HE DISTRIBUTION OF OppOfR -ruNI-rIES Improving the allocation of public expenditure in education is a key. Despite making efforts to this end, many countries have not been able to concentrate public investment on primary and secondary education. Inap- propriate allocations of public expenditures have led to low average attain- ment per dollar spent on students, which affects mostly the poor. Govern- ments need to reallocate public expenditure toward basic education, while at the same time enabling the private sector and public-private partnerships to increase efforts in higher education. Countries have compelling reasons to strengthen education at all levels. It can augment the poverty-reducing aspect of growth, in addition to improving welfare directly. It enables coun- tries to participate effectively in the global economy. Investing in education alone will not guarantee successful develop- ment or poverty reduction. Thus, this chapter went beyond education to issues related to the use of human capital, namely, the distribution of land and other productive assets and economywide policies. To reduce pov- erty, countries need a multidimensional strategy centered on people. There is the need to ensure access to education and health services and distribute them well; to facilitate fuller use of the human capital of the poor; and to empower the poor with land, equity capital, training, and job opportunities made possible by opening to international trade, invest- ment, and ideas. Notes 1. On the importance of asset distribution, see, for example, Ahluwalia (1976); Birdsall and Londofno (1997); Chenery and others (1974); Deininger and Squire (1998); Kanbur (2000); Knight and Sabot (1983); Lam and Levison (1991); Lanjouw and Stern (1989, 1998); Li, Squire, and Zou (1998); Ram (1990); Ravallion and Datt (1999); and Sen (1980, 1988). See annex table A3.5 for additional evidence. 2. Some arguments here apply to health, but due to space limits, this chapter fo- cuses only on education. 3. Certain assumptions apply here. This conclusion holds if there is a competitive market and two factors of production: physical and human capital. It is also true if human capital is decomposed to skilled and unskilled labor. 4. These measures, however, are sensitive to national promotion policies. Scores on intemationally comparable tests represent an improvement over traditional indicators, but they are available for only a few developing countries, and they are not comparable over time. Due to these problems they are not used here. 5. The same is true for industrial countries. A study estimated the cost of different kinds of national class size reduction policies in the United States and found 81 T iI E QU A L I T Y 0 F G R O W T 1 i the operational costs could be as large as US$2 billion to US$7 billion dollars a year (Brewer and others 1999). 6. There was heated debate over the "equity of what?" Sen (1980) sees individuals' levels of functionings, such as literacy and nutrition, as attributes to be equal- ized. Others see the opportunities people face as the attribute to be equalized (Arneson 1989; Cohen 1989; Roemer 1993). Yet others consider the amount of resources as the attribute to be equalized (Dworkin 1981). 7. Many studies have compared income, land, and wealth Gini coefficients (for ex- ample, Leipziger and others 1992 for Korea). However, no study has compared education Gini coefficients with those of income and land. Income Gini coeffi- cients are available only for selected years (Deininger and Squire 1996): 1970 1977 1983 1990 1992 1970 1976 1980 1985 1988 India 0.30 0.32 0.31 0.30 0.32 Korea 0.33 0.39 0.39 0.35 0.34 8. The cross-country, project-level data set included variables on education, per capita income, openness, government expenditure, and project performance. The project data covered 3,590 lending projects in 109 countries evaluated by the Operations Evaluation Department for 1974-94, with a rating of overall performance (satisfactory/not) and economic rates of return. 9. For more discussion on labor market and social protection issues, see Basu, Genicot, and Stiglitz 1999; Kanbur (2000); World Bank (1994) on old age cri- sis; and World Bank (2000i). 82 C H A P T E R 4 SUSTAINING NATURAL CAP ITAL If we truly care about the future of our planet, we must stop leaving it to "them" out there to solve all the problems. It is up to us to save the world for tomorrow; it's up to you and me. -Jane Goodall, Reason for Hope N t atural capital contributes enormously to human development and welfare. The term natural capital encompasses the sink functions, that is, air and water as receiving media for human-generated pollution and source functions, that is, production based on forests, fisheries, and mineral ores. Protecting sink functions is essential for human health. Protecting the productive or source functions is critical to the economic security of many who depend on these resources for their livelihoods. High-quality natural capital contributes to welfare indirectly as an essential part of the sustained production of economic goods and services. It also contributes to welfare directly as people derive enjoyment from pristine surroundings, old growth forests, and clean lakes and rivers in which to swim and fish. Chapter 2 demonstrated the importance of human, natural, and physical capital to economic growth and welfare. Because of imperfect substitution, these assets need to grow at undistorted or fairly balanced rates to achieve sustainable economic growth. Distorted or unbalanced growth-signaled by especially rapid accumulation of physical capital, slow accumulation of hu- man capital, and a draw down of natural capital-increases the volatility of growth, disproportionately hurting the poor. An economy fostering unbal- anced growth is likely to suffer long-run stagnation (see annex 2). Economies that derive much of their income from natural resources cannot sustain growth by substituting physical capital accumulation for 83 T H E QU A L I T Y OF G RPO W T H deteriorating natural capital (L6pez, Thomas, and Thomas 1998). Envi- ronmental degradation is likely to be most devastating for the poor, who often depend on natural resources for their income, with few possibilities for substituting other assets. Especially in the long run, growth ap- proaches that pay attention to environmental quality and resource use ef- ficiency contribute to accumulation, investment, economic growth, and human welfare (Munasinghe 2000). Yet countries throughout the world have overexploited their forests, fish- eries, and mineral wealth and polluted their water and air to accelerate short- term economic growth, with policymakers noting that their approach would increase the welfare of their citizens. While much natural capital has been sacrificed through deforestation, loss of biodiversity, soil degradation, and air and water pollution, access to safe water and sewage treatment and sanitation facilities has often shown improvements as economies grow. This chapter ex- amines the reasons why natural capital tends to be abused and overexploited, especially during rapid economic growth, and what measures can be taken to correct the negative spiral of environmental decline. The suitability of corrective actions will depend on the nature of the problem and the economic and institutional setting. For example, air quality may be improved by levying a pollution tax on industrial emissions, while production efficiency based on natural resources may be enhanced by such measures as assigning clear property rights to land or giving transferable quo- tas to fishers. Successful outcomes require active, yet selective, intervention by the state in collaboration with the private sector and civil society. Extensive Losses Air pollution from industrial emissions, car exhausts, and fossil fuels burned in homes kills more than 2.7 million people every year, mainly from respi- ratory damage, heart and lung diseases, and cancer (UNDP 1998). Of those who die prematurely, 2.2 million are rural poor exposed to indoor air pollu- tion from the burning of traditional fuels. Air pollution also reduces eco- nomic output because of the loss of productive workdays. (See table 4.1 to understand the magnitude of the losses due to air pollution in different parts of the world. The numbers are simply meant to illustrate the possible impacts of environmental pollution; they are far from being noncontrover- sial estimates of environmental damage.) Health costs associated with waterborne diseases and water pollution are also profound. In 1992, more than 2 million children under the age of five died from diseases caused by dirty water. Table 4.2 reports the findings of some studies on the health burden of water- and sanitation-related defi- ciencies and pollution effects.' 84 SUSTAINING NATURAL CAPITAL Table 4.1. Annual Health Costs Associated with Air Pollution Region and city ympact Cost China: 11 major cities Economic costs of premature More than 20 percent of urban mortality and costs of illness income East Asia: Bangkok, Jakarta, Number of premature deaths due 15,600 Seoul, Kuala Lumpur, to air pollution above WHO- Manila defined safe limits East Asia: Bangkok, Jakarta, Economic costs of premature More than 10 percent of urban Kuala Lumpur mortality and costs of illness income Newly independent states: Number of premature deaths due 14,458 Russian Federation (Volgograd); to air pollution above WHO- Armenia (urban areas); Azerbaijan defined safe limits (national); Kazakhstan (national) Note: The estimates are based on different studies applying different methodologies and are not comparable. In many cases, excess mortality is estimated using dose-response functions estimated for industrial economies for marginal changes in pollution, but then applied to nonmarginal changes, which tends to overestimate mortality reductions. Some studies use PPP-adjusted willingness to pay data from industrial economies; others use the cost of illness approach. Sources: World Bank (1997a, 1999f). Table 4.2. Annual Health Costs Associated with Waterborne Diseases and Pollution Region or country Impact Cost Vietnam Infant deaths avoided yearly by providing access 50,000 to clean water and sanitation China Premature deaths due to water-related diseases such as 135,000 diarrhea, hepatitis, and intestinal nematodes East Asia Cost of waterborne diseases US$30 billion a year Moldova Premature deaths 980-1,850 Loss of workdays due to illness 2-4 million a year Source: World Bank (1997a,c; 1999f). Toxic effluents, such as dioxins, pesticides, organochlorines, grease, oil, acids, alkalis, and heavy metals such as cadmium and lead, from factories, mines, and chemical plants have contaminated major bodies of water in all parts of the world. Workers, farmers, and others who come in contact with the contaminants face severe health hazards. As with air pollution, the poor are hurt the most. As many as 25 million poor agricultural workers in the de- veloping world ( 11 million in Africa alone) are poisoned by pesticides every year, and hundreds of thousands die (UNDP 1998). Fisheries, which provide a main source of protein for the poor, are also being destroyed by industrial 85 1 II E QU A I I T Y 0 F G, R 0 WV r II discharges and water pollution. In Manila Bay, fish yields have declined 40 percent in the last 10 years (UNDP 1998). For a disturbing description of en- vironmental degradation in India, as reported in the press, see box 4.1. Recent estimates using DALYs suggest that premature death and illness due to major environmental health risks account for approximately one- fifth of the total burden of disease in the developing world (Murray and L6pez 1996).2 Of the main environmental risks, which include poor water supply, inadequate sanitation, indoor air pollution, urban air pollution, ma- laria, and agro-industrial chemicals and waste, 14 percent of the total dis- ease burden is caused by poor water supplies, inadequate sanitation, and in- door air pollution. They affect predominantly children and women in poor families (Lvovsky and others 1999). Overexploitation and degradation of natural resources are also enor- mous concerns. Soil degradation is a problem everywhere, especially in Asia and Africa. In China the costs may run as high as 5 percent of GDP (ADB 1997) and for several African countries the annual costs are 1-10 percent of agricultural GDP (Bojo 1996). While the annual losses are worrisome, the cumulative effects are alarming. Desertification, a direct consequence of soil degradation, is estimated to cost US$42 billion a year 0 (1999)000S000 0 reporte the folwn ifraiote, and the0E *es have\0 none\ atal g f0 ~~~~~~~~~~~~~~~~~~~~~~7 the Oomti6River. matri8oeta6wc h aelmtsei *Ms ee ie aet ooiltms hs SUS-FAINING NATURAL CAPI'TAL in lost agricultural productivity alone. It directly puts some 250 million poor people at risk of starvation from reduced crop yields (UNDP 1998). At least 10-12 million hectares of forest land disappear each year. Unsus- tainable logging practices and forest conversion to agriculture and pasturelands account for the bulk of the losses (Brown, Flavin, and French 1998; World Bank 1999d). Declining yields of timber and nontimber forest products, reduced soil and water conservation services, and loss of carbon se- questration functions translate into net economic losses of US$1 billion to US$2 billion a year to the global economy (calculated from World Bank data). In 1997, forest fires caused haze and smoke-related damage of US$4 billion in Indonesia and extensive damage in neighboring Malaysia and Singapore (EEPSEA 1998). Future generations will bear the cost of the asso- ciated, if hard to quantify, loss of biodiversity. Like the forest fires in Indonesia, the impacts of local environmental ne- glect are not confined to political boundaries. Witness the increasing deserti- fication, coastal zone degradation, global climate change, transboundary acid rain, and depletion of the ozone layer (GEF 1998; Watson and others 1998). Global climate change during the 21st century could result in increases in the intensity and frequency of floods and droughts, inundation of low-lying coastal areas, more frequent outbreaks of infectious diseases, and accelerated dieback of forests. Climate change will also hurt food security by reducing ag- ricultural output in developing countries and poses a threat to human health and safety. It could cost the world economy as much as US$550 billion a year and developing countries are likely to bear an inequitable proportion of the burden (Furtado and others 1999). "Genuine savings" provides a useful concept to capture the degradation of natural capital that can be used to assess the environmental health of countries. Genuine savings equal gross domestic savings minus depreciation of physical capital, minus depletion of minerals and energy, minus net depletion of forests, minus pollution damage, plus investments in human capital. For the developing world as a whole, in 1997 gross domestic savings were 25 percent of GDP. Net domestic savings (after correcting for the de- preciation of physical capital) were about 16 percent of GDP, but corrected further for depletion of natural capital (such as forest, energy, and minerals) and damage from carbon dioxide emissions, the domestic savings were a little more than 10 percent of GDP (World Bank 1999e). After including human capital investment, the genuine savings rose to about 14 percent. This includes Nepal, where forest depletion alone was estimated at 10.3 percent, overwhelming the country's gross domestic savings of 10 percent, and the Russian Federation, where depletion of energy resources (oil, coal, and natural gas) reduced savings by more than 9 percent per year. 87 T H E QU A L I T Y O F G R O W T H Significant Benefits of Environmental Action From an economic point of view, not all pollution is to be fully controlled nor all natural resource degradation is to be fully reversed. Pollution and natural resource degradation are to be controlled to the point where the marginal (social) damages equal the marginal (social) costs of abatement or control, that is, the optimal level of environmental protection. The present discounted cost of providing everyone in China with access to clean water within 10 years, for example, is US$40 billion, and the present value of benefits is US$80 billion to US$100 billion (World Bank 1997a). Doing the same in Indonesia would cost an estimated US$12 billion to US$15 billion, with corresponding benefits of US$25 billion to US$30 bil- lion. In Moldova piped water of adequate quality would cost US$23 million to US$38 million a year to provide, but would bring benefits of US$70 mil- lion to US$120 million (World Bank 1999f). Controlling air pollution in China would cost an estimated US$50 billion, but would yield benefits of about US$200 billion in reduced illness and death (World Bank 1997a). With the payoffs so large, why do environmental degradation and de- struction continue?3 The main reason is that private returns on invest- ment in environmental protection are significantly smaller than private costs (Dasgupta and Maler 1994; Hammer and Shetty 1995). Many of the benefits are distributed to society at large, now and in the future, rather than to the private agent making the investments. Thus, indi- viduals, who see only their own private, short-term gains, seldom factor the cost of degradation-spread inequitably across the current genera- tion and affecting future generations as well-into their decisionmaking. This classic case of externalities and market failure provides strong justi- fication for public policy actions to create markets or marketlike condi- tions that align private incentives with the social costs and benefits of providing environmental services. Policy distortions reflecting undervaluation of the environment con- tribute to pollution and degradation (Dasgupta and Maler 1994). For ex- ample, agricultural subsidies on inputs and price supports for outputs make forest management noncompetitive and create pressures to convert forests to pastures. Energy subsidies to keep consumer prices low contrib- ute to overconsumption and excessive pollution. Tax exemptions, subsi- dies to firms operating on frontier land, road construction in ecologically fragile areas, and a host of other short-sighted policies also lead to degra- dation and mismanagement of resources and threaten vulnerable popula- tions who live in these areas (Chomitz and Gray 1996; Cropper, Griffiths, and Mani 1997). By removing subsidies and imposing environmental 88 SUSTAINING NATURAL CAPITAL taxes, policy reforms can alleviate distortions and allow prices to reach their optimal level. Several other contributory factors and false notions stand in the way of efficient management of natural capital: the grow-now-and-clean-up- later mind-set, corruption, ill-defined property rights, and inadequate funds for environmental management. Often, information gaps prevent a full understanding of the causes and consequences of environmental deg- radation, and public indifference hinders their resolution. Fostering inter- national action is difficult despite existing momentum for environmental protection (box 4.2). The relative contribution of these factors differs from country to country and needs to be assessed before effective public actions can be determined. The Growth-Natural Capital-Welfare Nexus After dispelling the idea that environmental degradation can wait for re- dress after other pressing reforms have been addressed (the grow-now-and- clean-up-later mind-set), we will explore empirical evidence that links growth to the quality of natural capital. Grow Now and Clean Up Later The growing evidence to the contrary has not dispelled the perception that the environment is a luxury good that will be demanded as incomes rise with economic growth. As a result, developing countries tend to ignore en- vironmental concems as policymakers focus almost exclusively on acceler- ating economic growth. They support their position by citing examples of industrial countries that paid scant attention to environmental degradation in the early phases of their growth and arrested and reversed the problem later. However, they ignore the potential enormity of economic, social, and ecological costs and the reality that sometimes the damage is irreversible. While air and water pollution levels appear to be reversible, their im- pacts on human well-being often are not. Promises of future remedial action can hardly compensate for welfare losses by the present generation. Only a policy of clean growth is consistent with intergenerational equity. Further- more, investing in pollution control up-front will yield positive returns in other areas. For example, improved health outcomes can lead to a more fa- vorable accumulation of human capital and more sustained growth.4 A grow-now-and-clean-up-later approach also tends to be inequitable; the poor and disadvantaged suffer the brunt of environmental pollution and resource degradation. For instance, when industrial toxic effluents and 89 T H E QU A L I T Y O F G R O W T H other pollutants degrade water quality, the poor often lack access to puri- fied municipal water supplies and the resources to invest in water filters and other purification systems. Air pollution also disproportionately hurts the poor, as they tend to tive closer to roads where pollution levels are highest and cannot afford to switch to cleaner fuels for indoor use (UNDP 1998). These distributional impacts aggravate income inequalities and can lead to serious social conflicts. Thus, paying attention to the environment while accelerating growth is fully consistent with a poverty reduction strategy. The irreversible loss of genetic material and the potentiat threat of eco- system colsapse provide other compelling reasons for rejecting a grow-now- and-clean-up-later approach. Some damage can never be undone. Habitat 90 SUSTAINING NATURAL CAPITAL destruction has resulted in the irreversible loss of terrestrial and aquatic biodiversity worldwide. Marine pollution and destructive fishing techniques have damaged a large proportion of coral reefs in East Asia and threaten much of the ocean's animal and plant life (Loh and others 1998). The experiences of high-income countries show that the health costs of delayed pollution control can exceed the prevention costs, although in comparing them, the difference in the time of their occurrence and result- ing uncertainty should ideally be accounted for. For example, the cost of cleanup and compensation to victims of itai-itai disease, caused by cad- mium poisoning; Yokkaichi asthma, the result of excessive exposure to sul- fur emissions; and Minamata disease, or mercury poisoning, are from 1.4 to 102 times the cost of prevention (Kato 1996). In addition to the impacts on human health, the high cleanup costs of widespread dumping of toxic wastes by U.S. industrial firms illustrate another limitation of the grow- now-and-clean-up-later approach (Harr 1995). Does Faster or Slower Economic Growth Ensure Protection of Natural Capital? Both fast- and slow-growing economies have experienced environmental degradation, but to different degrees. Analysis of GDP growth and an index of natural capital quality show a negative correlation coefficient (see figure 1.5). By looking at the link between rapid growth and various components of natural capital degradation at a more disaggregated level, one can get a better idea of the strength and direction of the relationships. Belying East Asia's phenomenal record of economic growth and poverty reduction is its poor environmental record. In 1995, China was home to 15 of the 20 most polluted cities in the world, as measured by the concentra- tion of total suspended particulates (World Bank 1999e). Air pollution, es- pecially high levels of total suspended particulates, has resulted in prema- ture deaths and severe health damage in urban areas such as Bangkok, Jakarta, Manila, and several cities in China (see table 4.1). Countries that experienced rapid growth in the context of economic reforms in the 1980s-China, Korea, Malaysia, and Thailand-saw carbon dioxide emis- sions per capita that doubled or tripled after the reforms and growth accel- eration (table 4.3). Natural resources fared poorly as well. Deforestation rates have been high and remain so in most countries (table 4.3). About 20 percent of vegetated land in East Asia suffers soil degradation from waterlogging, erosion, and overgrazing. Severe land degradation in China, Thailand, and Vietnam threatens several ecosystems with irreversible damage 91 T iI HL Q U A I I T Y 0 F G R 0 WV T II Table 4.3. Trade, Growth, Poverty, and Environmental Degradation, Selected Years (percent, unless otherwise indicated) Trade Growth Poverty Indicators of natural capital Percentage of Total suspended Percentage Annual growth Annual growth population living Annual particulates in increase in of merchandise of GNP on less than US$1 deforestation capital cities carbon dioxide Region and export volume, per capita, a day (PPP) (percent change) (micrograms per emissions per economy 1980-94 1970-95 various years 1990-95 cubic meter) capita, 1980-96 East Asia China 12.2 6.9 29.4 (1993) 0.1 377 86.7 Hong Kong, China 15.4 5.7 < I 0.0 - 15.6 Indonesia 9.9 4.7 14.5 (1993) 1.0 271 100.0 Korea, Rep. of 11.9 10.0 < 1 0.2 84 172.7 Malaysia 13.3 4.0 5.6 (1989) 2.4 85 180.0 Philippines 5.0 0.6 27.5 (1988) 3.5 200 12.5 Singapore 13.3 5.7 < I 0.0 - 63.6 Thailand 16.4 5.2 < 1 2.6 223 277.8 Latin America Argentina 1.9 -0.4 - 0.3 97 (Cordoba city) -2.0 Bolivia -0.3 -0.7 7.1 (1989) 1.2 - 62.5 Brazil 6.2 - 28.7 (1989) 0.5 86 (Rio = 139) 13.3 Chile 7.3 1.8 15.0 (1992) 0.4 - 36.0 Costa Rica 6.6 0.7 18.9 (1989) 3.0 - 27.3 Mexico 13.0 0.9 14.9 (1992) 0.9 279 2.7 Peru 2.4 -1.1 49.4 (1994) 0.3 - -21.4 Uruguay 0.9 0.2 - 0.0 - -15.0 Venezuela,RB 1.1 -1.1 11.8(1991) 1.1 53 10.7 - Not available. Sources: World Bank (1997a, 1999e); see also annex 4. (World Bank 1999b). Biodiversity in 50 to 75 percent of coastlines and protected marine areas in East Asia is classified as highly threatened. Not all indicators show worsening environmental conditions among the fast-growing economies in Asia. Access to clean water and sanitation increased rapidly in China, Korea, Malaysia, and Thailand. In 1995, the share of the population with access to safe water rose from 71 percent in 1982 to 89 percent in Malaysia, from 66 to 89 percent in Thailand, from 39 to 65 percent in Indonesia, and from 65 to 83 percent in the Philip- pines. Sanitation service availability rose from 46 to 96 percent in Thai- land, from 30 to 55 percent in Indonesia, and from 57 to 77 percent in the Philippines (World Bank 1999e). Though still at low levels in Cambodia, the Lao People's Democratic Republic, and Vietnam, access to safe water 92 SUSTAINING NATURAL CAPITAL and sanitation has been steadily increasing with economic growth (World Bank 1999b,e). However, it is not just fast growth that leads to problems of natural capital degradation. Like East Asian countries, slow-growing Latin American countries have seen improvements in access to clean water and sanitation (World Bank 1999e), but have also suffered environmental de- terioration. Most have experienced extensive deforestation, especially of ecologically sensitive areas and steep slopes; widespread soil degradation; overfishing and water pollution in coastal zones; water contamination from agrochemicals; and pesticide poisoning of people and livestock. While air pollution is not as widespread a problem as in Asia, in part be- cause of the relatively low growth of industrialization (table 4.3), it is a serious concern in Mexico City, Rio de Janeiro, and Santiago. Because of low growth, highly skewed income distributions, inadequate investments in education and health, and political instability, poverty has remained stubbornly high, creating vicious cycles of increasing natural resource degradation and further loss of income (see also box 4.3). Thus, neither rapid nor slow growth is an automatic ally of natural capi- tal (Thomas and Belt 1997). For example, in the 1980s, differences in air pollution and traffic congestion between slow-growing Manila and fast- growing Bangkok were minimal (Hammer and Shetty 1995). However, fast growth, with increasing urbanization, industrial expansion, and exploita- tion of renewable and nonrenewable resources, places pressure on the envi- ronment such that many indicators show a decline in the quality of natural capital during growth periods. Yet, growth creates conditions for environmental improvement by creating demand for better environmental quality and making resources available for supplying it. Does this imply the existence of an environ- mental Kuznets curve? As incomes rise, does environmental quality first deteriorate and then start to improve? If environmental goods are normal consumer goods with positive income elasticity of demand that is greater than unity at certain income levels, then quality will improve beyond the threshold income level. L6pez (1997) suggests that environmental goods, such as clean air and water and sewage treatment, which directly affect health and generate local externalities, are likely to be normal goods with a high income elasticity of demand. Thus, they are likely to improve after a period of decline during a growth period. Most empirical studies have focused on indicators of the sink function of environmental quality, such as the concentration of suspended particulates in air, the biochemical oxygen demand of water, the level of carbon dioxide and sulfur dioxide emissions, and the prevalence of inorganic industrial pollutants 93 I 11 E 0-t l A I, I I Y (! F: (; i\0) I j I Box 4.3. Population, Poverty, and the Enviroent Analysis of the poverty-population-environment or fuelwood for cooking. In addition,.members of nexus is complex. Population growth has often rural households devote time to growing food, car, been blamed for poverty and for environmental ing for livestock, and producing simple marketable degradation (Cropper and Griffiths 1994; Pearce products. Children are needed as extra workers, and Warford 1993). However, thli converse argu- even when parentsare in their prime. Small house- ment contends that poverty and environmental holds are simply not viable; each one needs many degradation are the causes of population growth, hands. As the community's resources are depleted, not the consequences of it. Both views are partial; more hands are necessary to gather fuel and water it needs to be recognized that the three factors are for daily use. More children are produced, further interlinked (Cleaver and Schreiber 1994; damaging the environment and providing an incen- Dasgupta 1915; Ekbom and Bojo 1999; Mink tive to enlarge the household even more. X I 1993).TDhe strength of these linkages will differ Factors that influence the parental demand fior from situat:ion to situation, and the policy recom- children can reverse this destructive spiral. The nmendations will depend on a host of factors, in- most potent policy will use many of the factors Ssi- cluding type of resource, density and rate of multaneously. Good economic policies, secure growth of population, institutional arrangements, tenure rights, and political stability can all allevi- and laws regulating the use of the resource (L6pez ate the population pressures. Providing cheap fuel 1998b). As a result, no general conclusions about and potable water will reduce the need for eextra the linkages between populatiton, environment, hands and lower the demand for children. Family and poverty are available. planning services allied to reproductive health ser- The following example from Dasgupta (1995) vices will help bridge t he unmet needs for contra- gives some insight into the complicated nature of ception, and a literacy and employment drive for the nexus. In rural settings much labor is needed women that empowers them in decisions f family even for simple tasks, suchas colecting clean water size becomes crucial. (Galeotti and Lanza 1999; Grossman and Krueger 1995; Ravallion, Heil, and Jalan 1997; Roberts and Grimes 1997; Selden and Song 1994; Shafik 1994; Stern, Common, and Barbier 1996), finding some support for an environ- mental Kuznets curve. Hettige, Mani, and Wheeler (1998), using international data, measured the relationship between water pollution from industrial discharges and per capita income. The study showed that pollution first increases with develop- ment, peaking at a per capita income of about US$12,000, and then levels off for all observable values. The study concluded, "economic development remains far short of a Kuznets-style happy ending in the water sector" (p. 26) and suggests that total emissions will remain constant with income growth unless other factors intervene.' The quality of natural resources is less likely to follow a Kuznets-curve pattern than pollution, because they are typically factors of production rather than consumer goods. Furthermore, the externalities associated with 9,4 SUSTAINING NATURAL CAPITAL the destruction of natural resources are mainly global and less likely to be internalized in local demand (L6pez 1997). As a consequence, a growing economy imposes even greater demands on natural resources and makes management interventions crucial.6 Countries do not need to wait until incomes reach the Kuznets-curve turning point. Sao Paulo curbed severe pollution within a generation, even while millions remained poor. Fast-growing Shanghai, the largest industrial base in China, has produced lower sulfur dioxide loads than slow-growing Sichuan (World Bank 2000d). These and other cases illustrate methods- which include characteristics such as appropriate regulatory regimes, central- ized and market-based instruments-for environmental conservation, policy and legislative frameworks, institutional capacity, and technological options that help prevent pollution and protect resources (Panayotou 1997). Evidence suggests that flexing of the Kuznets curve is both possible and necessary.7 Whether economies are fast or slow growing, many indicators of natural resources-deforestation, fisheries depletion, soil degradation, coastal zone pollution-have been deteriorating. Because natural resources are important as factors of production, increasing growth tends to put in- creasing demands on them. Many of the externalities associated with their overexploitation, such as carbon sequestration and biodiversity loss, are global. As a consequence, local governments often do not consider the re- percussions of misusing or exhausting their resources. Other components of natural capital, such as air and water quality and access to sewage and sanitation services, are typically normal con- sumer goods. For these goods with an income elasticity greater than unity, income growth is likely to be associated with improvements in quality. Though some empirical evidence suggests the existence of environmental Kuznets curves for a limited set of indicators, the costs of inaction can be extremely high, because many developing countries cannot reach the turnaround income level for decades. Two indicators, access to clean water and sanitation, appear to improve in both fast- and slow-growth scenarios and testify to the efficacy of inter- ventions, but a careful look at the benefits and costs is required to deter- mine whether the pace of improvements is optimum. Asset-Income Inequality and the Quality of Natural Capital A more equitable distribution of income and assets might be associated with improvements in key indicators of environmental quality, such as deforestation and water pollution. For example, if small-scale farmers must use marginal land because large-scale landowners occupy the best 95 T H E OU A L I T Y O F G R O W T H land, the inequitable land distribution may drive deforestation (Ekbom and Bojo 1999). Adoption of clean fuels and more energy efficient tech- nologies implies that the marginal propensity to emit carbon dioxide de- clines as income rises. Thus income redistribution can speed emissions re- duction (Holtz-Eakin and Selden 1995). In a study of 42 countries, Ravallion, Heil, and Jalan (1997) estimated a large positive coefficient between carbon dioxide emissions per capita and the Gini coefficient of income inequality. This study suggests that growth that reduces income inequality and poverty could lead to a decline in emission rates. Growth Can Complement Protection of Natural Capital A diagram can help to show that growth and protection of natural capital complement each other (figure 4.1). Consider a preindustrial economy with a low rate of growth and a pristine environment, represented as point A. The country attempts to accelerate economic growth by investing in in- dustry and exploiting the potential of globalization. In an ideal situation, it would seek to balance accelerated growth with high environmental quality, which can be graphically represented as a vertical move toward point E or to one to its right. However, even a well-managed environmental strategy, which could be shown as a move from point A to point F, may not totally Many combinations of growth and environmental quality are possible Figure 4.1. Growth Paths and Environmental Quality Per capita income F W, E Environmental quality Source: Author. 96 SUSTAINING NATURAL CAPITAL eliminate quality deterioration in both sink and source functions, although the negative impact on natural capital would be relatively small and revers- ible.' While countries embarking on a path of sustainable development can incorporate environmental policies directly into their economic strategy at any time, most countries have followed the grow-now-and-clean-up-later approach (table 4.4). The fast growers among developing countries, such as China, Indonesia, Korea, and Thailand, experienced a situation that could be represented by a move from point A to point B, where they paid severely in terms of dete- riorating environmental quality. Many of the slower growers, such as Ghana and Nepal, showed a move from point A to point C on figure 4.1 and also suffered considerable environmental damage. Still others in Cen- tral and South America and Africa followed policies that failed to stimulate growth while still harming the environment; their actions would be simu- lated by an arrow from point A to point D.9 The economies represented at points B, C, and D have incurred serious losses from ecosystem damage: disease, death, degraded forests and bodies of water, polluted air, and more. Industrial and developing economies that ig- nored the degradation of their natural capital have learned that the grow- now-and-clean-up later strategy created costs that are difficult or impossible to recoup. For example, the United States will need to spend tens of billions Table 4.4. Classification of Selected Countries by Growth-Environment Trajectory A to B: High growth and A to C: Medium growth A to D: Low growth with A to F: Growth with environmental degradation and environmental degradation environmental degradation environmental protection Carbon Carbon Carbon Carbon dioxide dioxide dioxide dioxide Deforestation emissions Deforestation emissions Deforestation emissions Deforestation emissions Indonesia China El Salvador El Salvador Algeria Mexico Botswana Malaysia India Ghana Pakistan Cameroon Sri Lanka Indonesia Guatemala Panama Haiti Thailand Korea, Mozambique Mexico Republic of Nepal Nicaragua Malaysia Pakistan Zambia Thailand Panama Note: Deforestation rates depicted by annual averages for 1990-95, carbon dioxide emissions are from 1980 through 1996. High growth is defined as per capita income growth of more than 2.3 percent a year in both the 1980s and the 1990s; medium growth includes countries that maintained a positive growth of per capita income in both decades, or improved growth of at least 2 percentage points a year higher in the 1990s than the 1980s; the rest are classified as low-growth countries. Source: World Bank (2000c). 97 T I F QU A L I T Y O F G R O W T H of dollars to restore the damage caused to the Florida Everglades by the devel- opment of irrigation canals for sugarcane cultivation. Incorporating Environmental Sustainability in Growth Policies Many countries have integrated environmental concerns and growth poli- cies. The following four cases show how this can be done. The cases were chosen to illustrate success stories in pollution management and natural re- source preservation and to highlight the types of interventions required to achieve specific environmental goals. For more case studies of instruments successfully suited to particular settings, see Thomas, Kishor, and Belt (1998), World Bank (1997e, 2000d). Costa Rica: Conserving Forests and Mitigating Climate Change The rich biodiversity of Costa Rica attracts wealthy ecotourists from around the world. Yet in the 1980s deforestation rates climbed to more than 3 percent a year. To protect this valuable natural resource, Costa Rica designed one of the most innovative and workable systems of forest protec- tion in the world. It surveyed the marketed and nonmarketed benefits of the environmental services of the forests and identified who bears the costs and who derives the benefits (table 4.5). It calculated annual values per hectare of US$29 to US$87 for primary forest and US$21 to US$63 for secondary forest. The survey showed that landowners or the government (in the case of national parks) bore the costs of habitat preservation, while Table 4.5. Environmental Services of Costa Rican Forests and Their Beneficiaries Beneficiaries Type of benefit Landowner Country World Sustainable wood production X Hydropower production potential X Water supply purification X Soil stabilization and regulation of hydrologic flow X Scenic beauty, ecotourism uses, and "existence" value X X Carbon sequestration X Biodiversity preservation X Source: Castro and others (1997). 98 SUSTAINING NATURAL CAPITAL substantial benefits went to foreign interests. Thus, Costa Rica decided to create markets for some environmental benefits (Castro and others 1997). Of the several environmental benefits forests provide, Costa Rica has been most successful in capturing carbon sequestration and watershed protection. The government acts as an intermediary in the sale of these services to international and domestic buyers. Funds from the sales (and from an earmarked fuel tax of 5 percent) go to the landowners for pre- serving the forest cover on their lands. Contracts for more than 50,000 hectares of forest protection were established in 1997. Before that time, cumulative protected areas had encompassed only 79,000 hectares. Costa Rica has also attracted international investments to compen- sate landowners who promote carbon sequestration by maintaining for- ests. Certifiable tradable offsets can be used to sell greenhouse gas offsets in the international marketplace. The first batch was sold in July 1996. Between 1996 and 1998, sales were negotiated bilaterally, but recently Costa Rica started working with brokerage firms in Chicago and New York to establish the certifiable trade offset as a freely tradable commod- ity, similar to emissions trading in sulfur dioxide in the United States (Chomitz, Constantino, and Brenes 1998). Costa Rica demonstrates the practical success of creating green markets and implementing green taxes to reduce environmental damage. The po- tential for replication elsewhere is good, but it depends to a large extent on completion of an international agreement on carbon offsets and acceptance by all parties of the Clean Development Mechanism of the Framework Convention on Climate Change. China: Controlling Water Pollution with a Levy TIhe pollution levy, an emissions charge covering thousands of factories in China, is one of the few economic instruments with a long, documented history in a developing country. Although the levy has been used for sev- eral decades, serious studies of its efficacy have appeared only recently (Wang and Wheeler 1996). The system, which has been expanded since 1982, has been imple- mented in most of China's counties and cities and includes 300,000 factories that are charged for their emissions. The National Environment Protection Agency's regulations specify variations in effluent standards by sector and fees by pollutant. Any enterprise whose effluent discharge exceeds a legal stan- dard must pay a levy. Levies are charged only on the most polluting emission from each source. The levy differs from a tax, which would cover each unit of pollutant, not just those in excess of a certain standard. 99 T H E QU A L I T Y F G R O W T H Between 1987 and 1993 total organic water pollution fell for the state- regulated industries that reported discharges to the National Environment Protection Agency (Wang and Wheeler 1996). With output growth of 10 percent a year, China experienced an especially impressive decline in pol- lution per unit of output. While provincial total discharges declined at a median rate of 22 percent, pollution intensities fell at a median rate of 50 percent. Econometric analysis shows that much of the decline was attrib- utable to the levy. Significant variation in effective levy rates across provinces is explained by local assessments of pollution impacts and local capacities to enforce na- tional standards. Allowing for such regional differences enhanced the feasi- bility and effectiveness of the pollution levy system. Since 1991, the authorities have collected more than US$240 million per year in levies.'° Approximately 60 percent of the funds finance indus- trial pollution prevention and control and represent some 15 percent of to- tal investment in these activities, providing an additional incentive for firms to abate. The remainder goes to local agencies for institutional devel- opment and administrative costs (Wang and Chen 1999). In addition to helping regulate pollution, the levies help build the monitoring and regula- tion capacities of local enforcement agencies and reinforce the incentives for effective regulation. Indonesia: Fighting Pollution with Information The government can set standards for maximum allowable pollution levels relatively easily, but monitoring and enforcing compliance can be difficult. Indonesia's Environment Impact Management Agency faced compliance challenges in the late 1980s and resorted to voluntary agreements, out-of- court settlements, and other ad hoc approaches that had a limited impact on pollution control. Seeking a more sustainable approach, the agency developed the Pro- gram for Pollution Control, Evaluation, and Rating (PROPER), which receives pollution data from factories, analyzes and rates their environmen- tal performance, and disseminates the ratings to the public (Wheeler and Afsah 1996; World Bank 2000d). The agency hoped that publicizing the performance ratings would encourage local communities to pressure nearby factories that scored poor ratings to clean up their operations. It also hoped to influence polluters through financial markets, which were expected to react to the ratings. To encourage firms to improve their performance, it also established a program for recognizing excellent environmental pollu- tion control practices. 100 SUSTAINING NATURAL CAPITAL The agency decided to focus on water pollution first. It gathered data on water pollution from factories through questionnaires and rigorous on-site in- spections. The govemment compiled information on 187 highly polluting factories and ranked the companies by level of emissions. The data were com- bined into a single performance rating keyed to five categories: gold, green, and blue signified compliance; red and black represented noncompliance. The agency disclosed the results in stages, first publicly recognizing the best performers and giving the others six months to clean up before their bad ratings would be revealed. This phased approach gave factories time to adjust to the program and increased the likelihood of compliance. Between June and September 1995, half the firms that had been rated noncompliant observed the new mandates. This suggests that PROPER created powerful incentives for pollution control. In many instances firms have realized that better envi- ronmental management reduces the costs of production, creating additional incentives for them to clean up their production practices. Encouraged by these initial efforts, the agency planned to rate 2,000 plants by 2000. Indonesia's PROPER program goes beyond the command-and-control approach traditionally used with limited success to regulate polluters. The rating system is unique in that it allows for multiple outcomes. The final choice lies with the firm and depends on its resources for pollution control, the perceived benefits to cleaning up, and its overall corporate strategy. PROPER relies on public disclosure and public pressure to bring polluting firms in line with environmental regulations. By involving the people af- fected by pollution, PROPER ensures a bargaining equilibrium between firms and stakeholders. The government's role is to set the rules of the game, monitor the level of discharges, take punitive action when necessary, and act as the final arbiter. Europe's Blue Flag Campaign: Increasing Awareness of the Coastal Environments The European Blue Flag Campaign, operating through a network of na- tional organizations, is coordinated by the Foundation for Environmental Education in Europe (Thomas, Kishor, and Belt 1998). It encourages citi- zen understanding and appreciation of the coastal environment and the in- corporation of environmental concems into the decisionmaking of coastal authorities. The European Commission finances approximately 25 percent of the campaign's budget, which now amounts to more than US$1 million, and private sponsors finance the remainder. A beach or marina receives a Blue Flag if it meets three sets of criteria relating to the environmental quality of the locality, management and 101 If E Q U A l I TY O F C. R O W T H safety, and environmental education and information. Recipients must meet mandatory and guideline criteria. Based on maps, photographs, water samples, and a completed question- naire, a national jury nominates sites for a European jury, which makes the final selection of Blue Flag recipients by unanimous vote. The results are announced at the beginning of June before the main holiday season com- mences. The campaign has attracted several commercial sponsors in addi- tion to schoolchildren, who have cleaned local beaches to maintain the high standards required by the Blue Flag judges. Over the years the stan- dards of environmental quality needed to win the award have been succes- sively raised to provide dynamic incentives for better environmental man- agement. More than 1,000 coastal localities, most of them in Denmark, Greece, and Spain, have been awarded the Blue Flag. Governments see the Blue Flag program as an efficient way to pro- mote environmental awareness and increase tourism revenues. Private sponsors view it as an opportunity to attract more tourists. The initiative brings the govemment, the private sector, and the public into partnership and generates competition between jurisdictions, which pushes environ- mental standards higher. Rethinking the State's Role Attributing environmental degradation to distorted policies, damaging subsi- dies, missing markets, extemalities, and incomplete public knowledge puts the state in the position of a catalyst for environmental protection and man- agement. However, the mixed record on govemment interventions has moti- vated extensive rethinking of the policies the state should foster. To achieve the greatest impact, govemment should intervene selectively (see box 4.4). Streamline Subsidies and Implement Environmental Taxes In principle, subsidies support the incomes of the poor; in practice, they often increase inequalities, drain the public budget, accelerate the deple- tion of natural resources, and degrade the environment. The global cost of subsidies in agriculture, energy, road transport, and water is estimated at US$800 billion a year, with about two-thirds of the expenses incurred in OECD countries (de Moor and Calamai 1997). In recent years subsi- dies have been coming down remarkably rapidly, especially in developing countries. In China, coal subsidies fell from US$750 million in 1993 to US$250 million in 1995 (UNDP 1998). Subsidy rates fell from 61 percent in 1984 to 11 percent in 1995 (World Bank 1997e). Removal of perverse 102 SLISTAIN ING NAT URAL CAPITAL Box 4.4. Development and tho Environment World Developmt Report 1992 (World Bank 1992, accelerating the provision of sanitation; providing p. 2) addressed the challenge of finding the right clean water; providing education, especially for balance between development and environment. girls; and empowering local people. The second The protection of the environment is an essenual seeks to break the negative links between environ- . ~~~mernt and development by such means as establish- part of development. Witout adeuate enironmental poecton,deopme t i sdeed ide ung standards, usuig market-based instruments such opmdn, resopre is uZ f eor ueeLa Jnvet- as green taxes and taking collaborative approaches opmnt em unrnronn b oe kg&jww for &dE inves- .toapolution management. menrs. -nd -nvironmental -rotection -ill fa h -The report emphasized that although the costs of bhigs ith it the risk ofapingh environmental 4m-- age.w te i c n i t .etoer e nviron- adequately protecting the environment were large, age. Akematively, it osdd bring -wf t better nvirn- the costs of inaction would be monumentalc It is r- met protection, cleaner ar and wae, and the w tional to act sooner rather than later. eirnatm ofae ortyoyoes' Eight years later, the prescriptions of the report male the difference, are still valid. Experience in natural resource man- The report highlighted two sets of policies for agement and environmental protection shows that sustaiable development. The first builds on posi- fmding the right mix of policies for environmen- tive, win-win links, such as removing environmen- taly sustainable development is even more feasible tally damaging subsidies; clarifying property rights; today if it is given a high priority. subsidies has three benefits: reducing environmental degradation, pro- moting equality, and conserving budgetary resources. Not all subsidies result in bad outcomes, and good subsidies should be encouraged. Inaccessibility to water and sanitation services accounts for a significant loss of life, especially among women and children in poor households. Studies show that these households have a high willingness to pay for reliable and adequate supplies of these services. Subsidies targeted to these households to purchase water and sanitation services, perhaps from the private sector, is probably a cost-effective intervention to miti- gate negative health impacts and reduce poverty by promoting human capital accumulation. On the other side of the ledger, green taxes on activities that cause envi- ronmental degradation provide a powerful way to fight pollution and resource depletion. Green taxes can be particularly useful in managing emissions that contribute to air and water pollution. Taxing the use of coal by industry or the use of gasoline in motor vehicles will reduce excessive resource use and emissions and raise tax revenues; thus, green taxes can provide a win-win ap- proach to managing environmental quality with growth (World Bank 1997d). Pollution taxes are most effective when a well-established regulatory framework with emissions norms and an efficient system of monitoring and 103 T H E OU A L I T Y G F G R O W T H enforcement are in place. Effective green taxes also encourage the use of cleaner energy sources, such as solar power. A switch from income to consumption taxes can also benefit the en- vironment and growth. The production and consumption of luxury goods often make heavy demands on environmental and natural resources. Consumption taxes can curb the overexploitation of these goods. Progres- sive consumption taxes also promote equity, and by encouraging savings, they promote economic growth (Frank 1998). In addition, green taxes can generate the funds needed to promote en- vironmental management. The public sector needs money for its facilitat- ing role, but funds often fall short of needs. Any strategy for environmental management must identify sources of adequate financing. Many developing countries now rely more on green taxes to create funds for environmental improvements than they have done in the past (World Bank 1999f). Move from Central Control to Partnerships In the past, governments relied too much on central control, which required extensive monitoring of compliance, for environmental management. The combination of command-and-control policy and inadequate resources for monitoring and enforcement ensured program failure. Policymakers are learning that community members affected by pollution can complement regulation. The involvement of local communities and civil society has other advantages. In rural areas, especially, they are both the source of key environ- mental information and the custodians of traditional environmental know- how. Thus, they can identify and implement strategies that balance growth with environmental protection. Where undervaluation of a resource can lead to its degradation, proper valuation of its economic and social benefits can ensure that its contribution is fully considered in decisionmaking (Dixon and Sherman 1990; Pearce and Warford 1993; Ruitenbeek 1989). Measurements of "green gross national product" and genuine savings are gaining prominence as a way to incorporate sustainability in traditional economic planning (Hamilton and Lutz 1996; World Bank 1997d). The gap between estimated total economic value and actual private valuation cannot be easily bridged, but evidence suggests that the state can do it by creating markets or by setting up appropriate institu- tions and legal statutes that create marketlike conditions and by generating adequate financial flows (see the illustrative case of Costa Rica). Based on data for 77 developing countries, a significant positive associa- tion between educational spending and increase in forest cover was found (correlation matrix available upon request). This suggests that added 104 SUSTAINING NATURAL CAPITAL impetus to environmental sustainability can come from state and private sector cooperation to raise the educational attainment of the population. Recognizing the limitations of state intervention and the need for ac- tive partnerships in environmental management, govemments are looking for new ways to promote environmental management. Spreading knowl- edge about the full consequences of environmental neglect to all stakehold- ers, together with a clear framework of accountability and environmental liabilities, can have a powerful impact (Thomas, Kishor, and Belt 1998). Alliances between state regulatory agencies and industrial firms are helping to control pollution in many countries (Hanrahan and others 1998; Schmidheiny and Zorraquin 1996). In Zimbabwe, the CAMPFIRE program promotes alliances between provincial govemments, the private sector, and local inhabitants in managing wildlife for profit and biodiversity conserva- tion within a legal framework laid down by the central government (Tho- mas, Kishor, and Belt 1998). In Africa, East Asia, and Latin America the conventional wisdom that the slash-and-bum agricultural practices of the poor are a cause of large- scale deforestation has given way to a realization that macroeconomic change, commercial enterprises, and infrastructure development often have far greater impacts on deforestation (Chomitz and Gray 1996; Deininger and Minten 1996; Mamingi and others 1996). As unsustainable timber ex- traction by large commercial logging companies leads to deforestation, poor indigenous communities lose their sources of fuelwood, fodder, medicinal plants, and even their means of livelihood. The poor communities that de- pend on forests should be the focus of public action that ensures improved and sustainable use of forests. The best chance of a negotiated agreement lies with a three-party partnership of the state, local communities, and the logging companies. The challenge for developing countries is to scale up such partnerships as quickly as possible. Clarify Property Rights, Resource Ownership, and Environmental Liabilities The empirical relationship between clear property rights and environmen- tal quality is strong (Dasgupta and others 1995). Farmers with secure title to land are more likely to invest in soil conservation, sustainable cultiva- tion techniques, and other environmental protection practices (Feder 1987). Vested with ownership rights, local communities have reforested de- graded lands in India and Nepal (Lynch and Talbott 1995). Establishing use rights for water, fisheries, and logging provides a clear incentive and means for resource management (World Bank 1997e). 105 T H E QU A L I T Y O F G R O W T H Without enforced property rights to natural resources, outside inter- ests take advantage of open access and, with no accountability for their actions, overexploit the natural capital by overfishing, overgrazing, over- using village woodlots, and extracting excessive amounts of groundwater. While experience varies, vesting communal property rights in these re- sources seems to relieve the pressures for overexploitation. The commu- nal group develops mechanisms for restricting access by outsiders, distrib- uting management responsibilities, allocating use rights among group members, and monitoring and compliance. Examples of communal man- agement systems include those for forests in Japan; fisheries in Turkey; ir- rigation water in south India; and pastures in the Swiss Alps, the Himalayas, and the Andes (World Bank 1992).11 Security of tenure for urban dwellers can also improve the quality of the environment by simplifying the identification and enforcement of en- vironmental liabilities for air and water pollution and solid, hazardous, and toxic chemical disposal (World Bank 1997e). A study of the relationship between property rights and the urban environment found that when people moved from squatter status to moderate security, the probability of purchasing garbage collection services increased by 32 percent, while mov- ing to high tenure security (high tenure security is characterized by land tenure accompanied by a certificate of legal title) increased the probability of buying trash removal by 44 percent (Hoy and Jimenez 1997). Thus, es- tablishing clear property rights and ownership and identifying environ- mental liabilities could be the state's most important contributions toward achieving environmental sustainability. Improve Governance and Reduce Corruption Rent-seeking and corruption hurt economic efficiency and prevent desir- able outcomes even when good policies for environmental management exist on paper (Bhagwati 1982; Krueger 1974; Rose-Ackerman 1997a). Corrupt officials undermine efforts to monitor and enforce environmen- tal measures, from industrial effluent discharges and automobile emis- sions inspections to allowable cuts for timber (box 4.5). It is found that controlling corruption is significantly associated with, for example, a reduc- tion in water pollution (annex 1, figure AL.1). Collecting knowledge and sharing it widely can combat corruption and foster good governance, with beneficial outcomes for economic growth and environmental man- agement. Specifically, the corruption diagnostics approach has shown promise to reduce corruption and promote integrity in several countries (chapter 6 in this volume; Kaufmann, Pradhan, and Ryterman 1998). 106 SUSTAINING NATURAL CAPITAL Box 4.5. Pivate P_roft nt!"Pbcxpense: Coruption hi the tor Corruiptlion is rampant in the logging nd timber ostb _ trade at all levels o, forfest-related' di-son, ,:. ,-- -US$L bi,llion a. ytar Xfom .lle loggg. making. -Most damaging t forest resources-i the * I-9-n 199t n-h n - misuse of public resources for private gain b ,the estry a t 'the ntrW political elite. -I additiot tothe degradnand a -$3. ' a - s misuse of -fiorests; corruption d govenms enal ee e- e . and local'communities of resources th-a could-re g--- ,, used for- development or impro4ved frt I 4 ku-ia gnv>etnoke management. Corrupt -practices inclue the o 3-2 e t t eoncealed or: secret saleofrvesting - petmits,- tal --re-Venu, lgg- tat-, i 'illegal underpricting of -wod by F,-co-mpanies US$184 mil-iupsteao, S$900 mlion (transfer pricinigl, false cer£tifcation-of s r, -ou4S$5,5 hlllloti-j volumes cut in- pubic forests, and illegal ging.on and Sm- ExamXples fromn around the r a,re prevalent t= h need f * Transfe,r pricingasi -so'praent minPapa : o p- New Guinea th-until 1986 not a singleo- r n toe posr g - pany declared a: profit despite tne boomiig -naize zfferdiig c ti&-canpitdiidul timber trade., An o Eth;is rcgiti W ord- I* n Gh-ana, It foreigncompanries were lmpit', - B a Arest law e cated in fraud aid other- pra ct, costing-foc m theast Ai t d the econmy about US5 mmiin corruption. Source: W -orl Commissionon Forestry ain Sustainable D Thus, developing countries should give high priority to stamping out corruption and improving governance. Global Environmental Issues Must Be Confronted Many environmental management issues are global in scale, though local in cause.12 The greenhouse effect and global climate change are clearly con- nected to human activities (see box 4.6). Fossil fuel combustion is the biggest source of greenhouse gases. Deforestation contributes to the problem because of the loss of forest sink functions, which transform gaseous carbon dioxide into biomass. Farm activities, coal mining, and leakages from natural gas transmission pipes also add to greenhouse gases by releasing methane. Because the problem originates with a large number of economic activi- ties considered essential to growth, controlling them raises difficulties. Most developing nations depend on fossil fuel combustion for economic production and are unlikely to switch to cleaner, but more expensive, fuels. 107 T HI E QU A L I T Y O F G R O WV F The first Wotid Climate Conference, held in 1979, ultim aority. It held its first session in Berlin in recgnied limte hane a a eriusprolemand 199. Te scon, hldin Geneva in 1996, took eaplored how it might affect human activities. The stock of progress and other issues,,Participating offi- conference declaration called on the world's gov-daIs stressed the need toa accelerate talks on how to erments to predict and prevent potential man- strengthen the ClimateChange Convention. The mad cns lithmhhevs G vDathe second assessment governmental Panel on Climate Change, estab- Change as the most compGneSrive Pand anthoita- firmed the scientific evidence for climate change. once of Pa*rties in December1997 is acknowledged as The Second World Climate Conference in 1990 a4 historic step towardbinding emissions limitations caled or frmewrk rety n cimae cang. i 3 inustialandtrasiion economies. these The nite NaionsFraewor Coventon o patiesagred t enure hattheir greenhouse gas Climate Change, which was opened for signature at emissions are reduced by at least 5.2 percent below the Rio Earth Summit in June 1992 and entered 1990 levels in the commitment period 2008-12. intofore inMarh 194, rovdesthe ontxt fr Tis s a ignficnt dvelpmet, because projec- a concerted international effort to respond to cli- tions for the$ UnitedStates, for example, indicate matechane. hereare166 ignaoris an 16 tha witoutsuchbining ommtments its emis- partiesto theconventon. sins coud be 30percen above1990 level~sby 2010. The Cnfernce f Paties whih repacedthe n- Dspit thesignificant prges the details of join rergoernmetal Ngotiaing Cmmitte forthe iplemetatin,eissions trading and developin Frameork onvetion becae th conentin's cuntr oblgatins siremantob reslved. Yet, switching to cleaner fuels can lead to better health outcomes, which is good for national economic objectives. As a consequence, a natural tension exists between the two goals, and many countries opt for more growth over better health (Munasinghe 2000). Financial and technical assistance from the international community, in return for reaping the benefits of switch- ing to cleaner fuels, can enable the joint realization of both national and global interests. Cooperation between rich and poor countries can also help control de- forestation. Despite the externalities that it generates, developing coun- tries see deforestation as an unavoidable consequence of their economic development. As with clean fuels, the international community needs to deal with the threat of global climate change by transferring resources, in- cluding technology, to control deforestation (Kishor and Constantino 1994; L6pez 1997). Under the joint Implementation Initiative of the 108 SUSTAINING NATURAL CAPITAL United Nations Framework Convention on Climate Change, several bilateral schemes of forest conservation are being tested in different parts of the world. Successful pilots will be replicated on a larger scale. The Global Environment Facility is the main institution addressing glo- bal environmental concerns. As the interim financial mechanism of the Convention on Biological Diversity and Climate Change, it addresses global environmental problems through collaboration between industrial and devel- oping countries that benefit both parties. For example, industrial countries can mitigate greenhouse gas emissions cheaply, and developing countries can benefit from financial and technological transfers in protecting their resource base and promoting economic development. Preventing global climate change and managing its consequences will be one of the biggest challenges of the 21st century. Global conventions, treaties, and agreements have been important in identifying common prob- lems, developing solutions, and allocating responsibilities. National aware- ness and commitment are increasing, and implementation must be encour- aged to secure national and global objectives. Successful examples of this marriage of national and global objectives need to be replicated widely (Castro and others 1997; Watson and others 1998). Conclusions For the developing world, depletion of natural capital (forest, energy, and minerals) and damage from carbon dioxide emissions is estimated to be 5.8 percent of GDP. Environmental health risks account for 20 percent of the total global burden of disease. In addition, the huge costs of global environ- mental problems need to be factored into domestic development policies. The poor, especially women and young children, often bear much of the burden of environmental degradation. Thus, natural capital is crucial for sustained growth, and its conservation and augmentation are crucial to na- tional and international development strategies. Three key findings emerge from the evidence presented in the chap- ter, namely: * Several indicators of the quality of natural capital, with the no- table exception of access to clean water and sanitation facilities in some countries, tend to worsen in both slow- and fast-growth economies, imposing heavy costs and diminished prospects for fu- ture growth. However, faster growth makes more resources avail- able to invest in the improvement of natural capital. Thus, the 109 F i F QL U A L. I T Y O F G R 0 W T H grow-now-and-clean-up-later approach espoused by many indus- trial and developing countries needs to give way to one of growth with sustainability of natural capital. * The state performs a crucial role in environmental management, but it needs to be selective and efficient in its interventions. It should focus on collaborative approaches with local communities and the private sector. * Global environmental problems are huge, but they offer opportuni- ties to simultaneously address national problems if international co- operation can be secured. The development of transfer mechanisms for resources to pay for global extemalities is key. Countries need to focus on strategies to achieve high-quality growth that is sustainable and compatible with domestic and extemal financial stability, that is, growth that supports the poor and vulnerable and does not excessively degrade the atmosphere, rivers, forests, oceans, or any other part of humanity's common heritage. Cost-benefit estimates support a strategy of growing clean, and the examples in this chapter show that such a strategy is feasible. Notes i. Some World Bank studies attributed 100 percent of all waterborne illnesses to the lack of piped water connections and sanitation facilities. However, epide- miological studies rarely showed declines of more than 40 percent in sickness due to water access interventions (Esrey and others 1990). Thus, health ben- efits related to a cleaned up water supply and improved sanitation services may be overstated in table 4.2. 2. Work done under the Global Burden of Disease Initiative uses a standardized measure of health outcomes, DALYs, across various causes of illness and death, giving a standard way to quantify some of the losses described here (Murray and L6pez 1996). 3. Most estimates use discount rates in the range of 6-10 percent to calculate the present value of benefits. If actual discount rates are higher, say 20-25 percent, as evidence from some developing countries suggests, then the present value of benefits will be much lower. Similarly, the opportunity cost of capital available to finance environmental improvements is assumed to be much lower than that developing countries actually face. The net result of applying the "true" values would be to reduce the gap of benefits over costs, thereby reducing the invest- ment required for optimal environmental management, or even reduce benefits below costs, making such investments unprofitable. This points to the need to carry out a sensitivity analysis with respect to discount rate changes to reliably identify the priority areas for intervention (Kishor and Constantino 1994). 110 SUSTAINING NATU RAL CAPITAL 4. Protecting the environment while accelerating growth can also have beneficial impacts on the accumulation of physical capital. If the authorities announce more stringent environmental standards in advance of the date when they be- come binding, investments embodying the improved standards can be made over a period of time, thereby reducing capital obsolescence or the need for costly retrofitting to meet environmental standards, for example, the experi- ence with emissions standards and catalytic converters for cars. 5. Even where an environmental Kuznets curve seems to be supported, this does not imply that environmental management is unnecessary. Take the case of sul- fur dioxide emissions, where Grossman and Krueger (1995) estimated the turn- ing point for emissions to decline at a per capita income level of US$4,053. Even with a high growth rate of 5 percent a year, India, for example, will take several decades to reach this income level. India and much of the rest of the developing world cannot continue to suffer the consequences of this type of pollution while hoping to "grow out of the problem." 6. Deforestation appears to follow a Kuznets curve path (Cropper and Griffiths 1994), but with a turning point of US$5,420 per capita income for Latin America, proactive polices are absolutely necessary. 7. Clearly, the links between pollution and growth depend on many factors, and a case-by-case analysis is necessary. In China, for example, the development of private town and village industrial enterprises was the prime engine of growth in the 1990s, lifting more than a 100 million people out of poverty. These en- terprises are often more efficient, with better pollution control technologies, than state-owned enterprises. Therefore, as a result of the expansion of private enterprises, accelerated growth is likely to be associated with declining pollu- tion intensities. 8. Moving from A to F implies that environmental quality deteriorated from its pristine state. This corresponds to the optimal rate of environmental protection referred to earlier. For the sink function of environment, this may be justified on the ground that "small" amounts of air pollution, water pollution, and so forth neither pose health risks nor impair the ability of the resource to "renew" itself; and the economic gains resulting from the pollution-generating activities are large. For the source function, a certain amount of forest clearing, for ex- ample, is justified as long as the alternative use of land provides greater social returns and forest clearing does not occur in the "wrong" places, such as steep slopes, along river basins, and so forth. 9. East Asia provides an interesting case. The recent economic crisis has plunged Thailand and Indonesia from B to C. As a consequence, they have the tough task of implementing policies to clean up the environment as they raise eco- nomic growth, that is, move from C to F. 10. Some Y 2 billion, converted at the rate of Y 8.3 to the dollar. China unified its dual exchange rate regime in 1994; hence, this amount should be viewed as ap- proximate. tIl T H E QU A L I T Y O F G RPO W T H 11. Protected areas, national parks, and other public lands that provide critical en- vironmental services typically do not enjoy the advantages of community man- agement. As a result, migration, encroachment, illegal extraction, and other forces continue to degrade govemment-managed lands in many areas. 12. Watson and others (1998) classify global environmental issues into two catego- ries: those involving the global commons (atmosphere, water, and so on) and those of worldwide importance, but not directly involving the global commons (biodiversity, land degradation, and so forth). On the basis of current scientific assessment, the most important global environmental issues for this century re- quiring urgent action are global climate change, stratospheric ozone depletion, loss of biological diversity, deforestation and unsustainable use of forests, deser- tification and land degradation, freshwater degradation, marine environment and resource degradation, and persistent organic pollutants. The interlinkages among these issues and the need to address them simultaneously is also empha- sized. Without downplaying the importance of other global environmental is- sues, this section focuses on global climate change mitigation and forest man- agement to illustrate the challenges that we face in this area. 112 C H A P T E R 5 DEALING WITH GLOBAL FINANCIAL RISKS Times of trouble prompt us to recall the ideals by which we live. -Michael J. Sandel, Democracy's Discontent: America in Search of a Puiblic Philosophy T he financial crisis of 1997-99 that affected most severely Brazil, the Russian Federation, and several countries in East Asia underscored the importance of financial stability as a contributor to the quality of growth. As with environmental sustainability, education, and good governance, managing the risks of financial instability, especially those of cross-country capital flows, can stimulate sustainable growth by reducing economic inequality, enhancing social stability, and strengthening democratic trends and institutions. Without social and political stability, "no amount of money put together in financial packages will give us financial stability" (Wolfensohn 1998). Global financial integration has undeniable benefits for developing and industrial countries, but it also exposes countries to the vicissitudes of international capital markets, such as volatility in currency values, inter- est rates, liquidity, and volumes of capital flows, with important macro- economic and growth consequences. These risks are pronounced and costly, as demonstrated recently by lost output and jobs, corporate and banking distress, and increased poverty in crisis-hit countries, especially in countries where the institutional and regulatory frameworks for open capital markets are not fully in place. The high social and economic costs associated with financial instability are unacceptable and make a strong case for devising better ways to deal with financial risks and to ensure stable growth. Chapter 2 shows how policy distortions, subsidies, and unnumerated guarantees can cause 113 T H E QU A l I T Y O F G RPO W T 11 overinvestment in certain physical and financial capital, but underinvest- ment in other assets. This chapter turns to the factors that influence the volatility in capital flows to developing countries and the associated subop- timal investments that could lead to an increased vulnerability to financial turbulence. After a brief review of the benefits and risks of financial market integration, the chapter examines the causes and consequences of capital flow volatility and its implications for the poor. It then reviews the evolu- tion of policy and institutional arrangements for managing risk and suggests a broad framework for risk management that integrates insights from the theory and practice of modem financial risk management with the political economy of open capital markets. For growth to be relatively stable, govemments can consider a spectrum of actions as follows: * Remove distortionary policies and implicit or explicit subsidized guarantees that provide incentives for short-term foreign capital in- flows, which may accentuate vulnerability to financial shocks. * Strengthen domestic regulation and supervision of banks and other financial intermediaries and improve corporate governance and transparency. * Build a broad framework for risk management, based on an orderly opening of capital markets combined with measures for controlling short-term capital flows. * Maintain public support for open capital markets by providing cush- ions against risks, either through the marketplace or through redis- tributive policies and a social safety net. Expansion of Capital Markets and Volatility of Capital Flows By any measure, the growth in intemational financial markets throughout the 1990s was astounding. International lending in new medium- and long- term bonds and bank loans reached US$1.2 trillion in 1997, up from US$0.5 trillion in 1988 (BIS various years). World trade in goods and ser- vices, though growing significantly since the early 1970s, is now dwarfed by international financial transactions of more than five times the value of world trade (figure 5.1). OECD cross-border transactions in bonds and equities, less than 10 per- cent of GDP in 1980, reached more than 100 percent of GDP in 1995. Av- erage daily turnover in foreign exchange markets reached US$1.6 trillion in 1995, up from US$0.2 trillion in 1986, and annual trade in goods and 114 DEALING WITH GLOBAL FINANCIAL RISKS Figure 5.1. Global Financial Market Size and World Trade, 1980-96 Percentage of world GDP 100 80 U Finance 60 E Trade 40 20 uii4 0- 1980 1982 1984 1986 1988 1990 1992 1994 1996 Note: Financial market size refers to world stock market capitalization plus stock of intemational bonds and loans outstanding. Trade figures are averages of imports and exports. Sources: BIS (1997, 1998); International Finance Corporation (various years). services reached US$6.7 trillion in 1998. Global market capitalization of stock markets relative to world GDP rose from 23:1 in 1986 to 68:1 in 1996, while derivative markets expanded from US$7.9 trillion in 1991 to US$40.9 trillion in 1997 (table 5.1). The net flow of foreign private capital to developing countries also rose dramatically, from US$43.9 billion in 1990 to US$299 billion in 1997. Most of the capital came from foreign direct investment (FDI) and interna- tional capital markets, which include portfolio equity flows, commercial bank lending, and equity and bond issues in offshore markets. Flows of FDI to developing countries increased more than sixfold between 1990 and 1998, and the share of global FDI flows to developing countries rose from 18 percent in the mid-1980s to 24 percent in 1991 and 36 percent in 1997. However, when the financial crisis struck Asia beginning in 1997, capital flows from international capital markets to emerging market economies took a heavy hit, falling to their lowest point since 1992 at US$72.1 bil- lion, while FDI remained resilient (figure 5.2) (World Bank 1999c). Causes and Consequences of Capital Flow Volatility The large expansion in private capital flows to developing countries from 1990 to 1997 was positively affected by advances in communication and ll5 T H E QU A L I T Y 0 F G R 0 W T H Table 5.1. Growth of Derivatives Markets, 1991-97 (notional values in billions of US$) Instruments traded on exchanges Over-the-counter (OTC) insrruments Stock narket Interest Interest Currency index Total Interest rate rate futures futures exchange rate Interest Currency Year futures options and options and options traded options rate swaps swaps Total Total 1991 2,157 1,073 81 109 3,420 577 3,065 807 4,449 7,869 1992 2,913 1,385 98 238 4,635 635 3,851 860 5,346 9,980 1993 4,959 2,362 110 340 7,771 1,398 6,177 900 8,475 16,246 1994 5,778 2,624 96 366 8,863 1,573 8,816 915 11,303 20,166 1995 5,863 2,742 82 502 9,189 3,705 12,811 1,197 17,713 26,901 1996 5,931 3,278 97 574 9,880 4,723 19,171 1,560 25,453 35,333 1997 7,489 3,640 85 993 12,207 5,033 22,116 1,585 28,733 40,940 Source: BIS (various years). Figure 5.2. The Rise and Fall of International Capital Flows, 1990-99 US$ billions 160 - 140 - 120 - 80 -Portfolio equity ~ - 60- 20 ~~~~~~~~~~Bond 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Note: International capital market flows to developing countries (including Korea) consist of portfolio equity, international private bonds, and loans. Source: World Bank (2000g). information technologies, which reduced cross-border transaction costs.' The advances facilitating cross-border capital flows have included the creation of the Euro-currency money market, the spread of derivatives, and the rapid expansion of hedge funds. In addition, both industrial and developing countries have opened their financial markets by removing 116 DEALING WITH GLOBAL FINANCIAL RISKS barriers to cross-border capital flows.2 However, various implicit or ex- plicit government guarantees provided to banks, corporations, and inves- tors in liberalized, but inadequately regulated, financial sectors, fueled overinvestment in certain industrial sectors in East Asian countries, on the one hand, and creating moral hazard and excessive risk-taking behav- ior among investors on the other hand.3 The accumulation of contingent government liabilities and overleverged corporate indebtedness knay have contributed to vulnerability, to the loss of investor confidence, and to the eruption of the recent financial crisis. Historically, shifts in the supply of foreign capital to developing countries have been caused by exogenous factors, such as the increase in oil prices in the 1970s, low interest rates, deregulation of institutional investors in industrial countries, and institu- tional innovation and competition in the 1990s. The dismantling of barriers to capital flows across national boundaries, such as capital controls and foreign exchange restrictions, accelerated in OECD countries in the 1980s and spread to emerging markets. OECD coun- tries liberalized almost all capital movements, including short-term transac- tions by enterprises and individuals, in compliance with the OECD Code of Liberalization of Capital Movements. The United Kingdom achieved full capital account convertibility in 1979, and in 1992 Greece, Ireland, Portugal, and Spain became the last OECD countries to fully abolish their capital con- trols (OECD 1990). By the early 1990s, the capital accounts of OECD coun- tries were open to a wide range of cross-border financial transactions in capi- tal market securities, money market operations, forward operations, swaps, and other derivatives.4 Many emerging market economies have also reformed their financial markets and liberalized cross-border capital movements. Based on an index of financial openness constructed for 96 countries, as of 1977, 46 can be classified as open and 10 as semi-open (box 5.1 and annex 5). As countries liberalized, banks and corporate borrowers gained access to a broader menu of foreign financing. The desirability of long-term capital, especially for funding infrastructure projects, provided a strong competitive advantage for foreign capital, particularly in countries with an exchange rate pegged to the U.S. dollar. Greater access to foreign capital in developing countries opens up possibilities for financing a broader set of investment projects, both sound and risky ones. Despite the positive potential of foreign capital, weaknesses in domes- tic policy and liberalization measures, including subsidized guarantees, cre- ated incentives for imprudent behavior by banks, corporations, and inves- tors that led to overinvestment in physical capital (for examples, see Demirgu,c-Kunt and Detragiache 1998; Williamson and Mahar 1998). 117 T H E QU A l I T Y OF G RPOWV T H :=~~~~~~~~~~~~~F Evidence about the open n megntguarketNoo sang fmesure f opennes exisanye iables economies to cross-bo rectal flowsis scanty and metasue po financi,a dop ss needt sutancr Ipoate fragentdnfomatonrn mexphodologypoblermsn distinctiesons libeltwesencthe rse erit ofxcnt veroiskan impeetedvelpmetaofinope qnfuentitaiv mbaothe typesti oftandacions.thern fionanl pnv estos(e contols(se, fr istac,Aeinnnand Grill, 97fo andt ansacytion. atlse mdesa)grentedsmeasursuch ceapk- Misi-errtt14;aznes andsRosed i94) How taeriin controlsk ased lonerihe clssfcatgions ond ifor-ig ever notall rans ctios rrensbct teomialcotrols,b mto conaiedg in theke Ineronatiosnal Monetary and ost easuesae finene toe beinniuence the in Fndst Annian Rcrisnixs.g rrnemns n cenive fo cetai acivTies CotolBangko nemafromnE lBankeingFacilons, Dr tawingoshed coing 199thin diret quntiativli itonancrain tibranszactions orabodog deveope bya Quin and frstoyborra (n9or7), assoiatdtansfrs,to cuchidrencyt mesuores mastaxithes measur is a coproesitaled otinde ofrlesdng Becausa- witholingoreseveeqiremengtoason b ethena as-thhionsg tand andmienistativer proedures d tha Taffc setanlabiitis.Suc Bcntrolsco Inealsonaply toBapkitgaclfltw foru7transactinsding The frincurrentcan tranfer of und as oiated o w Thafinan mcial trna b apita acounse tofU 3 balaince of payments ofopri96 tion orto he usiessactives tesetrlvens, cyteedountries. ndotes 99) G h ora overnment guarateesdookemnyifoms,s chas egged ex ichanged rtes takinginnflumencing baothd idomstriessch andinternaiconalctinvesatorsbi(see forelan the pbeginning. of th reEas thAsianvcrisis. -oeuiyrtiftho 30cmanebos. This infusioan of Jpaesen mony rhesulend n196 rapid growthr of B I Bing ThGora overnment direcednlndin too canyefors, which lspged exhnertos tverinvetnflentcing faored idumstriessch andinernaionaulinvesatomobiles, steelan shipbugiding. In Kore a thet Asaveag det-oeqiysatoofte.o 30omhanebos. wais moreuthan 500 Jpaeren bony thesulend of196 ranid grewtur on 18~ ~ ~ ~~ago nentonlBnigFclt u-nledn:Tefrincrec DEALING WITH GLOBAL FINANCIAL RISKS invested capital was below the cost of capital for two-thirds of the top chaebols (Park 2000). With excessive foreign short-term borrowing induced in part by subsi- dized guarantees, government contingent liabilities accumulated. Once investors realized that the government was no longer able to meet its ob- ligations, they took the exit route. Once a crisis began in one country, in- vestor herdlike behavior and contagion spread through international trade and financial linkages, which resulted in private capital flow rever- sals and considerable widening of spreads for almost all emerging market economies (see Calvo 1999; Reinhart and Kaminsky 1999; Van Rijckeghem and Weder 1999). Benefits and Risks of Open Capital Markets The benefits of open capital markets are indisputable; the policy debate concerns whether the benefits outweigh the risks. Governments can also consider employing instruments to minimize such risks. Open capital markets bring many benefits to both borrowing and credi- tor countries. They offer developing countries broader sources of investment finance to complement domestic savings. They also result in increased effi- ciency in domestic financial institutions and more disciplined conduct of macroeconomic policy. In addition, by easing financing constraints, open capital markets provide time for countries to make payment adjustments to correct imbalances that were created in response to extemal shocks.5 Open capital markets offer creditor countries broader investment- and risk- diversification opportunities, particularly as their aging populations with growing pension funds seek higher and safer returns on their investments. Open capital accounts also support the multilateral trading system, ex- panding the opportunities for portfolio diversification and for the efficient al- location of global savings and investment (Fischer 1998). An important property rights issue with regard to intemational finance also has attracted the attention of scholars and policymakers. Cooper (1998, p. 12) notes the view that embodies the thinking behind a liberal world order: "Individuals should be free to dispose of their income and wealth as they see fit, provided their doing so does not harm others." Others argue that openness to intema- tional capital flows is highly correlated with measures of political and civil liberty. The empirical evidence on the importance of financial openness and democratic governance is compelling, although the direction and nature of the link need study (see figure A5.1 in annex 5). Openness also brings with it increased risks. Volatility in capital flows creates uncertainty in economic conditions, raises the cost of capital, may 119 T H E QU A L I T Y O F G R O W T H adversely affect long-term investment and growth, and slows poverty reduction efforts. Based on data from 90 developing countries, a strong cor- relation exists between the volatility of capital flows and the volatility of growth, as measured by the standard deviation of annual growth rates in real GDP (figure 5.3). In addition, using data from 130 countries between 1960 and 1995, Easterly and Kraay (1999) found that growth volatility, based on the standard deviation of GDP growth, has a negative effect (- 0.18) on average per capita growth. Two broad risk categories, those related to distorted domestic policies and those associated with external factors, can create economic problems for for- eign investors and policymakers. Distorted domestic policies and weak regu- latory and institutional environments provide incentives for banks and cor- porations to build up excessive short-term external liabilities relative to their short-term assets or unhedged foreign exchange positions. Examples of do- mestic sources of such risks include explicit and implicit government guaran- tees, pegged exchange rates, directed lending to investment projects, and mounting contingent liabilities. Dooley (1996) argued that the adoption of fixed exchange rates and deposit guarantees, in the context of a liberalized but poorly regulated financial sector, may induce foreign investors to reap Figure 5.3. Relationship between Economic Growth Variability and Volatility in Private Foreign Capital Flows, 1975-96 Votality in annual GDP growth rates 14 - 12 - IQ - 6 * 0 % 00 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Volatility in annual capital flows Note: y = 2.02x + 2.15, r = 0.57. Source: See annex 5 for definitions and sources. 120 DEALING WITH GLOBAL FINANCIAL RISKS high private rates of return that do not benefit the borrowing countries. This underpricing of the true risks of underlying investments must be addressed to ensure balanced investment that could stimulate long-term growth and pov- erty reduction in borrowing countries. The second category of risk relates to the functioning of international financial markets, extemal factors, and foreign lender and investor shifts in sentiment, belief, and confidence that are not necessarily related to a country's long-term creditworthiness. Thus, Calvo, Leiderman, and Reinhart (1994) found that external factors, such as U.S. interest rates and volatility in OECD growth could explain 30-60 percent of the variance in capital flows to Latin America. Shifts in investors' sentiment and beliefs, as reflected in a sharp turnaround in capital flows and/or a spike in emerging market economies' borrowing costs, can be caused by coordination failures on the part of creditors. This coordination problem could happen because of incomplete information between creditors that could render their deci- sion to run or flee from a particular country dependent on the behavior of others. This dependence can generate a run, analogous to a bank run in do- mestic settings, adding a noncooperation premium on top of other country risk premiums (see Haldane 1999 for further elaboration of this point). Countries need to be prepared to deal with the risks associated with finan- cial integration and capital flow volatility. Because countries' preparedness var- ies and institution building takes time, governments can consider a spectrum of policy and regulatory actions as they open up to intemational capital flows in an orderly fashion. Growth Volatility and the Poor Financial crises are extremely costly. Latin America lost a decade of eco- nomic and social progress following the debt crisis of the early 1980s. East Asian countries lost an estimated US$500 billion, based on 1996 prices and exchange rates, in aggregate domestic output between 1997 and 1999, as measured by deviation from historical trends, or nearly 1.3 times the external debt of those countries in 1996 (see annex 5 for the method of calculation). In addition, the international financial community ex- tended substantial financial assistance through multilateral and bilateral rescue loans to crisis-affected countries in the 1990s. In particular, growth volatility has severe consequences for the poor, who lack assets to smooth their consumption during economic downturns.6 The social costs associated with the crises in emerging market economies have been substantial. In just one year, unemployment doubled in Thailand 121 T H E QU A L I T Y O F G R O W T H and tripled in Korea, while standards of living fell 14 percent and 22 per- cent, respectively. Indonesia also experienced a 25 percent decline in its standard of living (Stiglitz and Bhattacharya 1999) and a sharp increase in the number of poor. By the third quarter of 1998, Thai workers' real-wage income had fallen 24.8 percent from the noncrisis, trend rate (Krongkaew 1999). Levinsohn, Berry, and Friedman (1999) examined the impact of price hikes on the cost of living of poor households and found that in Indo- nesia, the poor were indeed hit the hardest compared with other groups. Because of a dramatic rise in food prices, the cost of living for the poorest income decile rose more than 130 percent after the crisis. The urban poor, who lack access to land and do not own their homes, were the most ad- versely affected by the crisis. Thus, as a result of the crises, East Asian countries experienced sharp reversals of their previous achievements in poverty reduction (see World Bank 2000a). Past and Present Risk Management To protect their growth and gains in poverty reduction, developing coun- tries must be better prepared to deal with the risks associated with finan- cial integration and capital flow volatility. Global financial risk and the strategies for managing it have changed substantially over the past 50 years, and new approaches are needed to deal with the new risks. Early Mechanisms and Arrangements for Managing Risk Viewed from the perspective of financial risk management, the Bretton Woods period (1945-73) exhibited a high degree of stability by judiciously combining fixed exchange rates with capital controls on the external side, and Keynesian macroeconomics and welfare state positions on the domestic side.7 The Bretton Woods approach gave priority to fixed exchange rates and national policy autonomy. Capital controls were an accepted norm of the in- ternational monetary system in the 1950s and 1960s. Not until September 1997 did the Interim Committee of the International Monetary Fund agree that the Fund's Articles of Agreement "should be amended to make the pro- motion of capital account liberalization a specific purpose of the Fund and to give the Fund appropriate jurisdiction over capital movements" (Fischer and others 1998, p. 47).8 With economies relatively closed to capital flows, gov- ernments could exercise fiscal and monetary policy in pursuit of national ob- jectives, such as full employment and social equity, without fear of capital flight. This high degree of policy autonomy also served the cause of democ- racy, particularly in Western Europe.9 122 DEALING WITH GLOBAL FINANCIAL RISKS In the 1970s, once Western European countries had achieved currency convertibility in their current accounts, the free movement of capital across national boundaries began to emerge as an important policy priority. The collapse of the Bretton Woods system between 1971 and 1973, the move toward a floating exchange rate regime, rising oil prices, chronic inflation, and slumping global economic conditions intensified currency- and interest-rate risks in global financial markets. The responses were princi- pally market-based solutions, exemplified by the drive toward the interna- tional diversification of capital and the rapid expansion of derivatives mar- kets (interest and currency forwards, options, and swaps). Macroeconomic policy in OECD countries shifted from an emphasis on full employment to greater attention to macroeconomic stability, defined as smaller fiscal defi- cits and lower inflation and interest rates. Financial Risk Management in the 1990s In the 1990s, numerous liquidity and currency crises erupted in both indus- trial and developing countries: the European Monetary System during 1992 and 1993, Mexico between 1994 and 1995, East Asia in 1997, the Russian Federation in 1998, and Brazil and Ecuador in 1999. All these emerging mar- ket economies first experienced a surge in capital flows (from the early to mid-1990s) and then fell victim to sudden reversals; East Asia experienced reversals on the order of 10 percent of GDP. The crises of the 1990s exposed several weaknesses in intemational financial markets, namely: * World capital markets failed at several levels. Borrowing countries were not monitoring the high exposure of their domestic banks and corporations to foreign currency risk. Credit rating agencies and other major international players failed to properly assess country risk in the globalizing financial environment of the 1990s. Regula- tors failed because of weak regulatory and supervisory frameworks. Financial risk management specialists underestimated the positive correlations between the quality of private sector credit and the quality of sovereign credit, and so failed to identify the causes of contagion in emerging market economies. * Capital flows into many developing countries were channeled through short-term banking instruments because of implicit gov- ernment guarantees for banks. Many market participants suc- cumbed to the moral hazard in these perceived government guar- antees. Credit standards and prudent project appraisals were often compromised, leading to overinvestment in sectors with surplus 123 I F1 1 A 1L I T Y 0 F G R () W T Ii capacity or declining demand. The result was simultaneous domes- tic banking system collapse and foreign exchange liquidity crises in countries with fixed exchange rates. The primary sources of instability were in the capital account, not in the current account, a situation that the Bretton Woods institutions were designed to prevent. In today's global financial environment, a country's total balance sheet, defined by its assets and its debt and eq- uity liabilities, should be the measure of its extemal payments position. These weaknesses reflect major shifts in the global financial landscape, which can be characterized by the internationalization of the banking busi- ness; the breakdown of traditional boundaries between financial and insur- ance functions; the new investment opportunities in emerging markets; and the broader investor bases in emerging market economies, such as commer- cial banks, pension funds, hedge funds, and insurance industries. These shifts put new demands on risk management; governments must employ more judicious strategies at international, institutional (corporate and fi- nancial institutions), and national levels. A Broad Framework of Risk Management Much current activity in risk management concerns how best to handle, through better crisis prevention, the risks of capital flows to developing coun- tries, and through containment and orderly resolutions when crises do occur. We present a broad, two-part framework for risk management that favors a moderate view. An appropriate regulatory framework and related instru- ments for controlling short-term capital flows should accompany an orderly opening up of financial markets. Public support for openness should be maintained by government provision of cushions against risks, such as social safety nets and well-designed and cost-effective redistribution policies. International Policy and Regulatory Responses With the memory of the 1980s debt crisis and its prolonged resolution still fresh, governments promptly implemented international policy and regula- tory responses to the 1997-99 crises. In major industrial countries, they eased monetary policy, extended large rescue loans, developed interna- tional standards of good practice and disclosure, and established high-level committees to strengthen the soundness of banks and other financial insti- tutions (see Drage and Mann 1999 for more examples of crisis resolution). In February 1999, the G-7 (which comprises Canada, France, Ger- many, Italy, Japan, the United Kingdom, and the United States) finance 124 DEALING WITH GLOBAL FINANCIAL RISKS ministers and central bank govemors endorsed the creation of the Financial Stability Forum. At the new roundtable, the G-7 convened monetary au- thorities, principal regulatory agencies, and multilateral institutions to as- sess vulnerabilities in the global financial system and identify responses. Institutional Responses Financial risk management at the institutional level advanced significantly in the late 1990s. Today, financial and nonfinancial institutions use quantita- tive risk measurement techniques such as value at risk, volatility and beta measures, option pricing models, and Sharpe ratios. Using these tools, financial institutions have the ability to systematically measure and control market-related risk under normal volatility. In addition, the rapid expansion in the credit derivatives market is fundamentally altering the banking busi- ness by providing opportunities to trade credit risk. Risk management at the corporate level is moving toward a companywide integrated approach that encompasses credit, market, and liquidity risks. National Responses: Reconciling Financial Integration and National Policy Autonomy The integration of financial markets imposes a much more severe constraint on national policy choice than do other aspects of globalization, such as trade in goods and services, on which liberalization efforts had concentrated since World War II. Capital market integration reduces the ability of national gov- emments to conduct policy, especially macroeconomic policy, because of the risk of capital exit. Those holding this view, which is based on the Robert Mundel and J. Marcus Fleming model of an open macroeconomy, argue that countries can attain only two of the following three conditions: capital mo- bility, fixed exchange rates, and monetary policy autonomy. Redistribution to Mitigate Risk. Democratic societies need to resolve the tension between financial market integration and national policy au- tonomy to pursue their democratically defined economic and social goals. This tension relates to the ability of national governments to regulate, tax for redistributive purposes, and share risk while following the discipline needed in a global setting. In a world of high international mobility of capi- tal, open democratic societies must balance the threat of capital exit, made easier by the open capital markets, with political demands for voice and government intervention to cushion market dislocations. Investors dissatis- fied with the host countries' policies or prevailing investment climate find it easier to shift their financial resources to other countries and regions, 125 THE QU A L I T Y O F G R O W T H with a subsequent disproportionate distribution of costs borne by less mo- bile factors of production, that is, labor and land. Thus, the motivation for redistribution as income insurance-distinct from altruism and other mo- tives related to poverty reduction-is induced by volatility and insecurity in underlying economic conditions and when citizens are risk averse. The risk of capital exit intensifies economic insecurity and risk for a broad sec- tion of society. Because the rich are likely to benefit relatively from capital market liberalization, at least initially, while the poor may bear the costs, the political dimension of capital market liberalization is important and re- quires careful attention. Social Sector Spending, Openness, and Political Liberty. The counter- balance to the threat of exit of capital is the political voice of citizens de- manding protection against extemal risks through redistribution, social safety net programs, and other insurance-like measures.d In the absence of a market for such risk insurance, rational citizens will structure nonmarket institutions to reduce the welfare losses incurred from volatility in economic conditions. Thus, in this interpretation voice belongs to the political sphere, and how it is exercised is a function of the underlying political institutions and, in par- ticular, the strength of democracy and the corresponding degree of political and civil liberty: the higher the degree of democracy, the greater the need to balance the threat of capital flight with political demands, which include po- litical incentives to increase govemment intervention in cushioning market dislocation. It is fair to say that the political voice of citizens, who demand protection through redistribution, social safety nets, and other insurance-like measures, has been critical in easing the tension between politics and finan- cial openness in OECD countries. Govemment spending on health, educa- tion, social security, and welfare in high-income countries between 1991 and 1997 averaged about 25 percent of GDP, with relatively small, open Euro- pean countries such as Denmark, Norway, and Sweden spending as much as 30 percent." A positive association exists among redistribution, financial openness, and civil and political liberty for a large sample of countries (table 5.2). Statistical analysis confirms that financial openness, democracy (as de- fined by political and civil liberties),"2 and govemment social spending go hand in hand (table 5.3, figures 5.4, 5.5, and 5.6). Yet, because redistribution often needs to be financed through discre- tionary taxation, policymakers need to assess the associated fiscal and mac- roeconomic costs. Almost all modern, advanced democracies are open to international capital movements. The relationship between financial openness and de- mocracy appears to be primarily a function of per capita income: with few exceptions, rich countries have democratic governments and are 126 DEALING WITH GLOBAL FINANCIAL RISKS Table 5.2. Country Grouping by Financial Openness Largely Largely Category Open open closed Closed 1 Democracy indexa 0.81 0.71 0.63 0.48 2 Civil liberties' 2.28 3.30 3.38 4.55 3 GDP per capita, 1990-97 13,147 3,051 2,317 1,557 4 Social expenditure (percentage of GDP)' 22.30 23.50 12.50 6.70 5 Total government expenditure (percentage of GDP)d 26.00 19.90 23.40 27.70 6 General governtment consumption (percentage of GDP), 16.10 17.90 15.50 14.70 Number of countries 46 10 34 11 Note: Table displays the group averages computed for countries with data. Definition of variables: a. Ranges from 0 (lowest) to 1 (highest), computed on the bases of political rights and civil liberties indices, see endnote 11 for details. b. A measure of respect for and protection of a country's citizens' religious, ethnic, economic, linguistic, and other rights, including gender and family rights, personal freedoms, and freedoms of the press, belief, and association. c. Sum of health, education, and social security and welfare; average 1991-97. d. Average of central govemment and budgetary accounts plus state or provincial govemment, 1990-97. e. All current expenditures for purchases of goods and services by all levels of government, excluding most government enterprises, 1990-97. Source: Annex 5. Table 5.3. Estimation Results of the Binomial Logit Model on the Likelihood of Countries Belonging to the High Democracy and High Financial Openness Categories Independent variable Coefficient Standard error Marginal effects Constant -11.234** 2.7500 -2.0296 Log (ratio of social expenditure to GDP) 1.534* 0.6146 0.2772 Log of GDP per capita 0.795* 0.3156 0.1436 Actual number of countries in the target group 28 Predicted number of countries in the target group b 20 Actual number of countries in other groups 39 Predicted number of countries in other groups 32 Log likelihood -27.744 *p < 0.05. -P < 0.01. Note: The dependent variable is coded 1 if the country falls into the financial openness-high democracy, and 0 otherwise. a. Marginal change in the probability resulting from an infinitesimal change in the explanatory variable. b. Target group refers to countries with high level of political rights and high financial openness. Source: Annex 5. open to international capital movement because they have a high degree of financial sector development and enjoy macroeconomic stability, stable expectations of peaceful regime change, domestic rule of law, and stable institutions that guarantee civil and political liberties (for a more 127 Tr H E OU A L I T Y O F G R O W T H Transfers and social spending ease the tension between financial openness and politics Figure 5.4. Relationships among Democracy, Financial Openness, Capital Mobility, and Government Social Expenditures Financial Democracy r = 0.45 openness \\\ r = 0.47 r=0.53 r = 0 7 03 r= 0.58 Govenment social Transfers expenditures (precentage of GDP) (percentage of GDP) Note: The cross-country data, with sample sizes ranging from 70 to 140, show statistically significant results at I percent (except for the correlation between transfers and financial openness, which is siginificant at 5 percent) for all relationships. Source: Annex 5 (table ASS). Figure 5.5. Relationship between Financial Openness and Social Expenditure (controlling for per capita income) Financial openness index 2.0 1.8 * . . 1.6 - urn 1.4 -** 1.2 -* m U 1.0 I I I 0.5 1_5 2.5 3.5 4.5 Log (ratio of social expenditure as a percentage of GDP) Note: y = 0.17x + 1.17 R2 = 0.32 Source: See annex 5 for a description of the data, 128 DEALING WITH GLOBAL FINANCIAL RISKS Figure 5.6. Country Classification: Political Rights and Financial Openness Financial openness Low High Low 23 9 Democracy High 32 37 Note: The level of democracy is derived from the political rights and civil liberties indexes from the Freedom House survey of Freedom in the World. The level of financial openness is defined by a score of >1.6 in the financial openness index (see table A5.2 in annex 5). Source: Author's calculation. detailed discussion of the link between democracy and financial open- ness, see Dailami 2000). However, the link between democracy and financial openness proves to be more complex; analysis reveals that more than just income influ- ences this link. International policy coordination in macroeconomic policy and financial regulation and supervision is part of the answer. It has been instrumental in reducing payment imbalances, in stabilizing ex- pectations for currency and interest rate movements, and in lessening the volatility of capital flows across borders. The coordination of interna- tional banking regulation in industrial countries, such as the Basel Capi- tal Accord of 1992 and the subsequent Core Principles for Effective Banking Supervision, has also been a significant factor in providing eco- nomic stability to OECD democracies.'" Empirical investigations of country classifications along the two axes of democracy and financial openness support the view that distribution policy contributes to democracies and open markets (figure 5.6 and table 5.3). Us- ing logit analysis it can be shown that both per capita income and the ratio of social expenditures to GDP are statistically related to the likelihood that a country will be both financially open and democratic (see annex 5 for model specification and estimation, and Dailami 2000 for a more detailed analysis). After controlling for income in the analysis, redistributive policy, which in- cludes programs for public expenditure on social security, health, housing, welfare, education, and transfers, figures prominently in the link between de- mocracy and financial openness. 129 T H E QU A L I T Y 0 F G R 0 W r H Capital Controls as Instruments of Risk Management. Capital con- trols can be employed as an alternative approach for resolving the tension between capital market integration and national policy autonomy. Interest in this approach has been rekindled by the 1997-99 financial crises in Asia and Latin America. Capital controls, particularly on short-term flows, are desirable to reduce volatility under some circumstances, such as weakness in local financial markets, euphoric or panic behavior by foreign investors, and structural balance of payments problems. Many policy interventions are available for managing capital flows, includ- ing taxes and market-based instruments, such as contingent liquidity facilities and remunerated or nonremunerated reserve requirements on risky short-term flows. Argentina and Mexico have used contingent liquidity facilities and re- munerable liquidity requirement for banks, and Chile used nonremunerable re- serve requirements on risky short-term capital inflows between 1991 and 1998. Controls on short-term capital in Chile have attracted considerable in- terest, partly because they are market based, transparent, and easier to phase out than quantitative controls (box 5.2). The controls were effective 130 DEALING WITH GLOBAL FINANCIAL RISKS in changing the composition of debt by reducing short-term capital inflows while increasing long-term flows and allowing for a larger wedge between domestic and foreign interest rates. The measures were countercyclical; they were imposed in 1991 after a growing tide of capital inflows, between 1988 and 1990, and phased out in September 1998 when they were no longer needed during the global financial crises. Conclusions Countries face two challenges in integrating their capital markets. The first is the pace at which countries dismantle administrative controls over capi- tal flows and move toward capital account convertibility. The second is the incentive system and regulation of international financial flows to mini- mize risks and panics. Countries need suitable mechanisms to balance both the benefits and the risks of financial integration. Technological advances and the sheer size of financial markets make the risk of panic and crisis ever present. However, governments have various options for significantly re- ducing that risk. Pursuing sound macroeconomic policies is an obvious first step, but not sufficient. Recent experience shows that macroeconomic stability is not enough to guarantee enduring results and sustainable growth. To en- sure sustainable growth it must be reinforced by actions that remove distortionary policies that provide incentives for short-term foreign capi- tal inflows that could lead to heightened financial vulnerability. Domes- tic regulation and supervision of banks and other intermediaries need to be strengthened and corporate governance improved. With the move toward democracy worldwide, mechanisms for pro- viding citizens with insurance against the risks of capital mobility, through either the marketplace or redistributive policies, are equally im- portant if political pressure for capital controls is to be averted. In the long run, the globalization of capital requires an open institutional framework to ensure transparent accounts; secure property rights; and permit enforceable contracts, regulations, and mechanisms to manage risks. Establishing such a framework enhances assurance that open finan- cial markets will contribute fully to stable growth and poverty reduction. The remarkable economic turnaround in crisis-hit countries over the past months, reinforced by measures already taken at the international level to strengthen the architecture of international financial markets, bodes well for the prospects of greater financial stability and collective commitment to an open and liberal international financial system in the new millennium. 131 T HE O_J ALIT Y OF G R O W T H Notes 1. A large body of literature has developed over the last few years discussing the causes and consequences of recent financial crises in emerging market econo- mies. See Calvo and Mendoza (1996); Corsetti, Pesenti, and Roubini (1998); Krugman (1998); Obstfeld (1996); Radelet and Sachs (1998); and Sachs, Tornell, and Velasco (1996). On causes of capital flow volatility, see Dooley (1996), Lopez-Mejia (1999), Montiel (1998), and World Bank (1997f). 2. From a historical perspective, the globalization of finance in the 1990s is equivalent to the level reached during the gold standard period of 1870-1914. However, during the gold standard only a few industrial countries were in- volved in capital flows (see Verdier 1998). 3. Moral hazard is a key concept in the economics of asymmetric information. Moral hazard occurs when economic actors covered by some form of insurance take more risks than they would otherwise take. Typical examples include an insured driver driving recklessly or an insured banker engaging in imprudent lending practices. 4. See also Helleiner (1994) for an account of how in 1974 the United States lifted the temporary capital restrictions of the mid-1960s. 5. Markets will provide leeway only if lenders perceive that countries are under- taking adjustments that fundamentally address existing and prospective imbal- ances. Otherwise, markets will eventually exert discipline that may brutally shorten the time allowed for readjustment (Dailami and ul Haque 1998). 6. See, for example, Diwan (1999); Krongkaew (1999); Levinsohn, Berry, and Friedman (1999); Lustig (1999). 7. This policy mix was referred to by Ruggie (1983) as "a compromise of embed- ded liberalism." It connotes a commitment to a liberal order different from both the economic nationalism of the 1930s and the liberalism of the gold standard. For further elaboration, see Garrett (1998). Sally (1998) also referred to embed- ded liberalism as "mixed system thinking." Also see Dailami (2000). 8. Reflecting the understanding of the time, Keynes expressed the issue succinctly in his often quoted 1944 speech to Parliament, stating: "Not merely as a feature of the transition, but as a permanent arrangement, the plan accords to every member govemment the explicit right to control all capital movements. What used to be heresy is now endorsed as orthodox... It follows that our right to control the domestic capital market is secured on firmer foundations than ever before, and is formally accepted as a part of agreed international agreements" (Gold 1977, p. 11). 9. However, in the Bretton Woods era there were periodic balance of payments crises, exchange rate devaluations, and stop-go growth episodes. 10. The idea of distribution as insurance has a long tradition in welfare eco- nomics going back to Harsanyi (1953), Lerner (1944), and Rawls (1971). 132 DEALING WITH GLOBAL FINANCIAL RISKS More recently this issue has been analyzed from the perspective of constitu- tional political economy (see Mueller 1998; Wessels 1993). 11. Focusing on globalization through trade, Rodrick (1997b) also emphasizes the relationship between redistribution and openness. 12. More precisely, one measure of democracy, following the recent literature ex- ploring the role of democracy on economic growth, income levels, and wages, defines democracy as a composite index and draws on the Freedom House mea- sures of political and civil liberty; that is: 14 - civil rights - political rights Democracy = 12 This index will be defined from 0 to 1, with 0 indicating low democracy and 1 indicating high democracy. Political and civil liberty indexes are from the Com- parative Survey of Freedom that Freedom House has provided on an annual basis since 1973. 13. See Bryant and Hodgkinson (1989) and Webb (1994) for a discussion of inter- national policy coordination in macroeconomic terms and Kapstein (1989) for information on intemational coordination of banking regulation. For selected readings in the voluminous literature on the need for better regulation and su- pervision, see Alba and others (1998); Caprio and Honohan (1999); Claessens, Djankov, and Klingebiel (1999); and Stiglitz 1993. 133 C H A P T E R 6 GOVERNANCE AND ANTICORRUPTION Just as it is impossible not to taste honey or poison that one may find at the tip of one's tongue, so it is impossible for one dealing with government funds not to taste, at least a little bit, of the King's wealth. -Kautilya, The Arthashastra L A 7 ritten in ancient India more than 2,000 years ago, the j / j / Arthashastra is a detailed vision of a society that weaves together socioeconomic, institutional, and political T V T variables. In contemporary development writing, such notables as Hirschmann, Myrdal, Coase, Stiglitz, North, Olson, and Williamson provided a broad view of the interplay of institutions and conventional economic variables. In recent years, increasing attention has tumed to corruption, starting with Rose-Ackerman and Klitgaard, in some measure due to the growing awareness of its dire consequences for development. However, most contemporary economic development work The work in this chapter draws from a number of collaborative initiatives between the chapter's author and World Bank staff on governance issues, including Aart Kraay, Sanjay Pradhan, Randi Ryterman, and Pablo Zoido; as well as collaboration with Joel Hellman and Geraint Jones at the European Bank for Reconstruction and Development and Luis Moreno Ocampo of Transparency International, and the invaluable inputs of the World Bank Institute's Govemance Team and the Bank's Public Sector Group. The data used in this chapter originate from various research projects and surveys (as well as outside expert rating agencies) and are subject to a margin of error. Their purpose is not to present precise com- parative rankings across countries, but to illustrate characteristics of government perfor- mance. Hence, no ranking of countries is intended by either the chapter author, the World Bank, or its Board of Directors. For further details on the empirical "unbundling" of gover- nance and corruption, the data, and methodological issues, see annex 6 and visit http:// www.worldbank.org/wbi/governance. 135 T H E QU A L I T Y O F G R O W T H has underestimated the primacy of governance, broadly defined, for growth and development. Missing too often is the recognition that an effective and transparent government, operating within a framework of civil liberties and good governance, is vital for sustained welfare gains and poverty alleviation. Also missing is an integrated view of governance and corruption. Indeed, corruption ought to be seen as.a symptom of the state's fundamental weaknesses, not some basic or single determinant of society's ills. This chapter does not present an all-encompassing approach to the study of governance and corruption. Instead, we dissect the notions of governance-and of corruption and state capture-and present aspects relevant to the growth and development of nations to derive insights for strategies to improve governance. We lack many of the answers; the emerging lessons of success and failure are being distilled. Nonetheless, advances have been made in the conceptual, empirical, and practical un- derstanding of these issues. Some of this progress pertains to the sharpen- ing and "unbundling" of the notions and measurements of governance and corruption. This unbundling permits a better understanding of the causes and consequences of misgovernance, helping provide improved policy advice. Governance Affects the Quality of Growth Worldwide evidence suggests that a capable state with good and transparent government institutions is associated with higher income growth, national wealth, and social achievements. Higher incomes, investment, and growth, as well as longer life expectancy, are found in countries with effective, hon- est, and meritocratic government institutions with streamlined and clear regulations, and also where the rule of law is enforced fairly, where the poli- cies and legal framework has not been captured by the vested interests of the elite, and where civil society and the media have an independent voice en- hancing the accountability of their govemments. International and histori- cal experience also tells us that capable and clean government does not first require a country to become fully modernized and wealthy. The experience of such industrializing countries as Botswana, Chile, Costa Rica, Estonia, Po- land, and Slovenia, as well as evidence over the past 20 years from economies like Singapore and Spain, illustrate this lesson. Previous chapters emphasized the need for policies, regulations, and public resources to promote market-oriented development and to mitigate the negative impacts of externalities and market failures. With an emphasis on poverty and income distribution, they examined those factors that ad- versely affect human capital and the environment. A key role for the state involves delivery of public services and goods vital to achieving sustained 136 GOVERNANCE AND ANTICORRUPTION growth and thus reducing poverty. Also, govemments need to set up effec- tive policymaking structures, market-friendly policies, and efficient and streamlined regulatory frameworks, as well as eliminate unnecessary regula- tions on enterprises, such as price controls, trade restrictions, enterprise li- censing, and bureaucratic harassment. Often, however, governments have paid insufficient attention to regula- tions governing child labor, worker safety, infrastructure monopolies, finan- cial sector supervision, and the environment. Moreover, in many settings there has been a bias in the size, composition, and delivery of public expendi- tures and investments to benefit elite interests, frequently resulting in underinvestment in human capital and in outcomes harming the poor. Such elite interests often also lead to legal, regulatory, and policy anticompetitive capture. The study of governance and inadequate institution building is es- sential to understand these outcomes. A political process determines public policies and the allocation of pub- lic benefits and expenditures. Its success depends on accountable govern- ment, community participation, and a strong voice for people and competi- tive enterprises. Effective adoption and use of policies and expenditures requires good governance. Enterprises need to operate within a legal and contractual framework that protects property rights, facilitates transactions, deters attempts by elite enterprises to capture the state, allows competitive market forces to determine prices and wages, and lets firms enter and exit the market. The public sector can do much to lower the transaction costs for farms and firms by supporting them with information and institutions and by rooting out misgovernance and corruption. Defining and Unbundling Corruption and Governance Corruption is commonly defined as the abuse of public office for private gain. Notwithstanding the debates about whether certain activities can be classified as corrupt or not and the need to unbundle corruption, the vivid day-to-day illustrations in the press and in conversation circumscribe the discussion of what constitutes corruption. However, governance is a much broader concept than corruption. We define governance as the exercise of authority through formal and informal traditions and institutions for the common good. Governance encompasses the process of selecting, monitor- ing, and replacing governments. It also includes the capacity to formulate and implement sound policies, and the respect of citizens and the state for the institutions that govern economic and social interactions among them. From this notion, we can divide governance into six components, or- ganized around three broad categories as follows (a) voice and accountabil- ity, which includes civil liberties and freedom of the press, and political 137 THE QU A L I T Y O F G R O W T H stability; (b) government effectiveness, which includes the quality of policymaking and public service delivery, and the lack of regulatory bur- den; and (c) rule of law, which includes protection of property rights and independence of the judiciary, and control of corruption (Kaufmann, Kraay, and Zoido-Lobat6n 1999a,b). Therefore, in unbundundling governance, we posit that corruption is one among six closely intertwined governance components. Governance affects welfare and the quality of life through complex direct and indirect channels that we do not yet fully understand. An improvement in one gov- ernance component, such as civil liberties, directly enhances the quality of life for a country's people even when all other socioeconomic factors re- main constant. Thus, governance can be a direct input into the well-being of the population. However, important indirect effects are also at play. For example, misgovernance can hurt the growth rate of incomes and human capital, and increase the rate of natural resource depletion-often the result of the vested interests of politicians and the elite. Furthermore, misgoverned states tend to exhibit a distorted set of economic and institutional policies that blunt factor productivity, growth, and poverty alleviation. Therefore, via complex direct and indirect mechanisms, effective and clean government is vital for imple- menting and sustaining sound economic and institutional policies and for promoting human capital development and poverty alleviation. Empirical Measures of Governance Recent empirical studies suggest the importance of institutions and gover- nance for development outcomes. Knack and Keefer (1997) found that the institutional environment for economic activity determines, in large measure, the ability of poor countries to converge to industrialized coun- try standards. In turn, La Porta and others (1999) investigate the determi- nants of the quality of governments and, inter alia, find that the type of le- gal regime matters, as well as other historical factors. The definition of governance, as presented in the previous section, is broad enough that a wide variety of cross-country indicators might shed light on its various aspects. Applying such a broad definition, Kaufmann, Kraay, and Zoido-Lobat6n analyzed hundreds of cross-country indicators as proxies for various aspects of govemance. These indicators came from a variety of or- ganizations, including commercial risk-rating agencies, multilateral organiza- tions, think tanks, and other nongovernmental organizations (NGOs). They are based on surveys of experts, firms, and citizens and cover a wide range of topics: perceptions of political stability and the business climate, views on the efficacy of public service provision, opinions on respect for the rule of law, 138 GOVERNANCE AND ANTICORRUPTION and reporting on the incidence of corruption.' (See annex 6 for a description of Kaufmann, Kraay, and Zoido-Lobat6n's methodology.) Skeptical reactions naturally arise regarding the wealth of data on gov- ernance. Are the data informative? What can business analysts on Wall Street possibly know about corruption in Azerbaijan, Cameroon, Moldova, Myanmar, or Niger? Are the data coherent? Do reported ratings by enter- prises about pressures from civil servants and their waiting times for cus- toms clearances tell us something about the government's effectiveness in general, or do they measure totally different things? Are the data compa- rable? Can a score of 3 (out of 4) in transition economies be compared with a score of 7 (out of 10) in Asian countries? In addition to meeting these criteria, can the data be useful for rigorous econometric analysis of corrup- tion or for policy advice purposes? These questions, addressed in detail in the two references and in annex 6, motivate the empirical strategy for measuring governance: the data are mapped to the six subcomponents of govemance and expressed in common units. The data are informative, within measurable limits, but the impreci- sion in the estimates requires care in their presentation and use for policy advice. These six distinct aggregate governance indicators are then devel- oped, imposing some structure on available variables and improving the re- liability of the measured governance component, which significantly ex- ceeds the accuracy of any single governance measure. For illustration, we first consider the measurement issues for one of the six composite governance components: rule of law. In figure 6.1, the vertical bars depict country-specific confidence intervals for the esti- mated ("point estimate") levels of governance. The confidence intervals (vertical lines) reflect the disagreement, or margin of error (among the original individual sources provided by the various external organiza- tions) about the application of rule of law.2 The differences among more than 160 countries are large for rule of law, as well as for the other five measures. Countries are ordered along the horizontal axis according to their (admittedly imprecise) rankings, while the vertical axis indicates the estimates of governance for each country. The margins of error for each country, depicted by each thin vertical line, can be considerable. Thus, it is misleading having countries "run" in seem- ingly precise worldwide "horse races" to ascertain their ranking on various governance indicators. Instead, the following approach that groups coun- tries into three broad categories akin to a traffic light for each governance dimension is more appropriate and statistically warranted: * Red light: Countries in this category could be considered to be in governance crisis in that particular component. In fact, despite the 139 T HE 0QUA L IT Y O)F G ROW T H z ..... CL~~~~~~~~~~~~~~~~~~~~~~~~~~~- C 0 ------------- -- Z~~0 C ----- ----- a.C C~~~~~~~~~~~~~~~~~~~~~~~~~>C o - CC~~~~~~~~~~~~~~........ , -~~~~~~ 0 ~ ~ ~~I 61. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ . CU CC CC~~~~~~~~~~~~~~~~~~~~~~~~~~~~....... g 0 C-- ~ CC~~~~~~~~~~~~~~~~~~~~~~~~~~~ It ~ C ~~~~~~ 0~~~~~ C- C-. V ~~~~~~~~~~~~~~~~ -~~~~t - C)0 ~~~~~~~~~~~~~~~~~~~~~ 140~~~~~~~~~~~~~~~~~~~~~~~~~~~~ GOVERNANCE AND ANTICORRUPTION margins of error in the available data, it is still the case that a group of approximately 30 to 40 countries exhibit an extremely high probability of being in crisis where rule of law (or other gover- nance measures) is assessed. * Yellow light: Countries are vulnerable or at risk of falling into a gover- nance crisis in a particular governance component. * Green light: Countries have better governance and are not at risk In moving away from a false sense of precision common in indexes that rank countries internationally (which are subject to considerable margins of error), this altemative approach of broad categorical groupings can flag vul- nerabilities where a country falls into the red light or yellow light groups. For another governance component, in this case measuring control of corruption (also based on data from the late 1990s), selected countries are presented within such an illustrative traffic light framework in figure 6.2. Effects of Governance The cross-country data indicate a significant simple correlation between governance and socioeconomic outcomes. To explore the effect of gover- nance on socioeconomic variables, we estimated a two-stage least squares regression of a socioeconomic variable (for example, per capita income) on a constant and on the governance component, using historical indicators as instruments (following the approach of Hall and Jones 1999). Within such an approach, concerns about measurement error and omitted variables were also addressed (see Kaufmann, Kraay, and Zoido-Lobat6n 1999b for de- tails). The evidence challenges the argument that only rich countries can afford the luxury of good governance. The empirical analysis suggests a large direct effect going from better governance to better development outcomes. Consider an improvement (of one standard deviation) in the rule of law from the low levels in the Rus- sian Federation today to the middling levels in the Czech Republic or a similar reduction in corruption from that in Indonesia to that in Korea. In this framework, it increases per capita incomes two to four times, it reduces infant mortality by a similar magnitude, and it improves literacy by 15 to 25 percentage points in the long run. And consider that the differences in govemance for these two pairs of countries are not very large. Much larger improvements in government effectiveness from the levels in Tajikistan (in the red light group) to those in Chile (in the green light group) in this framework would nearly double the development impacts just mentioned. The relationships between development outcomes and four measures of governance are illustrated in figure 6.3. The heights of the vertical bars show 141 T H E O_U A L I T Y 0 F G R 0 W T H .2 .......... ........... ........... ..... ................. --------------------- 2 JJ u u ------------- - Lu - ---------------- v m Ej -a CL ------------------------------------ ----------- --------- ----------- ................ ------ 0 T; 72 8 -------------- 7j, 0 C) .C vu LI u m 4- IE 1 Lvl CZ .2 ............ E --------------- ............. ------------------ ------------------ CL -------------- 't .0 0 0 at ---------- ZE -- - -------- ----------- '--,44:i 4, 0 El E 0 ------------------ F- 2 j bL CZ 142 GOVERNANCE AND ANTICORRUVTION Figure 6.3. The Development Dividend of Good Governance Infant mortality and corruption Per capita income and regulatory burden Infant mortality (deaths per 1,000) Per capita income (US$) 90 - 12,000 - 80 - 70 - K10,000 60 -8,000- 50- 40 - K < 6,000 - 30 -4,000- 20- 10 - 2,000Q Q Weak Average Strong High Medium Low Control of corruption Regulatory burden Literacy and rule of law Life expectancy and voice and accountability Literacy (percent) Life expectancy (years) 90 80- 100 - 70- 80 -60- 50- 60- 40- 40 -30- 20 - ~~~~~~~~~~~~20 20~~~~~~~~~~~~~~~1 10 Weak Average Strong Weak Average Strong Rule of law Voice and accountability Note: The heights of the vertical bars show the differences in average development outcomes in countries with weak, average, and strong governance. The solid lines show the estimated effect of governance on development outcomes. See endnote 3 in this chapter and table A6.1 in annex 6 for details on econometric tests (synthesized in the solid lines). Source: Kaufmann, Kraay, and Zoido-Lobat6n (1999b, 2000); http://www.imf.org/fandd. the differences in development outcomes in countries with weak, average, and strong governance and illustrate the strong correlation between good outcomes and good governance. After controlling for reverse causality and for the effects of other, nongovernance factors on development, the solid 143 T H E QU A L I T Y O F G R O W T H lines represent the estimated impact of governance on development out- comes: the "development dividend" of improved governance.3 Composite indicators of governance, based on multiple, outside sources of data, powerfully draw attention to governance issues. They are also indispensable for cross-country research into the causes and consequences of misgovernance. For instance, this large data set debunks the notion that larger countries are more corrupt (a statistical construct resulting from tests with a smaller number of countries). However, these new governance indicators provide only a first and rough benchmark of where countries stand relative to each other on governance issues and make a blunt tool for informed action to improve governance. To make composite indicators more specific and use- ful within a country, one needs to know much more about how perceptions of and data about misgovernance are reflected in policy and institutional failures. In-depth governance diagnostic tools are needed within a country to provide meaningful data and information for formulation of governance reforms. Against this background, the rest of the chapter addresses the folb lowing questions: How do corruption and misgovernance undermine devel- opment? What are the underlying causes of corruption? What kind of in- sights can be derived by unbundling corruption into distinct components? What kinds of diagnostic tools and strategic approaches can best serve a country intent on making progress toward good and clean government? Corruption Undermines Growth and Development Many studies have shown the pernicious effect of corruption on development. Mauro (1997) showed that corruption slows the growth rate of countries. He found that if Bangladesh reduced corruption to equal the level in Singapore and the growth rate was 4 percent a year, Bangladesh's average annual per capita GDP growth rate between 1960 and 1985 would have been 1.8 percent- age points higher, a potential gain of 50 percent in per capita income. The following are some of the many channels whereby corruption can weaken economic growth: * Misallocation of talent (Murphy, Shleifer, and Vishny 1991), in- cluding underutilization of key segments of society, such as women * Lower levels of domestic and foreign investment (Mauro 1997; Wei 1997) * Distorted enterprise development and growth of the unofficial economy (Johnson, Kaufmann, and Zoido-Lobat6n (1998) * Distorted public expenditures and investments and deteriorated physical infrastructure (Tanzi and Davoodi 1997) 144 C, O V E R>N A N C I A N 1) A N I I (: I' 1, 1 N * Lower public revenues and less provision of the rule of law as a pub- lic good (Johnson, Kaufmann, and Shleifer 1997) * Overly centralized government (Fisman and Gatti 2000) * State capture by the corporate elite of the ("purchased") laws and polices of the state, thereby undermining growth of output and in- vestment of the enterprise sector (Hellman, Jones, and Kaufmann 2000a; see annex 6 for details). Lower Investment Evidence from a large cross-section of countries suggests that corruption significantly reduces domestic and foreign investment. If the Philippines could reduce its corruption to the much lower level of that in Singapore, it would raise its investment-to-GDP ratio by 6.6 percentage points (Mauro 1997). By looking at bilateral FDI in the early 1990s from 14 source coun- tries to 41 host countries, Wei (1997) found evidence that corruption dis- courages investment. Reducing corruption to the low level in Singapore would have the same effect on foreign investment for a corrupt country as reducing the marginal corporate tax rate by more than 20 percentage points. Many countries afflicted by corruption also offer substantial tax in- centives to lure multinational firms. By controlling corruption, they could attract at least as much foreign investment without such tax incentives. Misallocation of Public Expenditures Some of the pioneers in the study of the economics of corruption have highlighted corruption's effect on the allocation of public finances (Klitgaard 1988; Rose-Ackerman 1989). Tanzi and Davoodi (1997) found that corruption increases public investment because it creates opportuni- ties for manipulation by dishonest high-level officials. It also skews the composition of public expenditure away from needed operations and maintenance spending and directs it toward new equipment purchases, thereby reducing the productivity of public investment, especially in in- frastructure. Under a corrupt regime, public officials shun health pro- grams because they offer less scope for rent-seeking. Corruption may also reduce tax revenue, because it compromises the government's ability to collect taxes and tariffs. Drawing on the findings of Tanzi and Davoodi, Wei (1997) showed that an increase in corruption, comparable to the corruption level of Singapore rising to that in Pakistan, would increase the public expenditure- to-GDP ratio by 1.6 percentage points and reduce the government revenue-to-GDP ratio by 10 percentage points. In addition, an increase in 145 T H E OQU A L I T Y O F G R O W T H corruption would reduce the quality of roads and increase the incidence of power outages, telecommunication failures, and water losses. Johnson, Kaufmann, and Zoido-Lobat6n (1998) also showed that cor- ruption reduces tax revenue, mainly through the growth of the unofficial economy. Overburdened by red tape and associated rent-seeking in the official economy, firms move to the unofficial economy and pay fewer taxes. Such reduced tax revenue is associated with a lower provision of key public goods, such as rule of law, further increasing the unofficial economy and impairing public finances. Impact on the Poor Where corruption prevails, growth is impaired, and this has an enormous ef- fect on poverty. Furthermore, the poor receive fewer social services, such as health and education. Corruption biases infrastructure investment against projects that aid the poor and impairs the use of small-scale entrepreneurial means to escape poverty. Worse, corrupt regimes often prefer defense con- tracts over rural health clinics and schools, a policy bias that worsens income distribution and diverts resources from the countryside to the cities. Gupta, Davoodi, and Alonso-Terme (1998) show that corruption in- creases income inequality and poverty through channels such as lower growth, regressive taxes, less effective targeting of social programs, unequal access to education, policy biases favoring inequality in asset ownership, re- duced social spending, and higher investment risks for the poor. As suggested in figure 6.3, Kaufmann, Kraay, and Zoido-Lobat6n (1999b) also found that corruption increases infant mortality and reduces life expectancy and literacy. Furthermore, when analyzing the UNDP's Human Poverty Index, the data suggest that it is negatively associated with the various indexes of governance and corruption even after controlling for GDP per capita. The mechanisms through which governance affects poverty are varied, complex, and still not fully understood. The matrix in table 6.1 suggests some of the complex effects of corruption on poverty through a variety of channels. Country analyses using new governance diagnostic tools illustrate how regressive corruption is as a tax. For instance, poor households in Ecuador must spend three times more in bribes as a share of their incomes than higher income households for access to public services (figure 6.4). Simi- larly, in various diagnostic surveys of public officials in Latin America in the late 1990s, bureaucrats in those agencies rife with corruption and lack- ing in meritocracy were found to discriminate against the poor by limiting access to basic services and by failing to pursue poverty alleviation-in contrast with the better access to the poor by agencies with less corruption and meritocracy (figure 6.5). 146 GOVERNANCE AND AN-FICORRUPTION Table 6.1. A Synthesis Matrix: Corruption and Poverty "Immediate" causes of poverty How corruption affects "immediate" cause of poverty Lower investment and growth Unsound economic/institutional policies due to vested interests Distorted allocation of public expenditures/investments Low human capital accumulation Elite corporate interests capture laws and distort policymaking Absence of rule of law and property rights Governance obstacles to private sector development Poor have smaller share in growth State capture by elite of govemment policies and resource allocation Regressiveness of bribery "tax" on small firms and the poor Regressiveness in public expenditures and investments Unequal income distribution Impaired access to public services Bribery imposes regressive tax and impairs access and quality of basic services for health, education, and justice Political capture by elites of access to particular services Lack of health and education Low human capital accumulation Lower quality of education and health care Source: Author. Impact of Corruption on Commerce and Corporate Influence on National Governance A common argument found in the literature contends that bribes to circum- vent bad government controls are like unofficial deregulation and can have positive effects, such as promoting enterprise development (Huntington 1968; Leff 1964; Liu 1985). This view-bribery as grease on the wheels of com- merce-may hold conceptually only in a very narrow sense if bad regulations are fixed independently of the behavior of public officials. Yet in reality, offi- tials often have discretion in the type and amount of harassment and regula- tions inflicted on individual firms. Tax inspectors can overreport taxable in- come (Hindriks, Keen, and Muthoo 1999), and fire inspectors can decide how many times to check a firm for safety "violations." Using data from two inde- pendent surveys on more than 6,000 firms in 75 countries, Kaufmann and Wei (1999) showed that firms that pay more administrative bribes waste more time with bureaucrats than those firms that do not pay bribes. Thus, the empirical evidence suggests that a firm engaging in petty or administrative bribery (for example, for licenses or red tape) does not nec- essarily benefit from paying bribes; neither does the business community or society more generally. Research evidence on the costs of corruption for overall business development is growing. For example, Fisman and 147 T FH E 0_ U A L I T Y O F G R O W T H Figure 6.4. Corruption Is Regressive: Results from Diagnostic Surveys Bribery paid by enterprises in Ecuador, 1999 Bribery paid by households in Ecuador, 1999 Ratio of firms' bribe costs in total revenues (percent) Bribe costs to household income ratio (percent) 6 6 4 4 2 2 Micro Small/ Large Low Medium High medium Firm size Household income Note: Estimates subject to a margin of error. Source: World Bank (2000e). Svensson (1999) found that in Uganda, administrative corruption reduces a firm's propensity to invest and grow, and Hellman, Jones, and Kaufmann (2000a) found that in those economies in transition where "grand" corrup- tion is more prevalent, the growth and investment rate of the enterprise sector is much smaller, while security of property rights is impaired. Corruption not only hobbles dynamic enterprise development, but it af- fects smaller firms and new entrants in particular. Newer and smaller firms tend to bear the brunt of the bribery "tax," as evidenced by a recent analy- sis of 3,000 enterprises in transition economies.4 Accordingly, smaller firms are prepared to pay significantly more taxes than their larger counterparts for their bribes to be reduced. This research on transition economies also provides insights into the link between political influence, grand corruption (more specifically, state capture), and enterprise performance. In a number of countries in the former Soviet Union, the survey finds that firms (including many with FDI) that purchased parliamentary laws, presidential decrees, and influence in central banks do benefit in the short run (in terms of revenues and the firm's own investment). Yet as stated earlier, their actions inflict a large in- direct cost on the development of the rest of the enterprise sector. These findings demonstrate that while individual firms engaging in state capture may benefit privately (in contrast with administrative corruption-figure 148 GOVERNANCE AND ANTICORRUPTION Figure 6.5. Corruption and Absence of Meritocracy in Public Agencies Impair Access to Services to the Poor: Results from Public Officials' Diagnostic Surveys Corruption control in central agencies in Civil service meritocracy in municipalities in Paraguay and access to services by the poor Paraguay and access to services by the poor Public service accessibility by the poor (percent) Public service accessibility by the poor (percent) 100 - 100 r = -0.41 r = 0.58 80 - 80 - 60 ..*.60- 40 40 - 20 20- E 0 20 40 60 80 100 20 40 60 80 100 Extent of bribery in central agency (percent) Meritocracy index for municipality (percent) Corruption control in municipalities in Bolivia Civil service meritocracy in municipalities in Bolivia and access to public services by the poor and poverty alleviation impact of public services Public service accessibility by the poor (percent) Public service accessibility by the poor (percent) 100 100 r= -0.72 r = 0.85 80 - 80 - *. 3 60 - 60 - 40 - * 40 - 20- I I E 20- I 0 20 40 60 80 100 20 40 60 80 100 Extent of bribery in municipality (percent) Meritocracy index for municipality (percent) Notes: Each depicted observation (dot) represents a central agency or municipality in the pertinent country. Based on public officials survey on governance, "Poverty alleviation impact" represents the percentage of cases in which the public services delivered are helpful to reduce poverty and "Accessibility to the poor" represents the percentage of cases in which the public services delivered are accessible to the poor, as reported by public officials in the diagnostic survey. Sources: World Bank (2000e). 149 T H E QU A L I T Y O F G R O W T H 6.6), such a form of grand corruption imposes a particularly pernicious social cost on enterprise development. (See annex 6 for details on the unbundling of the measurement of corruption into state capture, public procurement kickbacks, and administrative corruption). Causes of Corruption Empirical studies of the causes of corruption are fairly new. Yet, the emerging evidence suggests that some determinants are important. The available research supports the notion that corruption is a symptom of deep institutional weaknesses. Absence of Political Rights and Civil Liberties Political rights, which include democratic elections, a legislature, and oppo- sition parties, and civil liberties, which include rights to free and indepen- dent media and freedom of assembly and speech, are negatively correlated Figure 6.6. "Petty Bribery" versus State Capture: Does Engaging in Corruption Benefit the Firm? Administrative corruption does not benefit the firm State capture through purchase of policies (averages from all transition countries) and laws benefits the captor firm Percentage growth Percentage growth 20 - 30 - 15- 20- 10 ~ ~ ~ ~ ~ ~ ~ ~ ~ 1 0 ~~~~~~~~~~~~~~~0 Sales growth Investment Sales growth Investment growth growth * Firms paying a high level of administrative bribery * "Captor" firm (of laws/policies) [ Firms paying a low level of adminstrative bribery 3i "Noncaptor" firm Source: Hellman, Jones, and Kaufmann (2000a). Details in annex 6 and at http://www.worldbank.org/wbi/govemance. 150 GOVERNANCE AND ANTICORRUPTION with corruption. Figure 6.7 shows the high correlation between civil liber- ties and freedom of the press with corruption. Increasing evidence points to the empowerment of civil society in effectively addressing corruption (fig- ure 6.7). Furthermore, the enterprise survey evidence from transition economies suggests that the capture of state policies and laws by corporate interests is associated with the absence of full civil liberties (Hellman, Jones, and Kaufmann 2000a). The worldwide empirical evidence also sug- gests that the inclusion of women, whether measured in terms of parlia- mentary representation or social rights, does help achieve such empower- ment (Kaufmann 1998). Devolution, such as fiscal decentralization (Collier 1999; Fisman and Gatti 2000), under the right circumstances may also help control corruption. In addition, evidence points to a significant correlation between corruption and the rule of law. Public Finance and Regulation Corruption is higher in countries with a high degree of state ownership in the economy, excessive business regulation and taxes, arbitrary application of regulations, and trade restrictions. Monopolized economies also tend to have more corruption. Civil liberties and a free press can help control corruption Figure 6.7. Corruption and Civil Rights Civil liberties and bribery Press freedom and control of corruption Bribery Control of corruption High Correlation: -0.67 High Correlation: 0.66 *'-:: . a Low Low Low High Controlled Free Civil liberties Freedom of the press Source: Kaufmann (1998). 151 T H E QU A L I T Y O F G R O W T H Civil Service Civil service professionalism, which includes training, hiring, and promotion systems, is also associated with less corruption. Contrary to conventional wis- dom, the evidence on civil service pay is often ambiguous. Better public sec- tor salaries on their own may not explain a significant reduction in corrup- tion. For example, Ecuadorian public sector agencies that offer better pay to employees do not have a lower incidence of corruption. In many settings, a few of the more senior politicians or government officials often cause the most damaging corruption. While in some countries raising salaries of se- lected key civil service personnel may be warranted, this is unlikely to bear fruit without complementary measures. Among these, meritocracy in hiring, promoting, and firing within an agency is associated with less corruption (fig- ure 6.8). The contrasting results between the low impact of higher salaries, on the one hand, and the significant effect of meritocracy, on the other, ex- emplify the need to conduct in-depth, empirical diagnostics within countries intent on formulating serious anticorruption programs.5 A Multifaceted Anticorruption Strategy Given what is known about the main determinants of good governance and corruption, what kinds of programs may have an impact?6 Improving governance requires a system of checks and balances in society that restrains arbitrary action and bureaucratic harassment by politicians and bureaucrats, promotes voice and participation by the population, reduces incentives for the corporate elite to engage in state capture, and fosters the rule of law. Futhermore, the ongoing research on state capture high- lights the need to place checks and balances on the "elite" corporate sec- tor through promoting a competitive market economy and an active civil society. A meritocratic and service-oriented public administration is an- other salient feature of the strategy. Key Reforms Figure 6.9 synthesizes the strategy of key reforms for improving governance and combating corruption. However, how to combine and sequence these reforms to achieve the greatest impact on corruption is a particularly daunt- ing challenge, as is the task of detailing and adapting a strategy to each country-specific reality. For instance, a country that has been subject to state capture by the corporate elite will require a different strategy than a country where the main source of misgovernance originates in political structures or in bureaucracy. Specific questions about governance reforms 152 GOVERNANCE AND ANTICORRUPTION Figure 6.8. Meritocracy Can Reduce Corruption: Evidence for each Agency from Surveys of Public Officials in Three Countries Corruption in some agencies in Ecuador is Corruption in some municipalities in Paraguay is associated with lack of meritocracy associated with lack of meritocracy Extent of bribery in public agency (percent) Extent of bribery in municipality (percent) 5 80 r =-0.71 r =-0.59 4 4-3 * .N. ..6 $.,w ~~~~ ~~~~~60- 2 * ** * z40 3 1 5 U 0 - I I I 20- I l l 20 40 60 80 100 0 20 40 60 80 100 Meritocracy index for central agency and municipality Meritocracy index for municipality Corruption in some agengies in Bolivia is Corruption in some municipalities in Bolivia is associated with lack of meritocracy associated with lack of meritocracy Extent of bribery in central agency (percent) Extent of bribery in municipality (percent) 70 70 r =-0.64 * r =-0.64 60 - * * ~~~~~60- 50 - 40 - *50- 30 - 40- 20- 10- * 30- 10 0 - E I 20- - . I 20 40 60 80 100 20 40 60 80 100 Meritocracy index for central agency Meritocracy index for municipality Note: Each depicted observation (dot) represents a central agency or municipality in the pertinent country. Extent of bribery is measured by the reported percentage of public services and contracts that are affected by bribery in a central agency or municipality. The meritocracy index (0-100) is constructed using the survey questions related to personnel management in a central agency or municipality, as reported by public officials. Sources: World Bank (2000e); the contributions of Ed Buscaglia, Maria Gonzalez de Asis, Turgul Gurgur, AkikoTerada, Youngmei Zhou, and Pablo Zoido-Lobat6n to this line of research on public officials are acknowledged. 153 T H E tU A L I T Y O F C; R O W T H Figure 6.9. Multipronged Strategies for Combating Corruption and Improving Governance: Recognizing the Political Economy Political economy: * Political leadership accountability * Political will of leadership r-1 * Parliamentary reform Economic policy: * Addressing elite vested interests * Deregulation, entry, and competition and state capture * Tax simplification * Political party and * Public expenditure policies and composition campaign finance reform * Macroeconomic stability and fiscal discipline Institutional reforms: * Customs * Transparent privatization Financial controls: • Govemment reform0* Procurement reform * Government reform _ -~~mzt S_ *~~~~ Audit/financial * Decentralization/ _ A /n municipal reform management 4 * Corporate govemance/ethics Civil service: * Financial sector regulation * Pav and incentive reform * Budget control and 4 ~~~~~~~~~~~~~~~~~~treasury development * Restructuring of agenciest . Meritocracy * IT/Internet computing * Meritocracy * Transparency Legal-judicial: Civil liberties, public oversight, * Judiciary independence and civil society: * Meritocratic judicial * Civil society participation appointments * Freedom of the press * Altemative dispute * "Power of data"/IT/empirical resolution mechanisms/ surveys/"scorecard" NGO altematives a * Parliamentary oversight * Enforcement in visible grand * Coalition building and corruption cases collective action * Reducing legal/judiciary * Community-level/women capture involvement * Intemational agency and FDI corporate responsibility IT Information technology. Source: Author, in collaboration with the World Bank's Public Sector Group. 154 GOVERNANCE AND ANTICORRUPTION therefore include what types of changes are feasible under what political conditions and how reforms should be prioritized within the political, civil society, and corporate realities of each country setting. Competition and Entry. In some transition and developing countries, a source of grand corruption is the concentration of economic power in mo- nopolies that then wield political influence on the government for private benefits. The problem is particularly acute in countries rich in natural re- sources, where monopolies in oil, gas, and aluminum, for instance, wield considerable economic and political power that leads to different forms of corruption: nonpayment of taxes, nontransparent offshore accounts, pur- chasing licenses and permits, and purchasing votes and decrees that restrict entry and competition. Demonopolization, deregulation, facilitation of en- try and exit (through liquidation of assets and effective bankruptcy proce- dures), and promotion of competition are vital. Accountability of the Political Leadership. Measures are being imple- mented in various countries that provide checks and balances for the politi- cal leadership and senior officials in their commitment to good governance and anticorruption through public disclosure and transparency of their own actions, finances, income, and assets. In several countries this has entailed the following: * Public disclosure of votes in parliament * Repeal of unconditional parliamentary immunity * Public disclosure of sources and amounts of political party financing * Public disclosure of incomes and assets of senior public officials and their key dependents * Regulations against conflicts of interest for public officials * Protection of personal and employment security for public officials who reveal abuse of public office by others (whistleblower statutes). Meritocratic and Service-Oriented Public Administration. Recruit- ing and promoting on merit, as opposed to political patronage or ideologi- cal affiliation, is positively associated with both government effectiveness and control of corruption. Reforms in this area have included creating in- dependent, professional institutions with checks and balances (for ex- ample, a civil service recruitment commission) and introducing a compre- hensive performance management system with pay and promotion linked to performance. In Malaysia and Thailand, this led to increased recruit- ment and retention of managerial and professional staff and to increased 155 ; a) f- (K K) \V T . effectiveness in civil service performance. In addition, allowances and noncash benefits often need to be simplified, monetized, and made trans- parent. Care needs to be exercised in avoiding wholesale salary increases as a panacea. Transparency and Accountability in Public Expenditure Manage- ment.7 Basic systems of accountability in the allocation and use of public expenditures constitute a fundamental pillar for good governance. Ac- countability in public expenditure management requires the following: (a) a comprehensive budget and a consultative budget process, (b) transpar- ency in the use of public expenditures, (c) competitive public procurement, and (d) an independent external audit. The budget must first have comprehensive coverage of a government's activities. Many countries face problems of budgetary transparency, where major areas of budget expenditure do not pass through the treasury system, and there is substantial recourse to extrabudgetary funds and no effective system of controlling expenditure commitments. Several countries in tran- sition, such as Hungary and Latvia, have made progress in addressing these problems with comprehensive treasury reform programs. Second, disclosure matters. Many industrial countries (for instance, Aus- tralia and the United Kingdom) publish frameworks for public expenditure strategies, which are both a major tool for clarifying strategic choices and a means of enhancing the transparency of the policy objectives and the output targets underpinning annual budgets. More recently, South Africa has devel- oped the Medium-Term Expenditure Framework, revised annually and pub- lished on the web, as a means of clarifying strategic choices and establishing publicly accountable objectives for public expenditure. Third, transparent and competitive public procurement is key to clean government. Reducing corruption requires adhering to strict discipline in terms of transparent and competitive bidding of major contracts, maximiz- ing the scope of public oversight and scrutiny. The information technology revolution is proving to be a catalyst. Indeed, to make the process of gov- ernment procurement more efficient and curb corruption, three Latin American countries (Argentina, Chile, and Mexico) have recently adopted electronic govemment acquisition systems. All procurement notices and their results are placed on a publicly available web site. Other important in- novations related to activist extemal monitoring are taking place as well. NGOs are increasingly playing a role in spearheading public audiences to have a greater voice in establishing rules of the game for large-scale pro- curement projects (such as in the subway system in Buenos Aires) and throughout the transparent bidding process itself (where NGOs such as GOVERNANCE AND ANTICORRUPTION Transparency International have innovated). The World Bank has also taken an active role in aggressively pursuing firms engaged in misprocurement in projects; for instance, delisted firms that have been barred from bidding in Bank-funded projects for having engaged in corrupt procurement are publicly available on the Bank's web site. Fourth, establishing independent external audits matters. Several transition and emerging economies, such as the Czech Republic and Po- land, have established supreme audit institutions, which are genuinely in- dependent and have a constructive impact on public financial manage- ment systems. In the Czech Republic, audit reports are published, presented to the legislature, and discussed in the cabinet, along with a proposed plan for corrective actions, in the presence of the supreme audit institution and relevant ministers. Promoting the Rule of Law. According to the New Palgrave Dictionary of Economics and the Law, the rule of law is defined by opposing it to the rule of powerful men or women. This summarizes the challenge in many countries, where powerful politicians, leaders, elite interests, or oligarchs often influ- ence the practical operations of the parliament, judiciary, and legal enforce- ment institutions such as the police. These countries often have an adequate set of laws on the books, yet the failure is in their effective application and enforcement. And in some countries such laws have been captured by elite interests. The evidence from a vast array of worldwide data (synthesized in figure 6.1) suggests that there is a rule of law crisis in many countries of the former Soviet Union and in Africa, as well as in some in Latin America. The institutional dysfunction in such countries stands in sharp contrast to others, where, however imperfectly, the capacity of legal and judiciary institutions is improving. Illustrating the performance of courts in different countries, figure 6.10 shows how honest, reliable, and fair courts are reported to be by the enterprise sector in Estonia and Hungary. By contrast, in countries like the Russian Federation and Ukraine, they are seen as corrupt, very partial and unfair, unreliable, and nonenforcing. Misgovernance in the judiciary and legal institutions does not always originate solely in the public sector. In some countries elite corporate in- terests exert corrupt pressures as well, as also gleaned from the recent en- terprise survey in transition countries and illustrated in figure 6.11, which suggests the extent of capture (by enterprises including FDI) of the legal and judiciary system in some countries. Thus, even if the legal institutions are fully staffed by trained judges and personnel, they can be subject to capture by politicians or corrupt corporate interests. In this context, the public sector legal institutions are an integral 157 T H E QU A L I T Y O 1: G RPO W I H Figure 6.10. High Variation in the Quality of Courts in Selected Economies (view of the corporate sector on four dimensions; a larger inner diamond means a better performing court) Kazakhstan Russian Federation Ukraine A A A D B D B D B C C C Estonia Hungary "Benchmark" A A A: Uncorrupt D B D B D: B: Fair C C ~~~~~~~~~~~~~~~~~Enforce 0 decision C C C: Reliable Quality dimensions of courts measured: A Uncorrupt. B Fair. C Reliable. D Enforces decisions. Note: Four-pronged diamond on a scale of 0-1, where 1 indicates 100 percent of firm respondents giving highest rating in each relevant court quality dimension. The lower right corner panel is a hypothetical ideal benchmark, if 100 percent of firms would give perfect scores. Source: Hellman and others (2000); see also annex 6. Based on a 1999 enterprise survey in transition economies. part of the governance problem, and not part of the solution. This reduces the relevance of conventional advice on improving governance through the creation of institutions within the public sector (such as an ethics office and anticorruption department) passing anticorruption laws, providing technical assistance in the form of computers or other hardware, or sending sitting judges to training or "study tours." Instead, innovative mechanisms to improve governance are often needed, such as alternative dispute resolu- tion mechanisms, providing for more systematic involvement by NGOs and other alternative institutional arrangements, dissemination strategies through the media, and exploiting more fully and transparently the power 158 GOVERNANCE AND ANTICORRUPTION Figure 6.11. Legal and Judicial Capture by the Corporate Sector in Selected Transition Economies Proportion of enterprise sector affected by legal/judiciary corporate capture 50 - 40 - Percentage of firms affected by the corporate purchase of * Parliamentary legislation 30 -* Criminal court decisions * Commercial court decisions 20 -U Legal/judicial capture index 10 0 Azerbaijan Russian Estonia Hungary Federation Note: The legal/judicial capture index is the simple average of the firms reporting the effect of corporate purchase of parliamentary legislation, of criminal court decisions, and of commercial court decisions. Estimates subject to margin of error. Source: Hellman, Jones, and Kaufmann (2000a,b); see also for details columns 1, 4, and 5 in table A.6.1 in annex 6. Data from 1999. of data and information within and outside the public sector. Equally im- portant, the major challenge of addressing legislative capture in many countries would often require parliamentary and political reforms, such as public disclosure of all parliamentary votes, scaling back immunity laws for members of parliament, and political finance reform. Survey Diagnostic Tools for In-Country Governance Assessment The collection, analysis, and dissemination of country-specific data on corruption are altering the policy dialogue on corruption and empowering civil society through collective action. Yet important challenges remain, including the ongoing refinement of methods that transform survey evi- dence into reform priorities and how best to complement in-depth em- pirical diagnostic surveys with in-depth focus group methodologies-fully involving stakeholders on the key governance challenges within the country. A key challenge involves the development of an effective strat- egy for implementing the reform agenda. Once survey data and their analysis are available, countries where political will is present must begin the more difficult task of prioritizing measures according to the country reality and introducing reforms to root out the sources of corruption. 159 T HE QU A L I -1- Y OF GRO C W I II Sustaining the reform effort with broadly based participation involving all govemment branches, civil society, and the business community constitutes another challenge to the anticorruption and governance-improvement agenda (box 6.1). With cooperation from the private sector and NGOs, the govem- ment can leverage reform by allowing private competition alongside public provision of some services; for example, the adoption of private forms of alter- native dispute resolution to compete with the judiciary or private provision of garbage collection at the municipal level. In-depth governance and anticor- ruption survey diagnostics (and their concomitant data analysis) need to be institutionalized, so that statistics on agency-specific corruption can be moni- tored and acted upon periodically. Broad dissemination of the vast amounts of statistics being generated through diagnostic surveys and studies of govemance and capture can further empower stakeholders to strengthen and sustain insti- tutional change. The design and implementation of agency-specific, in-depth diagnostic surveys for public officials (figures 6.5 and 6.8), households or users (figure 6.4), and enterprises (figures 6.4 and 6.11) constitute an innovation that provides tangible inputs for countries committed to implementing capacity building and institutional change programs. New survey instruments can collect detailed information on behavior in even the most dysfunctional govemment agencies and on the delivery of specific services. For example, comparisons of the price of saline purchased by different hospitals, after ac- counting for transport and other idiosyncratic costs, can show whether cor- ruption exists in public hospitals. Used with other empirical devices, such diagnostic surveys can focus the political dialogue on concrete areas for re- form and rally civil society behind reform efforts. Such country self-diagnostic data, used by a variety of in-country stake- holders and disseminated through participatory workshops, have mobilized broader support for consensus building and collective action for institu- tional reforms. Countries such as Albania,8 Bolivia, Georgia, and Latvia have progressed from using diagnostics to taking concrete action. Bolivia is emphasizing civil service and procurement reforms. Latvia has given prior- ity to tax and customs reforms. In Georgia, following the abysmal survey results regarding the state of the judiciary, President Shevardnadze decided that all judges had to be retested, which was broadcast live on television. Two-thirds of the judges failed the examination and were replaced. In other countries, similar governance improvement efforts are taking place at the municipal level. For instance, in a number of Ukrainian cities specific actions to improve the effectiveness of local government in delivering public services are being carried out following diagnostic surveys. Pioneered in Bangalore, India, in the early 1990s, the now well-known 160 GOVERNANCE AND ANTFICORRUPTION Box 6.1. Governance and Survey Diagnot-ic Tooil The Power of < Empi The first set of in^depth goverance- and corrption frequenc ct br paid Iby, r diagnostic surveys of public officials, fims, and citi- regulators in different agencies as well as thesort- zens was carried out in Albania, Georgia, and coming of b de h- p - Latvia in 1998. More recently, implemnentation of: mance -and effctiveness iicatosA, ultipicit refined and expanded versions of these diagnAostic -of gn it e - surveys has been carried out in other countries, fo- -nostics, pemiing a epth a o cusing more broadly on the complex governance of such as mertocacy;, discretioni bdar - key agencies in the country and assessing the main transparency-,and pove'ty allev n f a - institutional determinants of misgovernance and pact. The anlysis cof thes St e as a corruption there. Challenging conventional wis- vital-ipupt ar prioritn - th dom, the new surveys of public officials, enterprises, governce improaevem f rm pr--r- . and citizens find respondents willing to provide de- A significant share :ofadinisve, bib -is < tailed information on misgovernance that they paid to officias -to avoid:taxes, -customs -dut, and have observed and experienced (as opposed to other liabilities-to the- state. -Some bri-suc a merely indicating their vague perceptions aborut kickback payments to -public-offi, o countrywide corruption, for instance). ig and Judicial -deciions, or fi>r public pKrcre-: Survey respondents report on embezzlement of ment-are found to -be- pculary . T-hs e ' - public funds, theft of state property, -bribery to vey results indica-te esaactiv,ities shorten processing time, bribery to obtain mo- viewed by public officias -as-particular - nopoly power, and bribery in procurement. For in- command the highest price for secuwing b, stance, in 1998 in Georgia embezzlement of public gesting- that se-cutin-g --sbc b -is funds and judiciary corruption, inter alia, was iden- viewed as a pr investment- witha s- tified as a serious problem. At that timne theft of expected private-return, - - - state property was identified as a particular problem When. the: data were-presentd irn wkhops t.O in Albania. Bribery in procurement and customs is members of the 'busfnea-onimunity, major civil-so- - a common challenge in most settings where these ciety, ard the executi,ve and legislative ranches,- diagnostic surveys have been implemented. Weak- thepolicydebate abruptlychange from - --un- ness in the judiciary was identified as one of the pri, substantiated, and oftn piersonalized accusatio to mary causes of corruption in Albania, while regula- one focused on, etmpicatl -evidence nd ssem tory failures are much less important there than in weaknesses that neede to be addressed. Action Georgia and Latvia, for instance. In these diagnos- programs, were formulatd iplemaro of tic surveys, detailed statistics are collected on the - institutiotal reformis began. Source: Kaufinann, Pradhan,andRyterman (1998). Fotradetailedgovernance n.mdantiorptuiondiagnostic implementation guide see http-//www.worldbank.org/wbi/govemance. - citizen "scorecard" user surveys allow citizens to evaluate the quality of lo- cal government services (box 6.2). In Campo Elias, Venezuela, thanks to the leadership of the mayor, a courageous woman who believes in the power of governance data to inform and mobilize for action, the reported incidence of corruption has been halved (Gonzalez de Asis 2000). 161 THE QU A L I T Y G E G R O W T Hl Thus, data are powerful in mobilizing support for reforms, but the ob- stacles presented by grand corruption and state capture by vested interests Cresisting such reforms are also powerful. Therefore, poaitical leadership, civil society, private sector investors, and the donor community need to build on the insights and momentum generated by the diagnostics and uti- tize and disseminate statistics in conjunction with promoting civil tiberties and media involvement, and resulting in higher accountability and actions against corruption. Transparency through Voice and Participation CSorr-uption can yield to knowledge and an informed citizenry. Indeed, the empowerment of civil society with more rigorous and reliable information is a key pillar of reform. Transparency is an important component of public empowerment and voice. As a result, policymaking and large public projects should be priatedite on incorporating the voice and participation of stakeholders in development (see box 6.2 for a discussion of transparency and governance). Indeed, Wortd Bank research shows that the greater the participation of beneficiaries in project design and impnementation, the better the project and service performance. 162 GOVERNANCE AND ANTICORRUPTION Civil Liberties Matter In previous sections in this chapter, we presented the close association be- tween civil liberties and freedom of the press on the one hand, and control of corruption and state capture on the other (figure 6.7). Yet the para- mount importance of civil and political liberties transcends its worth in lowering corruption, or merely as an "input" to a developmental outcome: it is a basic good that enhances welfare per se. At the same time, assessing whether civil liberties do matter as an input in developmental and financial outcomes is of relevance within the debate in the aid community regarding fiduciary responsibilities to make aid effective. Evidence from more than 1,500 World Bank-financed projects suggests that civil liberties and citizen participation are important factors for devel- opment outcomes. Researchers focused on measuring the impact of partici- patory and civil liberties variables on project performance and found con- sistent, statistically significant, and empirically large effects of civil liberties on project rates of return. Depending on the measure of civil liberties used, if a country were to improve its civil liberties from the worst to the best, the economic rate of return of projects could increase by as much as 22.5 per- centage points (table 6.2). Because these civil liberty indexes use different scales, a more standard method of comparison is to calculate how much the economic rate of return would increase if each index category were im- proved by one standard deviation. As seen in the last column of table 6.2, this still gives significant results, suggesting an impact of citizen voice on government performance. Moreover, the report Assessing Aid (World Bank 1998a) found that while both civil liberties and electoral democracy have beneficial effects on government performance, the main channel of influ- ence is likely to be the availability of civil liberties. In Rajasthan, India, a people's organization called Mazdoor Kisan Shakti Sanghathan held a public hearing where it exposed misappropriation by local governments of development funds intended for local workers. This gener- ated village demand for further investigation into the government. Local governments, being under public and press scrutiny, were compelled to oblige. Corruption was reduced. The government of Rajasthan recognized the people's right to official documents and enacted landmark legislation (Bhatia and Dreze 1998) (see box 6.3). Governments and citizen groups can elicit voice through surveys and data collection in more systematic ways. Client surveys can cast light on citizens' experiences with government services and identify suggestions for performance improvement. Follow-up surveys can be used to ensure ac- countability and ensure that improvements are in the desired direction. 163 T HEI QU A L I T Y O F G R 0 W T H Table 6.2. Impact of Civil Liberties on Project Socioeconomic Rates of Return Effect on economic Specification on independent variables rate of return Civil With exogenous With regional of one standard liberties control variables With regional With policy dummies and deviation increase variable only dummies variables policy variables in civil liberties Freedom House Civil liberties (1978-87), 1.81 1.16 1.71 1.07 1.57 (N = 649) (0.0005) (0.079) (0.002) (0.114) Humana (1982-85) 0.290 0.299 0.296 0.289 5.19 (N = 236) (0.003) (0.007) (0.002) (0.013) Media pluralism (1983-87) 4.61 4.45 3.66 3.43 3.12 (N = 448) (0.0001) (0.002) (0.001) (0.026) Freedom to organize (1983-1987) 3.17 1.81 2.41 -0.26 2.70 (N = 448) (0.0001) (0.184) (0.006) (0.854) N = number of observations. Nore: Standard error is in parentheses. Average economic rate of return on projects is in the range of 12-16 percent. Source: Isham, Kaufmann, and Pritchett (1997). Generating data and disseminating them widely are potent instruments to mobilize civil society and apply pressure on political structures. For example, simple comparative charts illustrating findings on corruption can help mobi- lize and give voice to previously silent and disparate citizenry groups. Toward a Social Contract Facilitating Civil Society Oversight and Participation Civil society oversight and participation over the decisionmaking and func- tioning of the public sector has been a crucial counterweight and instru- ment for combating corruption and improving governance. This involves making the state transparent to the public and empowering the citizenry to play an active role. While a few OECD countries have been in the fore- front in transparency reforms, in many of the transition and emerging economies the public sector culture is still one of secrecy of decisionmaking. Often, parliamentary votes are not publicly disclosed, pub- lic access to government information is not assured, and judicial decisions are typically not available to the public. Moreover, despite a growing civil 164 GOVERNANCE AND ANTICORRUPTION Box 6.3. Millions of 0Auditor.'W Improve lransparency and Governance in Budgetary Accounts and Beyond Transparency:means empowering the citizenry to asymmetries between the regulators- and the regu- -become millions of auditors in society, providing lated entities. rpt affect:s all majeor areas of voice and access ,to a free press. It enables thef fow public adminis'tration: revenue collect-nion a-s a, of tim£1ely and reliable eonomic, social, and politi- mneans of raising public- fns and ,revenue alloca- cal information about private investors' use of tion as a means of providing public goods. -It af_- loans and the creditworthiness of botrowers, gov- fects public regulation as a means of mitigating i-- ernment service provision, monetary and fiscal formation failures in markets, particularly capita policies, and the activities of international-inseitu- markets. Recent empirical research of financial tions, By contrast, a lack of transparency means crisis episodes indicate that the likelihood of such that someone, such as a government minister, crises was significantly larger where transparency public institution, corporation, or bank, is deliber- was absent. In-depth governance diagnostics of ately withholding access to or misrepresenting in- public agencies discussed earlier within a country formation. also suggest that agencies with transparent flows of In general, a lack of transparency increases the information tend to- exhibit lower corruption. and: scope for corruption by creating informational better govermance and performance overall. society, the government typically does not involve NGOs in the monitor- ing of its decisionmaking processes or performances. Concentrated media ownership and recent restrictions on reporting have weakened the ability of the media to ensure accountability of the public sector. Consequently, changing the culture to one of transparency involves a fundamental change in the way decisions in the public sector are made. The types of transparency reforms that have been demonstrated interna- tionally to be effective include the following: e Ensuring public access to government information (freedom of in- formation) e Requiring certain types of government meetings to be open to pub- lic observation e Conducting public hearings and referenda on drafts, decrees, regula- tions, and laws * Publishing judicial and legislative decisions and keeping a registry * Ensuring freedom of the press by prohibiting censorship, discouraging use by public officials of libel and defamation laws as a means for in- timidating journalists, and encouraging diversity of media ownership * Involving civil society to monitor its performance in areas such as anticorruption and large-scale public procurements bidding 165 T H E QU A L I T Y OF G R O W T H Utilizing the new web-based tools on the Intemet for transparency, disclosure, public participation, and dissemination. Civil society's role ought to be seen as dynamic and providing an op- portunity to political leaders intent on building the credibility of the state through its potential in coalition building and collective action. For in- stance, new activities in many countries where the World Bank provides assistance involve supporting the collective teamwork of civil society, the media, experts, the private sector, and the reformists in the executive and legislative branches in formulating govemance and anticorruption reform programs. The process of involvement by the key stakeholders in civil so- ciety creates a momentum toward ownership and sustainability of the re- forms and builds credibility, as is taking place in some countries in Eastern Europe, Africa, and Latin America, for instance. Conclusions Governance needs to be understood in a broader context than merely ad- dressing corruption, which is a key symptom of more fundamental institu- tional weaknesses. Both govemance and corruption need to be rigorously un- bundled and understood analytically and empirically. Misgovemance distorts policymaking and the allocation of factors of production, which in tum slows income and welfare growth and increases poverty. The many failed capacity building approaches in the past did not pay enough attention to fostering good govemance, to controlling corruption, to improving the bureaucracy and civil service, to promoting civil liberties and participatory approaches, to understanding the origins and consequences of state capture, or to furthering knowledge about the political economy of institution building. Govemance needs to enter center stage in capacity building and institutional change strategies. Understanding of particular vested interests by different influential groups is needed-including the corporate sector (both domestic and FDI)- as is the recognition that incentives, prevention, and systemic-change challenges within institutions vitally affect govemance, and are at least as important as traditional law enforcement aspects. Govemance, voice, and participation will be key for an improved ap- proach to technical assistance and capacity building in the future. Improv- ing governance should be seen as a process integrating three vital compo- nents: (a) knowledge, with rigorous data and empirical analysis, including in-country govemance diagnostics and transparent dissemination, utilizing the latest information technology tools; (b) leadership in the political, civil 166 GOVERNANCE AND ANTICORRUPTION society, and international arena; and (c) collective action via systematic participatory and consensus building approaches with key stakeholders in society (for which the technology revolution is also assisting). Collective responsibility also implies that transnational corporations, the domestic pri- vate sector, and international agencies need to collaborate with national governments and leadership intent on improving governance. The evidence points to the need for a more integrated and comprehen- sive approach to provide a climate for successful development. Economic institutions and policy measures, such as the budget and the nature of pub- lic investment programs, are important, as are civil liberties and participa- tion, with which they interact. This underpins the case for a more holistic approach to development that links economic, institutional, legal, and par- ticipatory variables. Participation and voice are vital in increasing transparency, providing for the necessary checks and balances, and ameliorating state capture by the elite's vested interests. It is not enough to get basic economic policies right on paper; the political economy forces at play must also be recognized. These forces will vary from country to country. In some countries address- ing legal, regulatory, and procurement reform will be necessary to improve governance and control corruption. In others, where state capture by the corporate elite and when there is weaker political will to reform, civil soci- ety oversight, enterprise competition, and working to improve property right protection could be key. For an enhanced focus on poverty alleviation, a concerted approach that integrates rigorous empirical understanding of the governance chal- lenges within a country, encourages active involvement by all key stake- holders, tailors it to the country's own realities, and is championed by the country's leadership is likely to bear fruit. Notes 1. The plethora of indicators measuring various aspects of governance are ordi- nal; that is, they have a qualitative or subjective element. The data are none- theless relevant. First, for some aspects of governance, these are the only kind of data available (and it is now possible to disentangle the "noise" from the "signal"). Almost by definition, hard data (numerical-cardinal) have until now been virtually impossible to obtain in a systematic format, and for those few governance dimensions where such cardinal data exist, they are accompa- nied by a large margin of error and/or methodological questions. Second, for many aspects of governance, survey results (even if they contain an element of perception) do matter at least as much as official data. For instance, if a 167 T H E QU A L I T Y ) F (_ R 0 WV 1 l country's business sector regards the judicial system as an arm of the govern- ment and avoids using the courts, it will think twice about investment deci- sions. See annex 6 for details. 2. The asymmetry of the horizontal bars is explained by the differences in the vari- ance inside each quartile. While differences in between countries are small in the first two quartiles, the differences are larger in the third and fourth. 3. The econometric methodology and its empirical application suggest that gover- nance variables do affect different socioeconomic variables such as infant mortal- ity, literacy, and income per capita in a causal way. However, having established that governance variables jointly do matter significantly for socioeconomic out- comes, care needs to be exercised in disentangling the independent causal impact on developmental variables of each single subcomponent of governance. Given the existence of multicolinearity among the various subcomponents of gover- nance, it is possible that the observed impact of voice on infant mortality, for in- stance, is picking up by proxy other govemance deterninants such as corruption or rule of law. See also annex 6 for methodological details. 4. Hellman and others (see http://www.worldbank.org/wbi/governance). 5. Other factors in the empirical work on causes of corruption also appear to be important. As expected, income per capita and education are correlated with lower corruption when other factors are held constant. General developmental variables are often proxies for more specific determinants of corruption, such as the quality of public sector institutions or the rule of law (see Ades and Di Tella 1999 for a useful review). 6. Much of this section is owed to collaborative work with Sanjay Pradhan, Randi Ryterman, and the Public Sector Group. See also World Bank (2000h). 7. We owe much of this section on Public Expenditure Management to the work of Allistair Moon, Sanjay Pradhan, and Gary Reid. 8. In 1998 the head of government, cabinet members, and hundreds of civil soci- ety stakeholders participated in Albania's national governance workshop, which took place at the same time as the semifinals of soccer's World Cup in France. The workshop featured the main findings of in-depth diagnostic results and a debate on the priorities for action. It concluded with a commitment by the leadership to a progovernance program. Exemplifying the importance as- cribed to it by the nation, on the next day the front pages of all the newspapers in Tirana featured charts showing the results of governance diagnostics, while the World Cup soccer results were relegated to the back pages. Albania is carry- ing out an anticorruption program, including judicial and customs reform, with support from the World Bank. 168 C H A P T E R 7 SEIZING THE OPPORTUNITIES FOR CHANGE We must use time creatively, and forever realize that the time is always ripe to do right. -Nelson Mandela, Higher Than Hope: The Authorized Biography of Nelson Mandela his book has revisited the development experiences of recent decades, with a focus on the 1990s. The final decade of the 20th century saw striking progress in parts of the world, but also stagnation and reversals, even in countries that had previously enjoyed rapid growth. As prosperity spread and the quality of life improved for many in society, poverty is estimated to have persisted stubbomly or worsened for others. Greater population pressures, little access to education, and degraded natural resources made the poor increasingly vulnerable to the volatility of growth. Between 2000 and 2010, the population of developing (including transi- tion) economies is projected to grow from 5 billion people to 6 billion. If countries follow a business-as-usual scenario, the number of people in the de- veloping world (excluding China) living below the poverty line could swell by some 130 million. This book addresses ways to improve future outcomes. The Framework and Themes Human, natural, and physical capital assets are a country's main resources for growth and welfare improvements. Their distribution, growth, and pro- ductivity largely determine people's income and welfare. The poor rely on 169 T H F QU A L I T Y OF G R O W T H human and natural capital in addition to physical capital, so the accumula- tion and productivity of these assets have a strong impact on poverty. Sur- veys show that "the poor rarely speak of income, but focus instead on man- aging assets-physical, human, social, and environmental-as a way to cope with their vulnerability" (Narayan and others 2000). Growth based on relatively undistorted (or balanced) accumulation is likely to be less volatile and to be sustained over the long term. This theme is supported by the country evidence in chapter 2. First, a compari- son of reformers and nonreformers showed that reforms helped to acceler- ate growth in the 1990s; however, this growth (in many instances) was based on a sharp rise in physical capital accumulation, while investments in human and natural capital lagged behind. Second, an econometric analysis of 20 mostly middle-income countries showed that the rate of economic growth declines as the stock of physical capital increases for given levels of human and natural capital, but the accumulation of hu- man capital assets, by increasing access to education and health care, can arrest this decline. Third, an econometric analysis of 70 developing coun- tries confirmed the previous findings about capital accumulation and showed that when natural capital is also considered as a factor of produc- tion, human capital can substitute for natural capital to some extent and reduce dependence on it as a source of growth. Improving the Distribution of Opportunities For growth to reduce poverty effectively, the assets of the poor need to be augmented. Their main asset is their human capital. Yet, inequality in edu- cation is staggering. If capabilities are normally distributed across the popu- lation-regardless of whether people are rich or poor-the inequality in ac- cessing basic education and jobs would represent one of the greatest welfare losses to society. When the quality of schooling is low and the inequality in schooling is high, the poor are hurt most by inadequate education. The underinvestment in the human capital of the poor is attributable to wealth gaps, gender gaps, market failures, and policy distortions. In addition, many countries have not adequately focused public invest- ment in basic education. A reallocation toward basic education is needed to improve the efficacy of public spending. Education at all levels, includ- ing higher education, needs to benefit from private investments and public- private partnerships. Decentralized decisionmaking and community- managed schools hold great promise for improving education outcomes. However, to make education more productive for the poor, they need to be empowered with land, equity capital, training, and job opportunities in open and competitive markets (chapter 3). 170 SE I Z ING T HE O P PO RTU N I TIES FOPR CHANGE Sustaining Natural Capital Several indicators of the quality of natural capital, with the notable excep- tion of access to clean water and sanitation, have tended to deteriorate in both slow- and fast-growth economies. For the developing world as a whole, depletion of natural capital (forest, energy, and minerals) and dam- age from carbon dioxide emissions is estimated at 5.8 percent of GDP. This deterioration of natural capital imposes significant current costs and dimin- ishes prospects for future growth. Faster growth has the potential to make more resources available to invest in natural capital accumulation, but ac- tions are needed to ensure the quality of the growth process. Thus, the grow-now-clean-up-later approach needs to give way to an environmental policy that is integrated with growth policies. The book documents successful initiatives that simultaneously incorpo- rated actions to stimulate growth and protect natural capital. These mea- sures often involve selective state interventions and focus on collaborative approaches with local communities and the private sector. Global and na- tional problems can be simultaneously addressed through international co- operation, including transfer mechanisms for payments to compensate for global externalities. As a consequence, the pursuit of high-quality growth is possible, and desirable, without extensive degradation of the atmosphere, forests, and rivers, or any other aspects of natural capital (chapter 4). Dealing with Global Financial Risks Integration with the global financial system has undeniable technological and economic benefits for countries, but it also exposes them to shocks and to great volatility in currency values, interest rates, and capital flows. The shocks can bring about important output and job losses, corporate and banking distress, and increased poverty. Thus, countries need suitable mechanisms to balance the benefits of globalization with its risks. They must reduce the risks of panic and crises, while maintaining their commit- ment to market openness. A sound macroeconomic policy environment is essential to sustained growth, but recent experience shows that macroeconomic stability cannot do it alone. It must be complemented by actions to remove the explicit or implicit government guarantees that provide incentives for short-term capi- tal inflows, to strengthen domestic regulation and supervision of banks and other intermediaries, to rebuild the information infrastructure of financial markets, and to improve corporate governance and transparency. Countries must also maintain public support for open capital markets. In democratic countries this entails providing insurance to citizens-either through the 171 T fH E QU A t. I T Y O F G R O W T II marketplace or through redistributive public expenditures on education, health, and transfer payments (chapter 5). Improving Governance and Fighting Corruption Government has a key role in delivering the public goods essential for achieving balanced and sustained growth and for reducing poverty. It also needs to have effective and streamlined regulatory regimes to address exter- nalities and market failure. Misgovemance and corruption distort policy- making and the allocation of key production factors, consequences that slow income and welfare growth and increase poverty. Many development projects and investments have failed because too little attention was paid to fostering good governance and civil liberties, controlling corruption, im- proving the bureaucracy, and building institutional capacity. Participation by beneficiaries, attention to the voices of people and competitive enterprises, and accountability and transparency in govern- ments are vital to controlling corruption and improving governance. New approaches to coalition building and the integration of the latest methods of governance and corruption surveys with new technologies for timely analysis and dissemination are producing encouraging results in some countries. Collective action from such a participatory consensus-building process, coupled with the power of information, disclosure, transparency, and knowledge and capacity building, can nurture the political will and technical ability to address misgovernance and sustain institution build- ing (chapter 6). Actions to Ensure Quality The emerging themes help clarify four dimensions that shape the quality of the growth process: distribution of opportunities, sustainability of the envi- ronment, management of global risks, and governance (box 7. 1). These ele- ments directly contribute to development. They have a two-way relation- ship with growth, they add to the impact of growth on welfare, they help make growth more sustained, and they address the conflicts that growth might pose to sustainability. The book provides evidence from a range of areas and sources showing that a focus on quantity will not in itself ensure quality. So public spending levels alone cannot give an adequate indication of impact. In a WHO study that ranks 191 countries on the quality (including fairness and broad cover- age) of their health systems, the United States is 1st in health expenditure per capita but 37th in overall health system performance. France-4th in 172 S EI ZING THE OP PO RTU N I TIES FOR CHANGE Box 7.1. Actions for Quality What might be the policy itmpication for, ensuring land, credit, skills training, and job oppor- quality growth? The book has presented several, tunities in open mtaetk. which can be organized under three prtincples- * Eure effective regularory frameworks and anticorrtipni rneasures to accompany fi- Policies for Undistorted Growth: of nanrcii openness and privazarion. Physical, Human, and Natural Capital * Align reforms and -restructuring:to mecha- nisms for mitigating the costs of crises, * Avoid direct or -indirect subsidies to capital, which will likely b bone disproporion- such as tax breaksi alloca-tion of monopoly t - =- t por powers and subsidies, special privileges t-hat atey b te poor. feed corruption, and implicit guarantees on u t G r - ~~~~~~~Build -the GOVIrnnse F01ramework rates of return. - ,e-e- -e- * Invest efficiently in human capital and ensure access for the poor through incentives and al- - Involve-all stakeholders-the private sec- location of public investments in education. tor, including transnational fi and the * Sustain natumal capital by clarifying property domesti privat-e- e trors NQOs, civil soci- rights, avoiding unrealistically low -levels of ety, and the govermnent.in implement- royalties for natural resources, and-enforcing ing a commonly shared -development environmental taxes. agenda. Empower people through voi,ce, participa- Aftend to the Stability and Distibutive otin and greater civil and political liber- Aspects of Growth ties. - - :f: ~~~~~~~~support economic Liber-alizaLtion by 'pro-: * Ensure that the poor can access education, - Sort econmliberalionm by pd-b technology, and health services, as well as o i d a ,, ter goverance. per capita expenditure-is 1st in performance, Colombia is 22nd, Chile 33rd, Costa Rica 36th, and Cuba 39th (WHO 2000). The book also shows the effects of market failures. On the ability of the poor to build human capital. On the undervaluation and subsequent overexploitation of natural capital. And on undue instability in financial markets. A focus on quality highlights the role of regulatory policies and public expenditures in dealing with these market failures. The answer is not necessarily to increase the regulatory burden of the economy or to increase public expenditures. Instead, it is to reallocate pub- lic expenditure following new priorities and to change the nature of regula- tion-eliminating regulations that are counterproductive and improving those that address market failures. What should be the new public expendi- ture priorities? Do more to promote the buildup of human capital, espe- cially among the poor. Invest more to prevent further degradation of 173 T HE Q_l A L I T Y o F G RPO W T H natural capital and reduce regressive subsidies benefiting physical capital. Target regulations at failures in financial markets and in markets that affect the use of environmental resources. Where Have the Policies for Quality Worked-or Not? There is no single setting in which the attributes of quality have been em- phasized with uniform success. However, the following four country experi- ences illustrate the pursuit of quality aspects of growth, with varying de- grees of effectiveness. Investing Efficiently in Basic Education: The Republic of Korea Starting with a war-torn economy and a poor natural resource base, Korea had an average annual GDP per capita of just over US$500, based on 1980 PPP dollars, in the late 1950s. GDP per capita then doubled in each of the next three decades, driven by relatively broadly based, export-oriented growth. The growth was accompanied by rapid poverty reduction and rela- tively equitable income distribution (Leipziger 1997). Korea spent an average of 3.4 percent of GNP on public education in the 1980s, which was in line with the regional average. However, unlike other developing countries, Korea spent two-thirds of its education budget on compulsory basic education in the 1960s and early 1970s. In the 1990s, public subsidies to primary school students were two to three times those to college students. Tertiary education was financed mainly by private spend- ing. Korea was able to expand basic education rapidly and reduce educa- tional inequality, as measured by a drop in the education Gini coefficient from 0.55 in 1960 to 0.22 in 1990. The Korean government supported favored industries through di- rected lending, subsidies, and guarantees. In liberalized but inadequately regulated environments, these measures led to excessive foreign borrow- ing and wasteful investment by the corporate sector and intensified finan- cial fragility. Both positive and cautionary lessons can be drawn from the Korean experience. Broad-Based Growth within an Incomplete Agenda: Kerala, India On such dimensions of social development as education, health, the gender gap, civil and political liberties, poverty reduction, and inequality, the de- velopment performance of Kerala, India, is comparable to that of much 174 SE I Z ING THE OP PO RTU NITI ES FOPR CHANGE richer economies. A baby bom in Kerala can expect to live longer than one bom in Washington, D.C. However, Kerala's measured economic growth has, until recently, been lower than average among Indian states. Kerala attended to the qualitative aspects of development while ne- glecting first-generation, growth-oriented policies. For balanced asset growth, market-friendly economic policies need to complement social ini- tiatives. The lack of progress in implementing an open and competitive en- vironment for economic activities has hindered economic growth in Kerala. Once these first-generation policy reforms are implemented, the high level of social development should provide a basis for sustained high- quality growth. Ravallion and Datt (1999) found that the poverty-reducing impact of growth varies with initial literacy, farm productivity, and standard of living in rural areas relative to that in urban areas. In states with high literacy rates and equitably distributed basic education every additional percentage point of growth has a stronger impact on poverty reduction than is the case in other states. The elasticity of poverty reduction to nonfarm growth in Kerala was the highest of all states in India. If all Indian states had Kerala's elasticity of poverty reduction, the share of their people in poverty would have fallen nearly three times as fast -at 3.5 percent a year, not 1.3 percent. Balancing Economic Growth and Sustainable Natural Capital: Costa Rica Costa Rica has a high literacy rate, economic and political stability, and no military budget. Income distribution and social indicators are among the best in Latin America. Yet the country needed to address several environ- mental problems, from urban pollution, excessive agrochemical use, and overexploitation of fisheries to biodiversity loss and a deforestation rate in the 1980s estimated at 3 percent a year. Costa Rica responded with an innovative and comprehensive system of forest protection. A system of compensation through markets for carbon se- questration, watershed conservation, and biodiversity protection has helped protect forests. The system generates its own resources through a dedicated fuel tax on domestic consumers, contracts with hydropower utilities, and payments by international parties for carbon emissions offsets. The Ministry of Environment and Energy has supported research-based policy reforms for environmental protection and environmental education programs for schoolchildren. Environmental impact assessments are manda- tory for most projects, including commercial and residential construction and mining. The law sets strict guidelines on protection of water resources, 175 T H F Q_U A L I T Y 0 F C R 0 W T HL wetlands, natural monuments, protected natural areas, and marine and coastal resources. It also sets strict guidelines on all types of pollution and on abuse of land and improper waste disposal. The plan now is to increase the effectiveness of the laws, to strengthen the capacity of institutions re- sponsible for enforcement, and to enter into partnerships with the private sector and civil society (Thomas 1998). Balancing Openness, Risk Management, and Social Protection: Chile After a decade of rapid opening of markets and volatile growth in the 1980s, Chile took measures for risk management in the 1990s. First, Chile implemented a highly targeted system of social assistance through health, education, and housing programs and income transfers. The government's social investments increased by 75 percent between 1987 and 1994, which, complementing robust economic growth, made solid contributions to pov- erty reduction. Second, as capital flows became more volatile, the indepen- dent central bank implemented selective capital controls on short-term capital inflows between 1991 and 1998. Arguably, they have helped to in- crease the wedge between domestic and foreign interest rates and change the composition of capital inflows toward longer maturities (Gallego, Hernandez, and Schmidt-Hebbel 1999). These and other policies have contributed to rapid economic growth with a significant decline in poverty. The incidence of poverty fell from 41 percent of the population in 1987 to 23 percent in 1994, while the inci- dence of severe poverty (based on a lower poverty/indigence line) fell from 13 to 5 percent. Income inequality seems to have stabilized since 1987, af- ter rising during most of the 1960-85 period (Ferreira and Litchfield 1999). However, natural resources have been underprotected and overexploited. In addition, the distribution of education has become less equitable, as re- flected by a widening gap in years of schooling between the rich and poor (World Bank 1997b). Political Economy of Quantity versus Quality The varied experiences of economies suggest that an emphasis on quality is essential on three counts. First, quality directly promotes well-being by in- fluencing a more even distribution of education and health care and an im- proved environment. Growth and quality aspects-linked to each other in a two-way relationship-need joint attention. 176 S El /I Nu I1-i I iPO I " N i;, iX C H ANG E Second, the pace of growth is less volatile and more sustainable when quality aspects are considered. Where growth rates are highly variable over time, the negative impacts are especially pronounced for the poor. Third, economies that focus on quality can deal better with difficult tradeoffs. One tradeoff mentioned in this book is the temptation to subsi- dize physical capital or overexploit natural capital in an effort to promote growth. In this and similar cases, a focus on the qualitative aspects of growth helps to manage the tradeoffs. Most countries-and much policy advice-have stressed macroeco- nomic stabilization and liberalization first. Meanwhile, actions on the qualitative aspects, such as the distribution of education and the sustain- able use of natural capital, are postponed. The evidence presented here shows the fundamental limitations of this approach and the benefits of joint action. Reformers have sometimes found it necessary to take advantage of win- dows of opportunity for liberalization, when vested interests and opposition to liberalization have been muted. Whether the qualitative dimensions also receive priority may depend on other conditions, including the spread of democratic institutions. In an increasingly mature and participatory setting, a country would not want to postpone the important qualitative aspects of growth to a time when the costs of addressing them will have multiplied. Sometimes, political difficulties can impede progress, even when the importance of the qualitative aspects is clear. Interest groups may drive a wedge between policy design and implementation. Collusion between poli- ticians and the elite may distort the allocation of public resources to reward the owners of physical capital. For example, tax holidays, implicit guaran- tees on infrastructure, monopoly powers, and easy access to natural re- sources often benefit the rich, but hurt the poor. The political economy of reforms, less explored than other aspects of growth, is an area difficult to evaluate. Yet some initiatives seem unambigu- ously to be worthwhile. Fostering beneficiary participation, encouraging country ownership of reform programs, and promoting political representa- tion of the poor are a good beginning. Going Forward How can countries accord greater priority to the quality dimensions of growth? And how can they finance and support such objectives in practice. Several observations drawn from the discussion in this book can guide ef- forts in that direction: 177 T H E QU A L I T Y O F G R O W T H * Explicit attention to ensuring transparency and reducing corruption and rent-seeking will not only raise national saving and investment and promote sustained growth, but will also help to distribute its fruits more equitably. * Some quality dimensions lend themselves to full-cost pricing or taxation, both of which generate public resources. * Other measures to ensure quality require a reallocation of public ex- penditures-reducing subsidies and distortions in some areas and in- creasing public investment in others. * Attending to quality does not have to mean more government in- tervention, but rather may mean more involvement of the private sector, nongovernmental organizations, and civil society in imple- menting shared goals. * Broader civil outreach can nurture civil liberties and participatory processes that can help to sustain policy changes. * All this in turn would call for a much greater focus on skills, tech- nological development, and capacity building and the effectiveness with which this is done. The evidence in this book provides strong motivation for focusing on the quality aspects marking the ability of people to shape their lives-for ex- ample, the equality of opportunities for human development, sustainability of the environment, management of global risk, and the manner of governing- along with the traditional facets of growth. Governments do not have to as- sume the entire burden of giving greater priority to the quality dimensions- nor should they. Rather than calling for more government intervention, the evidence in this book calls for greater voice and participation by the private sector, NGOs, and civil society. A broader involvement by all can move the em- phasis of development beyond measured GDP growth to include social and environmental progress, greater empowerment and voice, and better gover- nance. This reallocation of priorities will refine the contribution of the quali- tative aspects of the growth process and focus the spotlight on what develop- ment truly means. 178 A N N E X 1 BROAD OBJ ECTIVES AND INSTRUMENTS C n ontinuous assessment of development progress and policies requires a broader than usual framework even if all elements are not quantifiable. Here we look at development goals and policy instruments and formulate some hypotheses about the links between them. A formal model and discussion of the empirical results are in chapter 2 and annex 2. People value at least three dimensions of life in current and future time periods. They gain direct satisfaction from education and other as- pects of human capital, such as life expectancy or literacy; from clean air and water and other stocks of natural capital; and from flows of consump- tion goods, such as food and shelter. They also care about the welfare of future generations and their enjoyment of all these aspects of life (at some discount rate). A society will try to get the most from human, natural, and physical capital, subject to the total resource constraint. Together, increases in these dimensions signify quality growth. The simple associations among the goals and policies, in the form of correlations and scatter charts, are presented here. Annex 2 presents an econometric analysis, which complements a vast literature in this area. Goals and Policy Measures We constructed composite indexes for human development and environ- mental sustainability, enabling us to focus on three measures of quality of growth rather than a large number: human development, economic growth, and environmental sustainability. The policy instruments include those emphasized in the World Development Report 1991 (World Bank 1991) along with several others. The correlations between policies and 179 ANNEX I goals are shown in table Al.l. These correspond to the summary figures 1.2 and 1.5 in chapter 1. The associations between policies and compo- nent indicators forming the indexes are not shown. Scatter plots for se- lected combinations of goals and policies are shown in figure A1.1, after controlling for the effects of initial period income. The goals and the proximate variables used are *Human development indicators. We constructed a human develop- ment index from data on reductions in infant mortality, reductions in the illiteracy rate, and increases in life expectancy. The period over which changes were computed was from the early 1980s to the Table A1.1. Relationships between Development Objectives and Policy Instruments, 1981-98 Goals Hurnan development GDP growth Sustainable development Correlation Significance Number of Correlation Significance Number of Correlation Significance Number of Instruments coefficient level countries coefficient level countries coefficient level countries Educational spending/ GDP 0.04 0.72 87 -0.02 0.84 88 0.17 0.21 56 Health spending/GDP (1990-98) -0.01 0.95 70 -0.28 0.02 71 0.18 0.23 49 Budget surplus 0.12 0.40 55 0.27 0.05 55 0.01 0.97 39 Trade/GDP ratio 0.07 0.50 89 0.07 0.50 90 -0.05 0.69 56 Change in mean tariff 0.05 0.82 26 -0.09 0.65 26 -0.10 0.65 25 Capital account openness index (1988) 0.21 0.22 36 0.00 0.99 36 -0.22 0.23 31 Financial repression index (1996) -0.16 0.50 21 0.26 0.26 21 -0.35 0.12 21 M2/GDP 0.36 0.00 89 0.29 0.01 90 -0.08 0.58 56 Domestic environmental action (dummy variable) -0.16 0.15 80 0.24 0.03 8 1 -0.10 0.47 56 Intemational environmental action (dummy variable) 0.11 0.35 80 0.08 0.49 81 -0.24 0.08 56 Rule of law index (1997-98) 0.34 0.00 86 0.41 0.00 87 0.18 0.19 55 Government effectiveness index (1997-98) 0.35 0.Q0 81 0.27 0.00 82 0.05 0.73 55 Notes: Improvement in human development is defined as the Borda index of reduction in infant mortality, reduction in illiteracy, and increase in life expectancy between the 1980s and 1990s. Improvement in sustainable development is defined as the Borda index of decreases in carbon dioxide emissions, deforestation, and water pollution between the 1980s and 1990s. Correlations that are significant at least at the 10 percent level are shown in bold italics. Source: World Bank (2000c); authors' computations. 180 BROAD OBJECTIVES AND INSTRUMENTS Relationships between objectives and instruments Figure A1.1. Development Objectives and Policy Instruments GDP growth (percent change per year) Increase in literacy (percent change) -8 -20 15- - 10 C* 4 - .~~~~ ~ ~ ~~~~~~~ * 2* * 46- ........ ---32--.. ........ -------- -15 -4 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~-20- I I -40 -20 0 20 40 60 -50 -25 0 25 50 75 100 Financial depth (M2/GDP, percent) Trade-GDP ratio (percent) GDP growth (percent change per year) Increase in forest cover (percent change) 6 3 4 *2 2 I 0 -**-i--*-- 1 ~~~~ ~~~0 --------. V -2 - -1 -4 -2 a -6- I I3~ -6 -4 -2 0 2 4 6 -4 -2 0 2 4 Budget surpius (percentage of GOP) Public spending on education (percent of GOP) GDP growth (percent change per year) Reduction in water pollution (percent change) 8 ~~~~~~~~~~~~30- 6 ~~~~~~~~~~~~~20- -10 -2 ~~~~~~~~~~~-20 -2 -1 0 1 2 -1.0 -0,5 0 0.5 1.0 1.5 Rule of law (index) Control of corruption (index) Note: The scatter plots are constructed using residuals from the regressions of the respective variables-pertaining to both the axes-against GODP per capita in 1981. Source: World Bank (2000c); authors' computations, 181 ANN EX 1 late 1990s. We wanted to incorporate in the index variables reflect- ing income distribution, reduction in the incidence of poverty and the gender gap in educational attainment, but did not because data were unavailable for many countries. * Sustainable development indicators. We constructed a composite in- dex of the negative annual rate of deforestation, reductions in car- bon dioxide emissions per capita, and reductions in water pollution per capita. The period used was again from the early 1980s to the late 1990s. We wanted to include in the index a measure of air pol- lution in the major cities in developing countries, but comparable data were available only for recent years. * Income growth. We used growth rates between 1981 and 1998 of GDP and intermediate indicators such as capital stock and TFP growth, as used in many empirical studies, such as those employed by Barro (1990), Easterly (1999a), Easterly and others (1993), Nehru and Dhareshwar (1993), Pritchett (1998), World Bank (1991), and Young (1992). Policy instruments were represented by the following: • Social spending on education and health. They were expressed as per- centages of GDP, averaged over available values for the period: 1981-97 for education spending and 1990-98 for health spending. Due to data limitations, we could not include allocation of spending on basic education and preventive health services (Filmer, Ham- mer, and Pritchett forthcoming; L6pez, Thomas, and Wang 1998). * Environmental commitment. We used two dummy variables to repre- sent environmental commitment: one for domestic action, based on environmental strategy formulation and profiling, and another for international action, based on the signing of the Global Treaty on Climate Change. Unfortunately, more refined indicators of govern- ment policies toward sustainable development are not yet available. * Macroeconomic policy. We used budget surplus as a percentage of GDP (Barro 1990; Fischer 1993). * Openness. We used the trade-to-GDP ratio, parallel-market pre- mium, change in mean tariff, and a measure of capital controls based on Quinn (1997) and Quinn and Toyoda (1997) (a higher value of the index represents a greater degree of openness to capital inflows; see annex 5). * Financial depth, prudence, and risk management. We used financial depth as measured by the ratio of M2 to GDP and an index of financial 182 BROAD OBJECTIVES AND INSTRUMENTS repression (based on Williamson and Mahar 1998; a lower value of the index represents a more liberal financial system). * Governance indicators. Six governance indicators-rule of law, gov- ernment effectiveness, control of corruption, voice and accountabil- ity, regulatory burden, and political instability and violence-have been examined in this book. Of this list, the first three have been used in this annex. For further details and analysis see chapter 6 and annex 6. Composite Indexes of Human and Sustainable Development We used the Borda ranking technique to construct a single index for hu- man development and one for sustainable development. The indicators for the index of human development are * Reduction in infant mortality between the 1980s and 1990s * Reduction in adult illiteracy between the 1980s and 1990s * Increase in life expectancy between the 1980s and 1990s. The indicators for the index of sustainable development are * Decline in emissions of carbon dioxide per capita between the 1980s and 1990s * Decreases in emission of organic water pollutants (kilograms per day per worker) between the 1980s and 1990s * The negative of annual average rate of deforestation measured over 1980-95. The Borda ranking procedure involves assigning each country a point equal to its rank in each component criterion. Each country's points over all the components are averaged, and the averages are used to re-rank the countries. The procedure allows the aggregation of indicators with different units of measurement and different periods and country coverages; that is, it allows comparisons among countries across categories even when the number of countries studied varies by category. For more details regarding the Borda ranking techniques see Fine and Fine (1974a,b), Goodman and Markowitz (1952), Smith (1973), and Thomas and Wang (1996). We also tried an alternative method of aggregation that transformed each component variable into a standardized score, with a mean of 0 and a standard deviation of 1, and then averaged them. In the significance of cor- relation coefficients, the results were substantially similar. 183 A N N E X 2 FRAMEWORK AND EVIDENCE T his annex provides a framework and empirical evidence for chapter 2. A Welfare Function Define an additive and separable welfare function, U, for a society that consists of N individuals N N (A2 I) U = I u(c,) + X, v(h.; R), where c. is the consumption of individual i, hi is the human capital of indi- vidual i, and R is the (aggregate) level of environmental assets. R is assumed to be a pure public good, and hence its distribution among the population is irrelevant. Also, u( ) and v( ) are increasing and strictly concave in their arguments. A second-order approximation of U evaluated at the mean or average values of c and h yields N N (A2.2) U Nu(c) + Xu'(J)(c - ) + 2 U (C)C- C) N _i1N +Nv(h; R) + I v'(h; R)(h.- h) + v"(h; R)(h,- h)2, where c is average or per capita consumption, h is average or per capita hu- man capital, u'(c), v'(h; R) are first derivatives with respect to c and h, re- spectively, evaluated at mean values c and h and u'(c), and v"(h; R) are second derivatives. Taking expectations we obtain the average welfare per individual i, 185 ANN FEX 2 (A2.3) E(U) =u(-c) + 2 u(c)C2 + v(h; R) + 2h; R)a, where a2 is the variance of consumption across the population and & is the c h variance of the distribution of human capital across the population. By strict concavity of u( ) and v( ), we have that u" < 0 and v''() < 0. Thus, aggregate or expected welfare is increasing in c and h and decreasing in c and & . Moreover, because v( ) is increasing in R, Dv"IaR 0 is sufficient to obtain that E(U) is also increasing in R. From the definition in the text, sustained growth requires that the ex- pansion of physical capital through time be accompanied by positive growth of human capital without worsening its distribution. Also, sustained growth is likely to diminish poverty and is not consistent with a worsening of income distribution. Sustained growth increases c and h and reduces, or at least does not increase, C2 and 2. Thus, sustained growth is likely to in- crease welfare, E(U) in equation (A2.3), as long as R does not fall or falls at a sufficiently slow pace. Private Sector Optimization As indicated in the text, human capital (h) and natural capital (R) are subject to two possible externalities associated with consumption and production. Consumption externalities stem from the fact that the posi- tive direct effects of h and R on the welfare function may be only partially considered by the private sector in its resource allocation decisions. Pro- duction externalities arise because much of the positive technological spillovers associated with h may not be considered by the private sector. In addition, part of the value of R as a productive resource may also be ignored by the private sector, particularly in cases where natural capital property rights are not well defined. Here we make an extreme assumption: that all the direct consumption values of h and R on the welfare function (as well as the distributional ef- fects represented by ul and a& ) are ignored by the private sector's produc- tion decisions. Moreover, we assume that production externalities establish a wedge between the private marginal products of h and R and the true marginal products of these resources. That is, the private sector only con- siders a fraction of the contribution of h and R to production. In addition, we assume that a minimum subsistence consumption level, c, exists. The representative household needs a consumption level of c5 to survive and will not allow consumption to reach levels below ce. That is, we impose a subsistence constraint, c - cl > 0. 186 FRAMEWORK AND EVIDENCE Under these assumptions, the relevant problem is maximization of the discounted present value of u( c )-as opposed to that of E(U)-subject to the following constraints: (A2.4) (i) k = G(k, h; R; A(k, h); p) - c - Ih - Ihf- IR (ii) C-C 20 (iii) h If+ Ih (iv) R = o(R) + PIR - W[G( )], (v) k(O) = kc; h(O) = ho; R(O) = Ro where k is per capita physical capital, G( ) is the economy's per capita GDP function, A() is a productivity index, p stands for policy variables and ex- ogenous factors, lh is government investment in human capital, IP is private investment in human capital, P is a parameter, IR is government investment in natural capital, O(R) is a growth function of the renewable resources through time, and i( ) is an increasing function of GDP that reflects the possibly negative direct impact of increased economic activity on natural capital. We assume that population, N, is fixed, so that by using appropriate units it can be normalized to 1. Hence, the distinction between total and per capita variables in equation (A2.4) becomes irrelevant. Also, for alge- braic simplicity we assume a zero rate of depreciation of k and h. Assuming a constant logarithmic depreciation rate for these assets, as is usually done, does not affect any of the results. Several comments about equation (A2.4) are in order: It is assumed that I9 and IR are policy variables. We assume that the effect of GDP on natural capital is not at all in- ternalized by the private sector, and that, as a consequence, the pri- vate sector will not invest in natural capital. Thus, equation (A2.4) (iv) is only used as an accounting identity and is not directly (and ex ante) taken into consideration in decisions by the private sector, even though the evolution of R will affect its future decisions. - The effect of h on G( ) is only partially incorporated into the deci- sions of the private sector. The government may fill a part or the full extent of the possible human capital underinvestment gap left by the private sector. * We allow k and h to affect knowledge represented by the produc- tivity function A(-). It is assumed that knowledge is a public good that any firm can access at zero cost. In line with the "learning by doing" hypothesis, we follow Arrow (1962) and Romer (1986) and assume that learning by doing works through each firm's invest- ment in k. However, we specify that learning by doing requires 187 human capital, or that human capital facilitates and increases the effectiveness of this process. Thus, the function A(-) is assumed to be increasing in its arguments and the marginal effect of k on A increases with h, that is, a2Alakah > 0. * Equation (A2.4) (i) implies that public investment in human capi- tal is financed out of total savings via lump sum taxes. An alterna- tive approach is to assume that public investments are financed via an income tax proportional to GDP, as in Barro (1993). * Production of human capital is assumed to be generated through the same productive process as physical capital and consumer goods. This assumption has often been used in the literature (see, for ex- ample, Barro and Sala-l-Martin 1995). Alternatively, one may pos- tulate a separate production function of h as in Lucas (1988) or Rebelo (1991). Although the latter is a more realistic approach, the assumption of a common production function for consumer and all investment goods considerably reduces the algebra and does not al- ter the basic conclusions. The Case of a Middle-Income Economy with Initial Consumption far above Subsistence First we assume that constraint (A2.4) (ii) is not binding; the economy is sufficiently rich to allow c > c at all times. We will analyze the role of the subsistence constraint in the case of the poor economy. It can be shown that the private sector in this model invests only in k if the marginal product of physical capital, G,(-), is higher than the marginal product of human capital as perceived by the private sector, GP (').t It will invest in both k and h if G P= Gk and will only invest in h if G P > G . Thus, assuming that k is initially relatively low, IP = IP = 0 and k > 0. Of course, the main reason why the private sector only invests in one factor is our as- sumption that all factors are produced out of a common production func- tion. If we allow for a different production function for h, the private sector may be shown to invest in both k and h even outside the long-run equilib- rium. However, the essential point is that the private sector tends to underinvest in human and natural capital relative to physical capital. That is, the private sector tends to have too narrow an investment portfolio as long as the positive external effects associated with h and R are larger than those associated with k regardless of whether h or R have separate produc- tion functions. In a sense, the extreme specification (apart from simplifying the algebra) helps to highlight the fact that the market economy tends to overspecialize its investment choice. 8 1S FRAMEWORK AND EVIDENCE From the first-order conditions of the above problem one can derive the growth rate of the economy in the usual way if Gk > G P. Economic growth is an increasing function of the gap between the marginal return to capital and its marginal cost, b(.). Under the usual assumption of constant risk aversion-for example, that -u'(c) clu'(c) _ H > 0, is a constant-and where u(c) is defined in equation (A2.3), the rate of economic growth is 1 (A2.5) & Ic /= I Gk(k, h; R, A; p) - b(r; p)], where c Ic is the rate of growth of consumption per capita (we have sup- pressed the bar over c) Gk( ) is a function reflecting the marginal product of physical capital for a given level of A, and r is the discount rate.- There are four possible cases: i. Sustained growth requires absolute balanced asset growth. This case occurs if the aggregate production function G(') is subject to constant re- tums to scale (CRS); for example, the spillover effects of k and h on A(-) are negligible. Therefore, Gk is a function only of factor ratios. Assume that h and R remain constant as c Ic > 0 and k /k > 0. In this case the private sector does not invest in R and h. Thus, growth will be unbalanced relying exclusively on the accumulation of k. Because of CRS, Gk( ) declines as k increases. As a result, the expression in square brackets in (A2.5) declines and the "Solow curse" applies. A positive rate of growth cannot be sustained unless the govemment in- vests in h and/or R. (The growth decline is, of course, more rapid if R falls as a consequence of growth). So in this case, sustained growth can only be achieved by the govemment investing in h and R, so that k /k = h /h = R /R. Absolute balanced growth of the three assets is re- quired to sustain a positive growth rate. ii. Sustained growth can be achieved with unbalanced asset growth. This case may occur if large technological spillovers associated with capi- tal accumulation exist. In this case it is possible that the marginal product of k does not decline because A is increasing in k. Now even if h and R do not increase or if they decline at a sufficiently low rate, the growth rate can still be sustained. So in this case we can have sustained yet unbalanced growth based purely on physical capital growth and technological spillovers. iii. Sustained growth can be achieved with semibalanced asset expansion. This could happen if there is a high degree of substitution between h and R in the Gk function. Substitution between h and R allows for two pos- sible subcases as h > hi, where h' is a critical level of human capital: 189 ANN EX 2 a. Under CRS with no spillover effects growth can be sustained if h and k grow at identical rates, that is, the k/h ratio remains constant. Absolute semibalanced asset growth is necessary to pro- duce this scenario. b. Spillover effects that effectively imply that the production func- tion exhibits increasing returns to scale in k and h, but that the net marginal product of k is decreasing in k. In this case, h may grow at a pace slower than k, that is, relative semibalanced asset growth is needed. In this case aCkG/R decreases as h increases and DGJ/R = 0 as h > h', where ho is at a certain critical level. That is, as h increases over hc, economic growth becomes independent of R, although R still has a positive marginal product. Note that the rel- evant substitution is for the marginal product of k function, not for the production function, as is usually assumed. This implies that the relevant substitution between h and R relates to third- order effects and not to second-order effects as the usual Hicksian or Allen elasticities of substitution imply. iv. Sustained growth can be achieved with relative asset growth balance. This case may occur if technological spillovers are dependent on both k and h, with a strong complementary relationship in the A function and h < ho. We argue in the text that the technological spillovers as- sociated with physical capital are not likely to be large in developing countries that do not have a sufficiently high and increasing level of general education. That is, the elasticity of substitution between h and k in the A(.) function is small. If h is too low, the effect of k on A will be small. In this case, sustained growth can be achieved only if h and R increase so that Gk( ) does not fall as k increases. This implies that sustained growth can be achieved with relative, rather than ab- solute, balanced asset growth. An economy can sustain a positive rate of growth when the public sector invests in h and R at a rate generally lower than the rate of physical capital accumulation. The empirical results presented in the text allow us to rule out the first and second cases. That is, although complete or absolute asset balance is not necessary for sustained growth, growth based only on physical capital accumulation is not sustainable either. According to the empirical findings, the last two cases are empirically the most relevant. Poor countries that do not have large levels of human capital require that human and natural capi- tal grow at a certain rate, which is generally lower than that of physical capital, to sustain growth. That is, the last case reflects best the situation 190 FRAMEWORK AND EVIDENCE for poor economies that have not yet developed a solid human capital base. The third case, especially subcase (iii)b, is the most relevant to middle- income countries that already have a significant level of human capital. Figure A2.1 illustrates balanced and unbalanced growth processes under the assumption that there are no economies of scale or technological spillovers associated with capital accumulation and that h > h', which im- plies that changes in R play no role on economic growth. The marginal product G,(Gh) is decreasing (increasing) in the physical capital to human capital ratio. In the figure, Gh is the true marginal product of human capi- tal, x is the marginal contribution of human capital to welfare as a con- sumer good, and therefore G + x is the total and true marginal social con- tribution of human capital. GP is the marginal contribution of human capital as perceived by the private sector. For an economy that grows from a low k/h ratio, the marginal product of k falls along the Gk schedule as k/h increases. In the absence of interven- tion, a laissez faire economy will continue accumulating physical capital until it reaches point B, at which juncture no further growth occurs; k/h does not increase. At this point, Gk = b where b is the marginal cost and, hence, growth stops. In the lower quadrant of figure A2.1 we relate growth of consumption, c /c , to the level of k/h. In the absence of intervention, c /c continuously declines until it reaches point L at (k/h)° where c Ic = 0. (This case represents growth pattern 1 discussed in chapter 2.) If the public sector invests in human capital, however, long-run growth is possible. An optimal intervention would entail a public sector invest- ment in human capital once the economy reaches (k/h)* or point D, where the marginal product, Gk, of physical capital is equal to the social marginal product of human capital G h + x. At this point, Gk = Gh + x > b, so the economy is still growing. However, as growth is now balanced with k/k = h/h, k/h remains constant at (k/h)*. In the lower quadrant, the optimal intervention implies that the k/h stops growing at (kWhV at point d. Here we have a positive and sustainable growth rate of consumption equal to (c/c)*. (This situation reflects growth pattem 3 in chapter 2.) Alternatively, the government may choose to subsidize investors in physical capital by reducing b or increasing Gk over time (see equation A2.5). However, these subsidies must be financed. Assuming that they are financed through lump sum taxes, the budget constraint, equation A2.4 (i), implies that the government must reduce Ihg and/or IR. However, this means that the economy becomes even more dependent on subsidies as a means to sustain growth. In figure A2. 1 this pattem of growth can be shown by a shift to the right of the Gk schedule due to the capital subsidies (or by a fall of b). But the budget constraint implies that the govemment has less resources to 191 ANNEX 2 Figure A2.1. Constant Returns to Scale and No Technological Spillovers $' B b O k/h (k/h)* (k/h)° (k/h)°° dIc Social growth function Private growth function (c/c* ,,,..-------- O ~~(k/h)* (klh)° \ fh)°/*k/ Source; Author. invest in human capital. Thus to preserve growth (to maintain a positive gap between Gk and b) subsidies must be continuously increased over time. That is, schedule Gk should be constantly shifting to the right by permanent and increasing subsidies. Economic growth becomes dependent on ever increas- ing subsidies to capital owners with the consequent negative impact on in- come distribution and human and natural capital. (This is growth pattern 2 discussed in chapter 2.) 192 FRAMEWORK AND EVIDENCE The Case of a Poor Economy Here we consider a poor economy where the initial level of consumption is only slightly above subsistence and is in the process of growing toward a steady-state level of growth. We call this a semisubsistence economy. The idea is that the poor constitute by themselves a subeconomy where most of the growth arises from their own efforts to save and invest. The semisubsistence economy does have contacts with the modern sectors be- cause the poor sell some of their products to the modern sectors and be- cause some of the poor are able to migrate into the modem sectors. For the sake of brevity and simplicity we do not explicitly model either of these processes. We simply postulate that the GDP function of the poor is dependent on shocks arising from the rich sectors through the variable p in the G() function. For example, a recession in the modem economy is translated into a fall in p, which in turn causes both the G(-) and the Gk( ) functions to be displaced downward. Another possible shock arises from the degradation of R caused by an expansion of the modem sector into areas where the poor live.3 We assume that the economy is initially growing by investing primarily in k. The growth of h depends 100 percent on government expenditures in human capital. We define two limiting cases. The first is where income minus deprecia- tion of the asset stocks is exactly sufficient to cover the level of subsistence consumption: (A2.6) cs = ho{G[(k), 1; A; Pi- _(k() - ahk ' where ho is the initial level of human capital and R, is the initial level of natural capital, and we now assume a postive depreciation rate for k and h (5k and a h)' That is, for a given level of R., h. and exogenous policy variables p, there is a unique level (k/h)5 that permits the economy to. exactly satisfy its minimum subsistence consumption. If k/h > (k/h)5, the economy is above subsistence with potential for positive net savings and growth. If k/h < (k/h)5, the economy is not able to cover the depreciation of its stocks of capital, and therefore, with actual consumption equal to c5, stocks are be- ing reduced. That is, the economy is running down its capital. This causes negative growth as k/h falls. The other limiting case is when the economy is barely able to satisfy its subsistence consumption only if it uses its total output without allowing any replacement of stocks: (A2.7) cs = hoG1( ) . 1; h° X A; p] . 193 ANNEX 2 Once k/h = (k/h)", households need to use all their output for con- sumption. At (k/h)" the economy becomes infeasible. Note that both (k/h)' and (k/h)" are dependent on the levels of h0, Ro, A, and p. It can easily be seen that (k/h)' and (k/h)" are both decreasing in h0, R0, A, and p (assuming that G( ) is increasing in p, that is, p represents positive exogenous factors). Thus, a negative shock due, for example, to a recession in the modem economy that reduces the terms of trade of the poor or the level of R due to intrusion of commercial interests on the natural resources owned by the poor (which ironically is more likely to happen during boom times in the modem sector), will increase (k/h)'. Suppose that the economy is initially at (k/h)o greater than (k/h)5. That is, it is growing toward (k/h)* (see figure A2.2). Suppose now that a reces- sion occurs in the modem sector reducing p. This will cause (k/h)' to in- crease. If the new (k/h)' is now greater or equal to (k/h),, then the semisubsistence economy is thrown into a subsistence trap that could lead to negative growth driving k/h toward (k/h)". Consider the case where the initial shock occurs at time t and is even- tually reversed and p is brought back to its original level at time t + 1. Here there are two possibilities: - The fall of k/h between times t and t + T is not too large and (k/h), > (k/h), + , > (k/h)t + T That is, at t + t, when the policy is returned to its original level, the critical (k/h) + T (which is equal to the origi- nal level (k/h)'0 ) is still below the level of (k/h), + T (which is lower than the initial level (k/h)0). In this case, the shock had only a tem- porary negative effect on growth. But once p is returned to its origi- nal level, the semisubsistence economy retakes its growth path. * The fall of k/h between t and t + I is large so that (k/h), > (k/h) +T> (k/h), + That is, as p returns to its original level at time t + 1, the level of (k/h)T has fallen so low that it is now lower than the original critical level. In this case we have what is re- ferred to as "hysterisis": the temporary shock has a permanent effect on the economy, and even if the shock disappears, the economy does not return to its original level. The effect of the shock causes an irre- versible retrocess of the poor economy. The economy falls in a pov- erty cycle leading k/h to continuously fall towards (k/h)", at which point it ceases to exist as a viable economy. Figure A2.2 may help clarifying these points. The figure shows possible paths for the poor economy. If initially the k/h ratio is above the critical (k/h)', the economy is in a path of accumulation following the FJ line in the middle panel in figure A2.2 toward (k/h)*. Throughout this path per capita 194 FRAMEWORK AND EVIDENCE Figure A2.2. Subsistence, Growth, and Poverty Traps among the Poor: The Case of Constant Returns to Scale and No Spillovers $ G, o * e *k/h ((k/h) (k/h) ' (k/h) c/c o k/h *(k/h) (k/h)' (klh)* Iz cS vsN . *k/h (k/h)" (k/h)' (k/h)* Source: Author. consumption is continuously growing, although at a decreasing rate. Eventu- ally, the economy reaches (klh)*, at which point it grows at a constant rate indefinitely. The lower panel shows the evolution of the consumption level, which increases permanently. Suppose now that a negative shock takes place while the economy is along the FJ path. This causes both the schedules Gk and Gh +x to shift downward with a new intersection at a point lower than 195 AN N EX 2 D, which implies a lower rate of long-run growth. But the most important consequence is that (k/h)' will move to the right as it increases because of the reduced level of G implied by a negative shock. The key issue is whether the initial (k/h) ratio is now below or above the new critical (k/h)' after the shock. If (k/h) is below the new (k/h)W, the path of the economy reverses along a stagnation path such as the path NM in the lower panel of the figure. That is, the economy, which was originally growing, becomes stagnant and eventually, as it reaches point M, becomes infeasible. Suppose the initial k/h is sufficiently low so that the economy enters a stagnation phase, with a declining k/h ratio, but that after a period of time the level of p is reestablished at its original level. The question is whether or not the new (k/h)' is above the current k/h ratio. If it is above, then the economy does not retake its original growth path. It continues into a downward spiral, further reducing its wealth. That is, the reversal of the growth process becomes permanent and a purely temporary shock has had a permanent, irreversible effect, triggering a vicious cycle of poverty and asset disaccumulation. Econometric Specification Used to Estimate Growth Functions A basic behavioral equation arising from both neoclassical and endogenous growth models is the following: (A2.8) g = O[FK(K, H, R; A, p) - CK (r, 5, p)], where g is the rate of GDP growth per capita; K, H, and R are per capita physical, human and natural capital, respectively; p is a vector of policy vari- ables and prices; A is a productivity factor; FK is the marginal return to K; CK is the marginal cost of capital that typically depends on the discount rate (r), depreciation rate (3), and presumably on policy variable p (such as subsidies to investment); and p( ) is a monotonic and increasing function. Equation (A2.8) indicates that growth is dependent on the gap be- tween the marginal returns to capital and its marginal cost. If such a gap is positive, growth is also positive, and growth comes to a halt if such a gap disappears. Moreover, under certain commonly assumed conditions, growth is directly proportional to this gap. Thus, this basic behavioral expression relates economic growth to the level of asset stocks, total factor productivity, discount rates, and policy variables. Most empirical growth studies, however, do not use this behav- ioral approach and instead rely on various forms of growth accounting identities that relate growth to changes in asset stocks instead of their lev- els as a growth theoretic model suggests. 196 FRAMEWORK AND EVIDENCE Our empirical analysis is based on equation (A2.8). If one includes time in a discrete form, it is natural to postulate that growth in one period de- pends on asset stocks at the end of the previous period. Thus a more opera- tional expression for equation (A2.8) is, (A2,9) g, = [FK(l, _' Hi ,I R.t ,; Ar, p,t, al) - CK (r, 8,t P't; a)I, where i stands for country and t is time. We assume that r' and a. are coun- try fixed characteristics that influence technology and costs. That is, coun- tries differ in terms of their discount rates, r', and technological or institu- tional characteristics, a (for example, property rights and rule of law). We note that in equation (A2.9) growth at time t is dependent on lagged asset stock levels instead of current flow of asset changes as assumed in many em- pirical studies. That is, this growth theoretic equation provides for natural "instrumental" variables by postulating growth as a function of last period stock levels. This goes some way in decreasing biases arising from contempo- raneous correlation between explanatory variables and the error term due to endogeneity of such variables. Lagged stock levels are much less likely to be endogenous to growth rates than contemporaneous stock changes. Since we are relating current growth to lagged stocks of assets, we have that K itl = (1 - 3)K, 2 + I,t ,, where Iit _ I is investment per capita in period t - 1. So replacing this in equation (A2.9), we find that using lagged stock levels in the growth regression is equivalent to regressing growth on twice lagged stocks and lagged investment. If we repeat this process by substituting K. 2' using a similar expression we can go back to the first year asset stock. So estimating equation (A2.9) is equivalent to estimating growth on the per capita lagged investment levels of each asset and the "initial" level of each asset. Hence, this specification implicitly uses the initial level of income (given that the initial income level is a function of all initial assets) as an ex- planatory factor. That is, we could, in principle, relate the estimated coeffi- cients of assets to analyses of convergence of growth rates across countries. We also assume that the unobserved total factor productivity is related to assets stocks and other country characteristics, for example, A,,(K, - H,' I, R1,-,, a.). That is, even if FK is declining in K,,_,, the growth rate can be increasing or nondecreasing in K_,_ if the technological and scale spillovers are powerful enough. That is, if the partial effect of K, 1 on A. is positive and of sufficient magnitude so that dFKIdKi, - 1 = dFKIdK, - I + (dFKIdA,,)(dAjd/K, - [)> 0. Thus, we estimate a reduced form of equation (A2.9) allowing for fixed country effects, (A2.10) g1, = y4K11 , H, 1, R,, , P,j + J3 + f, + po where V'(.) is a general well-defined function, ,B is a coefficient capturing the country fixed effect related to the effects of a, and r' in equation (A2.9), 197 ANNEX 2 and ji,, is a random disturbance. The coefficient fL corresponds to time effects. The empirical estimation uses various functional forms for y(K), including a logarithmic one and a translog form to allow for interasset and policy interac- tions. The use of country fixed effects deals with biases arising from omitted variables corresponding to possibly large numbers of country-specific vari- ables that are not observed. Thus, the specification in equation (A2.10) helps to reduce biases due to both endogeneity of explanatory variables by using lagged asset stock variables as instruments, and omitted variables by using fixed effects. Evidence from Developing Countries Table A2.1 presents the empirical evidence for the section "Econometric Evidence: 20 Middle-Income Countries." Tables A2.2 and A2.3 show empirical results for the section "Econo- metric Evidence: 70 Developing Countries." Table A2.4 shows some empirical studies on the impact and size of capital subsidies. Notes 1. Consistent with the discussion above, Gbh < Gh, where Gh is the true marginal product of human capital. 2. The marginal cost function, b, is equal to r + 8, where 6 is the depreciation rate of capital. Here we allow for policies, p, to affect the marginal cost of capital. 3. This is consistent with one stylized fact that is valid for several tropical countries, especially in Latin America and Asia: though the poor are most dependent on natural resources, most destruction of these resources is caused by large commer- cial interests that intrude into resources owned by the poor (see the ample empiri- cal evidence on these issues provided by Kates and Haarmann 1992). 198 FRAMEWORK AND EVIDENCE Table A2.1. The Growth Equation under Various Specifications (dependent variable: GDP growth per capita) Fixed effects Random effects Variables Equation I Equation 2 Equation 3 Equation 4 Average schooling 0.005 0.004 -0.012 -0.013 (0.025) (0.020) (0.009) (0.009) Schooling x reform dummy variable 0.084** 0.084** 0.049** 0.049** (0.024) (0.024) (0.018) (0.018) Per capita capital stock -0.021* -0.021** -0.012** -0.009** (0.012) (0.010) (0.005) (0.004) Capital x reform dummy variable -0.016** -0.016** -0.008** -0.008** (0.005) (0.005) (0.004) (0.004) Dummy 1982-85 -0.019** -0.019** -0.017** -0.018** (0.005) (0.005) (0.005) (0.005) Labor force -0.001 n.a. -0.006 n.a. (0.067) n.a. (0.006) n.a. Standard deviation of log of schooling -0.018 -0.018 -0.034** -0.033** (0.019) (0.016) (0.012) (0.012) Homoscedasticity (Breusch-Pagan test) Rejected at Rejected at Not rejected Not rejected 5 percent 5 percent at 5 percent at 5 percent White test of specification Rejected at Rejected at n.a. n.a. 5 percent 5 percent Hausman test: Fixed vs. random effects n.a. n.a. Not rejected Not rejected at 5 percent at 5 percent n.a. Not applicable. * Significant at the 10 percent level. ** Significant at the 5 percent level. Note: All variables are in log form. All explanatory variables are lagged by one period. Standard errors of the coefficients are in parentheses. Data from 20 countries are presented. White's heteroscedasticity consistent standard errors are reported under fixed effects. Source: L6pez, Thomas, and Wang (1998). 199 ANNEX 2 Table A2.2. GDP Growth Rates Regressed on Stocks Per Worker, Using All Countries with Available Data from 1965 to 1990 No cross-products: Translog: fixed No cross-products: fixed effects (with effects (with fixed effects (with country and time country and time Variables country dummies) dummies) dummies) Observations 335 335 335 Countries 67 67 67 Log likelihood -631.70 -606.30 -605.80 In (capital/labor) 10.34 11.36 13.21 (4,79) (5.67) (3.19) In (forest area/labor) -1.31 -0.54 8.86 (-0.68) (-0.31) (2.15) In (education) -19.56 -21.41 -12.32 (-5.68) (-6.60) (-2.42) [In (capital/labor)]2 -0.74 -0.95 -1.11 (-6.34) (-6.93) (-4.88) [In (forest area/labor)]2 0.31 0.36 0.09 (2.74) (3.25) (0.62) [In (education)]2 1.36 1.44 0.84 (5.52) (6.20) (1.64) In (capital/labor) x In (forest area/labor) n.a. n.a. 0.108 n.a. n.a. (-0.54) In (capital/labor) x In (education) n.a. n.a. 0.467 n.a. n.a. (0.78) In (forest area/labor) x In (education) n.a. n.a. -0.596 n.a. n.a. (-2.03) n.a. Not applicable. Notes: 1. t-statistics are in parentheses. 2. The dependent variable is annual per capita GDP growth computed over a five-year interval using annual data. The regression is ln(GDP) = a + bt + e, where e is the residual. Growth rate equals 100*[exp(b).1l. 3. Parameters were computed by iterated feasible generalized least squares (FGLS), and therefore should be equivalent to maximum likelihood estimation. 4. The correction for AR( I) selected a single parameter for all countries together. 5. The correction for groupwise heteroscedasticity was done by computing a group variance for each country. 6. Measures of per capita GDP and labor were taken from Summers' and Heston's Penn World Tables Mark 5.6. Measures of education were taken from Barro and Lee (1997). They represent average years of education for people 25 years and older. Measures of per capita capital were taken from King and Levine (1993). Measures of forest area (resource stock) were taken from World Resources 1996-97 Data Disk, and is originally from FAOSTAT. The Penn World Tables may be downloaded from http:// ww-wnuff.ox.ac.uk/Economics/Growth/. The Barro-Lee and King-Levine datasets may be downloaded from the World Bank's Web page at http://,-w,wworldbank.org/html/prdmg/grthweb/ddkile93,htm. The forest data may be downloaded from the Food and Agriculture Organization of the United Nations web page at http://apps.fao.org/. Source: Lopez, Thomas, and Wang (1998). 200 FRAMEWORK AND EVIDENCE Table A2.3. Elasticities for Stocks Per Worker on GDP per Capita Growth Rates Elasticity Variables Minimum value Maximum value Average No cross-products allowed Capital/labor 0.038 -0.081 -0.040 (0.019) (0.022) (0.009) Forest area/labor 0.007 0.071 0.047 (0.046) (0.027) (0.022) Education (average schooling of labor force) -0.056 0.056 0.018 (0.011) (0.020) (0.011) Translog functon Capital/labor 0.046 -0.093 -0.045 (0.022) (0.026) (0.012) Forest area/labor 0.034 0.050 0.044 (0.049) (0.029) (0.023) Education -0.031 0.035 0.012 (0.028) (0.022) (0.013) Notes: 1. Elasticities are computed by converting the percentage growth rate to the log of the growth rate by dividing the percentage by 100. 2. Marginal effects are computed using the fixed effects regression with country and time dummies, corrected for groupwise heteroscedasticity for all countries, and a common AR(1) term for autocorrelation. Data are for all countries, 1965-90. 3. Marginal values (dy/dx) computed for the unlogged x's are simply the exponential of their respective logged values. This means that x bar is not the true mean. 4. Marginal values for the translog formulation utilize the mean values of the log of the crossed term. 5. Standard errors are in parentheses. They are based on variability in the parameter estimates only (including covariances between parameters) and not on any variability in the minimum or maximum variable mean. 6. The elasticity of labor is computed as the negative of the sum of the elasticities for capital/labor and resources/labor. Source: L6pez, Thomas, and Wang (1998). 201 AN N EX 2 Table A2.4. Selected Empirical Studies on the Impact and Size of Capital Subsidies Authors Methods Major findings Studies on the impact of subsidies Bergstrom (1998). "Capital The study examines the "In many countries, governments grant different capital Subsidies and the effects on TFP of public subsidies to the business sector in order to promote growth... Performance of Firms." capital subsidies to firms in The results suggest that subsidization can influence growth Sweden between 1987 and (in a short run), but there seems to be little evidence that the 1993. Panel data were used. subsidies have affected productivity" (p. 1). Bregman, Fuss, and Regev It uses a time-series cross- "An industrial policy of subsidizing physical capital (1999). "Effect of Capital section microdata set for investment has been utilized in many countries.... We Subsidization on 620 firms for Israel. estimated that for the years 1990-94, this policy has resulted Productivity in Israeli in production inefficiencies ranging from 5 to 15 percent for Industry," subsidized firms" (p. 77). Harris (1991). "The Uses CES (constant "The results indicate that, since manufacturing industry in the Employment Creation elasticity of substitution) province tends to operate with a labor-intensive technology Effects of Factor Subsidies." production function and a and, its price elasticity of demand for output is very low, the simulation model for the employment-creating effects of capital subsidies are strongly Northem Ireland negative" (p. 49). manufacturingindustry, 1955-83. Lee (1996). "Government Uses four-period panel data "Industrial policies, such as tax incentives and subsidized Interventions and for the years 1963-83. credit, were not correlated with total factor productivity Productivity Growth." growth in the promoted sectors" (p. 391). Lim (1992). "Capturing the Uses firm-level data from "Most developing countries provide fiscal incentives to Effects of Capital Subsidies." 3,900-4,900 firms in Malaysia, encourage domestic and foreign investment. This study shows from 1976 to 1979. that these schemes subsidize significantly the use of capital and produce greater capital intensity in Malaysian manufacturing" (p. 705). Oman (2000). "Policy The study addresses three "Incentive-based competition for FDI is a global Competition for Foreign sets of questions: (a) to what phenomenon: govemments at all levels in both OECD and Direct Investment," OECD extent do govemments non-OECD countries engage in it worldwide... The Development Centre. compete for FDI, (b) the distortionary effect of incentives.. can be significant... It can effect of competition, and be counterproductive for govemments to offer costly (c) implications for investment incentives" (p. 7-9). Investment incentives in the policymakers. automobile industry are shown in a table on page 73. Studies on the size of subsidies Gandhi, Gray, and Estimated subsidies that "Estimated subsidies to energy, roads, water, and agriculture in McMorran (1997). damaged the environment. developing and transition economies total over $240 billion per year in the 1990s. Cutting these subsidies in half would free over $100 billions of finance for sustainable development" (p. 10). Moore (1999). "Corporate A testimony before the "Corporate welfare is a large and growing component of the Subsidies in the Federal Budget Committee, U.S. federal budget" (p. 1). In 1997, the Fortune 500 corporations Budget." House of Representatives. are estimated to have received nearly US$75 billion in govemment subsidies. (table continues on following page) 202 FRAMEWORK AND EVIDENCE Table A2.4 continued Authors Methods Major findings de Moor and Calamai A report to the Earth "In OECD countries, total annual subsidies in four sectors, (1997). Subsidizing Council, which estimated energy, road transport, water and agriculture, amounted to Unsustainable public subsidies in four $490 -$615 billion, and in non-OECD countries, Development: sectors. $217-$272 billion. Total subsidies in all four sectors are Undermining the Earth estimated at $710-$890 billion worldwide" (p. 93). with Public Funds. Gulati and Narayanan This paper estimates the "Broadly half of the huge agricultural subsidy on fertilizers and (2000). "Demystifying amount of subsidies and power... comprising 2 percent of GDP, is either going to Fertilizer and Power examines the real industry in the case of fertilizers or is being stolen by non- Subsidies in India." beneficiaries. agricultural consumers in the case of power" (p. 784). Note: A more detailed table is available on request from the authors. Source: Authors. 203 A N N E X 3 DISTRIBUTION OF EDUCATION, OPENNESS, AND GROWTH EF conomic theories suggest a strong causal link from education to -.a growth, but the empirical evidence has not been unanimous and conclusive. L6pez, Thomas, and Wang (1998) focus on two factors that explain why the empirical studies have not overwhelmingly supported the theories. First, the distribution of education affects economic growth. Second, the economic policy environment greatly affects the impact of education on growth by determining what people can do with their education. Reforms of trade, investment, and labor policies can increase the returns from education. Using panel data from 20 developing countries for 1970-94, we investigated the relationship between education, policy reforms, and economic growth and made the following observations: *T he distribution of education matters. An overly skewed distribution of education tends to have a negative impact on per capita income in most countries. Controlling for education distribution and using the appropriate functional form leads to positive and significant ef- fects of average education on per capita income, while failure to do so leads to insignificant or negative effects of average education. * The policy environment matters greatly. Results indicate that economic policies that suppress market forces tend to reduce the impact of edu- cation on economic growth. Moreover, the stock of physical capital is negatively related to economic growth for economies in the sample, implying a declining marginal productivity of capital. 205 ANNEX 3 The Extended Production Function with Distribution of Education We used a model in which physical capital is fully tradable, but human capi- tal is not. The level, as well as the distribution, of human capital enters the aggregate production function. If education matches the dispersion of ability, the marginal effect of education distribution on income vanishes. If the dis- persion of education is greater than the dispersion of ability, the per capita income can be increased by reducing the dispersion of education. If the dis- persion of education is less skewed than ability, then govemments should concentrate investment on a few people with greater ability to learn. The education Gini coefficient is calculated in two steps. First, an educa- tion Lorenz curve is constructed based on the proportions of population with various levels of schooling and the length of each level of schooling, which shows the cumulative years of schooling with respect to the proportion of population. Then the education Gini coefficient is calculated as the ratio of the area between the Lorenz curve and the 45 degree line (perfect equality) to the total area of the triangle. An alternative definition of the education Gini coefficient is the ratio to the mean schooling of half the summation of the absolute differences of school attainment between all possible pairs of in- dividuals in a country (Deaton 1997).' Table A3.1 presents education Gini coefficients for 20 countries, and preliminary data estimated for 85 countries are available from Thomas, Wang, and Fan (2000). Using quinquennial data from 20 mostly middle-income countries, aggre- gate production functions were estimated. Table A3.2 reports four estimates of the aggregate per capita production function for 1970-94. The first col- umn presents the traditional fixed effect log-linear model that ignores both of the above explanatory factors: education distribution and the policy environ- ment. As the first column shows, human capital has a negative and signifi- cant effect on production; this is where the "education puzzle" lies. The second column shows the fixed effect model in log-linear form, but the estimation allowed the distribution of education to play a role in the function. Column two allows no country-specific effect from education distri- bution and shows positive associations between human capital stock, its dis- tribution, and level of income. In this case, the coefficient of average educa- tion becomes positive and statistically significant at 5 percent. The effect of education distribution on the production function was statistically different across countries. This cross-country diversity of the effect of education disper- sion is consistent with the idea that the effect of education dispersion is likely to vary and change sign according to whether it is below or above its optimal level. The third column presents the results obtained by allowing for country- specific effects of education distribution. The coefficients of the variability 206 DISTRIBUTION OF EDUCATION, OPENNESS. AND GROWTH Table A3.1. Gini Coefficients of Education for Selected Countries, Selected Years Country 1970 1975 1980 1985 1990 Algeria 0.8181 0.7683 0.7080 0.6525 0.6001 Argentina 0.3111 0.3257 0.2946 0.3182 0.2724 Brazil 0.5091 0.4290 0.4463 0.4451 0.3929 Chile 0.3296 0.3327 0.3151 0.3120 0.3135 China 0.5985 0.5541 0.5094 0.4937 0.4226 Colombia 0.5095 0.4594 0.4726 0.4752 0.4864 Costa Rica 0.4106 0.3916 0.4059 0.4165 0.4261 India 0.7641 0.7429 0.7517 0.7238 0.6861 Indonesia 0.5873 0.5817 0.5051 0.4388 0.4080 Ireland 0.2488 0.2454 0.2364 0.2377 0.2498 Korea, Republic of 0.5140 0.3942 0.3383 0.2877 0.2175 Malaysia 0.5474 0.5150 0.4719 0.4459 0.4204 Mexico 0.5114 0.4990 0.4978 0.4695 0.3839 Pakistan 0.8549 0.8450 0.8170 0.8065 0.6448 Peru 0.5048 0.5028 0.4258 0.4371 0.4311 Philippines 0.4327 0.3578 0.3404 0.3360 0.3285 Portugal 0.4985 0.5142 0.4255 0.4350 0.4315 Thailand 0.4185 0.4257 0.3591 0.3891 0.3915 Tunisia 0.8178 0.7589 0.6935 0.6710 0.6168 Venezuela, RB 0.5789 0.5585 0.3919 0.3970 0.4209 Source: L6pez, Thomas, and Wang (1998). For data on additional countries, see Thomas, Wang, and Fan (2000). of education for the various countries are jointly significant at 1 percent. However, 7 of the 20 country-specific coefficients are not statistically dif- ferent from zero. The last column uses the standard deviation in logs as another measure of dispersion of education. This measure of dispersion exerts a much greater effect on per capita income. Most of these country-specific coefficients are negative, and 8 out of 20 coefficients are highly significant. Table A3.3 presents the results obtained by using the nonlinear specifi- cation suggested by the theoretical model. That is, this specification deals with both the omitted variable and the functional form specification prob- lems. In all three specifications, the coefficients of average education are positive and statistically significant at the 5 percent level. In this functional form, the distribution of education is positively associated with the level of income, which is still consistent with the model that states that a certain level of education dispersion is important for production, especially in con- sidering technological progress and innovation. 207 ANNEX 3 Table A3.2. Production Function: Linear Estimation (dependent variable: log of per capita GDP) Fixed effects, Fixed effects, allowing Fixed effects, log-linear allowing for education log-linear allowing for education distribution distribution effects Fixed effects, for education distribution effect using coefficient using standard excluding effect of effect using coefficient of of variability of deviation of the log of Variables education distribution variability of education education education Human capital -0.275** 0.491** 0.004 -0.380** (0.085) (0.106) (0.112) (0.131) Physical capital 1.108** 0.981** 1.066** 1.083** (0.033) (0.012) (0.022) (0.071) Dummy 1982-85 -0.063** -0.077** -0.063** -0.033** (0.012) (0.012) (0.011) (0.009) Education 1.187** distribution effects (0.133) Brazil 2.828** 0.423** (0.350) (0.196) Chile -0.020 -0.320 (0.309) (0.279) China 0.354** -1.197** (0.139) (0.225) Colombia 0.765 -0.300 (0.916) (0.269) India 0.012 0.015 (0.278) (0.299) Korea, Republic of 1.146** 0.012 (0.089) (0.148) Mexico 0.843** -0.475 (0.264) (0.306) Malaysia 2.494** -0.690** (0.196) (0.304) Peru 0.574 -0.409 (0.559) (0.344) Philippines -2.138 -0.861** (2.627) (0.275) Thailand -2.478** -0.541** (0.618) (0.175) Venezuela, RB 1.032** -0.109 (0.142) (0.330) Algeria -0.685* 0.818* (0.378) (0.471) Argentina 1.307** -0.367 (0.316) (0.269) Costa Rica -3.849** -0.666** (0.579) (0.222) (table continues on following page) 208 DISTRIBUTION OF EDUCATION, OPENNESS, AND GROWTH Table A3 .2 continued Fixed effects, Fixed effects, allowing Fixed effects, log-linear allowing for education log-linear allowing for education distribution distribution effects Fixed effects, for education distribution effect using coefficient using standard excluding effect of effect using coefficient of of variability of deviation of the log of Variables education distribution variability of educaton education education Indonesia 2.081** -1.004** (0.298) (0.157) Ireland 1.287** 0.251 (0.161) (0.284) Pakistan -0.024 -0.292 (0.165) (0.321) Portugal -0.001 0.027 (0.483) (0.238) Tunisia 0.654** -0.065 (0.188) (0.484) * Significant at the 10 percent level. ** Significant at the 5 percent level. Note: A first-order autoregressive coefficient was estimated by maximum likelihood for each country separately. This information was used to correct the data. Standard errors (in parentheses) reported are White's heteroscedastic consistent. All variables are in log forms, except for dummies. Source: L6pez, Thomas, and Wang (1998). Empirical Analysis on Education and Investment Returns Based on the World Bank's lending experience during the past 20 years, Thomas and Wang (1997) examined whether education and openness can improve the developmental impact of investment projects. The model is a country's production function separated into export production and pro- duction for domestic markets. The reduced forms are as follows: P(Sat= 1),= ct + -j+ + y-G,+ (p3-R + ERR.= acEc+ Xi+Y'G.+ p-R.+s. where P(sat = 1) is the probability of a project i being rated as satisfac- tory, ERR, is the economic rate of return for project i, Ei is the change in average level of schooling of the labor force for the country where the project is located and the period when the project is implemented, X, is the vector of variables indicating export growth or openness, G is the vector of variables indicating governance and institutional capability, and R includes exogenous variables and regional dummies. The first 209 ANNEX 3 Table A3.3. Production Function: Nonlinear Estimation (dependent variable: log of per capita GDP) Nonlinear, allowing for distribution effects to Nonlinear, allowing for vary across countries Nonlinear, allowing distribution effects to with different levels Variables for one distribution effect vary across continents of education variability Human capital 0.369** 0.272** 0.159** (0.049) (0.051) (0.056) Physical capital 0.842** 0.863** 0.897** (0.018) (0.019) (0.017) Dummy 1982-85 -0.066** -0.065** -0.061** (0.012) (0.12) (0.011) Education distribution effects (pa) Overall 7.532** (0.831) Latin America 13.040** (2.407) Asia 9.541** (1.611) Africa 3.720** (0.656) Europe 8.140** (2.362) Low variability 11.416** (3.624) Medium variability 32.595** (10.195) High variability 3.145** (0.533) * Significant at the 10 percent level. ** Significant at the 5 percent level. Note: A first-order autoregressive coefficient was estimated by maximum likelihood for each country separately. This information was used to correct the data. All variables are in log form except for dummies. Data from 20 countries were used in the analysis. Standard errors are in parentheses. Source: L6pez, Thomas, and Wang (1998). equation is estimated using Probit analysis because the dependent vari- able is a discrete (0/1) variable, and the second equation using Tobit procedure because ERRs are truncated at 5 percent. Project Data After each World Bank project is completed, a project completion report is written and two performance measures are calculated. Operations 210 DISTRIBUTION OF EDUCATION, OPENNESS, AND GROWTH Evaluation Department staff evaluate the project and assign an overall performance rating of satisfactory or unsatisfactory in achieving the project's development objectives. An ex post economic rate of return (ERR) is also calculated for projects in eight sectors-infrastructure, agri- culture, industry, energy, water, urban, transport, and tourism-where the stream of project benefits can be quantified. The ERR is the discounted stream of project costs and benefits over the life of the project, evaluated at economic prices. The ex post ERRs are calculated approximately two to three years after project completion, at which time the evaluators know the actual investment costs and the actual operating costs and demand, but they still need to estimate the future stream of benefits. Explanatory Variables No attempt is made to build a complete model of determinants of project success, which would require sector- and project-level information as well as country-level information. Four groups of explanatory variables were used: - Education, which can be measured by three variables. They include changes in the average years of schooling of the labor force between project approval and evaluation years; interaction of education and openness, measured by deviations in trade shares; and initial level of education, which was based on Nehru, Swanson, and Dubey (1995) and updated by Patel. * Indicators of openness, including the foreign exchange black market premium and deviations in trade shares, defined by actual trade share minus predicted trade share that were estimated by a simple gravity model. * Governance and institutional capability, which can be reflected indirectly by an index for corruption in government (Interna- tional Country Risk Guide 1982-95), by shares of government consumption in GDP, and shares of budget surplus/deficit in GDP. The second and third measures can reflect the government's ability to control its finances and implement strict fiscal prudence and discipline. Regression results are presented in table A3.4. The findings suggest the importance of trade openness and education for improving invest- ment project performance and the potential gains from outward- oriented learning. Good governance and strict fiscal discipline are also found to be conducive to higher project returns (see Thomas and Wang 1997). 211 ANN EX 3 Table A3.4. Education, Openness, and Lending Project Performance Dependent variable = economic rate of return Dependent variable = satisfactory or not Independent variables Tobit coefficient Prob > Chi Probit coefficient Prob > Chi Education vlanables Change in education levels between the approval and evaluation years 3.33 0.01 0.34 0.00 Education x trade openness (measured by deviations from predicted trade shares) 0.00 0.04 0.00 0.45 Lack of openness Log of foreign exchange black market premium (3-year moving average) -3.14 0.04 -0.23 0.01 Institution and governance Share of budget surplusideficit in GDP (3-year moving average) 0.26 0.05 Corruption in government (I = more, 6 = less) 0.06 0.04 Other controlling variables and dummies Initial level of GDP per capita in the project approval year 0.00 0.95 -0.06 0.02 Dummy for project complexity -4.27 0.00 -0.45 0.00 Sub-Saharan Africa 5.31 0.41 1.56 0.00 East Asia 9.13 0.15 2.56 0.00 South Asia 10.47 0.09 2.13 0.00 Latin America and the Caribbean 7.77 0.24 1.92 0.00 Europe, Middle East, and North Africa 10.80 0.09 2.20 0.00 Log likelihood -3,209.00 -1,032.00 Number of observations 830.00 1,826.00 Note: Prob = 0.05 means rejection of coefficient = 0 at 95 percent confidence. The regressions cover projects evaluated in 1974-92. Source: Thomas and Wang (1997). Selected Literature on Asset Distribution and Growth Table A3.5 includes a selected set of empirical studies on asset distribu- tion and economic growth, which provided some of the evidence used in chapter 3. Notes 1. The education Gini coefficient can be calculated using the formula below: r= IN( 1) ,-T I I x - Xj I . V'pN(N- 1) "i i I where y is the Gini index, ,u is the mean of the difference in school grade at- tained, and N is the total number of observations (see Deaton 1997). 212 DISTRIBUTION OF EDUCATION, OPENNESS, AND GROWTH Table A3.5. Selective Empirical Studies on Asset Distribution, Growth, and Poverty Authors Methodology Major findings Maas and Criel (1982) Calculated education Gini coefficients Inequality in the distribution of education based on enrollment data for 16 East opportunities varies enormously across African countries. countries. Ram (1990) Calculated standard deviations of education As the average level of schooling rises, for about 100 countries, educational inequality first increases, and after reaching a peak, starts to decline. The tuming point is about seven years of education. O'Neill (1995) 1. Assumed that the stock of human Among the developed countries, capital is the accumulation of the past convergence in education levels has resulted education, not sensitive to current in a reduction in income dispersion. income level. However, worldwide, incomes have 2. Used the variance of income and that diverged despite substantial convergence in of human and physical capitals in education levels. analysis. 3. Used both quantities and prices of human and physical capital. Ravallion and Sen (1994) Presented a country case study on assessment Land-contingent poverty alleviation of effectiveness of poverty reduction policy. schemes in Bangladesh made an impact on poverty reduction, "though the maximum gains tum out to be small" (p. 823). Deininger and Squire (1996) Land Gini coefficient Countries with more equitable land Average GDP growth (1960-90) distribution tend to grow faster. Ravallion (1997) Income Gini coefficient At any positive rate of growth, the higher Growth rate the initial inequality, the lower the rate at which income-poverty falls. Birdsall and London'o (1998) A cross-country analysis using a traditional Initial levels of educational inequality and growth model, after controlling for capital land Gini coefficient have strong negative accumulation, initial income and education impacts on economic growth and income levels, and natural resources. growth of the poorest. Deininger and Squire (1998) Provided cross-country data on income and "There is a strong negative relationship asset (land) distribution between initial inequality in the asset distribution and long-term growth; inequality reduces income growth for the poor, but not for the rich; and available longitudinal data provides little support for the Kuznets hypothesis." Li, Squire, and Zou (1998) Land Gini coefficient Income Gini coefficient is positively related Income Gini coefficient to log of land Gini coefficient IDB (1998) Regression using data from 19 countries, Income inequaliry (Gini) is negatively land Gini, income Gini, education, related to land Gini, and positively related standard deviation of education to standard deviation of education. (table continues on following page) 213 ANNEX 3 Table A3.5 continued Authors Methodology Major findings L6pez, Thomas, and Wang (1998) A production function with nontradable 1. The distribution of education matters education is estimated using quinquennial for income levels as well as for growth. data for 20 countries, after controlling for 2. Trade openness and reforms improved physical capital, labor, and so forth. the productivity of human capital in Education Gini coefficients were estimated growth models. by attainment data. Ravallion and Datt (1999) Used 20 household surveys for The growth process was more India's 15 major states in 1960-94 pro-poor in states with higher initial to study the issue of "when is literacy, higher farm growth pro-poor." Elasticities of productivity, and higher rural poverty to nonfarm output were living standards relative to urban residents. estimated. Kerala has the highest elasticity of poverty to nonfarm output. Source: Compiled by authors. 214 A N N E X 4 MEASURING NATURAL CAP ITAL ystematic quantitative analysis of the links between the state of the environment and economic growth has been hampered by a lack of reliable data on natural capital. Because of the increasing need for rigorous research on the subject, a number of initiatives have been directed at collecting more reliable and systematic data worldwide: data on deforestation compiled by the Food and Agriculture Organization of the United Nations; urban air pollution measurements compiled by the WHO; statistics on access to sanitation, clean water, and many other environmental aspects compiled by the UNDP; and the Global Environ- mental Monitoring System. The World Development Indicators (WDI) is an annual publication of the World Bank and includes a large body of environmental data taken from the sources mentioned above and many others as well. By bringing to- gether useful data from diverse sources in a user-friendly format, the WDI has rapidly become a one-stop shop for accessing environmental data. Most environmental data used in chapter 4 and elsewhere in this report have been drawn from WDI tables. In addition to being a rich source of data, the WDI is also a useful refer- ence on how to access other sources of data. The publication and its data tables may be accessed on the web at: www.worldbank.org/data/wdi. Two indexes for natural capital have been constructed on the basis of data available in the WDI, namely: * The sustainable development index to measure environmental outcomes. This is a change index constructed by giving equal weights to annual rates of deforestation between 1980 and 1995, decrease in water 215 ANNEX 4 pollution proxied by the emissions of organic water pollutants in kilo- grams per day per worker between the 1980s and 1990s, and decline in carbon dioxide emissions in per capita metric tons between the 1980s and 1990s. The environmental policy index as a measure of government commitment to protecting natural capital. This includes two dummy variables: one for domestic action based on the formulation of a national environ- mental strategy and environmental profiles and the other for intema- tional action based on signing the Convention on Climate Change. Both indexes are preliminary and work is ongoing to develop more com- prehensive and reliable indicators. The following web site gives an idea of World Bank efforts in this direction: http://www-esd.worldbank.org/eei/. 216 A N N E X 5 FINANCIAL OPENNESS O u r broad measure of financial openness incorporates controls and/or restrictions on both current and capital account transactions. It is constructed as the simple arithmetic average of quantitative measures of degree of controls or restrictions on 27 individual transactions related to import payments, export proceeds, invisible transactions, and capital account transactions as shown in table A5.1. This classification is based on the International Monetary Fund's annual report on Exchange Arrangements and Exchange Restrictions. The scoring draws on the methodology developed by Quinn and Inclan (1997). It is based on a five-tiered scale that ranges from 0 to 2 for each item, indicating the degree of openness (0 highly controlled, 2 highly lib- eral) defined as follows:' 0.0 Laws and/or regulations that impose quantitative or other regulatory restrictions on a particular transaction, such as licenses or reserve requirements, that completely forbid such economic transactions. 0.5 Laws and/or regulations that impose quantitative or other regulatory restrictions on a particular transaction, such as licenses or reserve requirements that partially forbid such economic transactions. 1.0 Laws and/or regulations requiring the particular transaction to be ap- proved by the authorities or subjecting it to heavy taxes when appli- cable, whether in the form of multiple currency practices or other taxes. 1.5 Laws and/or regulations requiring the particular transaction to be registered, but not necessarily approved, by the authorities and also taxed when applicable. 2.0 No regulations requiring the particular transaction to be approved or registered with authorities and free of taxation when applicable. 217 ANN EX 5 Table A5.1. International Transactions Category Type of transaction Imports and import payments Foreign exchange budget Financing requirements for imports Documentation requirements for release of foreign exchange for imports Import licenses and other nontariff measures Import taxes and/or tariffs State import monopoly Exports and export proceeds Repatriation requirements Financing requirements Documentation requirements Export licenses Export taxes Payments for invisible transactions and current transfers Controls on these payments Proceeds from invisible transactions and current transfers Repatriation requirements Restrictions on use of funds Capital account transactions Capital market securities Money market instruments Collective investment securities Derivatives and other instruments Commercial credits Financial credits Guarantees, sureties, and financial backup facilities Direct investment Liquidation of direct investment Real estate transactions Personal capital movements Commercial banks and other credit institutions Institutional investors Source: IMF (1998). Applying this coding methodology, the estimated financial openness index ranges from 1.93 for Ireland and Luxembourg to 1.12 for Ethiopia (see table A5.2). A more narrowly defined index, which captures the degree of openness to capital account transactions, can be arrived at in a similar manner. The nar- row index uses only the 13 transactions listed in the capital account transac- tions category in table A5. 1. 218 FINANCIAL OPENNESS Table A5.2. Financial Openness Index, Selected Countries, 1997 Open Largely open Partially closed Largely closed Country Index Country Index Country Index Country Index Argentina 1.78 Croatia 1.54 Bahamas, The 1.36 Bangladesh 1.21 Australia 1.77 Ecuador 1.54 Belhze 1.44 Barbados 1.28 Austria 1.92 Honduras 1.56 Benin 1.48 Bhutan 1.19 Bahrain 1.73 Israel 1.59 Botswana 1.48 Brazil 1.19 Belgium 1.88 Mongolia 1.56 Bulgaria 1.46 Ethiopia 1.12 Bolivia 1.79 Philippines 1.59 Burkina Faso 1.49 India 1.20 Canada 1.92 Poland 1.54 Burundi 1.39 Malawi 1.26 Denmark 1.92 Slovak Republic 1.58 Cameroon 1.41 Malaysia 1.34 Egypt, Arab Rep. of 1.81 Slovenia 1.50 Cape Verde 1.39 Morocco 1.27 El Salvador 1.91 Turkey 1.52 Chile 1.43 Pakistan 1.31 Estonia 1.88 China 1.37 Syrian Arab Rep. 1.20 Finland 1.83 Colombia 1.38 France 1.73 Congo, Dem. Rep. of 1.42 Germany 1.84 Costa Rica 1.48 Greece 1.91 Czech Republic 1.48 Guatemala 1.73 Dominican Republic 1.49 Guyana 1.72 Ghana 1.43 Iceland 1.74 Hungary 1.49 Ireland 1.93 Indonesia 1.46 Italy 1.84 Korea, Rep. of 1.42 Jamaica 1.76 Lesotho 1.41 Japan 1.73 Mali 1.49 Kuwait 1.77 Malta 1.40 Latvia 1.88 Moldova 1.46 Lithuania 1.85 Mozambique 1.41 Luxembourg 1.93 Namibia 1.33 Mauritius 1.82 Papua New Guinea 1.36 Mexico 1.69 Romania 1.48 Netherlands 1.87 Russian Federation 1.43 New Zealand 1.90 South Africa 1.44 Nicaragua 1.82 Sri Lanka 1.43 Norway 1.83 Thailand 1.46 Panama 1.90 Tunisia 1.39 Paraguay 1.81 Ukraine 1.36 Peru 1.90 Portugal 1.84 Singapore 1.78 Spain 1.82 Sweden 1.86 Switzerland 1.88 Trinidad and Tobago 1.67 (table continues on following page) 219 ANN EX S Table A5.2 continued Open Largely open Partially closed Largely closed Country Index Country Index Country Index Country Index United Kingdom 1.86 United States 1.85 Uruguay 1.77 Venezuela, RB 1.84 Zambia 1.79 Note: Open: none or minimal regulation for outward and inward transactions and a generally nondiscriminatory environment. Largely open: some regulations are exercised on outward or inward transactions with the need for documentary support but without the need for govemment approval. Partially closed: regulation and govemment approval is required for outward and inward transactions and usually granted. Largely closed: substantial restrictions and government approval is required and seldom granted for outward and inward transactions. Source: Author estimate. Note on Country Vulnerability and Volatility Measures The vulnerability classification is based on our estimates of volatility in pri- vate capital flows, based on the following forecasting equation: (A5.1) KF at = (Xi + P,KFi(,1 + ui where KF. denotes total net private capital flows for country i in year t and U. denotes the error term. Volatility in country i is defined as (A5.2) V. - S(ut) , GDP, 1996 where S(u) is the ordinary least squares estimate of the standard error of the residuals in equation (A5.1) using time series data from 1975 to 1996. See table A5.3 for the index. Note on Gross Domestic Product Gaps The present values of the difference between the potential GDP, extrapo- lated on the basis of historical growth rates of the real economy (1980-96) and the actual or estimated GDP from 1997 through 1999, is calculated as the economic cost due to the financial crisis. The present value was calcu- lated to 1996 values using a real discount rate of 3 percent per year. Ex- pressed as a percentage of the stock of debt in 1996, the estimated costs of the crises were 81 percent for Malaysia, 97 percent for Indonesia, 128 per- cent for Thailand, and 291 percent for Korea. A similar computation for Brazil yields an estimate of 21 percent. Note that for Korea, the debt stock figure used is for 1997. 220 FINANCIAL OPENNESS Table A5.3. Developing Countries Classified by Degree of Volatility to Private Foreign Capital Flows Highly volatile Volatile Moderately volatile Least volatile Country Index Country Index Country Index Country Index Jamaica 3.23 Mexico 1.54 Colombia 0.99 Uganda 0.63 Gabon 2.93 Ecuador 1.50 Tunisia 0.98 Brazil 0.60 Nigeria 2.07 Kenya 1.49 Indonesia 0.95 Paraguay 0.58 Venezuela, RB 2.04 Nicaragua 1.48 Turkey 0.93 China 0.54 Malaysia 1.97 Bolivia 1.44 Argentina 0.91 Sri Lanka 0.46 Jordan 1.82 Chile 1.38 Costa Rica 0.85 Pakistan 0.44 Panama 1.79 Ethiopia 1.31 Uruguay 0.83 Guatemala 0.43 Cameroon 1.60 Philippines 1.27 Egypt, Arab Rep. of 0.74 Dominican Republic 0.39 Zambia 1.59 Honduras 1.04 Tanzania 0.72 India 0.38 Zimbabwe 1.59 Thailand 1.01 Morocco 0.64 El Salvador 0.32 Nepal 0.29 Bangladesh 0.10 Sources: World Bank (1999c); authors' computations. Note on the Binomial Logit Model The binomial logit model is used to estimate the impact of the independent variables on the likelihood that a country would fall in the high democracy-high financial openness category. In this model, the dependent variable is defined by a dichotomous random variable y, which takes the value of 1 if country i belongs to the high democracy-high financial open- ness category, and 0 if it does not. This is given by y =p +e, where p is the probability that a given country belongs in the high democracy-high financial openness category, and specified as p = F(a'x), where x is a vector of independent variables, a is the corresponding vector of coefficients, F(a'x) is the cumulative distribution function, and e, is an error term assumed to follow the Bernoulli distribution. Expressing the probability for the i-th observation via the logit model, we will obtain F(a'x) exp(a 'x) 1+ exp(a'x,) Subsequently, the logit transformation would yield: log Pi = a xx. 221 ANNEX 5 In this chapter we focus on the explanatory variables defined as * Log of total social expenses as a percentage of GDP (average 1990- 96) (x,) * Log of GDP per capita, current U.S. dollars (average 1990-96) (x2). This binomial model was estimated by the maximum likelihood method using cross-country data for a sample of 67 countries for which consistent data on the two explanatory variables were available. Computationally, we obtained the maximum likelihood using the Newton-Raphson algorithm, which utilized the STATA logit procedure. The results were generated after five iterations. The estimated results are reported in the text in table 5.3. The results indicate that both per capita income and the ratio of social ex- penditures to GDP have a statistically significant impact in explaining the likelihood that a country falls into the high-high category. The model also performs well in predicting the percentage of countries that are correctly classified as belonging to the high-democracy high-financial openness group, that is, out of 27 countries in the high-high group, 19 were correctly predicted to be in that group (based on the threshold probability of 0.5), thus producing a 70.37 percent correct classification rate. Summary Statistics for Variables Used in Chapter 5 Several statistical analyses were performed during the course of this chap- ter. Table A5.4 provides summary statistics of the main variables and their sources. Relationships between these variables are explored in figure A5.1 and table A5.5. Notes 1. For some countries where data on certain transactions were unavailable, aver- age values of 1 were assigned. 222 FINANCIAL OPENNESS Table A5.4. Summary Statistics for Selected Industrial and Developing Countries Standard Number of Variables Mean deviation Minimum Maximum countries Political rights 3.42 2.18 1.00 7.00 123 Civil liberties 3.09 1.73 1.00 7.00 138 Capital account openness 15.39 3.62 8.50 22.00 97 Financial openness 38.21 6.48 21.50 48.50 99 Transfer payments 3.09 3.09 0.001 13.84 68 Social expenditure 17.82 11.77 3.35 49.11 51 Trade openness 70.60 45.90 15.21 378.67 141 Income per capita 5,803.00 8,645.00 91.59 37,198.00 68 Note: Political rights, from Freedom in the World, 1998, published by the Freedom House; a country grants its citizens political rights when it permits them to form political parties that represent a significant range of voter choice and whose leaders can openly compete for and be elected to positions of power in government. Civil liberties, from Freedom in the World, 1998, published by the Freedom House; a country upholds its citizens' civil lberties when it respects and protects their religious, ethnic, economic, linguistic, and other rights, including gender and family rights, personal freedoms, and freedoms of the press, belief, and association. Capital account openness is a measure of the degree of controls and/or restrictions applying to only capital account transactions (13 transactions as classified by the IMF AREAER) and are defined in table A5.1. Financial openness is a broader measure incorporating controls and/or restrictions on both current and capital account transactions (see table A5.1). Transfer payments, Government Financial Statistics: Average of central, state, and local expenditure as percentage of GDP. Transfers to other levels of national government, 1991-97; countries with the minimum for this variable are: Chile, Costa Rica, Dominican Republic, Greece, Ireland, Lesotho, Panama, Sri Lanka, and Thailand. Demark has the maximum value for this variable. Social expenditure, Government Financial Statistics and UNESCO: Average of central, state, and local expenditure as percentage of GDP, 1991-97; Pakistan has the minimum for this variable. Denmark has the maximum value for this variable. Trade openness, World Development Indicators: Average trade as percentage of GDP, 1980s. Income per capita, average income per capita, 1990-97; in the sample of countries the minimum is US$91.60 for Mozambique and the maximum of US$37,199 for Switzerland. Sources: Freedom House (1998); IMF (1999, various issues); UNESCO (1998); World Bank (19990. 223 AN N EX 5 Figure A 5.1. Correlations of Financial Openness with Political Rights and Civil Liberties Financial openness r = 0.38 r = 0.51 Political rights Civil liberties r = 0.92 Note: The estimated correlation coefficients are statistically significant at I percent (based on the z-test). Source: Author's computations. Table A5.5. The Relationship among Financial Openness, Democracy, and Social Expenditure Financial Capital Political Civil Trade Social Transfer Income Variables openness openness rights liberties openness expenditure payments per capita Financial openness 1.00 Capital account openness 0.91 1.00 Political rights -0.32* 0.22* 1.00 Civil liberties -0.55* 0.46** 0.91* 1.00 Trade openness 0.18 0.21* -0.06 -0.14 1.00 Social expenditure 0.52* 0.42"" -0.72* -0.74* 0.28** 1.00 Transfer payments 0.23 0.17 -0.19 -0.21 -0.14 0.31** 1.00 Income per capita 0.52* 0.44** -0.54* -0.68* 0.29* 0.67* 0.26 1.00 * Significant at the 0.01 level of confidence. ** Significant at the 0.05 level of confidence. Sources: Based on data from Freedom House (1998); IMF (1999, various years); UNESCO (1998); World Bank (1999f); see table A5.4 for variable explanation. 224 A N N E X 6 GOVERNANCE AND CORRUPTION INDEXES AGGREGATION METHODS, NEW EMPIRICAL MEASURES, AND ECONOMETRIC CHALLENGES R e ecent interest in the consequences of political economy factors, formal and informal institutions, rule of law, legal and judiciary capture, and corruption has been accompanied by a proliferation of data purporting to measure various aspects related to what may be broadly called governance. In this annex we summarize some recent research relating to methodological and empirical results on governance and corruption presented in the text of chapter 6. The first part of the annex summarizes the empirical challenges on governance indicators, while the second part summarizes the research project unbundling the measurement of corruption into administrative bribery, state capture (which includes legal and judiciary capture), and public procurement kickbacks. Defining and Unbundling Governance Inter alia, this report unbundles the concept of governance into six aggre- gate component indicators that were constructed by Kaufmann, Kraay, and Zoido-Lobat6n (KKZ) (1999a,b). Details about these aggregates and the no- tion of governance underlying them are given in the chapter text. The aggregate indicators are based on more than 300 measures pro- duced by 13 different organizations. The sources include published and un- published data from a number of private forecasting and business risk orga- nizations, think tanks, and other NGOs, and the results of surveys carried out by multilateral and other organizations. The database covers 170 coun- tries. Currently, the data are for only one period: 1997-98. The data and 225 ANN EX 6 further details on the econometric methodology are available at http:// www.worldbank.org/wbi/governance. Govemance, or the manner of goveming, encompasses the process of se- lecting, monitoring, and replacing governments, and refers to the govemment's capacity to formulate and implement sound policies and the re- spect of citizens and the state for its institutions. In chapter 6, we presented the following six component indexes of govemance are measured: (a) voice and accountability, (b) political stability, (c) govemment effectiveness, (d) regulatory burden, (e) rule of law, and (f) control of corruption. For each of the six aspects of govemance, a large number of individual indexes from dif- ferent sources were identified as relevant and aggregated to form one of the six composite measures. The aggregation uses an unobserved components model. Advantages that derive from the method include * A large number of single, and rather noisy, indicators is transformed into a smaller number of more reliable aggregate indicators. These aggregates reflect the statistical consensus of many different sources in a rigorous aggregation method that separates signal from noise. As a result, these aggregate indicators are more precise than more conventional indicators. * This method computes statistically sound margins of error around the estimates of governance for individual countries; that is, one can be relatively confident about the degree of uncertainty associ- ated with country-specific estimates of governance. The methodology used emphasizes a limitation of current governance indicators: they are unable to produce precise measures. In view of the mar- gins of error surrounding the estimated measures of governance, small dif- ferences in the estimates will not be statistically or practically significant. It would be misleading to offer very precise rankings of countries according to their level of governance. Instead, broad country groupings, along the lines of a traffic light approach, is statistically defensible and is presented in the chapter text. The above summarizes some of the methodological issues addressed in "Aggregating Govemance Indicators" (KKZ 1999a) and in their interpreta- tion. Furthermore, in this paper KKZ organize the data so that within each one of the six governance clusters, each individual indicator measures a similar underlying basic concept of governance. There are considerable ben- efits from combining these related indicators into an aggregate govemance indicator for each cluster, because (a) aggregate indicators span a much larger set of countries than any individual source, (b) aggregate indicators can provide more precise measures of governance than individual indicators, and (c) it is possible to construct quantitative measures of the precision of 226 GOVERNANCE AND CORRUPTION INDEXES both the aggregate governance estimates for each country, as well as their components. This allows formal testing of hypotheses regarding cross- country differences in governance. For each of these clusters, KKZ combine the component indicators into an aggregate governance indicator using an unobserved components model. This model expresses the observed data in each cluster as a linear function of the unobserved common component of governance, plus a disturbance term capturing perception errors and/or sampling variation in each indica- tor. Estimates are then generated of each of the six governance measures for each country, as well as measures of their precision. The choice of units for governance ensures that the estimates of governance have a mean of zero, a standard deviation of 1, and range from around -2.5 to around 2.5. Higher values correspond to better outcomes. Since the distribution of govemance conditional on the observed data is assumed to be independent across countries, it is possible to make probabilistic statements comparing gover- nance in pairs of countries. It is found that the underlying governance concepts in each cluster are not very precisely estimated, as depicted in figures 6.1 and 6.2 in chapter 6. The rather large size of these confidence intervals has important implica- tions for the use of these aggregate governance indicators. Small differences in point estimates of governance across countries are not statistically sig- nificant. As a result, users of this data should focus on the range of possible governance for each country as summarized in the 90 percent confidence intervals shown in figure 6.1 in chapter 6. For two countries at opposite ends of the scale of governance, whose 90 percent confidence intervals do not overlap, it is reasonable to conclude that there are significant differ- ences in governance. For pairs of countries that are closer together and whose 90 percent confidence intervals overlap, circumspection is in order and seemingly precise comparisons ought to be avoided. Despite the imprecision of these aggregate indicators, they are very useful, for several reasons. First, since each of these aggregate indicators spans a much larger set of countries than any individual indicator, it is possible to make comparisons-however imprecise-across a much larger set of countries than would be possible with any single indicator. Second, each aggregate indicator provides a more precise signal of its corresponding broader govemance con- cept than do any of its individual component indicators, as well as a consistent summary of the available evidence. Third, the measures of precision for each country are useful as well, because they enable formal statistical tests of cross- country differences in governance-instead of arbitrary comparisons. Fourth, it is possible to use information in the estimates of the precision of each aggre- gate to quantify the effect of measurement error in regression analyses that use governance indicators as right-hand-side (independent) variables. 227 ANNEX 6 In the companion paper ("Governance Matters," KKZ 1999b), the au- thors detail all external data sources, describe each individual variable, and analyze the relationship between the governance components and develop- mental variables. The cross-country data indicate a significant simple corre- lation between governance and socioeconomic outcomes (literacy, infant mortality, longevity, income per capita). To explore the effect of governance or socioeconomic variables, con- trolling for other factors, specific econometric tests were done, based on a two-stage least squares regression of a particular socioeconomic dependent variable on a constant and on the governance component, using indicators of colonial heritage as instruments and following the approach of Hall and Jones (1999). The model is well-specified in the sense that (a) the instru- ments have strong predictive power for governance, and (b) the null hy- potheses the these instruments affect incomes only through their effects on governance is not rejected. Concerns about measurement error and omitted variables are also addressed in detail in KKZ (1999b). Measuring and Unbundling Corruption Particular coverage to the challenge of addressing corruption was given in the text of chapter 6, given its importance within governance and the emergence of new empirical findings. Until recently, the measurement of corruption, where done, followed a unidimensional and generalized ap- proach to this complex problem. Recent empirical advances in the study of corruption through improved survey techniques and approaches permit a more in-depth and multifaceted unbundling of corruption. In the chapter's text we reported on the interface between corporate strategies and national governance, and also showed that while on balance administrative bribery does not "pay" for business, "grander" forms of corruption such as state cap- ture (figure 6.11) elicit significant private benefits to the captor firm (figure 6.6 in chapter 6), while resulting in large social costs. We detail below how such unbundling was carried out. The Business Environment and Enterprise Performance Survey: Unbundling Corruption The ability to distinguish between these various manifestations, causes, and consequences of different types of corruption stems from the concerted ef- fort to conceptually and empirically unbundle the problem of corruption initiated within a large-scale survey of transition economies. The Business Environment and Enterprise Performance Survey (BEEPS) was conducted 228 GOVERNANCE AN D CORRU PTION I N DEXES on the basis of face-to-face interviews with firm managers or owners in site visits during June through August 1999 in 22 countries and covered about 3,000 enterprises. In each country, between 125 and 150 firms were interviewed with the ex- ception of three countries where larger samples were used: Poland (250), the Russian Federation (550), and Ukraine (250). The survey questions examine corruption from a number of different angles providing consistency checks on each firm's responses. Moreover, tests were conducted to detect any systematic positive or negative bias among the firms' responses in any given country. In designing the survey, corruption was approached as a multifaceted phenomenon requiring rigorous unbundling, and on such a basis a typology of corruption to distiguish between the different country pattems and con- sequences was arrived at. Particular emphasis was given to three dimensions of corruption: administrative corruption, public procurement corruption, and state capture (Hellman, Jones, and Kaufmann 2000a ["Seize the State, Seize the Day"]; Hellman and others 2000 ["Measuring Governance, Cor- ruption, and State Capture"])-recognizing that different dimensions of corruption might have unique origins and consequences. Typology of Corruption: Definitions Administrative corruption refers to the distorted or arbitrary application and implementation of existing laws, rules, and regulations for illicit private gain by a public office holder, and is subject to a variety of quantitative measures in BEEPS (such as the percentage of administrative bribes paid by the firm as a share of their total revenues). Public procurement corruption, an important dimension of corrupt allocation of public finances and public re- sources, is measured through the percentage bribe fee paid to secure con- tracts. State capture refers to the actions of economic agents or firms both in the public and private sectors to influence the formation and formulation of laws, regulations, decrees, and other govemment policies (that is, the ba- sic rules of the game) to their own advantage-as a result of illegal pay- ments from private agents to public officials. For instance, an influential oligarch at the head of a powerful financial-industrial group may purchase the votes of legislators to erect barriers to entry in the energy sector. Unbundling State Capture and Calculating an Overall Capture Index Within BEEPS, firms were asked about their propensity to purchase legisla- tive capture influence, and they were asked in some detail to report on the 229 ANNEX 6 impact on the firm of different dimensions of state capture in the economy. For a selected group of transition economies (for full details and data, in- cluding measurement of margins of error, see Hellman, Jones, and Kaufmann 2000a; Hellman and others 2000), Table A6.1 presents the vari- ous dimensions of state capture that were measured, as well as the overall state capture index derived through the simple average of all subcompo- nents in the previous columns measuring the effects of the state capture component. In turn, the overall state capture index was used in the second panel of figure 6.6, while the measure of administrative corruption was used in the first panel of that figure. On the basis of the unbundled components of state capture, it is also possible to construct other subindexes of rel- evance. From table A6.1 a judiciary capture index can be calculated, for in- stance, based on the proportion of firms affected by the purchase of crimi- nal and commercial court decisions (columns 4 and 5). The analysis of the causes (including absence of civil liberties and economic reforms) and con- sequences (on output and investment growth and property rights protec- tion) of state capture is carried out through multivariate econometric analysis (including logit and ordinary least squares specifications). The in- depth analysis of the FDI links with state capture, public procurement kick- backs, and other forms of influence (including legal ones) is also based on the data from BEEPS and presented in detail in Hellman, Jones, and Kaufmann (2000b). From Cross-Country Surveys to In-Depth Country Specific Diagnostics For detailed action programming in a country, even much improved cross- country surveys cannot substitute for the need to carry out in-depth coun- try diagnostics on governance and corruption within a particular setting. Such country-specific governance diagnostic tools are discussed in the text of the chapter, where further website references are also provided. 230 GOVERNANCE AND CORRUPTION INDEXES Table A6.1. Percentage of Firms Affected by Different Forms of State Capture, and Overall State Capture Index, Selected Countries, 1999 Corporate purchase of Overall Criminal Commercial state Parliamentary Presidential Central bank courts courts Political party capture legislation decrees influence decisions decisions finance indexb Country (1) (2) (3) (4) (5) (6) (I + + 6) A2erbaijan 41 48 39 44 40 35 41 Bulgaria 28 26 28 28 19 42 28 Croatia 18 24 30 29 29 30 27 Estonia' 14 7 8 8 8 17 10 Georgia 29 24 32 18 20 21 24 Hungarya 12 7 8 5 5 4 7 Latvia 40 49 8 21 26 35 30 Moldova 43 30 40 33 34 42 37 Poland, 13 10 6 12 18 10 12 Romania 22 20 26 14 17 27 21 Russian Federation 35 32 47 24 27 24 32 Slovenia' 8 5 4 6 6 11 7 Ukraine 44 37 37 21 26 29 32 a. The state capture classification is medium for these countries. For all other countries listed the state capture classification is high. b. The state capture index is the simple average of the measured subcomponents in columns 1 through 6. Subgrouping of such components also permits calculations of a judiciary or legal capture index (columns 1, 4, and 5, which under an extended interpretation could also encompass the purchase of presidential decrees in column 2), a stand-alone judiciary capture index (columns 4 and 5), or a legal capture index (columns 1 and 2). Note: Individual estimates subject to margin of error. Such margins of error are significant, thus care ought to be exercised in the use of each individual estimate. Nonetheless, we have tested for country-specific respondent perception bias and did not find it to be significant (see Hellman and others 2000). Sources: Hellman, Jones, and Kaufmann 2000a; see also http://www.worldbank.org/wbi/govemance. 231 R E F E R E N C E S BIBLIOGRAPHY AND REFERENCES The word "processed" describes informally reproduced works that may not be commonly available through libraries. ADB (Asian Development Bank). 1997. Emerging Commodities and Policy Unit, Washington, Asia: Changes and Challenges. Manila. D.C. Processed. Aldeman, Irma. 1975. "Growth, Income Distribution, Alba, Pedro, Amar Bhattacharya, Stijn Claessens, and Equity-Oriented Development Strategies." Swati Ghosh, and Leonardo Hemandez. 1998. World Development 3(2 and 3): 67-76. "The Role of Macroeconomic and Financial Sector Linkages in East Asia's Financial Cri- Adelman, Irma, and Cynthia Taft-Morris. 1967. Soci- S."Ppr pres in the Centre frE- ety, olitis andEconoic Deelopmnt: Asis." Paper presented in the Centre for Eco- ety, Politics and Economic Development; A Quantitative Approach. Baltimore, Maryland: nomic Policy Research/World Bank confer- Th e JohnsAHopproach. Univerityress. Mrln:ence on Financial Crises: Contagion and The Johns Hopkins University Press.MaktVliiy,a,Lodn Market Volatility, May, London. Ades, Alberto, and Rafael Di Tella. 1997. "National Alesina, Alberto, and Dani Rodrik. 1994. "Distributive Champions and Corruption: Some Unpleasant Politics and Economic Growth." Quarterly Interventionist Arithmetic." The Economic Journal of Economics 108: 465-90. Joumal 107: 1023-42. Alesina, Alberto, and Roberto Peortti. 1994. "The Po- _.__ 1999. "Rents, Competition, and Corruption." ltclEooyo rwh rtclSre American Economic Review 80(4). 982-93. litical Economy of Growth: A Critical Survey of the Recent Literature." World Bank Eco- Ahluwalia, Montek S. 1976. "Income Distribution and nomic Review 8(3): 351-71. Development: Some Stylized Facts." American Alesina, Alberto, Vittorio Grilli, and Gian Maria Economic Review 66(2): 128-35. Milesi-Ferretti. 1994. "The Political Economy Akiyama, Takamasa. 1995. "Has Africa Turned the Cor- of Capital Controls." In Leonardo Leiderman ner?" World Bank, International Economics and Assaf Razin, eds., Capital Mobility: The 233 B I B LI OG RAP HY AND RE FE RE C ES Impact on Consumption, Investment, and and Lewis C. Solomon, eds., Economic Policy, Growth. Cambridge, U.K.: Cambridge Uni- Financial Markets, and Economic Growth. Boul- versity Press. der, Colorado; and Oxford, U.K.: Westview Press, in cooperation with the Milken Institute Amin, Sajeda, and Anne R. Pebley. 1994. "Gender In- for Job and Capital Formation. equality within Households: The Impact of a Women's Development Programme in Thirty- Barro, Robert, and J. W. Lee. 1994. "Losers and Win- Six Bangladesh Villages." Bangladesh Develop- ners in Economic Growth." In Proceedings of ment Studies 22(2-3): 121-54. the World Bank Annual Conference on Develop- ment Economics 1993. Washington, D.C.: Ariyoshi, Akira, Kal Habermeier, Bernard Laurens, World Bank. Inci Otker-Robe, Jorge Ivan Canales-Kiljenko, and Andrei Krilenko. 1999. Country Experi- . 1997. "International Measures of Schooling ences with the Use and Liberalization of Capital Years and Schooling Quality." American Eco- Controls. Washington, D.C.: International nomic Review, Papers and Proceedings 86(2): Monetary Fund, Monetary Exchange Affairs 218-23. Department. Barro, Robert, and X. Sala-I-Martin. 1995. Economic Ameson, Richard. 1989. "Equity and Equality of Oppor- Growth. New York: McGraw-Hill. tunity for Welfare." Philosophical Study 56: 77-93. . 1996. "Convergence," Economic Growth: Theory Arrow, K. 1962. "The Economic Implications of and Evidence, vol. 1. International Library of Learning by Doing." Review of Economic Stud- Critical Writings in Economics no. 68. ies 29(June): 155-73. Cheltenham, U.K.: Elgar Reference Collec- tion. Aslund, Anders. 1999. "Why Has Russia's Economic Transformation Been So Arduous?" Paper pre- Baster, Nancy, ed. 1972. Measuring Development: The sented at the Annual Bank Conference on De- Role and Adequacy of Development Indicators. velopment Economics, April 28-30, Washing- London: F. Cass (reprint). Journal of Develop- ton, D.C. ment Studies 8(3) (original). Aviation Week and Space Technology 151(6), August 9, Basu, Kaushik, Garance Genicot, and Joseph E. 1999. Stiglitz. 1999. "Household Labor Supply, Un- employment, and Minimum Wage Legisla- Banerjee, Abhijit. 1999. "Land Reforms: Prospects and tion." Policy Research Working Paper no. Strategies." Paper presented at the Annual WPS 2049. Development Economics, World Bank Conference on Development Economics, Bank, Washington, D.C. April 28-30, World Bank, Washington, D.C. Beason, R., and D. E. Weinstein. 1996. "Growth, Barro, Robert J. 1990. "Government Spending in a Economies of Scale, and Targeting in Japan Simple Model of Endogenous Growth." Joumal (1955-1990)." Review of Economics and Statis- of Political Economy 98: s103-125. tics 78(2): 286-95. . 1993. "Economic Growth, Convergence, and Becker, Gary S. 1964. Human Capital. New York: Co- Government Policies." In Benjamin Zycher lumbia University Press. 234 BIBLIOGRAPHY AND REFERENCES Behrman, Jere, and Nancy Birdsall. 1983. "The Quality Besley, Timothy, and Anne Case. 1994. "Unnatural of Schooling, Quantity Alone Is Misleading." Experiments? Estimating the Incidence of En- American Economic Review 73(5): 928-46. dogenous Policies." Working Paper no. 4956. National Bureau of Economic Research, Cam- Behrman, Jere, and Anil B. Deolalikar. 1988. "Health bridge, Massachusetts. and Nutrition." In Hollis Chenery and T. N. b Srinivasan, eds., Handbook of Development Eco- Bhagwati, Jagdish. 1973. "Education, Class Structure, nomics, vol. 1. Amsterdam: North-Holland. and Income Equality." World Development 1(5): 21-36. Behrman, Jere, and James C. Knowles. 1999. "House- hold Income and Child Schooling in Vietnam." . 1982. "Directly Unproductive Profit-Seeking World Bank Economic Review 13(2): 211-56. Activities." Journal of Political Economy 90(5): 998-1002. Behrman, Jere, Suzanne Duryea, and Miguel Szekely. 1999. "Schooling Investments and Aggregate Bhalla, Surjit S. 2000. "Growth and Poverty in India- Conditions: A Household-Survey-Based Ap- Myth and Reality." Oxus Research and Invest- proach for Latin America and the Caribbean." ment. Processed. Available at http:// Global Research Project Paper. Global Devel- www.oxusresearch.com. opment Network. Processed. Bhatia, Bela, and Jean Dreze. 1998. "For Development Benhabib, Jess, and Mark Spiegel. 1994. "The Role of and Democracy." Frontline. Chennai, India, Human Capital in Economic Development." March 6. An abridged version is available at In Mario Paganetto Baldassarri and Edmond S. http://www.transparency.de/documents/work- Luigi Phelps, eds., International Differences in papers/bhatia-dreze.html. Growth Rates: Market Globalization and Eco- Binswanger, Hans P. 1991. "Brazilian Policies That En- nomic Areas. Central Issues in Contemporary courage Deforestation in the Amazon." World Economic Theory and Practice Series. New Development 19(7): 821-29. York and London: St. Martin's Press and Macmillan Press. Binswanger, Hans P., and Klaus Deininger. 1997. "Explaining Agricultural and Agrarian Poli- Bergstrom, Frederik. 1998. "Capital Subsidies and the cies in Developing Countries." Journal of Eco- Performance of Firms." Working Paper Series nomic Literature XXXV(December): 1958- in Economics and Finance no. 285. Stockholm 2005. School of Economics, Stockholm, Sweden. Binswanger, Hans P., Klaus Deininger, and Gershon Berry, R. A., and W. R. Cline. 1979. Agrarian Structure Feder. 1995. "Power, Distortions, Revolt, and and Productivity in Developing Countries. Balti- Reform in Agricultural Land Relations." In more, Maryland: The Johns Hopkins Univer- Jere Behrman and T. N. Srinivasan, eds., sity Press. Handbook of Development Economics, vol. Illb. Besley, Timothy. 1995. "Property Rights and Invest- Amsterdam: Elsevier Science. ment Incentives: Theory and Evidence from Birdsall, Nancy. 1994. "Govemment, Population, and Ghana." Journal of Political Economy 103(5): Poverty: A Win-Win Tale." In K. L. Kiessling 903-937. and H. Landberg, eds., Population, Economic 235 B I B LI OG RAP HY AND RE FE RE NC ES Development, and the Environment. Oxford, Bonilla-Chacin, Maria, and Jeffrey S. Hammer. 1999. U.K.: Oxford University Press. "Life and Death among the Poorest." Paper Birdsall, Nancy, and Juan Luis Londofio. 1997. "Asset presented at the World Bank Economist Fo- Inequality Matters: An Assessment of the rum, May 3-4, Washington, D.C. World Bank's Approach to Poverty Reduc- Borensztein, E., J. De Gregorio, and J. W. Lee. 1998. tion." American Economic Review 87(2): 32-37. "How Does Foreign Direct Investment Affect . 1998. "No Tradeoff: Efficient Growth via More Economic Growth?" Journal of International Equal Human Capital Accumulation." In Economics 45(1): 115-35. Carol Graham, Nancy Birdsall, and Richard Borensztein, Eduardo, Mohsin S. Khan, Carmen M. Sabot, eds., Beyond Tradeoffs: Market Reform Reinhart, and Peter Wickham. 1994. "The Be- and Equitable Growth in Latin America. Wash- havior of Non-Oil Commodity Prices." Occa- ington D.C.: Brookings Institution Press in as- sional Paper no. 112. Intemational Monetary sociation with the Inter-American Develop- Fund, Washington, D.C. ment Bank. Bouis, Howarth E., Marilou Palabrice-Costello, Orville BIS (Bank for Intemational Settlements). 1997, 1998, Solon, Daniel Westbrook, and Azucena B. various years. Annual Report. Basel, Switzer- Limbo. 1998. "Gender Equality and Invest- land. ments in Adolescents in the Rural Philippines." Bloom, David E., and Adi Brender. 1993. "Labor and Research Report no. 108. Intemational Food the Emerging World Economy." Working Pa- Policy Research Institute, Washington, D.C. per no. 4266. National Bureau of Economic Bourguignon, Francois. 1999. "Crime, Violence, and Research, Cambridge, Massachusetts. Inequitable Development." Paper presented at Bloom, David E., and Richard B. Freeman. 1988. "Eco- the Annual Bank Conference on Develop- nomic Development and the Timing and ment Economics, April 28-30, World Bank, Components of Population Growth." Journal of Washington, D.C. Policy Modeling 10(1): 57-81. Bregman, Arie, Melvyn Fuss, and Haim Regev. 1999. Bloom, David E., and Jeffrey G. Williamson. 1998. "Effects of Capital Subsidization on Productiv- "Demographic Transitions and Economic ity in Israeli Industry." Bank of Israel Economic Miracles in Emerging Asia." World Bank Eco- Review 72: 77-101. nomic Review 12(3): 419-55. Brewer, Dominic J., Cathy Krop, Brian P. Gill, and Bojo, Jan. 1996. "The Costs of Land Degradation in Robert Reichardt. 1999. "Estimating the Cost Sub-Saharan Africa." Ecological Economics of National Class Size Reductions under Dif- 16(2): 161-73. ferent Policy Altematives." Education Evalua- tion and Policy Analysis 21(2): 179-92. Bojo, Jan, and Lisa Segnestam 1999. "Towards a Com- mon Goal: The Experience of NEAPs and Brown, Lester R., Christopher Flavin, and Hilary Their Relationship to NSDSs." Paper submit- French. 1998. "State of the World 1998." In A ted to the Regional Consultative Meeting on Worldwatch Institute Report on Progress Toward Sustainable Development in Africa, Septem- a Sustainable Society. New York: W. W. ber 7-9, Abidjan. Norton. 236 B I B LI OG RAP HY AND RE FE RE N CE S Brunetti, Aymo. 1997a. "Political Variables in Cross- York and London: St. Martin's Press and Country Growth Analysis." Journal of Eco- Macmillan Press in association with the Inter- nomic Surveys 11(2): 163-90. national Economic Association. _.___ 1997b. Politics and Economic Growth: A Cross- Caprio, Gerard, and Patrick Honohan. 1999. "Restor- Country Data Perspective. Paris: Organisation ing Banking Stability: Beyond Supervised for Economic Co-operation and Development, Capital Requirements." Journal of Economic Development Center Studies. Perspectives 13(4): 43-64. Brunetti, Aymo, Gregory Kisunko, and Beatrice Card, David, and Alan B. Krueger. 1992. "School Weder. 1997. "Credibility of Rules and Eco- Quality and Black-White Relative Earnings: A nomic Growth." Policy Research Working Pa- Direct Assessment." Quarterly Joumal of Eco- per no. 1760. World Bank, Policy Research nomics 107(1): 151-200. Department, Washington, D.C. Castaneda, Beatriz. 1997. " An Index of Sustainable Eco- Bryant, Ralph C., and Edith Hodgkinson. 1989. "Prob- nomic Welfare for Chile." Masters Thesis. Uni- lems of International Cooperation." In Can versity of Maryland, College Park, Maryland. Nations Agree? Issues in International Economic Cooperation. Washington, D.C.: The Castro, R., F. Tattenbach, N. Olson, and Luis Gamez. Brookings Institution. 1997. "The Costa Rican Experience with Mar- ket Instruments to Mitigate Climate Change Buckley, Robert. 1999. "1998 Annual Review of Devel- and Conserve Biodiversity." Paper presented at opment Effectiveness." World Bank, Operations the Global Conference on Knowledge for De- Evaluation Department, Washington, D.C. velopment in the Information Age, June 24, Burgess, Robin. 2000. "Land Distribution and Welfare World Bank, Toronto. in Rural China." Paper presented at the Annual Chenery, Hollis, Montek S. Ahluwalia, Clive Bell, Bank Conference on Development Economics, John H. Duloy, and Richard Jolly. 1974. Redis- April 20, World Bank, Washington, D.C. tribution with Growth. New York: Oxford Uni- Calvo, Guillermo A. 1999. "Contagion in Emerging versity Press. Markets: When Wall Street Is a Carrier." Uni- Chong, Albert, and Cesar Calder6n. 1997a. "Empirical versity of Maryland, College Park. Processed. Tests on Causality and Feedback between In- Calvo, Guillermo A., and Enrique G. Mendoza. 1996. stitutional Measures and Economic Growth." "Petty Crime and Cruel Punishment: Lessons World Bank, Washington, D.C. Processed. from the Mexican Debacle." American Economic . 1997b. "Institutional Change and Poverty, or Review, Papers and Proceedings 86(2): 170-75. Why Is it Worth it to Reform the State?" Calvo, Guillermo A., Leonardo Leiderman, and World Bank, Washington, D.C. Processed. Carmen M. Reinhart. 1994. "Capital Inflows . 1998. "Institutional Efficiency and Income In- to Latin America: The 1970s and 1990s." In equality: Cross-Country Empirical Evidence." Edmar L. Bacha, ed., Economics in a Changing World Bank, Washington, D.C. World: Proceedings of the Tenth World Congress of the International Economic Association, vol. 4, Chomitz, Kenneth M., and David A. Gray. 1996. Development, Trade, and the Environment. New "Roads, Land Use, and Deforestation: A Spatial 237 B I B LI OG RAP 1HY AND RE F ERE NC ES Model Applied to Belize." Working Paper no. 3. Collins, Susan M., and Barry P. Bosworth. 1996. "Eco- Research Project on Social and Environmental nomic Growth in East Asia: Accumulation Consequences of Growth-Oriented Policies. Versus Assimilation." Brookings Papers on Eco- World Bank, Policy Research Department, nomic Activity 2: 135-91. Washington, D.C. Washington, D.C. Commander, Simon, Mark Dutz, and Nicholas Stem. Chomitz, Kenneth M., Luis Constantino, and Esteban 1999. "Restructuring in Transition Economies: Brenes. 1998. "Financing Environmental Ser- Ownership, Competition, and Regulation." vices: The Costa Rican Experience and its Impli- Paper presented at the Annual Bank Confer- cations." World Bank, Development Economics ence on Development Economics, April 28- Research Group, Washington D.C. Processed. 30, Washington, D.C. Claessens, Stijn, Simeon Djankov, and Daniela Cooper, Richard. 1998. "Should Capital-Account Klingebiel. 1999. "Financial Restructuring in Convertibility Be a World Objective?" In Es- East Asia: Halfway There?" Financial Sector Dis- says in International Finance Series 207. cussion Paper 3. World Bank, Washington, D.C. Princeton, New Jersey: Princeton University, Cleaver, Kevin, and Gotz Schreiber. 1994. Reversing the Department of Economics, Intemational Fi- nance Section. Spiral: The Population, Environment, and Agricul- ture Nexus in Sub-Saharan Africa. Directions in Corbo, Vittorio, Stanley Fischer, and Steven B. Webb. Development. World Bank, Washington, D.C. 1992. Adjustment Lending Revisited: Policies to Restore Growth. Washington, D.C.: World COCHILCO. 2000. Annual Report of COCH-IILCO. Bank. Report from the Senate of Chile. Comia, Giovanni A. 1999. "Rising Income Inequality Cohen, G. A. 1989. "On the Currency of Egalitarian and Poverty Reduction: Are They Compatible?" JUNU/WIDER Research Project. Available at Collier, Paul. 1999. "On the Economic Consequences http://www.wider.unu.edu/wiid/wiid.htm. of Civil War." Oxford Economic Papers Corsetti, Giancarlo, Paolo Pesenti, and Nouriel 51 (January): 168-83. Roubini. 1998. "What Caused the Asian Cur- Collier, Paul, and Anke Hoeffler. 1998. "On the Eco- rency and Financial Crisis?" New York Uni- nomic Causes of Civil Wars." Oxford Economic versity, New York. Processed. Papers 50(October): 563-73. Cropper, Maureen, and Charles Griffiths. 1994. The Collier, Paul, David Dollar, and Nicholas Stem. 2000. Interaction of Population Growth and Envi- "Fifty Years of Development." World Bank. ronmental Quality. American Economic Review Washington, D.C. Processed. 84: 250-54. Collins, Chuck. 1996. "Aid to Dependent Corpora- Cropper, Maureen, Charles Griffiths, and tions." In Marc Breslow, Miller Reed, and Muthukumara Mani. 1997. "Roads, Population Betsy Reed, eds., Decoding the Contract: Pro- Pressures, and Deforestation in Thailand, gressive Perspectives on Current Economic Policy 1976-1989." Policy Research Working Paper Debates. Somerville, Massachusetts: Economic no. 1726. World Bank, Policy Research De- Affairs Bureau. partment, Washington, D.C. 238 BIBLIOGRAPHY AND REFERENCES Dailami, Mansoor. 1998. "Euphoria and Panic: Devel- De Gregorio, Jose. 1992. "Growth in Latin America." oping Countries' Relationship to Private Fi- Journal of Development Economics 31(1): 59-84. nance." EDI Forum 3(2): 1-3, 6. De Long, J. Bradford, and Lawrence Summers. 1991. . 2000. "Financial Openness, Democracy, and "Equipment Investment and Economic Redistributive Policy." Policy Research Work- Growth." Quarterly Journal of Economics ing Paper no. 2372. World Bank, Washington, 106(2): 445-502. D.C. de Moor, Andre, P. G., and Peter Calamai. 1997. Sub- Dailami, Mansoor, and Nadeem ul Haque. 1998. sidizing Unsustainable Development: Undermin- "What Macroeconomic Policies Are 'Sound?"' ing the Earth with Public Funds. San J6se, Costa Policy Research Working Paper no. 1995. Rica: Institute for Research on Public Expen- World Bank, Washington, D.C. diture and The Earth Council. Daly, H. E. 1997. "Georgescu-Roegen Versus Solow/ Deaton, Angus, 1997. The Analysis of Household Sur- Stiglitz." Ecological Economics 22(3): 261-66. veys: A Microeconometric Approach to Develop- ment Policy. Baltimore, Maryland: The Johns Dasgupta, Partha. 1990a. "Well-Being and the Extent m of Its Realization in Developing Countries." Hopkins University Press in association with EconomicJournal 100(4, supplement): 1-32. the World Bank. _ 1990b. "Well-Being in Poor Countries." Eco- Deininger, Klaus. 1999. "Making Negotiated Land Re- nomic and Political Weekly (India) 25: 1713-20. form Work: Initial Experience from Colombia, Brazil, and South Africa." Policy Research _.___ 1993. An Inquiry into Well-Being and Destitution. Working Paper no. 2040. World Bank, Policy New York: Oxford University Press. Research Department, Washington, D.C. . 1995. "Population, Poverty, and the Local En- Deininger, Klaus, and Bart Minten. 1996. "Determi- vironment." Scientific American (February): nants of Forest Cover and the Economics of 40-45. Protection: An Application to Mexico." Dasgupta, Partha, and Karl-Goran Maler. 1994. "Pov- Working Paper no. 10. Research Project on erty, Institutions, and the Environmental Re- Social and Environmental Consequences of source Base." Environment Paper no. 9. World Growth-Oriented Policies. World Bank, Policy Research Department, Washington, D.C. Bank, Environment Department, Washington, D.C. Deininger, Klaus, and P. Olinto. 1999. "Asset Distribu- Dasgupta, Partha, and Ismail Serageldin. 1999. Social tion Inequality and Growth." World Bank, Capital: A Multifaceted Perspective. Washing- D.C. Processed. ton, D.C.: World Bank. Dasgupta, Susmita, Ashoka Mody, Subhendu Roy, and Deininger, Klaus, and Lyn Squire. 1996. "A New David Wheeler. 1995. "Environmental Regu- Dataset Measuring Income Inequality." World lation and Development: A Cross-Country Bank Economic Review 10(3): 565-91. Empirical Analysis." Policy Research Working _ . 1998. "New Ways of Looking at Old Issues: In- Paper no. 1488. World Bank, Policy Research equality and Growth." Journal of Development Department, Washington, D.C. Economics 57(2): 259-87. 239 B I B LI OG RAP HY AND RE FE RE NCE S Demirguic-Kunt, Asli, and Enrica Detragiache. 1998. Drage, John, and Fiona Mann. 1999. "Improving the "Financial Liberalization and Financial Fragil- Stability of the International Financial Sys- ity." Policy Research Working Paper no. 1917. tem." In Financial Stability Review, issue 6. World Bank, Development Research Group; Bank of England. and International Monetary Fund, Research Drewnowski, Jan F., and Wolf Scott. 1966. "The Level of Living Index." Report no. 4. United Na- Denison, E. F. 1962. Sources of Economic Growth in the tions Research Institute for Social Develop- United States and the Alternative before Us. New ment, Geneva. York: Committee for Economic Development. Dr6ze, Jean, and Amartya Sen. 1995. Hunger and Public . 1967. Why Growth Rates Differ: Post-War Expe- Action. WIDER Studies in Development Eco- rience in Nine Western Countries. Washington, nomics. Oxford, U.K.: Clarendon Press. D.C.: The Brookings Institution. Dubey, Ashutosh, and Elizabeth M. King. 1996. "A Devarajan, Shantayanan, and Jeffrey S. Hammer. New Cross-Country Education Stock Series 1998. "Risk Reduction and Public Spending." Differentiated by Age and Sex." Journal of Edu- Policy Research Working Paper no. 1869. cational Planning and Administration 11(1): 5-24. World Bank, Public Economics, Development Duraisamy, P., Estelle James, Julia Lane, and Jee-Peng Research Group, Washington, D.C. Tan. 1998. "Is There a Quantity-Quality Diewert, Walter Erwin. 1986. The Measurement of the Tradeoff as Enrollments Increase? Evidence from Economic Benefits of Infrastructure Services. Ber- Tamil Nadu, India." Policy Research Working lin and New York: Springer-Verlag. Paper no. 1768. World Bank, Washington, D.C. Dikhanov, Yuri, and Michael P. Ward. 2000. "Towards Dworkin, Ronald. 1981. "What Is Equity: Part 2. a Better Understanding of the Global Distribu- Equality of Resources." Philosophy & Public Af- tion of Income." World Bank, Development fairs 10: 283-345. Economics Data Group, Washington, D.C. Promcs D Easterly, William. 1998. "The Joys and Sorrows of lProcessed. Openness: A Review Essay." Paper prepared Diwan, Ishac. 1999. "Labor Shares and Financial Cri- for the Seminar on Economic Growth and Its ses." Working Paper. World Bank Institute, Determinants, March 23-24, Ministry for De- Washington, D.C. velopment Cooperation, The Hague, The Netherlands. Dixon, John A., and Paul Sherman. 1990. Economics of Protected Areas: A New Look at Benefits and _. 1999a. "Life during Growth." Journal of Eco- Costs. Washington, D.C.: Island Press. nomic Growth 4(3): 239-79. Dollar, David, and Aart Kraay. 2000. "Growth Is Good _. 1999b. "The Lost Decades: Explaining Devel- for the Poor." Working paper. World Bank, oping Countries' Stagnation 1980-1998." De- Policy Research Department, Washington, D.C. velopment Economics Research Group, World Dooley, Michael. 1996. "A Survey of Literature on Con- Bank, Washington, D.C. trols over Intemational Capital Transactions." __. 1999c. "The Ghost of Financing Gap: Testing IMF Staff Papers 43(December): 639-87. the Growth Model Used in the Intemational 240 BIBLIOGRAPHY AND REFERENCES Financial Institutions." Journal of Development ages in SE Asia." Economy and Environment Economics 60(2): 423-38. Program for Southeast Asia, Singapore. Easterly, William, and Aart Kraay. 1999. "Small Eichengreen, Barry. 1996. Globalizing Capital: A History States, Small Problems?" Policy Research of the International Monetary System. Princeton, Working Paper no. 2139. World Bank, Devel- New Jersey: Princeton University Press. opment Research Group, Macroeconomics amnd th, WeseaschinGton, Dac.oeconomicsEichengreen, Barry, and Michael Mussa, eds. 1998. "Capital Account Liberalization: Theoretical Easterly, William, and Ross Levine. 2000. "It's Not and Practical Aspects." Occasional paper 172. Factor Accumulation: Stylized Facts and Intemational Monetary Fund, Washington, D.C. Growth Models." Working Paper. World Bank, Washington, D.C. Available at http:// Ehrlich, Paul R. 1968. The Population Bomb. New www.worldbank.org/html/prdmg/grthweb/ pdfiles/fact3.pdf. Ekbom, Anders, and Jan Bojo. 1999. "Poverty and Envi- Easterly, William, and Hairong Yu. 2000. "Global De- ronment: Evidence of Links and Integration into velopment Network Growth Database." Avail- the Country Assistance Strategy Process." Dis- able at http://www.worldbank.org/research/ cussion Paper no. 4. World Bank, Africa Region, growth/GDNdata.htm. Environment Group, Washington, D.C. Easterly, William, Roumeen Islam, and Joseph E. Stiglitz. Esrey, Steven A., James Potash, Leslie Roberts, and 1999. "Shaken and Stirred: Volatility and Mac- Clive Shiff. 1990. "Health Benefits from Im- roeconomic Paradigms for Rich and Poor Coun- provements in Water Supply and Sanitation: tries." Michael Bruno Memorial Lecture, Xllth Survey and Analysis of the Literature on Se- World Congress of the International Economics lected Diseases." WASH Technical Report no. Association, August 27, Buenos Aires. 66. U.S. Agency for International Develop- ment, Washington, D.C. Easterly, William, Michael Kremer, Lant Pritchett, and Larry Summers. 1993. "Good Policy or European Bank for Reconstruction and Development. Good Luck? Country Growth Performance 1999. Transition Report: Ten Years of Transi- and Temporary Shocks." Journal of Monetary tion. London. Economnics 32: 459-83. Evans, P., ed. 1996. "Government Action, Social Eastwood, Robert, and Michael Lipton. 1999. "The Capital, and Development across the Public- Impact of Changes in Human Fertility on Pov- Private Divide." World Development 24(6, spe- erty." Journal of Development Studies 36(1): 1- cial section): 1033-1132. 30. Fakin, Barbara. 1995. "Investment Subsidies during Edwards, Sebastian. 1999. "How Effective Are Capital Transition." Eastern European Economics Controls?" Journal of Economic Perspectives 33(5): 62-75. 13(4): 65-84. Fedderke, Johannes, and Robert Klitgaard. 1998. "Eco- EEPSEA (Economy and Environment Program for nomic Growth and Social Indicators: An Ex- South East Asia). 1998. "Interim Results of a ploratory Analysis." Economic Development and Study on the Economic Value of Haze Dam- Cultural Change 46(April): 455-89. 241 B I B LI OG RAl' HY AND RE FE RE NC ES Feder, Gershon. 1987. "Land Ownership Security and Fischer, Stanley. 1993. "The Role of Macroeconomic Farm Productivity: Evidence from Rural Thai- Factors in Growth." Journal of Monetary Eco- land."Journal of Development Studies 24(1): 16-30. nomics 32(3): 485-512. . 1993. "The Economics of Land and Titling in _ . 1998. "Capital Account Liberalization and the Thailand." In Karla Hoff, Avishay Braverman, Role of the IMF." In Essays in International Fi- and Joseph Stiglitz, eds., The Economics of Ru- nance Series, 207. Princeton, New Jersey: ral Organizations: Theory, Practice, and Policy. Princeton University, Department of Econom- New York: Oxford University Press. ics, Intemational Finance Section. Ferreira, Francisco H. G., and Julie A. Litchfield. 1999. Fischer, Stanley, Richard N. Cooper, Rudiger "Calm after the Storms: Income Distribution Dombusch, Peter M. Garber, Carlos Massad, and Welfare in Chile, 1987-94." World Bank Jaques J. Polak, Dani Rodrik, and Savak S. Economic Review 13(3): 509-38. Tarapore. 1998. Should the IMF Pursue Capital- Account Convertibility? Ffrench-Davis, Ricardo, Patricio Leiva, and Roberto Madrid. 1992. "Liberalizacion Comercial y Fishlow, Albert. 1995. "Inequality, Poverty, and Crecimiento: La Experiencia de Chile, 1973- Growth; Where Do We Stand?" In Michael 89." Pensamiento Iberoamericano 21: 33-55. Bruno and Boris Pleskovic, eds., Annual World Bank Conference on Development Economics Filmer, Deon, and Lant Pritchett. 1999a. "Educational 1995. Washington, D.C.: World Bank. Enrollment and Attainment in India: House- hold Wealth, Gender, Village, and State Ef- Fisman, Raymond, and Roberta Gatti. 2000. "Decen- fects." Journal of Educational Planning and Ad- tralization and Corruption: Evidence across ministration 13(2): 135-63. Countries." Policy Research Working Paper no. 2290. World Bank, Development Research _.___ 1999b. "The Effect of Household Wealth on Group, Washington, D.C. Available at Educational Attainment: Evidence from 35 www.wblnO0 18.worldbank.org/research/ Countries." Population and Development Review workpapers.nsf/. 25(1): 85-120. Fisman, Raymond, and Jakob Svensson. 1999. "The Ef- . 1999c. "The Impact of Public Spending on fects of Corruption and Taxation on Growth: Health: Does Money Matter?" Social Science Firm-Level Evidence." Processed. and Medicine 49(10): 1309-23. Foster, Andrew, and Mark R. Rosenzweig. 1995. "Leam- Filmer, Deon, Jeffrey Hammer, and Lant Pritchett. ing by Doing and Leaming from Others: Human Forthcoming. "Health Policy in Poor Coun- Capital and Technical Change in Agriculture." tries: Weak Links in the Chain." World Bank Journal of Polical Economy 103: 1176-1209. Economic Observer. Foumier, Gary, and David Rasmussen. 1986. "Targeted Fine, B., and K. Fine. 1974a. "Social Choice and Indi- Capital Subsidies and Economic Welfare." vidual Rankings L." Review of Economic Studies Cato Journal 6(1): 295-312. 41(3): 303-322 Fox, Jonathan. 1997. "The World Bank and Social Capi- _.___ 1974b. "Social Choice and Individual Rankings tal: Contesting the Concept in Practice." Journal II." Review of Economic Studies 41(4): 459-75. of International Development 9(7): 963-71. 242 BIBLIOGRAPHY AND REFERENCES Frank, Robert. 1998. Luxury Fever: Why Money Fails to GEF (Global Environmental Facility). 1998. Valuing Satisfy in an Era of Excess. New York: Free Press. the Global Environment: Actions and Invest- ments for a 21st Century, vol. 1. Washington, Freedom House. 1998. Freedom in the World: Political D.C. Rights and Civil Liberties. New York. Gilson, Stuart. 1995. "UAL Corporation." Case Study Furman, Jason, and Joseph Stiglitz. 1998. "Economic 9-295-130. Harvard University Business Crises: Evidence and Insights from East Asia." School, C ard Masachsetts. Brookings Papers on Economic Activity (2nd is- sue for 1998): 1-114. Gold, Joseph. 1977. International Capital Movements under the Law of the International Monetary Furtado, Jose, Nalin Kishor, G. V. Rao, and Catherine Fund. Washington D.C.: International Mon- Wood. 1999. Global Climate Change and etary Fund. Biodiversity: Challenges for the Future and the Way Ahead. Washington D.C.: World Bank Institute. Gonzalez de Asis, Maria. 2000. "Reducing Corruption: Lessons from Venezuela." PREM Note no. 39. Galeotti, Marzio, and Alessandro Lanza. 1999. "Des- World Bank, Washington, D.C. perately Seeking (Environmental) Kuznets." Paper presented at the 1998 World Congress Goodman, L. A., and H. Markowitz. 1952. "Social of Environment and Resource Economists, Welfare Functions Based on Individual June 25-27, Venice, Italy. Rankings." American Journal of Sociology 58. Gallego, Francisco, Leonardo Hemandez, and Klaus Gray, Cheryl W., and Daniel Kaufmann. 1998. "Cor- Schmidt-Hebbel. 1999. "Capital Controls in ruption and Development." Finance and Devel- Chile: Effective? Efficient?" Working Paper. opment 35(1): 7-10. Central Bank of Chile, Santiago. Greaney, Vincent, and Thomas Kellaghan. 1996. Galor, Oded, and J. Zeira. 1993. "Income Distribution "Monitoring the Learing Outcomes of Educa- and Macroeconomics." Review of Economic tion Systems." Direction in Development Pa- Studies 60: 35-52. pers. World Bank, Washington, D.C. Griliches, Zvi. 1971. Price Indexes and Quality Changes: Gandhi, V., D. Gray, and R. McMorran. 1997. "A Studies in New Methods of Measurement. Cam- Comprehensive Approach to Domestic Re- bries inssachuetts: Mearement. Cam- source Mobilization for Sustainable Develop- bridge, Massachusetts: Harvard University Press. ment." Paper presented to the Fourth Expert . 1997. "Education, Human Capital, and Group Meeting on Financial Issues of Agenda Growth: A Personal Perspective." Journal of 21, January 6-8, Santiago, Chile, United Na- Labor Economics 15(1, part 2): s330-42. tions, Department for Policy Coordination and Sutial eeopet e ok Grossman, Gene M., and Elhanan Helpman. 1989. Sustainabl Developent,Ne"Growth and Welfare in a Small Open Garrett, Geoffrey. 1998. "Global Markets and National Economy." Working Paper no. 2970. National Politics: Collision Course or Virtuous Circle?" Bureau of Economic Research. International Organization 52(4): 787-824. 1990. "Comparative Advantage and Long Run Gazetta Mercantil. 1999. An international weekly. New Growth." American Economic Review, 80: 796- York, May 21. 815. 243 B I B LI OG RAP HY AND REER: ELR NC ES Grossman, Gene M., and A. B. Krueger. 1995. "Eco- Hanushek, Eric A., and Dongwook Kim. 1995. nomic Growth and the Environment." Quar- "Schooling, Labor Force Quality, and Eco- terly Journal of Economics 112: 353-78. nomic Growth." Research Working Paper no. 5399. National Bureau of Economic Research, Gulati, Ashok, and Sudha Narayanan. 2000. "Demystifying Fertilizer and Power Subsidies Cambridge, Massachusetts. in India." Economic and Political Weekly, March Harr, Jonathan. 1995. A Civil Action. New York: Vin- 4, pp. 784-94. tage Books. Gupta, Sanjeev, Hamid Davoodi, and Rosa Alonso- Harris, Richard 1. D. 1991. "The Employment Cre- Tenne. 1998. "Does Corruption Affect Income ation Effects of Factor Subsidies: Some Esti- Inequality and Poverty?" Working Paper no. mates for Northem Ireland Manufacturing In- WP/98/76. Intemational Monetary Fund, Fis- dustry, 1955-1983." Journal of Regional Science cal Affairs Department, Washington, D.C. 31(1): 49-64. Haldane, Andrew. 1999. "Private Sector Involvement Harriss, John, ed. 1997. "Policy Arena: 'Missing Link' in Financial Crisis." Financial Stability Review or Analytically Missing? The Concept of So- (November) Bank of England. cial Capital." Journal of International Develop- Hall, Robert E., and Charles I. Jones. 1999. "Why Do ment 9(7, special section): 919-71. Some Countries Produce So Much More Out- Harsanyi, J.C. 1953. "Cardinal Utility in Welfare Eco- put Per Worker Than Others?" Quarterly Jour- nomics and in the Theory of Risk-Taking." nal of Economics 114(1): 83-116. Journal of Political Economy 61: 434-35. Hamilton, Kirk, and Ernst Lutz. 1996. "Green Na- Helleiner, Eric. 1994. States and the Reemergence of Glo- tional Accounts: Policy Uses and Empirical bal Reforn: From Brettan Woods to the 1990s. Experience." Environment Department Paper Ithaca, New York: Comell University Press. no. 39. World Bank, Environment Depart- Hellman, Joel, Geraint Jones, and Daniel Kaufmann. ment, WVashington, D.C. 2000a. "Seize the State, Seize the Day: An Hammer, Jeffrey S., and Sudhir Shetty. 1995. "East Empirical Analysis of State Capture and Cor- Asia's Environment: Principles and Priorities ruption in Transition Economies." Paper pre- for Action." World Bank Discussion Paper no. sented at the Annual Bank Conference on De- 287. World Bank, Washington D.C. velopment Economics, April 18-20, Washington, D.C. Available at http:// Hanrahan, David, David Wheeler, Michelle Keene, W Iww tondba k.or Avbilgove ancep. and David Shaman. 1998. "Developing Part- nerships for Effective Pollution Management." . 2000b. "Far from Home: Do Transnationals Im- Environment Matters Annual Review(Fall): port Better Governance in the Capture 62-65. Economy." World Bank, Washington, D.C. Hanushek, Eric. 1995. "Interpreting the Recent Re- Processed. Available at http:// search on Schooling in Developing Coun- www.worldbank.org/wbi/govemance/. tries." World Bank Research Observer Hellman, Joel S., Geraint Jones, Daniel Kaufmann, 10(2):227-46. and Mark Schankerman. 2000. "Measuring 244 B I BL IO G RAPHY AND RE FE RE NC ES Govemance, Corruption, and State Capture: _. 1981. Essays in Trespassing: Economics to Politics How Firms and Bureaucrats Shape the Busi- and Beyond, Cambridge, U.K.; and New York: ness Environment in Transition." Policy Re- Cambridge University Press. search Working Paper no. 2313. World Bank, Hoeffler, Anke. 1997. "The Augmented Solow Model Washington, D.C. and the African Growth Debate." University Hernandez, Leonardo, and Klaus Schmidt-Hebbel. of Oxford, U.K. Processed. 1999. "Capital Controls in Chile: Effective? Holtz-Eakin, Douglas, and Thomas M. Selden. 1995. Efficient? Endurable?" Paper presented at the "Stoking the Fires? CO, Emissions and Economic World Bank Conference on Capital Flows, Fi- Growth."Journal of Public Economics 57: 85-101. nancial Crisis and Policies, April 15-16, Washington, D.C. Hoy, Michael, and Emmanuel Jimenez. 1997. "The Im- pact of the Urban Environment of Incomplete Herrera, Alejandra. 1992. "The Privatization of the Property Rights." Working Paper no. 14. Re- Argentine Telephone System." CEPAL Review search Project on Social and Environmental 47: 149-61. Consequences of Growth-Oriented Policies. Hettige, Hemamala, Muthukumara Mani, and David World Bank, Policy Research Department, Wheeler. 1998. "Industrial Pollution in Eco- Washington, D.C. nomic Development (Kuznets Revisited)." Hughes-Hallet, A. J. 1989. "Econometrics and the Policy Research Working Paper no. 1876. Theory of Economic Policy: The Tinbergen- World Bank, Development Research Group, Theil Contributions 40 Years on." Oxford Eco- Washington, D.C. nomic Papers 41(January): 189-214. Hicks, Norman, and Paul Streeten. 1979. "Indicators Huntington, Samuel P. 1964. "Modernization and of Development: The Search for a Basic Needs Corruption." In Arnold J. Heidenheimer, ed., Yardstick." World Development 7(6): 567-80. Political Corruption: Readings in Comparative Hill, Kenneth, Rohini Pande, Mary Mahy, and Gareth Analysis. New York: Holt Reinehart. Jones. 1999. Trends in Child Mortality in the De- _. 1968. Political Order in Changing Societies. New veloping World: 1960-96. New York: United Haven: Yale University Press. Nations Children's Fund. IDB (Inter-American Development Bank). 1998. Fac- Hindriks, Jean, Michael Keen, and Abhinay Muthoo. ing Up to Inequality in Latin America. Economic 1999. "Corruption, Extortion, and Evasion." and Social Progress in Latin America. 1998-1999 Journal of Public Economics 74(3): 395-430. Report. Baltimore, Maryland: The Johns Hirschman, Albert 0. 1958. The Strategy of Economic Hopkins University Press. Development. New Haven, Connecticut: Yale IMF (International Monetary Fund). 1998. "Fiscal Re- University Press. forms in Low-Income Countries: Experiences . 1970. Exit, Voice, and Loyalty: Responses to De- under IMF-Supported Programs." Occasional cline in Firms, Organizations, and States. Cam- Paper no. 160. Washington, D.C. bridge, Massachusetts: Harvard University _ . 1999. Government Financial Statistics. Washing- Press. ton, D.C. 245 BIBLIOGRAI'HY AND REFERENCES . Various issues. Exchange Arrangements and Ex- Jimenez, Emmanuel, Bernard Kugler, and Robin Horn. change Restrictions: Annual Report. Washing- 1986. "An Economic Evaluation of a National ton, D.C. Job Training System: Columbia's Servicio India Today. 1999. "The Poisoning of India." Special Nacional de Aprendizaje (SENA)." EDT24. Indi Tody. 199. The oisningof Idia. SpeialWorld Bank, Education and Training Depart- collectors issue, January (Living Media India, New Delhi). ment, Washington, D.C. Johnson, Simon, Daniel Kaufmann, and Andrei International Country Risk Guide. 1982-95. Computer Shleifer. 1997. "The Unofficial Economy in file. Syracruse, New York: PRS Group. Avail- Transition." Brookings Papers on Economic Ac- able to World Bank and International Mon- ti.VIty (2). Washington, D.C. etary Fund staff at http://jolis. worldbankimflib.org/nldbs.htm Johnson, Simon, Daniel Kaufmann, and Pablo Zoido- Lobat6n. 1998. "Regulatory Discretion and the International Finance Corporation. Various years. Emerg- Unofficial Economy." American Economic Re- ing Stock Markets Factbook. Washington, D.C. view 88(2): 387-92. Isham, Jonathan, and Daniel Kaufmann. 1999. "The Jones, Charles 1. 1995. "Time Series Tests of Endog- Forgotten Rationale for Policy Reform: The enous Growth Models." Quarterly Journal of Productivity of Investment Projects." Quarterly Economics 110(2): 495-525. Journal of Economics 114(February): 149-84. Kakwani, Nanak. 1993. "Performance in Living Stan- Isham, Jonathan, Daniel Kaufmann, and Lant H. dards: An Intemational Comparison." Journal Pritchett. 1997. "Civil Liberties, Democracy, of Development Economics (Netherlands) and the Performance of Government 41(August): 307-336. Projects." World Bank Economic Review 11(2): 219-42. Kaminsky, Graciela, and Carmen Reinhart. 1999. "Twin Crises: The Causes of Banking and Bal- Islam, Nazrul. 1995. "Growth Empiries: A Panel Data ance of Payments Problems." American Eco- Approach." The Quarterly Journal of Economics nomic Review 89(3): 473-500. 110: 1127-70. Kanbur, Ravi. 1999. "Why Is Inequality Back on the James, Estelle, Elizabeth King, and Ace Suryadi. 1996. Agenda?" Paper presented at the Annual "Finance, Management, and Cost of Public Bank Conference on Development Econom- and Private Schools in Indonesia." Economics ics, April 28-30, World Bank, Washington of Education Review 15(4): 387-98. D.C. Jimenez, Emmanuel, and Vicente Paqueo. 1996. "Do . 2000. "Income Distribution and Development." Local Contributions Affect the Efficiency of In A. Atkinson and F. Bourguignon, eds., Public Primary Schools?" Economics of Educa- Handbook of Income Distribution. Amsterdam: tion Review 15(4): 377-86. North-Holland. Jimenez, Emmanuel, and Yasuyuki Sawada. 1999. "Do Kapstein, Ethan B. 1989. "Resolving the Regulator's Community-Managed Schools Work? An Evalu- Dilemma: International Coordination of ation of El Salvador's EDUCO Program." The Banking Regulations." International Organiza- World Bank Economic Review 13(3): 415-41. tion 43(2): 323-47. 246 BIBLIOGRAPHY AND REFERENCES Kates, Robert W., and Viola Haarmann. 1992. "Where Kaufmann, Daniel, Sanjay Pradhan, and Randi the Poor Live: Are the Assumptions Correct?" Ryterman. 1998. "New Frontiers in Anti-cor- Environment (U.S.) 34(May): 4-11, 25-28. ruption Empirical Diagnostics: From In-Depth Kato, Kazu. 1996. "Grow Now, Clean Up Later? The Survey Analysis to Action Programs in Transi- * ' ~~~~~~~~~~~tion Economies." Poverty Reduction and Eco- Case of Japan." In Effectve Financing of Environ- nomic Management Note no. 7. World Bank, mentally Sustainable Development. Environmen- tally Sustainable Development Proceedings Se- Washington, D.C. ries no. 10. Proceedings of the Third Annual Keefer, Philip, and Stephen Knack. 1997. "Why Don't Conference on Environmentally Sustainable Poor Countries Catch Up? A Cross-National Development. Washington D.C.: World Bank. Test of Institutional Explanation." Economic Kaufmann, Daniel. 1998. "Challenges in the Next Inquiry 35(3): 590-602. Stage of Anticorruption." In New Perspectves Keller, Wolfgang. 1995. "International R&D on Combating Corruption. Washington, D.C. Spillovers and Intersectoral Trade Flows: Do Transparency International and the World They Match?" Yale University, New Haven, Bank Institute. Connecticut. Processed. Kaufmann, Daniel, and Yan Wang. 1995. "Macroeco- _. 1998. "Are International R&D Spillovers nomic Policies and Project Performance in the Trade-Related? Analyzing Spillovers among Social Sectors: A Model of Human Capital Randomly Matched Trade Partners." European Production and Evidence from LDCs." World Economic Review 42(8): 1469-8 1. Development 23(5): 751-65. Kelley, Allen C. 1988. "Economic Consequences of Kaufmann, Daniel, and Shan-Jin Wei. 1999. "Does Population Change in the Third World." Jour- 'Grease Money' Speed up the Wheels of Com- nal of Economic Literature 26(4): 1685-1728. merce?" World Bank Policy Research Working . 1998. "The Impacts of Rapid Population Paper no. 2254. World Bank, Washington, D.C. Growth on Poverty, Food Provision, and the Kaufmann, Daniel, Aart Kraay, and Pablo Zoido- Environment." Working Paper. Duke Univer- Lobat6n. 1999a. "Aggregating Govemance In- sity, Chapel Hill, North Carolina. dicators." Policy Research Working Paper no. Kelley, Allen C., and Robert M. Schmidt. 1999. "Eco- 2195. World Bank, Policy Research Depart- nomic and Demographic Change: A Synthesis of ment, Washington, D.C. Available at http:// Models, Findings, and Perspectives." Working www.worldbank.org/wbi/govemance/. Paper no. 99-01. Duke University, Chapel Hill, _____ 1999b. "Governance Matters." Policy Research North Carolina. Forthcoming in Nancy Birdsall, Working Paper no. 2196. World Bank, Policy ed., Populaion Change and Economic Development. Research Department, Washington, D.C. Khan, Mohsin S. 1996. "The Implications of Interna- _____ 2000. "Governance Matters: From Measure- tional Capital Flows for Macroeconomic and ment to Action." Finance and Development. Financial Policies." International Journal of Fi- Available at http://www.imf.org/faudd/2000/ nance and Economics (U.K.) 1 (July): 155-60. 06/kauf.htm. International Monetary Fund, Kim, Jong-Il, and Lawrence J. Lau. 1994. "The Sources Washington, D.C. of Economic Growth of the East Asian Newly 247 B I B LI OG RAI' HY AND RE FE RE NC ES Industrialized Countries." Journal of the Japa- Klenow, Peter and Andres Rodriguez-Clare. 1997a. nese and International Economies 8(3): 235-7 1. "Economic Growth: A Review Essay." Journal Kim, Jooseop, Harold Alderman, and Peter F. Orazem. of Monetary Economics (40): 597-618. 1999. "Can Private School Subsidies Increase . 1997b. "The Neoclassical Revival in Growth Schooling for the Poor? The Quetta Urban Economics: Has It Gone Too Far?" In NBER Fellowship Program." World Bank Economic Macroeconomic Annual 1997. Cambridge, Mas- Review 13(3): 443-65. sachusetts: National Bureau of Economic Re- search. King, Elizabeth M., and M. Anne Hill, eds. 1993. Women's Education in Developing Countries: Klitgaard, Robert. 1988. Controlling Corruption. Berke- Barriers, Benefits, and Policy. Baltimore, Mary- ley, California, and London: University of land: The Johns Hopkins University Press. Califomia Press. King, Elizabeth M., Peter F. Orazem, and Darin Klitgaard, Robert, Ronald Maclean-Abaroa, and H. Wohlgemuth. 1999. "Central Mandates and Lindsey Parris. 2000. Corrupt Cities: A Practical Local Incentives: The Colombia Education Guide to Cure and Prevention. Oakland, Cali- Voucher Program." World Bank Economic Re- fornia and Washington, D.C.: ICS Press and view 13(3): 467-91. World Bank Institute. King, Robert G., and Ross Levine. 1993. "Financial Knack, Stephen, and Philip Keefer. 1995. "Institu- Intermediation and Economic Develop- tions and Economic Performance: Cross- ment." In Colin Mayer and Xavier Viver, Country Tests Using Alternative Institu- eds., Capital Markets and Financial Interme- tional Measures." Economics and Politics 7: diation. Cambridge, U.K.: Cambridge Uni- 207-27. versity Press. . 1997. "Does Social Capital Have an Economic King, Robert G., and Sergio Rebelo. 1990. "Public Payoff? A Cross-Country Investigation." Quar- Policy and Economic Growth: Developing terly Journal of Economics 112: 1251-88. Neoclassical Implications." Journal of Political Knight, J. B., and R. H. Sabot. 1983. "Educational Ex- Economy 98(5) part 2: s126-50. pansion and the Kuznets Effect." American ____. 1993. "Transitional Dynamics and Economic Economic Review 73(5): 1132-36. Growth in the Neoclassical Model." American Economic Review 83 (4): 908-93 1 Knight, John B., and Li Shi. 1991. "The Determinants of Educational Attainment in China." Oxford Kishor, Nalin M., and Luis Constantino. 1994. "Sus- Applied Economics Discussion Paper no. 127. tainable Forestry: Can it Compete?" Finance Oxford University, Oxford, U.K and Development 3 1(4): 36-39. Kornai, Janos. 2000. "Ten Years After 'The Road to a . 1996. "Voting for Economic Policy Reform." Free Economy.' The Author's Self-Evalua- Dissemination Note no 15. World Bank, Latin tion." Paper presented at the Annual Bank America Technical Environment Department, Conference on Development Economics, Washington, D.C. April 20, World Bank, Washington, D.C. 248 B I B LI OG RAP HY AND RE FE RE NC ES Krongkaew, Medhi. 1999. "A Tale of an Economic Quality of Govemment." Journal of Law, Eco- Crisis: How the Economic Crisis Started, De- nomics, and Organization 15(1): 222-79. veloped, and Ended in Thailand." Paper pre- Lam, David, and Deborah Levison. 1991. "Declining sented at the International Conference on ' i Economic Crisis and Impacts on Social Wel- Inequality in Schoolning Brazil and its Ef- fare, June 14-15, Taipei, China. fects on Inequality in Earnings." Journal of De- fare, June 14-15, Taipei, China. velopmental Economics 37(1-2): 199-225. Krueger, Anne 0. 1974. "The Political Economy of the Rent-Seeking Society. American Economic Lanjouw, Peter, and Nicholas Stern. 1989. "'Agricultural Review 64(3): 291-301. Change and Inequality in Palanpur 1957-84." Discussion Paper no. 24. London: London Krueger, Anne O., Maurice W. Schiff, and Alberto School of Economics and Political Science, De- Valdes, eds. 1991. The Political Economy of velopment Economics Research Programme. Agricultural Pricing Policy, vols. 1-5. Balti- . 1998. Economic Development in Palanpur over Five more, Maryland: The Johns Hopkins Univer- sity Press. Decades. New York: Oxford University Press. Lee, Jong Wha. 1996. "Government Interventions and Krugman, Paul. 1996. "The Myth of Asia's Miracle." rdciiy rwh"Juna fEooi Pop Internationalism, Cambridge, Massachu- Productivity Growth." Journal of Economic setts: MIT Press (reprint). Foreign Affairs Growth 1(3): 392-415. 73(6): 62-78, 1994 (original). Lee, Jong-Wha, and Robert J. Barro. 1997. "Schooling . 1998. "Fire-Sale FDI." Paper prepared for the Quality in a Cross-Section of Countries." National Bureau of Economic Research Con- Working Paper no. 6198. National Bureau of ference on Capital Flows to Emerging Markets. Economic Research, Cambridge, Massachusetts. Cambridge, Massachusetts. Processed. Also Leff, Nathaniel H. 1964. "Economic Development available at http://web.mit.edu/krugman/www/ through Bureaucratic Corruption." The Ameri- FIRESALE.htm. can Behavior Scientist 2: 8-14. Kuznets, S. 1955. "Economic Growth and Income and Leipziger, Danny M., ed. 1997. Lessons from East Asia. Income Inequality." American Economic Review Ann Arbor, Michigan: University of Michigan 45:1-28. Press. . 1968. Toward a Theory of Economic Growth: Leipziger, D. M., David Dollar, A. F. Shorrocks, and S. With Reflections on the Economic Growth of Y. Song. 1992. The Distribution of Income and Modern Nations. New York: Norton. Wealth in Korea. Washington, D.C.: World La Porta, Rafael, and Florencio Lopez-De-Silanes. BankInstitute. 1999. "The Benefits of Privatization: Evidence Lerner, A. P. 1944. Economics of Control. New York: from Mexico." Quarterly Journal of Economics Macmillan. 1 14(4):1193-1 242. Levine, Ross. 1997a. "Financial Development and Eco- La Porta, Rafael, Florencio Lopes-de-Silanes, Andrei nomic Growth: Views and Agenda." Journal of Shleifer, and Robert Vishny. 1999. "The Economic Literature 35(2): 688-726. 249 B I B LI OG RAP HY AND RE FE RE NC ES . 1997b. "Napoleon, Bourse, and Growth in Loh, Jonathan, Jorgen Randers, Alex MacGillivray, Latin America." Conference on the Develop- Val Kapos, Martin Jenkins, Brian ment of Securities Markets in Emerging Eco- Groombridge, and Neil Cox. 1998. "Living nomics: Obstacles and Preconditions for Suc- Planet Report 1998." World Wildlife Fund In- cess, October 28-29, Washington, D.C. ternational, New Economics Foundation, and World Conservation Monitoring Center, . 1999. "Law, Finance, and Economic Growth." Journal of Financial Intermediation 8(1-2): 8-35. Gland, Switzerland. Levine, Ross, and Sara Zevos. 1998a. "Capital Control Londofio, Juan Luis. 1990. "Kuznetsian Tales with At- tention to Human Capital." Paper presented at Liberalization and Stock Market Develop- the Third Inter-American Seminar in Eco- ment." World Development 26(7): 1169-83. nomics, Rio de Janeiro, Brazil. 1998b. "Stock Markets, Banks, and Economic L6pez, Ramon 1997. "Protecting the 'Green' Environ- Growth." American Economic Review 88(3): ment in a Context of Fast Economic Growth: 537-58. The Case for Demand-Based Incentives." Uni- Levinsohn, James, Steven Berry, and Jed Friedman. versity of Maryland, College Park, Maryland. 1999. "Impacts of the Indonesian Economic Processed. Crisis: Price Changes and the Poor." Working . 1998a. "Growth and Stagnation in Natural Re- Paper no. 7194. National Bureau of Economic source-Rich Economies." University of Mary- Research, Cambridge, Massachusetts. land, College Park, Maryland. Processed. Lewis, W. Arthur. 1955. The Theory of Economic . 1998b. "Where Development Can or Cannot Growth. New York: Harper Torchbooks. Go: The Role of Poverty-Environment Link- Li, Hongyi, Lyn Squire, and Heng-Fu Zou. 1998. "Ex- ages." Paper presented at the Annual Bank plaining International and Intertemporal Conference on Development Economics, April Variations in Income Inequality." Economic 30-May 1, World Bank, Washington, D.C. Journal 108: 26-43. L6pez, Ramon, and A. Valdes. 2000. Rural Poverty in Lim, David. 1992. "Capturing the Effects of Capital Latin America: Analytics, New Empirical Evi- Subsidies." Journal of Development Studies dence, and Policy. London and New York: 28(4): 705-16. MacMillan Press and St. Martin's Press. L6pez, Ram6n, T. Thomas, and Vinod Thomas. 1998. Lin, Justin Yifu. 1992. "Rural Reforms and Agricultural "economic Growth and the Tusraii of Growth in China." American Economic Review" 82(1): 34-51. Natural Resources." University of Maryland, Department of Agricultural and Resource Eco- Liu, Francis. 1985. "An Equilibrium Queuing Model of nomics, College Park, Maryland. Bribery." Journal of Political Economy 93(4): L6pez, Ram6n, Vinod Thomas, and Yan Wang. 1998. 760-8 1. "Addressing the Education Puzzle: The Distri- Lockheed, E. Marlaine, and Adriaan M. Verspoor. 1991. bution of Education and Economic Reform." Improving Primary Education in Developing Coun- Policy Research Working Paper no. 2031. tries. New York: Oxford University Press. World Bank, Washington, D.C. 250 B I B LI OG RAP H Y AND RE FE RE NC ES Lopez-Mejia, Alejandro. 1999. "Large Capital Flows-A Mamingi, Nlandu, Kenneth Chomitz, David Gray, and Survey of the Causes, Consequences, and Policy Eric Lambin. 1996. "Spatial Patterns of Defor- Responses." Working Paper no. WP/99/17. In- estation in Cameroon and Zaire." Working Pa- ternational Monetary Fund, Washington, D.C. per no 8. Research Project on Social Environ- Loury, Glenn C. 1999. "Social Exclusion and Ethnic mental Consequences of Growth-Oriented Groups: The Challenge to Economics." Paper Policies. World Bank, Policy Research Depart- ment, Washington, D.C. presented at the Annual Bank Conference on Development Economics, April 28-30, World Mankiw, Gregory N., David Romer, and David N. Bank, Washington D.C. Weil. 1992. "A Contribution to the Empirics Lucas, R. 1988. "On the Mechanics of Economic of Economic Growth." Quarterly Journal of Growth." Journal of Monetary Economics 22(1): Economics 105(2): 407-37. 3-42. Mauro, Paolo. 1995. "Corruption and Growth." Quar- Lucas, Robert E. 1993. "Making a Miracle." terly Journal of Economics 110(3): 681-712. Economerrica 61(2): 251-72. _. 1997. "The Effects of Corruption on Growth, Lundberg, Mattias, and Lyn Squire. 1999. "Growth and Investment, and Goverment Expenditure: A Inequality: Extracting the Lessons for Cross-Country Analysis." In Kimberly Ann Policymakers." World Bank, Washington, D.C. Elliot, ed., Corruption and the Global Economy. Washington D.C.: Institute for Intemational Lustig, Nora. 1999. "Crises and the Poor: Socially Re- Economics. sponsible Macroeconomics." Inter-American .1998. "Corruption and the Composition of Development Bank, Sustainable Development Government Expenditure." Journal of Public Department, Poverty and Inequality Advisory Economics 69(2): 263-79. Unit, Washington, D.C. Lvovsky, Kseniya, Maureen Cropper, James Listorti, A. McGranahan, Donald. 1972. Contents and Measure- Edward EmdrCnament of Socioeconomic Development. New York: Edward Elmendorf, Candace Chandra, Julian Pagr Lampiette, Ronald Subida, and Meghan Praeger. Dunleavy. 1999. "Environmental Health." McKinley, Terry. 1996. The Distribution of Wealth in Background Paper for the Environment Strat- Rural China. New York: M. E. Sharpe. egy. Draft. World Bank, Environment Depart- McKinnon, Ronald I., and Huw Pill. 1997. "Credible ment, Washington, D.C.' Liberalizations and International Capital Lynch, Owen J., and Kirk Talbott. 1995. "Balancing Flows: The Over-Borrowing Syndrome." Acts: Community-Based Forest Management American Economic Review, Papers and Proceed- and National Law in Asia and the Pacific." ings 87(2): 189-93. World Resources Institute, Washington D.C. Megginson, William J., and Jeffrey M. Netter. 2000. Maas, Jacob van Lutsenburg, and Geert Criel. 1982. "From State to Market: A Survey of Empirical "Distribution of Primary School Enrollments Studies on Privatization." Working Paper. in Eastern Africa." Working Paper no. 511. Norman, Oklahoma: The University of Okla- World Bank, Washington, D.C. homa Press. 251 B I B LI OG RAI' HY AND RE FE RE NC ES Mehrez, Gil, and Daniel Kaufmann. 2000. "Transpar- Morris, M. D. 1979. Measuring the Condition of the ency, Liberalization, and Banking Crises." World World's Poor: The Physical Quality of Life Index. Bank Policy Research Working Paper no. 2286. New York: Pergamon Press World Bank Institute, Washington, D.C. Mueller, Dennis C. 1998. "Constitutional Constraints Middleton, John, Adrian Ziderman, and Arvil Van on Governments in a Global Economy." Con- Adams. 1993. Skills for Producavity. New York: sntutional Political Economy 9(3): 171-86. Oxford University Press. Munasinghe, Mohan. 2000. "Towards Sustainomics: Milanovic, Branko. 1997. Income Inequality and Poverty A Trans-Disciplinary Metaframework for during the Transition from Planned to Market Making Development more Sustainable." In Economy. Washington, D.C.: World Bank. M. Munasinghe, 0. Sunkel, and C. de Miguel, eds., The Sustainability of Long-Term Mincer, Jacob. 1962. "On the Job Training Costs: Re- Growth: Socioeconomic and Ecological Perspec- tums and Some Implications." Journal of Political tives. London: Edward Elgar. Economy 70(October) supplement, part 2:50-79. Murphy, Kevin, Gary S. Becker, and Edward Glaeser. . 1974. Schooling, Experience, and Earmings. New 1999. "Population and Economic Growth." York: Columbia University Press. American Economic Review 89(May): 145-49. Mingat, Alain, and Jee-Peng Tan. 1998. "The Mechanics Murphy, Kevin M., Andrei Shleifer, and Robert W. of Progress in Education, Evidence from Cross- Vishny. 1989. "Industrialization and the Big Country Data." Policy Research Working Paper Push." Journal of Political Economy 97(5): no. 2015. World Bank, Washington, D.C. 1003-1026. Mink, Stephen D. 1993. "Poverty, Population, and the . 1991. "The Allocation of Talent: Implications Environment." World Bank Discussion Paper for Growth." Quarterly Journal of Economics no. 189. World Bank, Washington, D.C. 106(2): 503-530. Mishkin, Frederic S. 1997. "Understanding Financial Murray, Christopher J. L., and Alan D. L6pez, eds. Crises: A Developing Country Perspective." In 1996. The Global Burden of Disease and Injury Proceedings of World Bank Annual Conference Series, vol. 1. Cambridge, Massachusetts: on Development Economics. Washington, D.C.: Harvard School of Public Health for the World Bank. World Health Organization and World Bank. Montiel, Peter. 1998. "The Capital Inflow Problem." Narayan, Deepa, with Raj Patel, Kai Schafft, Anne Working Paper. World Bank Institute, Wash- Rademacher, and Sarah Koch-Schulte. 2000. ington, D.C. Voices of the Poor: Can Anyone Hear Us? New York: Oxford University Press. Moore, Stephen. 1999. "Corporate Subsidies in the Federal Budget." Testimony before the Budget Narayan, Deepa, and Lant Pritchett. 1999. "Cents and Committee, the U.S. House of Representa- Sociability: Household Income and Social tives, June 30, 1999. Available at http:// Capital in Rural Tanzania." Economic Develop- www.cato.org/testimony/ct-smO63099.html. ment and Cultural Change 47(4): 871-97. 252 BIBLIOGRAPHY AND Rl FERENCES Nehru, Vikram, and Ashok Dhareshwar. 1993. "A Movements and Financial Services in the OECD New Database on Physical Capital Stock: Area. Paris. Sources, Methodology, and Results." Revista de Sourcs,s MEtodoiogy, and Results." Revista ~ Oman, Charles P. 2000. "Policy Competition for For- Analhsis Economico 8(1): 37-59. eign Direct Investment." OECD Development Nehru, Vikram, Eric Swanson, and Ashutosh Dubey. Centre, Paris. Processed. 1995. "A New Database on Human Capital Stock in Developing and Industrial Countries: O'Neill, Donal. 1995. "Education and Income Growth: Sources, Methodology, and Results." Jourwl of Implications for Cross-Country Inequality." Sources, Methodology, and Results." Journal of Ju-o fPltclEooy136:18-31 Development Economics 46(2): 379-401. Journal of Political Economy 103(6): 1289-1301. Nelson, Richard R., and Howard Pack. 1998. "The Ostrom, Elinor. 1990. Goveming the Commons: The Asian Miracle and Modem Growth Theory." Evolution of Institutions for Collective Action. Policy Research Working Paper no. 1881 Cambridge, U.K.: Cambridge University Press. World Bank, Development Research Group, Owen, Ann L. 1995. "Intemational Trade and the Ac- Washington, D.C. cumulation of Human Capital." Board of Gov- emors of the Federal Reserve System, Finance and Economics Discussion Series 95/49 No- Nordhaus, William D., and James Tobin. 1972. Is Eco- vember, Washington, D.C. nomic Growth Obsolete? Fiftieth Anniversary Colqi V.owth Nationlet Bureautof Economiery Panayotou, Theodore. 1997. "Demystifying the Envi- Colloquium V. National Bureau of Economic romna unt ure unn lc Reerch Ne Yok Coumi Unvrst Press.. ronmental Kuznets Curve: Turning a Black Box into a Policy Tool." Environment and De- North, Douglass C. 1989. "Institutions and Economic velopment Economics 2(4): 465-84. Growth: An Historical Introduction." World Development 17(9):1319-32. Park, Jae Ha. 2000. "Korea's Crisis Resolution and Fu- ture Policy Directions." Paper presented at the . 1993. "The Ultimate Sources of Economic World Bank Institute senior policy seminar on Growth." In Adam Szirmai, Bart Vanark, and Managing Capital Flows in a Volatile Finan- Dirk Pilat, eds., Explaining Economic Growth: cial Environment, February 21-24, Bangkok, Essays in Honor of Angus Maddison. Thailand. Amsterdam: North-Holland. Pauly, Louis W. 1995. "Capital Mobility, State Au- Nurkse, Ragnar. 1953. Problems of Capital Formation in tonomy, and Political Legitimacy." Journal of Underdeveloped Countries. New York: Oxford International Affairs 48(2): 369-88. University Press. .vsPe 1997. Who Elected the Bankers? Ithaca, New Obstfeld, Maurice. 1996. "Models of Currency Crises York, and London: Comell University Press. with Self-Fulfilling Features." European Eco- noic Reviewu40(3-5): Features." European Eco-Pearce, David, and Jeremy J. Warford. 1993. World without End: Economics, Environment, and Sus- OECD (Organisation for Economic Co-operation and tainable Development. New York: Oxford Uni- Development). 1990. Liberalization of Capital versity Press. 253 B I B LIOG RAPHY AN D RE F E RE N C E S Persson, Torsten, and Guido Tabellini, eds. 1994. Radelet, Steven, and Jeffrey Sachs. 1998. "The East "Growth, Distribution, and Politics." In Mon- Asian Financial Crisis: Diagnosis, Remedies, etary and Fiscal Policy, vol. 2, Politics. Cam- Prospects." In William C. Brainard and bridge, Massachusetts: MIT Press. George L. Perry, eds., Brookings Papers on Eco- nomic Acrivity, No. 1. Washington, D.C.: The Pritchett, Lant. 1996. "Where Has all the Education Booin Intiitution. Gone?" Policy Research Working Paper no. Brookings Institution. 1581. World Bank, Policy Research Depart- Ram, Rati. 1982a. "Composite Indices of Physical ment, Poverty and Human Resources Division, Quality of Life, Basic Needs Fulfillment, and Washington, D.C. Income: A 'Principal Component' Representa- _____ 1998. "Patterns of Economic Growth: Hills, tion." Journal of Development Economics 11 (Oc- Plateaus, Mountains, and Plains." Policy Re- tober): 227-47. search Working Paper WPS 1947. World . 1982b. "International Inequality in the Basic Bank, Washington, D.C. Needs Indicators." Journal of Development Eco- Psacharopoulos, George, and Ana-Maria Arriagada. nomics 10(1): 113-17. 1986. "The Educational Attainment of the La- . 1990. "Educational Expansion and Schooling bor Force: An International Comparison." Re- Inequality: Intemational Evidence and Some port no. EDT38. World Bank, Washington, D.C. Implications." Review of Economics and Statis- Putnam, Robert D. (with Robert Leonardi and Raffaella tics 72(2): 266-74. Nanetti). 1993. Making Democracy Work: Civic Ramey, Garey, and Valerie A. Ramey. 1995. "Cross- Traditions in Modem Italy. Princeton, New Jer- Country Evidence on the Link between Vola- sey: Princeton University Press. tility and Growth." American Economic Review Qian, Yingyi. 1999. "The Institutional Foundation of 85(5): 1138-51. China's Market Transition." Paper presented Ranis, Gustav, Frances Stewart, and Alejandro at the Annual Bank Conference on Develop- Ramirez. 2000. "Economic Growth and Hu- ment Economics, April 28-30, World Bank, man Development." World Development 28(2): Washington D.C. 197-219. Quinn, Dennis. 1997. "The Correlates of Change in Ravallion, Martin. 1997. "Can High-Inequality Devel- International Financial Regulation." American oping Countries Escape Absolute Poverty?" Political Science Review 91(3): 531-5 1. Economics Letters 56: 51-57. Quinn, Dennis, and Carla Inclan, 1997. "The Origins Ravallion, Martin, and Shaohua Chen. 1997. "What of Financial Openness: A Study of Current Can New Survey Data Tell Us about Recent and Capital Account Liberalization." American Changes in Distribution and Poverty?" World Journal of Political Science 41(July): 771-813. Bank Economic Review 11(2): 357-82. Quinn, Dennis, and A. Maria Toyoda. 1997. "Measur- Ravallion, Martin, and Gaurav Datt. 1998. "Why ing International Financial Regulation." Have Some Indian States Done Better Than Georgetown University, Washington, D.C. Others at Reducing Rural Poverty?" Economica Processed. 65: 17-38. 254 B I B LI OG RAP HY AND RE FE RE NC ES . 1999. "When Is Growth Pro-Poor? Evidence Rodriguez, Andres Gerardo. 1993. "The Division of from the Diverse Experiences of India's Labor, Agglomeration Economics, and Eco- States." Policy Research Working Paper no. nomic Development." Ph. diss., Stanford Uni- 2263. World Bank, Development Research versity, Palo Alto, Califomia. Group, Washington, D.C. Group, Washington, D.C. Rodrik, Dani. 1994. "Getting Interventions Right: Ravallion, Martin, and Binayak Sen. 1994. "Impacts How South Korea and Taiwan Grew Rich." on Rural Poverty of Land-Based Targeting: National Bureau of Economic Research Work- Further Results for Bangladesh." World Devel- ing Paper no. 4964. National Bureau of Eco- opment 22(6): 823-38. nomic Research, Cambridge, Massachusetts. Ravallion, Martin, Mark Heil, and Jyotsna Jalan. 1997. . 1997a. "TFPG Controversies, Institutions, and "A Less Poor World, but a Hotter One?" Economic Performance in East Asia." Working Working Paper no. 13. Research Project on Paper 5914. National Bureau of Economic Re- Social and Environmental Consequences of search, Cambridge, Massachusetts. Growth-Oriented Policies.World Bank, Policy Research Department, Washington, D.C. . 1997b. Has Globalizaton Gone too Far? Wash- ington, D.C.: Institute for Intemational Eco- Rawls, John. 1971. A Theory of Justice. Cambridge, nomics. Massachusetts: Harvard University and Belknap Press. 1. 998. "Where Did All the Growth Go? Extemal Shocks, Social Conflict, and Growth Collapses." Razin, Assaf, and Andrew K. Rose. 1994. "Business- Discussion Paper Series 1789. Center for Eco- Cycle Volatility and Openness: An Explor- nomic Policy Research, London. Available at atory Cross-Sectional Analysis." In Leonardo www.cepr.demon.co.uk/pubs/papers.htm. Leiderman and Assaf Razin, eds., Capital Mo- bility: The Impact on Consumption, Investment, . 1999. The New Global Economy and Developing and Growth. Cambridge, U.K.: Cambridge Countries: Making Openness Work. Policy Essay University Press. no. 24. Washington, D.C.: Overseas Develop- ment Council (distributed by The Johns Rebelo, Sergio. 1991. "Long-Run Policy Analysis and Hopkins University Press). Long-Run Growth." Journal of Political Economy 99(3): 500-21. Roemer, John E. 1993. "A Pragmatic Theory of Re- sponsibility for the Egalitarian Planner." Phi- Reinhart, Carmen, and Graciela Kaminsky. 1999. "On losophy and Public Affairs 22: 146-66. Crises, Contagion, and Confusion." Processed. Rogers, Everett. 1983. Diffusion of Innovations. New Robboy, Rebeca. 1999. Today, Sept. 8. World Bank, York: Free Press. Washington, D.C. Romer, Paul. 1986. "Increasing Returns and Long Run Roberts, J. Timmons, and Peter E. Grimes. 1997. "Car- 'rowth." gourng of al Long Run bon Inensit and Eonomi DevelpmentGrowth." Journal of Political Economy 90(6): bon Intensity and Economic Development 1002-37. 1962-91: A Brief Exploration of the Environ- mental Kuznets Curve." World Development . 1990. "Endogenous Technological Change." 25(2): 191-98. Journal of Political Economy 98(5): s71-102. 255 B I Bf t. OG RA P HY AN D RE F E RE N C E S . 1993. "Two Strategies for Economic Develop- Sally, Razeen. 1998. "Classical Liberalism and Intema- ment: Using Ideas and Producing Ideas." In tional Economic Order: An Advance Sketch." Proceedings of the Annual World Bank Confer- Constitutional Political Economy 9(1): 19-44. ence on Development Economics 1992 Supple- Sandel, Michael J. 1996. Democracy's Discontent: ment. Washington, D.C.: World Bank. ment. Washington, D.C.: World Bank. America in Search of a Public Philosophy. Cam- Rose-Ackerman, Susan. 1989. "Corruption and the bridge, Massachusetts: The Belknap Press of Private Sector." In Arnold J. Heidenheimer, Harvard University. Michael Johnston, and Victor T. Levine, eds., Pical Corution: A H . new Schaffer, Mark E. 1995. "Govemment Subsidies to En- Brunswick,a New Co ersey: Aandboxford,U.K.:terprises in Central and Eastem Europe: Bud- TransaictioNew Books. y; and Oxford, U.K.-getary Subsidies and Tax Arrears." In David Transaction Books. M. B. Newberry, ed., Tax and Benefit Reform in . 1997a. Corruption and Government: Causes, Central and Eastern Europe. London: Centre Consequences, and Reform. Cambridge, U.K.; for Economic Policy Research. New York; and Melbourne: Cambridge Uni- versity Press. Schmidheiny, Stephan, and Federico Zorraquin. 1996. Financing Change. Cambridge, Massachusetts: . 1997b. "The Political Economy of Corruption." MIT Press. In Kimberly Ann Elliot, ed., Corruption and the GloKimbarly Econoy Washingt,edCorrupton D. nstie Schultz, T. Paul. 1998. "Inequality in the Distribution Global Economy. Washington D.C.: Institute of Personal Income in the World: How It Is for Intemational Economics. Changing and Why." Journal of Population Eco- Ruggie, John G. 1983. "International Regimes, Trans- nomics 11(3): 307-344. actions, and Change: Embedded Liberalism in acthPonstwar EConoe:EmiceOder. inbStephesn D Schultz, T. W. 1961. "Investment in Human Capital." the Postwar Economic Order." In Stephen D. Aeia cnmcRve 1l:-7 Krasner, ed., Intemational Regimes. Ithaca, New York: Cornell University Press. Selden, Thomas M., and Daqing Song. 1994. "Envi- Ruitenbeek, H. Jack. 1989. "Social Cost-Benefit ronmental Quality and Development: Is There Analysis of the Korup Project, Cameroon." a Kuznets Curve for Air Pollution?" Journal of Environmental Economics and Management World Wildlife Fund for Nature (London). 2() 4-2 Processed. ~~~~~~~~~27(2): 147-62. P'rocessed. Ruzindana, Augustine. 1997. "The Importance of Sen, Amartya K. 1980. "Equality of What?" In S. Leadership in Fighting Corruption in McMurrin, ed., Tanner Lectures on Human Val- Uganda." In Kimberly Ann Elliot, ed., Corrup- ues, vol 1. Cambridge, U.K.: Cambridge Uni- ton and the Global Economy. Washington D.C.: versity Press. Institute for Intemational Economics. _. 1988. "The Concept of Development." In H. Chener and T. N. Srinivasan, eds., Handbook Sachs, Jeffrey, Aaron Tornell, and Andres Velasco. ' ~~~~~~~of Development Economics, vol. I. New York: 1996. "Financial Crises in Emerging Markets: EfseveroScencnubiss. The Lessons from 1995." Discussion Paper no. 1759. Harvard University, Harvard Institute of _. 1994. "Economic Regress: Concepts and Fea- Economic Research, Cambridge, Massachusetts. tures." In Proceedings of the World Bank Annual 256. B I B LI OGRAP HY AND RE FE RENCES Conference on Development Economics. Wash- Smith, John H. 1973. "Aggregation of Preferences with ington, D.C.: World Bank. Variable Electorate." Econometrica 41(6): . 1997a. "Development Thinking at the Begin- 1027-41. ning of the 21st Century." Development Eco- Solow, Robert M. 1997. "Georgescu-Roegen Versus nomics Research Programme No. 2 (March), Solow/Stiglitz." Ecological Economics 22(3): London School of Economics. 267-68. . 1997b. "What Is the Point of a Development Srinivasan, T. N. 1995. "Long-Run Growth Theories Strategy?" Development Economics Research and Empirics: Anything New?" In Takatoshi Programme No.3 (April), London School of Ito and Anne Krueger, eds., Growth Theories in Economics. Light of the East Asian Experience. Chicago: Sengupta, Jati K., and Karl A. Fox. 1969. Economic University of Chicago Press. Analysis and Operations Research: Optimization . 1997. "As the Century Turns: Analytics, Techniques in Quantitative Economic Models. Empirics, and Politics of Development." Dis- Amsterdam: North-Holland. cussion Paper no. 783. Yale University, New Shafik, Nemat. 1994. "Economic Development and En- Haven, Connecticut. vironmental Quality: An Econometric Analy- . 2000. "Growth, Poverty, and Inequality." Yale sis." Oxord Economic Papers 46(5):757-73. University, New Haven Connecticut. Processed. Shafik, Nemat, and Sushenjit Bandyopadhyay. 1992. Steinherr, Alfred. 1998. Derivatives: The Wild Beast of "Economic Growth and Environmental Qual- Finance. New York: John Wiley. ity: Time Series and Cross-Country Evidence." Stern, David I., Michael Common, and Edward Policy Research Working Paper WPS 904. Barbier. 1996. "Economic Growth and Envi- Worl Bankc, WUashington, D.C. ronmental Degradation: The Environmental Shleifer, Andrei, and Robert W. Vishny. 1994. "The Kuznets Curve and Sustainable Development." Politics of Market Socialism." Journal of Eco- World Development 24(7): 1151-60. nomic Perspectives 8(2): 165-76. Stem, Nicholas. 1998. "The Future of the Economic =_.__ eds. 1998. The Grabbing Hand: Govemment Pa- Transition." Working Paper (International) thologies and Their Cures. Cambridge, Massa- no. 30. European Bank for Reconstruction and chusetts: Harvard University Press. Development, London. Sieh Lee, M. L. 1998. "Competing for Foreign Direct Stiglitz, Joseph E. 1975. "The Theory of 'Screening,' Investment: The Case of Malaysia." Processed. Education, and the Distribution of Income." Simon, Julian. 1976. "Population Growth May Be American Economic Review 65(): 283-300. Good for LDCs in the Long Run: A Richer . 1993. "The Role of the State in Financial Mar- Simulation Model." Economic Development and kets." In Proceedings of the World Bank Confer- Cultural Change 24(2): 309-337. ence on Development Economics, 1993. Wash- Slottje, Daniel J. 1991. "Measuring the Quality of Life ington. D.C.: World Bank. across Countries." Review of Economics and Sta- . 1997. "Georgescu-Roegen Versus Solow/ tistics 73(4): 684-93. Stiglitz." Ecological Economics 22(3): 269-70. 257 B I B L 1 OC( R API' HY AND RE FE RE NC ES _.___* 1998. "More Instruments and Broader Goals: Temple, Jonathan, and Paul A. Johnson. 1998. "Social Moving toward the Post-Washington Consen- Capability and Economic Growth." Quarterly sus." Paper presented at the 1998 World Insti- Journal of Economics 63(3): 965-90. tute for Development Economics Research Thomas, Vinod. 1998. "Economic Globalization and Annual Lecture, January, Helsinki, Finland. Sustainable Development in Costa Rica." Pa- . 1999. "Whither Reform? Ten Years of the Tran- per presented at the conference on Stability sition." Paper presented at the Annual Bank and Economic Development in Costa Rica: Conference on Development Economics, April The Pending Reforms, April 23-25, Acadamy 28-30, World Bank, Washington D.C. of Central America, Costa Rica. Stiglitz, Joseph E., and Amar Bhattacharya. 1999. Thomas, Vinod, and Tamara Belt. 1997. "Growth and "Underpinnings for a Stable and Equitable Environment: Allies or Foes." Finance and De- Global Financial System: From Old Debates velopment 34(June): 22-24. to a New Paradigm." Paper presented at the Thomas, Vinod, and Yan Wang. 1996. "Distortions, Annual Bank Conference on Development Interventions, and Productivity Growth: Is Economics, April 28-30, World Bank, Wash- East Asia Different?" Economic Development ington D.C. and Cultural Change 44(2): 265-88. Summers, Lawrence H. 2000. "A New Framework for . 1997. "Education, Trade, and Investment Re- Multilateral Development Policy." Remarks to turns." Working Paper. World Bank Institute, the Council on Foreign Relations, March 20, Washington, D.C. New York. Thomas, Vinod, Nalin Kishor, and Tamara Belt. 1998. Summers, Robert, and Alan Heston. 1991. " The Penn "Embracing the Power of Knowledge and Part- World Table (Mark 5): An Expanded Set of nerships for a Sustainable Environment." International Comparisons. 1950-88." Quar- Background paper prepared for the World De- terly Journal of Economics 106(2): 32 7-68. velopment Report 1998/99. Processed. Tan Jee-Peng, Julia Lane, and Gerard Lassibille. 1999. Thomas, Vinod, Yan Wang, and Xibo Fan. 2000. "Student Outcomes in Philippine Elementary "Measuring Education Inequality: Gini Coeffi- Schools: An Evaluation of Four Experiments." cients of Education." Working Paper. World World Bank Economic Review 13(3): 493-508. Bank Institute, Washington, D.C. Tanzi, Vito. 1998. "Corruption around the World: Townsend, Robert. 1999. "Agricultural Incentives in Causes, Consequences, Scope, and Cures." In- Sub-Saharan Africa: Policy Challenges." ternational Monetary Fund Staff Papers 45(4): Technical Paper no. 444. World Bank, Wash- 559-94. ington, D.C. Tanzi, Vito, and Hamid Davoodi. 1997. "Corruption, UNCTAD (United Nations Conference on Trade and Public Investment, and Growth." Working Pa- Development). 1994. Directory of Import Re- per no. WP/97/139. International Monetary gimes, Part 1: Monitoring Import Regimes. Fund, Washington, D.C. Geneva: United Nations. 258 B I B LI OG RAP HY AND RE FE RE N CES . 1997. "Income Inequality and Development." Working Paper no. 1644. World Bank, Policy In Trade and Development Report 1997. New Research Department, Washington, D.C. York and Geneva: United Nations. Warford, Jeremy J., Mohan Munasinghe, and Wilfrido UNDP (United Nations Development Program). Cruz. 1997. "The Greening of Economic 1998. Human Development Report 1998. New Policy Reform," vols.1 and 2. World Bank, En- York: Oxford University Press. vironment Department and World Bank Insti- _.___ 2000. Human Development Report 2000. New tute, Washington, D.C. York: Oxford University Press. Warford, Jeremy J., A. Schwab, W. Cruz, and S. Hansen. 1994. "The Evolution of Environ- UNESCO (United Nations Educational, Scientific, mental Concerns in Adjustment Lending: A and Cultural Organization). 1998. World Edu- Review." Working Paper no. 65. World Bank, cation Report. Paris. Environment Department, Washington, D.C. UNRISD (United Nations Research Institute for So- Watson Robert, John Dixon, Steven Hamburg, An- cial Development). 1970. "Studies in the thony janetos, and Richard Moss. 1998. Pro- Methodology of Social Planning." Geneva, tecting Our Planet, Securing Our Future: Link- Van Rijckeghem, Caroline, and Beatrice Weder. ages among Global Environmental Issues and 1999. "Sources of Contagion: Finance or Human Needs. Nairobi and Washington, D.C.: Trade?" Working Paper no. WP/99/146. Inter- United Nations Environment Programme, national Monetary Fund, Washington, D.C. United States National Aeronautics and Space Administration, and World Bank. Verdier, Daniel. 1998. "Domestic Responses to Capital Market Internationalization under the Gold Webb, Michael C. 1994. The Political Economy of Policy Standard, 1870-1914." International Organiza- Coordination: International Adjustment since tion 52(1): 1-34. 1945. Ithaca, New York; and London: Cornell University Press. Vishwanath, Tara, and Daniel Kaufmann. 1999. "To- wards Transparency in Finance and Govern- Wei, Shang-Jin. 1997. "How Taxing Is Corruption on ment." World Bank, Washington, D.C. Pro- International Investors." Working Paper no. cessed. Available at http:// 6030. National Bureau of Economic Research, www.worldbank.org/wbi/governance. Cambridge, Massachusetts. Wang, Hua, and Ming Chen. 1999. "Industrial Firms' _ . 1999. "Corruption in Economic Development: Pollution Control Efforts under a Charge-Sub- Beneficial Grease, Minor Annoyance, or Ma- sidy System: An Empirical Analysis of Chinese jor Obstacle?" Policy Research Working Paper Top Polluters." World Bank, Policy Research 2048. World Bank, Development Research Department, Washington, D.C. Group, Public Economics, Washington, D.C. Wang, Hua, and David Wheeler. 1996, "Pricing Indus- Wessels, J. H. 1993. "Redistribution from a Constitu- trial Pollution in China: An Econometric tional Perspective." Constitutional Political Analysis of the Levy System." Policy Research Economy 4(3): 425-48. 259 B I B LI OG RAP HY AND RE FE PE NC ES Wheeler, David, and Shakeb Afsah. 1996. "Going . 1996b. World Development Report 1996: From Plan Public on Polluters in Indonesia: BAPEDAL's to Market. New York: Oxford University Press. PROPER-PROKASIH Program International 1997a. Can the Environment Wait in East Asia? Executive Reports." World Bank, Washington, P f E D.C. Prorties for East Asia. Washmgton, D.c. .1997b. "Chile: Poverty Reduction and Income WHO (World Health Organization). 2000. The World Distribution in a High-Growth Economy: H-Iealth Report 2000. Geneva. 1987-95." Report no. 16377-CH. World Williamson, John, and Molly Mahar. 1998. "A Survey Bank, Latin America and the Caribbean Re- of Financial Liberalisation." In Essays in Inter- gion, Washington, D.C. national Finance Series, 211. Princeton, New . 1997c. Clear Water Blue Skies. China's Environ- Jersey: Princeton University, Department of ment in the New Century. China 2020 series. Economics, International Finance Section. Washington D.C. Wolfensohn, James D. 1998, October. "Address to the . 1997d. Expanding the Measures of Wealth: Indi- Board of Governors." World Bank, Washing- cators of Environmentally Sustainable Develop- ton, D.C. Processed. ment. In Environmentally Sustainable Devel- . 1999. "A Proposal for a Comprehensive Devel- opment Studies and Monograph Series no. 17. opment Framework." World Bank, Washing- Washington, D.C. ton, D.C. Processed. . 1997e. Five Years after Rio: Innovations in Envi- Woolcock, Michael. 1998. "Social Capital and Eco- ronmental Policy. In Environmentally Sustain- nomic Development: Toward a Theoretical able Development Studies and Monograph Se- Synthesis and Policy Framework." Theory and ries no. 18. Washington, D.C. Society 27: 151-208. 1. 997f. Private Capital Flows to Developing Coun- World Bank. 1990. World Development Report 1990. tries: The Road to Financial Integration. World New York: Oxford University Press. Bank Policy Research Report Series. Oxford, U.K.; and New York: Oxford University Press. 1991. World Development Report 1991: The . . Challenge of Development. New York: Oxford Chin1 . Sh sing Seoes. Disparities in University Press. ~~~~~China. China 2020 Series. Washington, D.C. I]niversity Press. _____ 1992. World Development Report 1992: Develop- 1997h. Trends in Developing Economies. Wash- ment and the Environment. New York: Oxford ington, D.C. University Press. 1. 997i. World Development Indicators. Washing- . 1994. Avertng the Old Age Crisis: Politics to Pro- ton, D.C. tect the Old and Promote Growth. Washington, . 1997j. World Development Report 1997. Oxford D.C.: World Bank and Oxford University University Press: New York. Press. . 1998a. Assessing Aid: What Works, What Doesn't, . 1996a. Global Economic Prospects 1996. Wash- and Why? A World Bank Policy Research Re- ington, D.C. port. Oxford University Press: New York. 260 BIBLIOGRAPHY AND REFERENCES 1998b. East Asia: The Road to Recovery. Wash- . 2000a. Global Economic Prospects and the Devel- ington D.C. oping Countries. Washington, D.C. 1998c. "The Business Environment and Corpo- _. 2000b. World Development Report 1999/2000: rate Governance." Discussion Draft. Private Entering the 21st Century. Washington, D.C. Sector Development Department, Business Senvironmentlopment Department,Business 2000c. World Development Indicators. Washing- Environment Group.to,DC ton, D.C. ___1999a. Pdlucorion sector Strategy. Washington. D.C. ___. 2000d. Greening Inldustry: New Rules for Com- munities, Markets, and Governments. New 1999b. "Environmental Implications of the York: Oxford University Press. Economic Crisis and Adjustment in East Asia." East Asia Environment and Social De- .Z00C. Bolivia, Ecuador, and Paragay Gover- velopment Unit Discussion Paper Series no.1. nance and Anticorruption Empirical Diagnostic Washington, D.C. Studies. Washington, D.C.: World Bank Insti- tute. Available at http://www.worldbank.org/ 1999c. Global Development Finance. Washing- wbi/governance. ton, D.C. _.vdc 2000f. East Asia: Recovery and Beyond. Wash- 1999d. "Poverty Trends and Voices of the ington, D.C. Poor." Poverty Reduction and Economic Man- agement, Human Development, and Develop- -_. 2000g. Engendering Development. New York: ment Economics, Washington, D.C. Pro- Oxford University Press. cessed. Also available at http:// _____. 2000h. Anticorruption in Transition: Confrontng www.worldbank.org/poverty/data. the Challenge of State Capture. Washington, 1999e. World Development Indicators. Washing- D.C. Forthcoming. ton, D.C. _ 2000i. World Development Report 200012001: 1999f. Transition toward a Healthier Environ- Attacking Poverty. Washington, D.C. ment: Environmental Issues and Challenges in the -__. 2000j. "Reforming Public Institutions and Newly Independent States. Washington, D.C. Strengthening Governance: A World Bank l999g. "Towards Collective Action to Improve Strategy." A Sectoral Strategy Paper. Public Govemance and Control Corruption in Seven Sector, Poverty Reduction, and Economic African Countries." Action programs prepared Management Network, Washington, D.C. by African countries for the ninth Annual Inter- natioal Cnfernce gaint Corupton, cro- . Various issues. Commodities Quarterly. Wash- national Conference Against Corruption, Octo- ington, D.C. ber 10-15, Durban, South Africa. Available at http://www.worldbank.org/wbi/govemance. __. Various issues. Global Development Finance. 1999h. "New Empirical Tools for Anticorrup- Washington, D.C. tion and Institutional Reform: A Step-by-Step World Commission on Forestry and Sustainable Devel- Guide to their Implementation." Washington, opment. 1999. "Our Forests, Our Future." In D.C.: Europe and Central Asia Public Sector. Summary Report of the World Commission on 261 BIBLIOGRA['HY AND REFtRENCES Forestry and Sustainable Development. . 1994. "Lessons from East Asian NICS: A Winnipeg, Manitoba, Canada. Contrarian View." European Economic Review Wyplosz, Charles. 1999. "Ten Years of Transformation: 38(3-4): 964-73. Macroeconomic Lesons." Paper presented at . 1995. "The Tyranny of Numbers: Confronting the Annual Bank Conference on Development the Statistical Realities of the East Asian Economics, April 28-30, Washington, D.C. Growth Experience." Quarterly Journal of Eco- Young, Alwyn. 1991. "Learning by Doing and the Dy- nomics 110(3): 641-80. namic Effects of International Tradei" Quar- 7anowirz, Victor. 1999. "Theory and History behind terly Journal of Economics 106(2): 369-405. Business Cycles: Are the 1990s the Onset of a . 1992. "Tale of Two Cities: Factor Accumula- Golden Age?" Joumal of Economic Perspecives tion and Technical Change in Hong Kong and 13(2): 69-90. Singapore." NBER Macroeconomics Annual 1992. Cambridge, Massachusetts; and London: MIT Press. 262 "'Economic growth,' properly understood, has many dimensions imperfectly captured by such single figures as GDP per capita. This World Bank study calls attention to the wider span of con- tributions to human welfare and carefully documents them. The result is a more sober and nuanced assessment of the conditions of life in developing countries." KennethJ. Arrow, Professor, Statnfor-d Univ'ersity and Nobel Laureate, Economics "The book. . . supports the argument that ensuring universal coverage of basic social services of good quality is one of the most effective ... ways of reducing poverty .... The book's message is one of hope, namely that a better world for all can be achieved .... with genuine participation by the poor people themselves." Carol Bellamy, Executive Director, 1NICEF "From the World Bank comes a refreshing new perspective on what development is really all about: improving the quality of people's lives. Here is essential reading that goes beyond growth and markets to other tough challenges: ensuring equal opportunity, reducing people's insecurity, addressing corruption and environmental abuse. Detailed analysis is careful, sophisticated and transparent - must 'reading for scholars and practitioners of development all over the world." Nancy Birdsall Senior Associate. Carnegie Endoitument for International Peace "Combining modern economic theory with a wealth of evidence on what enables some commu- nities to flourish while others regress, this is a terrific book. It can be read with profit by the general public and it can be used fruitfully in courses on economic development. The book has both depth and stvle." Partba Dasgupta, Professor, Unirersity of Camnbridge "The Quality of Grouth represents an important contribution to the broadening of development studies .... It takes account of institutional and political factors as key determinants of growth, and sees growth itself not as an end, but as a means to human welfare ...." Fr-ancis Fukuyanna, Professor, George Mason Unidrersity "Growth can reduce poverty. But the strength of its influence depends crucially on the quality of the growth process, in terms of the participation of the poor, and of its sustainability and sta- bility .... This book provides a lucid and structured synthesis of our current understanding of some of the key issues of development." Nick Stern. Professoi; Londonr School of Economlics , R L The World Bank 1818 H Street N.W. B ~~~~~Washington, D.C. 20433 USA K ~~~~~Telephone: 202-477-1234 3 t / Facsimile: 202-477-6391 I I. o =! Telex: MCI 64145 WORLDBANK .p 0 aMCI 248423 WORLDBANK * Our dream is a Internet: www.worldbank.org Oxford University Press Ourldreem is a woHddfree of poverty E-mail: books@worldbank.org ISBN 0-1 9-521 593-1